a ee i ain i - 
Pi aC NN he ee ea, A cals le EP CE Ee ASS 


/ Fe 2 VOLS f | 


PE 


Library of the Museum 
OF 
‘COMPARATIVE ZOOLOGY, 


AT HARVARD COLLEGE, CAMBRIDGE, MASS. 


Pounded by private subscription, in 1861. 


Deposited by ALEX. AGASSIZ. 


{ / Nowa 277 
/ 
/ os ~ ie) a 
WOd ) = ees |S | ia Mel 
/ } 


SPF 


LOLS 


QUARTERLY JOURNAL 


OF 


MICROSCOPICAL SCIENCE: ~ 


EDITED BY 


E. RAY LANKESTER, M.A., LL.D., F.RB.S., 


Honorary Fellow of Exeter College, Oxford ; Jodrell Professor of Zoology in University 
College, London; and Deputy Linacre Professor of Human and Comparative 
Anatomy in the University of Oxford. 


WITH THE CO-OPERATION OF 


EH. KULELN, MD. BLES; 


Lecturer on General Anatomy and Physiology in the Medical School of 
St. Bartholomew’s Hospital, London, 


AND 


ADAM SEDGWICK, M.A., F.RS., 
Fellow and Assistant-Lecturer of Trinity College, Cambridge. 


VOLUME XXXI.—New Szrizs. 
Mith Rithographic Plates und Engrabings on Wood, 


LONDON: 


J & A. CHURCHILL, 11, NEW BURLINGTON STREET. 
"1890. 


TAATUOL YA 


7 idle 
4 ‘ea aunaud i Sal LAGHOE age 


¢ 
i ii} Ba A: 


Til urs 


Mires rh ae) a ELA CA 
‘ ’ yy Wj ary ye? if wu ah te at Ft | 
Oy gh Stee A sand GR EHSL i 
LAA 
+ i AUT a OaEMe Bea 
TO \ iy of Ay i. teu' jabte indy ehh 


> 
7 
tT AY © 
7 - 
. 
ri a 
- EY 
fi ” 
A b 


* AVG Oud 
Tae VOUOALIG PW if. Hd ae 
Oger ; 


CONTENTS. 


CONTENTS OF No. CXXI, N.S., APRIL, 1890. 


MEMOIRS : 


On Phymosoma varians. By Artuur HE. Surrtey, M.A, 
Fellow and Lecturer of Christ’s College, Cambridge, and De- 
monstrator of Comparative Anatomy in the University. (With 

- Plates I, II, III, and IV) 


The Spinning Apparatus of Geometric Spiders. By Cecit War- 
BuRTON, B.A., Christ’s College, Cambridge. (With Plate Y) . 


On the Structure and Functions of the Cerata or Dorsal Papillee in 
some Nudibranchiate Mollusca. By W. A. Herpmay, D.Sc., 
F.L.S., Professor of Natural History in University College, 
Liverpool. (With Plates VI, VII, VIII, IX, and X) ; 


Further Observations on the Histology of Striped Muscle. By 
C. F. Marswatt, M.B., M.Sc., late Platt Sc Scholar 
in the Owens College. (With Plate XI) 

On Cheetobranchus, a New Genus of Oligochextous eneedatil 
By Atrrep Grszs Bourne, D.Sc.Lond., F.L.S., C.M.Z.S., Fellow 
of University College, London, and of the Madras University. 
(With Plate XII) : : 

The Presence of Ranvier’s Constrictions in the Spinal Cord of 
Vertebrates. By Dr. Wittiam TownsEenD Porter, of St. Louis. 
(With Plate XII dis) 


Prorsssor Birscuu’s Experimental Imitation of Protoplasmic 
Movement 


PAGE 


29 


41 


65 


83 


91 


99 


CONTENTS. 


CONTENTS OF No. CXXII, N.S., JUNE, 1890. 


MEMOIRS : 


The Embryology of a Scorpion (Euscorpius italicus). By 
Matcoitm Lavriz, B.S8c., Falconer Fellow of Edinburgh Uni- 
versity. (With Plates XIII—XVIII) 


On the Morphology of the Compound Eyes of Arthropods. By § 
Wartasz, Fellow of the Johns Hopkins ones (With Plate 
XIX) : ; 


On the Structure of a Species of Earthworm belonging to the Genus 
Diacheta. By Franx EH. Bepparp, M.A., Prosector to the 
Zoological Society of London. (With Plate XX) 


Hekaterobranchus Shrubsolii, a New Genus and Species of 
the Family Spionide. By Fiorence Bucuanay, Student of 
University College. (With Plates XXI and XXII) 


An Attempt to Classify Karthworms. By W. B. Benuam, D.Sc., 
Assistant to the Jodrell Professor of Sere in a 
College, London . 2 


CONTENTS OF No. CXXIII, N.S., AUGUST, 1890. 
MEMOIRS: 


On the Origin of Vertebrates from Arachnids. By WiL1i1Am Patten, 


Ph.D., Professor of Biology in the University of North Dakota, 
Grand Forks. (With Plates XXIII and XXIV) 


On the Origin of Vertebrates from a Crustacean-like Ancestor. By 
W. H. Gasxewt, M.D.,F.R.S. (With Plates XXV, XXVI, 
XXVII, and XXVIII) 


The Development of the Atrial Chamber of Amphioxus. By 
E. Ray Lanxester, M.A., LL.D., F.R.S., and AntHUR WILLEY, 
Student of University College. (With Plates XXIX, XXX, 
XXXI, and XXXII) : i : 


PAGE 


105 


148 


159 


175 


201 


317 


379 


445 


CONTENTS. Vv 


CONTENTS OF No. CXXIV, N.S., NOVEMBER, 1890. 
MEMOIRS: PAGE 


On the Structure of a New Genus of Oligocheta (Deodrilus), 
and on the Presence of Anal Nephridia in Acanthodrilus. 
By Franx E. Bepparp, M.A., Prosector of the Zoological So- 
ciety of London. (With Plates XXXIII and XXXIIJa) . 467 


Excretory Tubules in Amphioxus lanceolatus. By F. Ernest 
Weiss, B.Sc., F.L.8., University College, London. a 
Plates XXXIV and XXXV) : ; 489 


Studies in Mammalian Embryology. II.—The evalgoen of the 
Germinal Layers of Sorex vulgaris. By A.A. W. Huprecut, 
LL.D., C.M.Z.S., Professor of Zoology in the University of 
ipeeht (With Plates XXX VI—XLII) : : 499 


Terminations of Nerves in the Nuclei of the Epithelial Cells of 
Tortoise-shell. By Joun Berry Haycrart, M.D., D.Sc. (from 
the Physiological Laboratory of the Faia o yr 
(With Plate XLIIT) . ; : . 563 


TitTLE, ConTENTS, AND INDEX. 


e 
» 1 
= ¥ Re Ai) 
on 
- : 
on s 
_ 
a H . 1‘. 
7 wet Ait a ae im ¥e ne .: LAA sO fk ay) HM | 
e MAT ‘ A he 
Sandribehil) alaivo iO) li Ati earl ae eerie 
AMA Ou! A ond. albdullee kA Ve feasliina') Sie 
‘ i ri “ « s “yy 
ailts 1%) a LI aly Sel ; : SL. ch gdh PALU Eds Hi mau’ il 
Yue. (ARLLE AL Pe CLASS etal Uy) Simlincel iG ehiyG 
qgeann |. Evil. setilosed al @egol lens oF muda yee 


Q if? oat aloe Nera ohh Sa af ome 

2 OO ee (PURE Und, VOLES A 

: - ne hd inate jopstl alts ell Se og a ee ee alge . 
' ent: nite. Vie TU RA 39 fF ‘me tty a 
FS s, rye ur i ei ret) a." eeanl's ee idels whe 


UB . ‘a . Se 4 hie he y eos b3 itd iit 2 
a To MO leet a job awh = rr 
. ' ‘ i 
wen CY vitae be vitae ite | 21 cin adn tages 


- stint nok Ly wiaptsyy eae hy Fh et teh idea ots “alee 
; ald re 7 A © . . , iJ lek 4 bake 


4 s . z -F50HL AY get ve bbe 
‘ ‘ 
q 
a 
. J J " 
a 
A a 
a G- 
a 
“ D 
» La 
- i 7 
- 
7 i. 
than oul 4 


On Phymosoma varians. 
By 


Arthur E, Shipley, M.A., 


Fellow and Lecturer of Christ’s College, Cambridge, and Demonstrator of 
Comparative Anatomy in the University. 


With Plates I, Il, III, IV. 


Tue material which forms the basis of the following paper 
was collected and preserved by Mr. W. F. R. Weldon, of 
St. John’s College, Cambridge, during a visit to the Bahamas. 
On his return to England Mr. Weldon commenced to work at 
Phymosoma, and made many microscopic sections and 
drawings. When, however, he received the appointment which 
he now holds at Plymouth he handed the whole material, 
together with his drawings, to me, with a request that I would 
complete the work thus interrupted. This statement will serve 
to show how much I am indebted to Mr. Weldon, both for 
material and for many of the drawings ; but I have further to 
express my indebtedness to him for many suggestions and 
much help in completing the work he was unfortunately obliged 
to lay aside. 

The observations here recorded were made on a species of 
Phymosoma (Ph. varians, Selenka) collected in the Bahama 
Islands. 

This species was sufficiently common in the island of New 
Providence; but it occurred still more abundantly in the 
lagoon of the Bemini atoll. The specimens were obtained by 
breaking up soft masses of coral rock with a hammer. Pieces 

VOL. XXXI, PART I,—NEW SER. A 


2 ARTHUR E. SHIPLEY. 


of rock which were completely covered at low water contained 
many more specimens than those which were left dry by the 
tide. 

The species seems to be capable of much variation; and the 
descriptions hitherto published are incomplete in one or two 
important points. A detailed account of the external cha- 
racters may therefore be not altogether useless. 


EXTERNAL CHARACTERS AND EcTopERM. 


The length of fully extended specimens averages 50 mm., 
varying, however, from about 40 mm.to 55 mm. The greatest 
diameter of the trunk is from 4 mm. to 5 mm.; that of the 
introvert about 2 mm. The introvert is at least equal in 
length to the rest of the body. 

The head (figs. 1 and 5) bears a crown of about eighteen 
tentacles, arranged in the form of a horseshoe, with the open 
ends directed backwards; the whole structure lying far back 
on the dorsal region of the head (fig. 1). The ends of the 
teutacular horseshoe are connected with the lower lip; which 
is a thick vascular crescent enclosing considerably more than 
three fourths of the circumference of the head (figs. 2 and 6). 
The mouth is a narrow crescentiform slit, extending between 
the dorsal margin of the lower lip and the convex surface of 
the crown of tentacles. These relations of tentacular crown, 
mouth, and lower lip are shown in the diagram (figs. 1 and 32). 
It will be seen that in this species the condition of the head 
presents a marked resemblance to that which obtains in 
Phoronis. 

The tentacles themselves are short and simple, the surface 
directed towards the outer (convex) side of the lophophor 
being grooved, and the groove is ciliated; the opposite surface 
is covered with a deep brown pigmented epithelium (fig. 5). 

The space included within the concavity of the lophophor 
(the representative of the przoral lobe) is covered with a 
wrinkled, pigmented skin. In its centre lies a deep depression, 
similar to that of Sipunculus, at the base of which lies the 


ON PHYMOSOMA VARIANS, 3 


brain; while a sense-pit opens on to it on each side! (figs. 1 
and 7). 

The introvert is dividable into several regions. Immediately 
behind the head follows a narrow, perfectly smooth region, 
extending for about 2 mm. At the posterior edge of this 
region is attached a small but very extensile collar, its 
anterior margin being free (figs. 1 and 4.) Behind the 
attachment of the collar the introvert swells slightly, and 
there follows a region about 6 mm. in length, which bears 
about twenty rows of hooks. Then follows a region of variable 
length, bearing papille; and lastly a second region of hooks, 
which in our specimens bore from forty to between fifty and 
sixty rings. Among the hooks of the posterior region are 
many papille; and these in passing backwards get more and 
more conspicuous, at the expense of the rings of hooks. 
These papille also exhibit traces of a tendency to form rings 
round the base of the proboscis. The characters of the hooks 
have been well described by Selenka and by Keferstein:? it 
will be sufficient here to refer to description given by these 
authors, and to the drawing (fig. 21). 

The papille on the introvert have the form shown in fig. 15; 
they are hemispherical or hemielliptical, being often higher 
than broad, each having a central opening surrounded by 
three or four plates of chitin, which often fuse into a single 
piece; and surrounding this central piece are numerous small 
rounded plates covering at least the upper half of each papilla. 

The papille on the trunk (figs. 11, 14, and 16) have a some- 
what different appearance, being larger and flatter, and having 
no marked central plate. They are also surrounded by a much 
pigmented ring. These trunk papille agree with the de- 
scription given by Selenka, who, however, seems to have over- 
looked the difference between the papille in the two regions of 
the body. The papillz are large and conspicuous at the two 
extremities of the trunk, where they are present on all sides ; 


1 Cf., Speugel, “ Die Sipunculiden,” ‘ Reisen im Archipel der Philippinen,’ 


Bd. iv, 1883. 
2 Selenka, loc. cit., ‘Keferstein, Zeit. fiir Wiss. Zool.,’ Bd. xv, 1865, 


4, ARTHUR E. SHIPLEY. 


in the middle of the body they are, however, almost entirely 
confined to the dorsal surface. These papille are shown in 
fig. 11. 

The colour varies in different specimens. The ground 
colour is always yellowish-brown, with a peculiar iridescence, 
noticed by other observers: on this are patches of a black or 
deep brown pigment, which are generally so arranged as to 
form a few irregular rings in the middle of the introvert and 
smaller patches on the anterior dorsal part of the trunk. 
Individuals are, however, found in which the pigment is only 
very slightly developed; while in others the whole dorsal 
surface of the body is thickly mottled with dark patches. 

The body wall is everywhere covered by an ectodermal 
epithelium, one cell thick. The characters of the cells pre- 
senting marked differences in different regions. 

The ectoderm covering the lower lip and the outer 
grooved surface of the tentacles is columnar and covered with 
short thickly set cilia (figs. 4 and 8). 

The preoral lobe, together with the inner surface of the 
tentacles, is covered by a layer of cubical cells, the outer half 
of each cell in this region being loaded with granules of a dark 
brown pigment (figs. 4, 7, and 8). These cells are not ciliated. 

The epithelium covering the collar is formed of short 
cubical cells, which appear to become more flattened when this 
organ is extended (fig. 4). 

On the remainder of the introvert the ectoderm secretes, 
except in the region of the hooks and papille, a clear homoge- 
neous cuticle 0:02 mm. thick. 

Each hook is secreted by a raised papilla, which projects 
into the cavity of the hook. The cells covering the papilla 
being large and cubical, provided with conspicuous spherical 
nuclei (fig. 21). 

Behind each hook is a small organ, apparently sensory, 
which will be described below. 

The ectoderm of the trunk consists of lamellar, dome-shaped 
cells, secreting a thick cuticle almost ‘04 mm. in thickness 
(fig. 13). The outer surface of this cuticle is rough and 


ON PHYMOSOMA VARIANS. 5 


granular; and it absorbs staining fluids with a certain readi- 
ness, while the main body remains in all the preparations 
quite unstained. The cuticular substance appears in the 
greater part of the body to be arranged in wavy columns, 
running more or less regularly at right angles to the surface 
of the body, and resting each on a single ectoderm cell (fig. 
10). Each column exhibits a further tendency to a laminated 
structure, the layers composing it lying concentrically to the 
body of the animal. 

A result of the peculiar shape of the ectoderm cells in the 
trunk-region is the formation beneath them of a series of small 
cavities, containing a coagulum. By a kind of lifting up of 
several cells from the adjacent muscles, these cavities commu- 
nicate with one another and so attain a considerable size (fig. 
10). They communicate with the cavities, to be presently 
described, which lie between the two layers of the papille 
(fig. 16). 

The function of these channels is in all probability con- 
nected with the circulation of the nutrient fluids; but I have 
not succeeded in tracing a connection between these and any 
other of the cavities of the body. The analogy between these 
spaces and the dermal spaces of Sipunculus need hardly be 
pointed out. A surface view of the skin shows that the 
cuticle is broken up into a series of fusiform areas (fig. 11). 
These areas roughly correspond with the skin-papille, the 
lines limiting them being formed by thickened portions of 
cuticle. When the animal is in an expanded condition the 
areas become thicker and shorter. 

The papille of the introvert and trunk are entirely ecto- 
dermal. Their external appearance has already been de- 
scribed ; the arrangement seen in section is shown in figs. 14 
and 16. 

The cuticle seems, in the region round the base of each 
papilla, to contain irregular spaces, as if its inner and outer 
surfaces had been pulled apart, an appearance which may, of 
course, be due to the action of the knife used in cutting sec- 
tions. On the papilla itself, the plates seen in surface views 


6 ARTHUR E. SHIPLEY. 


are visible as local thickenings of the cuticle, and are often 
loaded with a bright yellow-brown pigment. 

The body of the papilla has the form of a double cup, as if 
it had been formed by the invagination of a spherical out- 
growth of the general ectoderm. The outer layer of the cup 
is composed of flattened cells, which are continuous with those 
of the general ectoderm at the base of the papilla, and with 
those of the inner cup at its apex. The inner layer of the cup 
consists of large cells, loaded with granules of a bright yellow 
substance, so that the remains of their protoplasm are seen as 
slender strings of stained material, separating masses of the 
yellow formed material. This inner cup contains a small 
cavity, which communicates with the exterior by the pore at 
the apex of the papille. Between the two cups is a cavity, 
continuous with the subepidermal system of spaces above 
mentioned. 

In the absence of a detailed knowledge of the habits of the 
living Phymosoma it would be rash to assign any function to 
these very curious organs, but it seems not improbable that 
the secretion they produce may assist in softening the coral 
rock in which the animals form long tubular passages. 


GENERAL ANATOMY. 


The arrangement of the internal organs is shown in fig. 3 
which represents a Phymosoma cut open longitudinally and 
the body wall turned back to expose the viscera. The intro- 
vert is invaginated to almost its full extent, the true anterior 
end of the body being at the point where the sense-pits lie. 

The longitudinal and circular muscles of the skin have been 
omitted for the sake of clearness ; a detailed description of them 
is given below. 

The retractors of the introvert are four in number. They 
fuse round the first half of the csophagus forming a muscular 
tube, and then separate into a dorsal and a ventral pair. The 
former are much the shorter pair; between them lies the 
dorsal blood-vessel, whilst the ventral pair have at their base 
the generative ridge and between them the nerve-cord. The 


ON PHYMOSOMA VARIANS. 7 


spindle muscle supporting the alimentary canal is shown 
running up the axis of the intestinal coil. The cesophagus is 
anteriorly surrounded by the retractor muscles, but the poste- 
rior half is free and ends in the coiled intestine. The number 
of coils varies, usually there are about fifteen. The intestine 
forms a thicker tube than the cesophagus, it ends in the 
rectum which passes straight to the anus in the dorsal middle 
line. 

The only part of the vascular system visible is the crumpled 
dorsal vessel. 

The brain is indicated through the walls of the introvert, 
and close behind it, at the sides, two black spots, the sense- 
pits, are visible ; the ventral nerve-cord is seen running down 
the body. 

The nephridia or brown tubes are conspicuous objects, vary- 
ing very much in size and shape in different individuals. 
Their external opening is at the anterior end and a little in 
front of the level of the anus. The opening is followed by a 
short neck which opens into the swollen portion or bladder 
which passes into the true secreting portion. The anterior 
half of the nephridia is attached to the body wall by muscle- 
fibres, the posterior is free (fig. 18). 

The generative ridge runs across the body at the base of the 
ventral retractors (fig. 22). It is sometimes V-shaped, the 
ridges slanting backward in the middle ventral line. 


Tue Muscurtar System, 


The muscular system is composed throughout of fusiform 
fibres with simple pointed ends. Lach fibre consists of an 
outer contractile and an inner granular portion, the outer por- 
tion being longitudinally striated. The elongated oval nucleus 
lies entirely within the inner layer, the nucleus and the con- 
tractile layer being easily stained, while the inner substance 
does not absorb staining fluids (figs. 13 and 21). 

The fibres of the retractor muscles are much larger than 
those of the body wall, their diameter being at least twice as 
great. 


8 ARTHUR E. SHIPLEY. 


The fibres of the general body wall are arranged in an 
external circular and an internal longitudinal layer, separated 
by an exceedingly delicate layer of oblique fibres. This latter 
can only be seen in surface views, as, owing to its extreme 
thinness, it is difficult to detect in sections. 

The circular muscles commence behind the collar fold, 
where they form a series of rings round the introvert, one 
lying beneath each ring of hooks (fig. 1). Posteriorly to the 
hook-bearing region the circular fibres form a continuous 
sheath, which extends to the posterior end of the animal 
(fig. 22). 

The longitudinal fibres form a complete sheath round 
the introvert, commencing anteriorly just behind the attach- 
ment of the collar. At the posterior extremity of the intro- 
vert these fibres separate into longitudinal bundles, generally 
about twenty-two in number, which run parallel with one 
another down the trunk. In passing backward these bundles 
gradually fuse with one another, and so become fewer and 
larger, till near the “tail” they form a series of projecting 
ridges, giving to a section of the body-cavity in this region a 
characteristic star-shaped appearance (fig. 18). At the poste- 
rior extremity of the body the bundles finally unite. The lon- 
gitudinal bands occasionally give off side branches, which pass 
into the adjacent bands (fig. 22). 

The retractor muscles of the proboscis arise by a common 
origin from a kind of dissepiment, stretching across the body 
at the level of the origin of the mantle fold, and just behind the 
skeletal tissue of the collar (fig. 9). Almost immediately after 
their origin they split into two bands, which pass backwards, 
one on each side of the ceesophagus, for about half its length. 
Each lateral band then again divides into two branches, a 
shorter dorsal and a longer ventral branch, which run to the 
body wall, where they fuse with the adjacent bands of longitu- 
dinal fibres. The ventral bands, being longer than the dorsal, 
are attached to the body wall behind these, lying one on each 
side of the nerve-cord, and being connected by the generative 
ridge. The posterior ends of the retractor muscles are fan- 


ON PHYMOSOMA VARIANS. 9 


shaped and split up into bundles of fibres, which pass into the 
adjacent longitudinal bundles. 

A special muscle accompanies the nervous system on each 
side (fig. 29), and is described in connection with the nerve- 
cord. Its purpose is probably to regulate the movements of 
this important organ during the eversion or retraction of the 
introvert. 

The spindle-muscle and the intrinsic muscles of the ali- 
mentary canal are described with the digestive organs, and 
the intrinsic muscles of nephridia with the account of these 
organs. 

Except along the generative ridge, the body wall is lined by 
a layer of flat epithelial cells, which is never ciliated, in this 
respect differing from that of Sipunculus. 


THe SKELETAL TISSUE. 


A curious form of tissue is found in the collar and the ten- 
tacular crown of Phymosoma. As it seems to subserve the 
purpose of supporting and stiffening the collar and tentacles, 
and as a support for the insertion of the retractor muscles, I 
propose to call it the skeletal tissue. 

The cells composing this tissue are large rounded cells, which 
lie close to one another, but are not so crowded as to become 
hexagonal. ‘The cell nucleus is large, and both it and the proto- 
plasm of the cell staindeeply. Running across the cell, usually 
in a radial direction, are a small number of wavy lines. 

This tissue forms a ring lying in the substance of the collar, 
which it seems to stiffen. The horseshoe-shaped blood-space 
lies internal to this tissue, which is thicker at some parts, 
and thus serves to break up the blood-space as indicated in 
figs. 4 and 6. It also sends extensions into the tentacles, a 
group of these skeletal cells heing formed on both sides of the 
tentacular nerve in each section of the tentacle (fig. 17). 

From the position of this skeletal ring in the collar it will be 
readily understood that it is just in front of the invaginable 
introvert, and consequently it affords a valuable hold for the 


10 ARTHUR E. SHIPLEY. 


insertion of the retractor muscles which are attached to this 
part of the body. 


THe ALIMENTARY CANAL. 


The digestive tube may be divided into three parts: (1) the 
cesophagus, which extends from the mouth to the beginning of 
the coiled intestine; (2) the intestine which forms a close, 
fairly regular coil with from ten to sixteen turns ; in its coiled 
state it is almost 10 mm. long; (3) the rectum, which is a 
straight tube passing from the anterior end of the coil to the 
anus. 

In spirit specimens the whole of the alimentary canal is 
white in colour, and is usually full of fine sand. A spindle- 
muscle serves to support and keep in position the coiled 
intestine and rectum. This muscle arises from the extreme 
posterior end of the body wall, and passes forward along the 
axis of the coiled intestine and then parallel with the rectum, 
to be inserted into the body wall a little in front of the anus 
(fig. 8). It gives off during its course numerous fibres, which 
are inserted into the walls of the intestine and rectum. In 
addition to the spindle-muscle the intestine is held in position 
by a thin muscle, which arises from the ventral surface of the 
body and is inserted into the anterior end of the coil. 

The position of the mouth has been described above. It is 
a crescentiform slit, lying between the lip and the convex side 
of the tentacular crown (fig. 6). It is lined with a continua- 
tion of the columnar ciliated cells which cover the inside of the 
lip and the ciliated grooves of the tentacles. The walls of the 
cesophagus are produced inwards into a series of from six to 
eight ridges, which reduce the lumen of the esophagus to a 
star-shaped tube. The grooves between these ridges are 
continuous with the grooves on the outside of the tentacles 
(fig. 9). The whole is beset with short thick-set cilia. 
Surrounding the cesophagus are a few muscle-fibres arranged 
circularly. For about half its length this first part of the 
alimentary canal lies between the retractor muscles, which in 
this region of the body have been reduced to two bundles of 


ON PHYMOSOMA VARIANS. 11 


fibres by the fusion of the anterior and posterior muscles of 
the left and right side respectively. These lateral bundles 
have fused with the cesophagus, a small amount of gelatinous 
connective-tissue containing branched cells being found be- 
tween them and the circular muscles of the esophagus. The 
dorsal blood-vessel lies between the lateral muscles in a 
groove, closely applied to the dorsal side of the cesophagus, 
and extending back almost to the beginning of the intestinal 
coil. 

Owing to the presence of very fine sand in the intestine and 
the delicacy of the tube which made it impossible to satis- 
factorily wash the sand out, I had considerable difficulty in 
studying the histology of this part. The intestine is lined 
throughout by a layer of columnar epithelial cells, one cell 
thick. The nuclei of these cells are situated near the base. 
Outside this layer is a thin membrane in which muscle-fibres 
are sparsely scattered. I do not think the intestine is uni- 
formly ciliated, but patches of cilia occur here and there. 
The arrangement of these ciliated patches I failed to make out. 
There is no groove with long cilia running the whole length of 
the animal, such as has been described by Keferstein in 
Sipunculus. 

The lumen of the rectum is almost occluded by the presence 
of numerous folds projecting into it. These folds are covered 
with a number of columnar cells some of which are ciliated, 
but the majority are crowded with large vacuoles containing 
minute granules; these are devoid of cilia. The rectum has no 
czeca opening into it, such as are found in Sipunculus. 

The external cuticle is folded into the anus for a little way, 
and the circular muscle-fibres of the body wall are thickened 
around the anus in this region, forming avery efficient 
sphincter. A number of radially arranged fibres also pass out 
all round the anus; these fibres are derived from the longi- 
tudinal muscles. Their action is obviously antagonistic to that 
of the sphincter. 


12 ARTHUR E. SHIPLEY. 


THe VASCULAR SYSTEM. 


There are two varieties of blood-corpuscle found in Phy mo- 
soma. The larger kind exist in great numbers in the body- 
cavity, together with the ripe generative products (fig. 30). 
They are oval, about 02 mm. long and two thirds as broad ; 
their protoplasm is very clear and transparent, but the nucleus 
stains well and they have a very definite outline. The ccelomic 
fluid, in which these corpuscles float, bathes all the internal 
organs of the animal, and when the contraction of the poste- 
rior circular muscles forces the fluid forward it would serve to 
evert the introvert, which is withdrawn again by the retractor 
muscles. 

The second variety of blood-corpuscle is much smaller than 
the first, being about half as long and as broad; the proto- 
plasm is not so transparent and stains more readily. These 
corpuscles are contained in a close space which is usually called 
the vascular system. This space may best be described as 
consisting of three parts, all communicating with one another. 
The first of these is a horse-shoe shaped space (figs. 2 and 7) 
at the base of the tentacles. From this space there runs up 
into each tentacle a series of three vessels which anastomose 
freely with each other and communicate at the tip. Asa rule 
sections of the tentacles show one vessel near the inner pig- 
mented surface of the tentacle, just external to the tentacular 
nerve and two near the outer surface, one each side of the 
ciliated groove (fig. 17). The free ends of this horseshoe- 
shaped space at the base of the tentacles, near the dorsal 
middle line, are continuous with the ends of another horseshoe- 
shaped space which lies in the collar. This forms the second 
of the above-mentioned spaces. As the diagram (fig. 2) shows, 
it is very irregular in form, breaking up and anastomosing 
into a number of spaces. This communicates only with the 
inner smaller horseshoe, between the two is the crescentiform 
space in which the mouth opens. The third space—usually 
termed the dorsal blood-vessel—is a very extensile sac running 
along the dorsal middle line of the esophagus between the 


ON PHYMOSOMA VARIANS. 13 


right and left retractor muscles (figs. 2, 3, and 9). It usually 
extends about 3 cm. behind the head, and it ends blindly behind. 
Anteriorly it opens in the middle ventral line into the smaller 
or tentacular horseshoe, and at the point of junction is a large 
sinus which surrounds about three quarters of the brain—in 
fact, all those parts which are not in contact with the epidermis 
(figs. 2,4, and 8). The nervous matter is thus in close contact 
with the blood, being separated only by a thin layer of con- 
nective tissue, and the endothelium of the blood-space (fig. 27). 

The walls of this third part or dorsal vessel are muscular, 
and in some specimens are much contracted and crumpled. 
This vessel appears to serve as a reservoir for the corpusculated 
fluid, and when it contracts and the fluid is forced forward, it 
would serve to evert the lip and extend the tentacles. The 
whole of this space is lined by flat epithelium. I have never 
seen cilia on the walls, and it is entirely closed. 


Tue NEFHRIDIA. 


The nephridia or the renal organs are in the form of a 
single pair of ‘‘ brown tubes,” as in other Sipunculide. They 
lie on either side of the middle ventral line at some little dis- 
tance from the nerve-cord. Their anterior extremities, near 
which are the external openings, being a little anterior to the 
level of the anus (fig. 3). 

Each nephridium is about 1 cm. long, the length in preserved 
specimens varying according to the space of contraction of its 
muscular coat; by means of this muscular layer the whole 
organ has the power of shortening and dilating, and also of 
throwing itself into a number of curious curves. 

At the anterior extremity is a dilated bladder, the diameter 
of which is from four to five times that of the posterior cellular 
portion of the organ. The internal opening is situated at the 
anterior extremity of the bladder and is provided dorsally with 
a prominent ciliated lip! (fig. 18). The external orifice is just 

1 The existence of this opening is doubted by Selenka, ‘ Die Sipunculiden,’ 


but it is sufficiently obvious in all the specimens. It was demonstrated in 
another species of Phymosoma by Dr. Spengel. 


14, ARTHUR E. SHIPLEY. 


behind the internal, and opens also into the bladder. The 
opening to the exterior is surrounded by a thickened ring of 
connective tissue with muscle-fibres intermingling, the latter 
forming a sphincter. The walls of the passage are folded and 
lined with cubical epithelial cells. The communication between 
the internal opening and the bladder is effected by means of a 
short passage, the epithelium of which is ciliated. The walls 
of the bladder itself are formed of a single layer of cubical 
cells, a middle coat of irregularly arranged muscle-fibres, and 
an external investment of peritoneum. The relations of the 
bladder and its openings will be evident from the diagram, fig. 
18. The walls of the bladder are very elastic, they contain 
many muscular fibres, and are lined with cubical epithelial 
cells. 

The tubular portion of the kidney is a backward prolonga- 
tion of the bladder, and is attached from the anterior half of 
its course to the body wall by a mesentery, its posterior half 
being free. The tube possesses anteriorly a simple lumen, 
which is broken up posteriorly by a number of septa, producing 
an appearance which reminds one of that presented by the 
interior of a frog’s lung, the transition between the two regions 
is very gradual. 

The epithelium lining the tubular portion of the kidney 
is generally one cell thick; it is produced internally into a 
series of long papille, which are separated from one another 
by a series of depressions (see figs. 19 and 20). 

The cells forming the papille are extremely long, and are 
loaded with fine, yellowish granules. In specimens killed 
during the functional activity of the organ these papilla-cells 
are furnished at their inner extremities with a series of large 
thin-walled vesicles, which appear to be thrown off from time 
to time into the lumen of the kidney (fig. 20). 

The granules, with which the kidney-cells are loaded, appear 
to decrease in number as the vesicles are approached ; and it 
seems possible that the excretory products of the nephridial 
cells are stored up in the vesicles before being thrown, together 
with the vesicles themselves, into the nephridial tube. The 


ON PHYMOSOMA VARIANS. 15 


whole process is very similar to what takes place in a mammary 
gland during the excretion of milk. Théel mentions that the 
excretory organs of Phascolion emitted yellow vesicles which 
resembled drops of oil when the living animal was disturbed.! 

Between the papille lie a series of hemispherical depressions 
lined by a flattened epithelium, the cells of which are usually 
loaded at their base with the yellow granules above men- 
tioned. ‘These cells seem to develop into the high columnar 
cells described above. 

The muscle-fibres form an irregular network outside the 
nephridial cells, lying chiefly at the bases of the papillae. The 
hemispherical depressions seem to pass through the meshes of 
the muscular coat, and to lie in direct contact with the perito- 
neal investment of the organ (figs. 19 and 20), forming a series 
of projections visible on the external surface. 

The peritoneal epithelium which surrounds the kidney is dis- 
tinguishable from the nephridial cells by the greater ease with 
which it absorbs staining fluids, and by the absence of secretion 
granules. In the region of the hemispherical depressions 
the peritoneal cells frequently form thick masses several cells 
deep. 

It is difficult to avoid the conclusion that the excretion pro- 
ducts are passed through the peritoneal cells to the cells of the 
hemispherical cups, and thence to the cells of the papille, 
the internal opening of the nephridium having relation chiefly 
to its function as a generative duct. 

The relative amount of the secreting epithelium to the cubical 
epithelium lining the bladder varies greatly ; in one specimen 
even the area between the external opening and inner end of 
the internal opening was lined with the former cells, thus 
reducing the bladder to a very small structure. 

The lumen of the nephridium contains nothing but the 
vesicles above described, together with ripe ova or spermatozoa. 
It is remarkable that the cwlomic corpuscles appear never to 
pass through the internal opening of the organ. 


1 Théel, “Recherches sur le Phascolion strombi,” ‘ Kongl. Svenska Ve- 
tenskaps-Akademiens Handlingar,’ Bandet 14, No. 2. 


16 ARTHUR E. SHIPLEY. 


Tue Nervous System AND SENSE-ORGANS. 


The brain is a bilobed organ, continuous by its anterior 
face with the ectoderm of the invaginated preoral lobe, and 
surrounded elsewhere by a process of the lophophoral blood- 
vessel, from which it is separated, not only by the endothelium 
of the vessel, but also by a connective-tissue capsule (see figs. 
2,4, 8, and 27). The groove between the two lobes is deepest 
and widest on the anterior surface, where the substance of the 
brain is continuous with that of the przoral ectoderm. 

In the brain, as in the ventral nerve-cord, the ganglion-cells 
are aggregated in the side nearest the skin; they are on the 
dorsal side of the animal in the brain, on the ventral in the 
nervous system. 

As the figs. 24, 25, and 26 show, there is a cap of ganglion- 
cells covering the anterior, dorsal, and posterior surfaces of the 
brain. The ventral surface is not invaded by the ganglion- 
cells; but here the fibrous tissue, which makes up the rest of 
the brain, comes in contact with the thin connective capsule. 
It is this region of the brain which projects into the blood- 
sinus. 

The majority of the ganglion-cells are small, with deeply 
stained nuclei, occupying about one half of the cell; they are 
either unipolar or bipolar. At the postero-dorsal angle of the 
brain, however, a certain number of giant ganglion-cells are 
found (fig. 27). These cells have a diameter of 02 mm., at 
least four times that of the smaller cells; their nuclei are rela- 
tively smaller, and they are unipolar. I was unable to trace 
what becomes of the fibres given off from these giant-cells. 
No such giant-cells occur in any other part of the nervous 
system. 

A pair of sense-organs, usually described as eyes, lie em- 
bedded in the substance of the brain. 

Each of these sense-organs has the form of a long tube bent 
upon itself, so that one limb is nearly at right angles to the 
other. The outer limb, the lumen of which is narrow, opens on to 
the surface of the przoral lobe (figs. 1 and 25), the opening lies 


ON PHYMOSOMA VARIANS. 1; 


at the dorsal lateral angle of the brain, just dorsal to where 
the circumcesophageal nerve-commissure leaves the brain ; the 
lumen of the inner limb dilates into a vesicular swelling in the 
substance of the brain (fig. 23); the whole tube has, therefore, 
nearly the shape of a retort, and lies entirely in the lateral 
part of the brain. The wall of the tube is everywhere formed 
by a layer of clear, nucleated cells. In the outer limb these 
cells form a fairly regular columnar epithelium one cell thick, 
which becomes less regular as the inner limb is approached. 
The cells bounding the inner limb are arranged irregularly, and 
they appear to send out processes from their peripheral extremi- 
ties, which may be supposed to communicate with the pro- 
cesses of adjacent nerve-cells. The cells of the inner limb also 
secrete a deep black pigment, which lies in that portion of each 
cell which is turned towards the lumen of the tube. A clear 
coagulum sometimes lies in the cavity of this sense-pit. These 
organs are visible as two black spots at the level of the brain 
in the dissected animal (fig. 3). 

No trace exists in this genus of the curious finger-like pro- 
cesses which project from the brain of Sipunculus into the 
body-cavity. 

Three pairs of nerves are given off from the brain: (1) 
dorsally, a small pair supplying the skin of the preoral lobe— 
these lie nearest to the middle line (fig. 26) ; (2) ventrally, a 
nerve on each side, going to the corresponding area of the 
lophophor, and supplying a branch to each tentacle (fig. 24) ; 
(3) and posteriorly on each side arises a nerve which passes 
round the cesophagus, and joins its fellow of the opposite 
side to form the ventral cord (fig. 24). The lophophoral 
nerve arises between the point of origin of the nerve of the 
preoral lobe and the exit of the circumcesophageal commis- 
sures. 

The ventral cord itself shows no trace either of a division 
into two halves, or of a segregation of its nerve-cells into 
ganglia, It runs along the ventral surface of the body as a 
perfectly uniform filament, terminating posteriorly without 
any ganglionic swelling such as that found in Sipunculus. 

VOL. XXXI, PART I.——NEW SER. B 


18 ARTHUR E. SHIPLEY. 


The fibres are on the dorsal, the cells on the ventral side of 
the cord. 

Along each side of the nerve-cord runs a longitudinal band 
of muscle-fibres, the cord and its pair of muscles being together 
enclosed in a special peritoneal sheath. The space between 
the sheath and the cord is filled with a peculiar connective 
tissue (fig. 29), which has been regarded by some observers as 
clotted blood, the cord being said to lie in a blood-vessel. My 
preparations afford no evidence in support of this view; and I 
am strongly of opinion that the substance lying between the 
nerve-cord and its peritoneal investment is, as above stated, 
connective tissue. 

By contraction of the muscles within the peritoneal sheath 
the nerve-cord may become crumpled, so that while the sheath 
is perfectly straight the cord within it presents the appearance 
shown in fig. 28. 

The nerve-sheath is attached to the ventral body wall by a 
series of mesenteric cords, each of which contains, not only a 
prolongation of peritoneal epithelium, but also a central axis 
of connective tissue (figs. 28 and 29). 

The peripheral nerves form, as in Sipunculus, a series 
of rings encircling the body, and lying between the circular and 
the longitudinal muscles. In the region of the introvert a 
nerve-ring lies beneath each ring of hooks, at the base of the 
circular muscle which supports them (figs. 1 and 2). 

Each nerve-ring is connected with the ventral cord by a 
single short nerve, which runs from one to the other in the 
middle ventral line. 

The lophophoral nerve runs along the base of the tentacles, 
one on each side of the lophophore. Lach gives off a series of 
small nerves, one of which passes up the axis of each tentacle, 
lying immediately beneath the ciliated groove (figs. 2, 5, 
and 17). 

In addition to the sense-pits on the brain there are a number 
of ectodermal structures on the introvert, which are probably 
sensory in function, and may well be described here. These 
bodies are arranged in circles parallel to the rows of hooks 


ON PHYMOSOMA VARIANS. 19 


running round the introvert (fig. 21). One of these organs 
is shown in fig. 12; the ectoderm-cells have multiplied and 
increased in size, forming a small heap; some of these cells 
have then formed stiff processes, which project beyond the 
level of the skin. These processes are gathered up intoa small 
brush by a chitinous ring which surrounds the base. 

The hooks (fig. 21) are very closely packed in a series of 
ridges formed by the circular muscle-fibres of the introvert. 
The point is directed backward, while the row of sense-organs 
lies immediately behind them, embedded in the muscular 
cushion. 


THe GENERATIVE ORGANS. 


Phymosoma varians is diccious; in no case are ova and 
spermatozoa found in the body of the same individual. 

The ovaries are formed by a fold of the peritoneal epi- 
thelium, elsewhere flat, which occurs at the base of the 
insertion of the long ventral pair of retractor muscles. This 
genital ridge extends beyond the inner edge of the muscle 
attachment across the ventral middle line lying between the 
nerve-cord and the skin; it does not extend beyond the outer 
or dorsal end of the muscle. The ridge is not quite con- 
tinuous, but it is interrupted from time to time; its free 
border is also irregular, and this gives it a puckered or frilled 
appearance (fig. 22). 

In transverse section—parallel to the long axis of the 
Phymosoma—the ovary is seen to be much thicker at its free 
border than at its base; the latter indeed is formed of but two 
layers of cells, thus giving the appearance of a simple fold of 
endothelium. These layers, however, thicken towards the free 
edge. Nearly all the cells have become ova, and are held 
together by a very scanty matrix. The organ is solid, and 
the ova dehisce from it into the body-cavity. 

In the ovary the ova increase in size towards the thickened 
free edge, where the oldest are. Those found free in the body- 
cavity also differ somewhat in size, and undoubtedly grow 
whilst suspended in the perivisceral fluid; but there is a very 


20 ARTHUR E. SHIPLEY. 


marked difference in size between the largest ovarian ovum 
and the smallest floating one—a difference I am quite unable 
to account for. 

The floating ova are oval in shape, the largest about 1 mm. 
long, with a thick zona radiata, in which the radial markings 
can only be detected with very high powers (fig. 30). 
This membrane stains deeply except its outermost layer, which 
does not absorb any staining fluid. The protoplasm is very 
granular, and stains well. The nucleus is very large, and 
sometimes reaches almost from one side of the cell to the 
other ; it does not stain at all. No micropyle was to be seen. 

The testis occupies in the male a position similar to that of 
the ovary in the female. The mother-cells of the spermatozoa 
separate from the testis before or whilst dividing. Whilst 
floating in the perivisceral fluid the nuclei of these cells com- 
menced to divide, and the whole floats about as a multi- 
nucleated mass of protoplasm. The stages which most com- 
monly occurred were those with eight or sixteen nuclei 
(fig. 8). The males were much rarer than the females, and 
none of them contained ripe spermatozoa. 


SUMMARY. 


The following is a brief summary of the more important 
points described in detail in the body of the paper. 

(1) The head of Phymosoma is surrounded by a stiffened 
vascular horseshoe-shaped lip, the dorsal ends of which are 
continuous with the ends of a hippocrepian lophophor. The 
lophophor bears a crown of about eighteen tentacles—the 
number is always even. In the hollow of the lophophor lies 
the brain, which is continuous with the ectoderm of the 
preoral lobe. The inner surface of the tentacles and the 
ectoderm above the brain is crowded with dark brown pigment- 
granules, and the ectoderm of the preoral lobe is curiously 
wrinkled. Between the hippocrepian lophophor and the 
vascular lip is the crescentiform opening of the mouth. 

(2) At some little distance behind the lip is a thin but very 


ON PHYMOSOMA VARIANS. 21 


extensile collar, which may be so extended as to entirely cover 
the head. 

(3) The ectoderm consists of a single layer of cells. This 
secretes outside a cuticle of varying thickness. The ecto- 
dermal cells are vaulted, so that spaces are left in which a 
nutrient fluid might circulate between the circular muscles 
and the ectoderm. The ectoderm of the lower lip and of the 
outside of the tentacles is ciliated. 

(4) The skin-glands are of two kinds; each is formed by 
the modification of ectoderm-cells, which results in the pushing 
in of certain of the cells to form a double cup. The inner 
layer of cells thus produced develops a number of granules, 
which are extruded through a median aperture. In one kind 
of skin-gland, those of the introvert, this aperture is sur- 
rounded by a chitinous ring, which is absent on those of the 
trunk. 

(5) Rows of hooks set very closely together are found 
in the introvert; these are each secreted by a small multi- 
cellular papilla. 

(6) A skeleton tissue is present in the lip and tentacles. 
This seems to stiffen these structures, and to form a firm hold 
for the attachment of the retractor muscles of the introvert. 

(7) The nephridia or brown tubes consist of two parts, the 
bladder and the secreting part. The former opens both to the 
exterior and to the body-cavity, the latter opening being 
shaped like a flattened funnel and ciliated. The secreting 
part opens only into the bladder. Its walls are lined with a 
columnar epithelium, the cells composing which are crowded 
with granules. From time to time a vesicle or bubble crowded 
with these granules is formed at the free end of the cell, and 
ultimately breaks off into the lumen of the nephridium, and so 
passes out of the body. The only other structures found in 
the cavity of these organs besides these vesicles, were the 
ripening generative cells. 

(8) The vascular system consists of a horseshoe-shaped 
plexus in the lower lip, a similar plexus in the lophophor 
which gives off branches into each tentacle, and a reservoir 


22 ARTHUR E. SHIPLEY. 


lying dorsal to the esophagus. This communicates with the 
lophophoral sinus in the dorsal middle line. Just at this point 
is a blood-sinus which surrounds all those parts of the brain 
which are not continuous with the ectoderm. This system of 
blood-vessels is closed. It contains numerous small oval 
corpuscles. In addition to these the ccelomic fluid contains a 
number of much larger corpuscles, as well as ova and sperm 
-morule. The celom is lined by a flat epithelium which is not 
ciliated. 

(9) The brain is a bilobed mass, partly connected with the 
ectoderm of the przoral lobe and partly surrounded by a 
blood-sinus. The relative position of the ganglion-cells and 
fibrous tissue is described above. There are a number of 
giant ganglion-cells arranged in the lateral and posterior 
parts of the brain. 

(10) The brain gives off three pairs of nerves: (1) the first 
pair pass to supply the pigmented tissue of the preeoral lobe; 
(2) the second pair run along the base of the lophophor, and 
send a branch into each tentacle; (3) the third pair pass 
round the csophagus, and unite to form the ventral nerve- 
cord. This is supported by a strand of muscle in each side, 
and by numerous connective-tissue strands which pass to the 
body wall. It has no trace of a double structure, and no seg- 
mentaily arranged nerve-ganglia. It gives off from time to 
time a median nerve, which soon splits, and each half runs 
round the body, these fuse together again in the dorsal 
middle line, thus forming a nerve-ring. 

(11) The sense-organs consist of two pigmented pits in the 
brain, and of certain structures in the introvert. The former 
pits open on to the preoral lobe, and then pass into the brain at 
each side. Hach pit is bent on itself, and expands slightly at 
its inner end. The cells lining the pit are crowded with black 
pigment. The sense-organs on the introvert lie in rows close 
behind the rows of hooks. Each consists of a number of ecto- 
dermal cells produced outwards into a stiff process. These 
processes are gathered up into a little brush by a chitinous 
ring which surrounds their base. 


ON PHYMOSOMA VARIANS. 25 


(12) The animals are diccious. The generative organs 
are in the form of ridges at the base of the ventral retractors. 
The flat coelomic epithelium is here modified to give rise to 
ova in the females and the sperm morule in the males. 


CONCLUSIONS. 


I do not propose to consider at any length the theoretical 
conclusions which might be drawn from the facts above indi- 
cated until I have worked out in detail other forms of the 
Gephyrea, which I hope to do in the immediate future. I 
should, however, like to say something in favour of maintain- 
ing the genus Phoronis in its old position—that is, as a form 
closely allied to the more normal Gephyrea inermia. 

This relationship is most easily seen by comparing a view of 
the head of Phymosoma as seen from above with a view of 
Phoronis (figs. 31 and 82). In both genera the mouth is sur- 
rounded by a pair of vascular horseshoe-shaped ridges, one of 
which is dorsal and the other ventral : the sole point of difference 
lies in the fact that while in the one case the tentacles of the 
lophophor extend along both the ventral and the dorsal horse- 
shoe, they are in the other case confined to the dorsal limb. 

Again, the preoral lobe of Phoronis bears two large sen- 
sory pits, one on each side of the middle line; these are 
obviously comparable to the similar pits which open into the 
area in the concavity of the Gephyrean lophophor which I 
have spoken of as the preoral lobe. Further, the nervous 
system of Phymosoma, like that of Phoronis, is permanently 
connected with the epidermis. 

I do not enlarge upon the resemblances in the position of 
the anus, and the lengthening of the ventral surface at the 
expense of the dorsal, or on the presence of two nephridia, as 
these points have been already emphasised by Lankester. 
But I would direct attention to two structures hitherto, I 
believe, undescribed in the Gephyrea, which in my opinion 
have homologues in Phoronis. 

The first of these is the skeletal tissue ; this, as the descrip- 
tion above shows, agrees in position and function with the 


24, ARTHUR E. SHIPLEY. 


mesoblastic skeletal tissue which supports the tentacles of 
Phoronis as described by Caldwell. The second structure I 
wish to refer to is the thin membranous fold which I have 
above termed the collar. This seems to me to correspond 
very closely with the calyx or web which surrounds the base 
of the head in Phoronis. 

The absence in the unarmed Gephyrea of mesenteric parti- 
tions in the post-oral body-cavity, similar to those which exist 
in Phoronis, may be accounted for by the twisting of the 
intestinal loop in the more normal genera. The radial 
muscles which extend from the visceral loop to the body wall 
are, in all probability, the remains of an ancestrally continuous 
mesentery. 

It will be remembered that in Phoronis the body-cavity is 
divided into an anterior and a posterior division by a trans- 
verse septum passing from the body wall to the cesophagus, at 
the level of the nerve-ring. The former division includes the 
cavity of the preoral lobe and tentacles, the latter the rest of 
the body-cavity. I am disposed to think that a similar dispo- 
sition of parts obtains in Phymosoma. The organ which is 
usually regarded as forming the blood-vessels in the Gephyrea 
occupies precisely the same position as the anterior body- 
cavity in Phoronis; it has, however, acquired a reservoir— 
the dorsal vessel—into which the fluid may pass when the 
head is retracted. As this involution is impossible in Pho- 
ronis no such reservoir has been developed. If this homo- 
logy holds, there is nothing in the Gephyrea homologous with 
the true blood system of Phoronis. In connection with this 
it is perhaps worth noticing that the so-called vascular system 
in the Gephyrea gives off no vessels or capillaries, but simply 
consists of a number of intercommunicating spaces. 


April, 1889. The Morphological Laboratory, 
Cambridge. 


ON PHYMOSOMA VARIANS. 25 


DESCRIPTION OF PLATES I, II, III, and IV,* 


Illustrating Mr. Arthur E. Shipley’s paper “On Phymo- 
soma varians.” 


PLATE I. 

Fic. 1.—A semi-diagrammatic view of the anterior end of Phymosoma 
varians. ‘The introvert is everted and the tentacular crown expanded. The 
collar is not extended and lies at the base of the head. Only two rows of 
hooks are shown. 

Fie. 2.—A semi-diagrammatic view of the closed vascular system and 
nervous system, showing their relation to the alimentary canal. The vascular 
system shows the three parts, the lophophoral, the lower lip, and the dorsal 
blood-vessel. The latter communicates with the lophophoral in the middle 
line, and just at this point the sinus round the brain is given off. The brain 
is relatively too small. The three main nerves are shown, and the circular 
nerves which run in the skin. The cesophagus is cut off abruptly in front in 
order to display the vascular ring. 

Fic. 3.—View of a Ph. varians which has been opened along the 
median dorsal line. The introvert is retracted, the true anterior end of the 
body being where the eye-spots lie. Here and there patches of skin are 
seen which bear papille. 

PLATE II. 


Fic. 4.—A median longitudinal section through the head. The introvert 
is retracted, and the collar expanded until it encloses the whole head. The 
section is not quite in the middle line, or the lip on the dorsal surface 
would not be shown, cf. Fig. 6. The brain is cut through that part which is 
continuous with the ectoderm. 

Fic. 5.—A transverse section through the tentacles: the introvert is re- 
tracted. The tentacles show the ciliated groove on the outer surface, the 
pigmented epithelium in the inner, and the vascular spaces and tentacular 
nerves. 

Fic. 6.—A transverse section through the base of the lophophor and lower 
lip, just where the two fuse dorsally, the introvert is retracted. The skeletal 
tissue is shown in the lip, which is ciliated all round. 

Fie. 7.—An oblique transverse section through the base of the lophophor, 
showing the blood-space ; and in the centre some of the wrinkled pigmented 
tissue of the preoral lobe. The introvert is everted. 


1 | am indebted to Mr. Weldon for the following figures :—Nos. 1, 7, 10, 
12, 14, 15, 16, 20, 21, 23, and 27. 


26 : ARTHUR E. SHIPLEY. 


Fic. 8.—A transverse section through the head in the region of the brain, 
The introvert is everted. This specimen had its body wall pushed upwards 
inside the lower lip in the ventral side into a kind of hernia, this accounts for 
the swelling containing blood-corpuscles and sperm-morule. The brain is 
shown in its sinus, also the depressions in the tissue of preoral lobe leading 
to the sensory pits. 

Fic. 9.—A transverse section through the cesopkagus. The dorsal and 
ventral retractor of each side have fused into a common lateral muscle, which 
almost fills up the body-cavity. The lumen of the cesophagus is occluded by 
ciliated ridges. 

Fic. 10.—A section through the ectoderm and cuticle. Below the ectoderm 
some fibres of the circular muscle may be seen. The ectoderm is vaulted 
leaving spaces which sometimes contain a coagulable fluid. The cuticle is 
traversed by numerous perpendicular lines, and the outer part only stains. 

Fic. 11.—A surface view of the skin, showing the longitudinal and circular 
muscle-fibres, the skin papille, and the ridges formed by thickenings of the 
cuticle. 

Fic. 12.—A section of one of the sense organs on the introvert, at the 
base of the ring of hooks. 

Fic. 13.—A transverse section through the posterior end of the animal. 
The longitudinal muscles have fused together and reduced the lumen of the 
body-cavity to a star-shaped mass. The skin papille are very numerous in 
this region, and the cuticle unusually thick. 


PLATE III. 


Fic. 14.—Section taken through one of the skin papille of the trunk. It 
shows the opening to the exterior, and the small cavity in the cup composed of 
enormous cells crowded with spherules. 

Fic. 15.—Surface view of the papille and hooks in the introvert. The 
chitinous plates round the orifice of these papillae are shown. 

Fic. 16.—An oblique section through a trunk papilla. This section shows 
the space between the two layers of the cup in communication with the sub- 
ectodermal spaces of the skin. 

Fic. 17.—Transverse section of a tentacle. At the base of the ciliated 
groove the tentacular nerve lies. Three blood-spaces are seen, and between 
them certain skeletal cells. The inner epithelium is crowded with pigment 
grains. 

Fic. 18.—A diagram showing the anatomy of the nephridium. The pos- 
terior blind diverticulum is the secreting part, the anterior thin-walled part is 
the bladder. The arrangement of the internal and external openings may 
also be seen. 

Fic. 19.—An oblique section through the secreting part of the nephridium, 
under a low power. This shows the peritoneal epithelium, then a dark 


ON PHYMOSOMA VARIANS. 27 


layer of muscle-fibres and internally the secreting epithelium. The breaking 
up of the lumen into numerous side chambers is also shown in this figure. 

Fie. 20.—A portion of the same under a high power. The secreting 

epithelium is seen crowded with granules ; at their free edges these cells form 
~ vesicles, which break off and fall into the lumen. 

Fic. 21.—A section through parts of several of the hooks on the introvert. 
The multicellular papillae which secrete the hooks are shown. One of these 
sense organs at the base of the hooks is also shown cut tangentially. 

Fig. 22.—A view of the base of the two ventral retractor muscles, showing 
the generative organ. The ventral nerve-cord lies between the muscles and 
dorsal to the generative ridge. The circular and longitudinal muscles are 
also shown, and the outline of the papille. 


PLATE IV. 


Fie. 23.—A section through the antero-dorsal corner of the brain, to show 
the blind end of the sense-pit. The cells lining the inner end of the pit are 
crowded with pigment. A few cells of the ectoderm of the preoral lobe are 
seen, and part of the blood sinus in which the brain lies. 

Fic. 24.—An oblique section through the lateral part of the brain, showing 
the origin of the circumcesophageal commissure and of the lophophoral nerve. 
This figure and the three succeeding ones show the arrangement of the 
ganglion-cells, the giant cells, and nerve-fibres. 

Fic. 25.—A section through the brain, transverse to its long axis, and 
nearer to the middle line than the preceding figure. It shows the fusion of 
the brain with the ectoderm of the preoral lobe, and the commencement of 
the preoral lobe nerve. 

Fic. 26.—A section in a place parallel to the preceding, but still nearer to 
the median line, it shows the origin of the preoral lobe nerve. 

Fic. 27.—A horizontal section through the posterior part of the brain at 
right angles to the preceding. This shows the histology of the giant-cells 
and their relative size. 

Fig. 28.—A longitudinal median section of the ventral nerve-cord, showing 
the arrangement of the ganglion-cells and fibres, and the mesenteries which 
attach the cord to the ventral body wall. 

Fic. 29.—A transverse section of the nerve-cord, showing a mesentery 
from the ventral body wall, the arrangement of ganglion-cells and nerve- 
fibres, the connective-tissue sheath, and the lateral muscles which run along 
each side of the nerve-cord. 

Fic, 30.—An ovum and some of the ccelomic corpuscles. The ovum shows 
the granular protoplasm, the large nucleus, and the zona radiata. 

Fic. 31.—A diagrammatic view of the head of Phoronis, seen from in 

ront. 

Fic, 32.—A similar view of Phymosoma. 


ae sane 


ue aww Brie 1h 


> i," « 


el 6 | 


t 
hha) alone 


. 
id 
¢ 
i ives 
yeu Rviei iPad 
, 
cee bar eee! 
he eel 
be , 
igs 
- » ty ae Wy 
oat i 
ri 
v Las : 7 
tims 
- 7 
Mi) fi eee 
hiiw Were & mr? ais ist 
Mi) thir rej 71% 
ih MB \ibad > ome 
‘ 
fi 
see Bory 4h 7 
sé 
&. Pen Ah | 
ae FPO f 
Aen he; j - 
Mia! & eto 
{ i ! ise y 
oT Wil itive rae @ tid 


= ) 
Bil uae Gi , 


ait ste Wi 
« fps Wi 
PG AIT) no MT WA 


ee 2 


meet - ~ 


vita 


A} 4 00) 1 Wii... 
ss 
_ 
‘4 6 (ere@h@) Ap were? Paese 
¢ q Wieg T Ue t= i. 4 


Tir My , ome 


(huik, ptlg ii plee — be 
1 
f 
e749 9 - ie ly 7 36 
_. nae & 2 = Ula = 
is & ' Bay Pv, a 
i i a | © é As 
é Vote , am «4 oot 
b voit , ai| | i) ae Pe La) | wd 
i 7 ’ » at wi i hada 
» JOP Pas 
al oe ae 
( . ? be i» 
i ‘ ‘ 48 
ri id s i) oe i] 
' td 4 eae ae 7 
J ' ‘ Th 
Soft Ae, itt 
=4 » ; ‘. 4 im if 
+4 ‘| 
: ‘ ; ae: 
7 e a 
? i MO 
® io erage ets ‘i 
utt ssa el pe 
i rH & wi il gun af 


hint gi ob ald ont } ih boned 
ete vt, aghe a Allan becesinnd A, 
4) oe deena its vey iat Liha 
| Pete etre omar eS 


7 i amapert: i oe of! 
whet aupolert ae 


ae eke een 


4" (a ee writ pileleigs wt ba c 


sci eagaeie er? 


Bay 


SPINNING APPARATUS OF GEOMETRIC SPIDERS. 29 


The Spinning Apparatus of Geometric Spiders. 
By 


Cecil Warburton, B.A., 
Christ’s College, Cambridge. 


With Plate V. 


Tue familiar circular snare constructed by the “ geometric ” 
spiders has always been an object of interest to naturalists, but 
it is remarkable how little has been known until lately of the 
highly complicated organs which compose the spinning apparatus 
of these animals. 

Thanks mainly to the labours of Blackwell,! Emerton,’ 
Bertkau,® and lastly Apstein, a tolerably complete knowledge 
has now been obtained of the structure and general arrange- 
ment of these organs. 

Apstein’s excellent paper,* recently published, contributes 
much that is new and valuable, and fairly represents our 
present knowledge of the subject. Recent researches, however, 
have led me to dissent from some of that author’s conclusions 
as regards the functions of the various spinning glands, con- 
clusions based upon evidence for the most part too indirect to 
be entirely satisfactory. 

Before discussing this matter, some description of the 


1 Qn the Mammale of Spiders in Spinning,’ ‘Trans. Linnean Soc. 
London,’ 1889, vol. xviii, pt. ii. 

2 «The Structure and Habits of Spiders,’ Boston, Cassino & Co., 18838. 

3 *Cribellum und Calamistrum,” ‘Archiv fiir Naturgeschichte,’? 1882, 
p. 316. 

4 « Bau und Function der Spinnendrusen der Araneida,”’ ‘Archiv fiir Natur- 
geschichte,’ 1889, p. 29. 


30 CECIL WARBURTON. 


morphology of the organs in question will be necessary. The 
large garden spider, Epeira diademata, is taken as the most 
convenient type of the family, but the following remarks apply 
in the main to all its orb-weaving congeners. 


External Spinning Organs. 

These occupy a small round area on the under surface of the 
abdomen towards the posterior end, where, when at rest, they 
present a bluntly conical protuberance (figs. 1 and 2, sp.). If 
this area be» examined under a low power, it is seen to be 
occupied mainly by four conical spinnerets, their bases form- 
ing a quadrilateral, and their apices meeting in the centre of 
the area (fig. 8). The narrow space which intervenes between 
the bases of the anterior (or inferior) spinnerets (a) is filled by 
a small tongue-like process (¢). The wider gap separating the 
posterior (or superior) spinnerets (p) is occupied by a terminal 
projection of the abdomen (z) containing the anus. Each of 
these spinnerets is two-jointed, and furnished at its extremity 
with a multitude of hair-like tubes containing the ducts of the 
spinning glands. 

They are possessed of a wonderful mobility, and can be 
widely separated, or energetically rubbed upon each other with 
a rotary motion at the will of the animal. Their separation 
discloses a third and smaller pair of spinnerets consisting of 
one joint only, and having their apices directed backwards and 
inwards, so as to lie immediately beneath the apices of the 
posterior spinnerets (fig. 10, 2). 

These again present a large number of glandular orifices. 
They will be referred to hereafter as the intermediate spinnerets. 
Thus we have, in all, three pairs of spinnerets capable of a great 
variety of movement, and bearing at their extremities, as will 
be presently seen, about 600 spinning tubes. 


Internal Spinning Organs. 
Apstein has shown that there are, in this group of spiders, 
five distinct kinds of glands, to which he assigns the names 
Ampullaceal, Aggregate, Tubuliform, Piriform, and 


SPINNING APPARATUS OF GEOMETRIC SPIDERS. 31 


Acinate. The first three kinds are few in number and of large 
size, extending throughout the greater part of the abdomen. 
The piriform and acinate glands are minute and numerous, 
and are closely grouped together immediately above the 
spinnerets. 

Their exact arrangement is important and may be summarised 
as follows : 

There are two pairs of Ampullaceal glands (fig. 3) debouch- 
ing on the anterior and intermediate spinnerets re- 
spectively on the inner side. 

There are three pairs of Aggregate glands, their three 
outlets on each side being situate upon the inner surface of 
the posterior spinneret. 

There are three pairs of Tubuliform glands, two opening 
on the inner side of the posterior spinnerets, and one upon the 
outer surface of the intermediate spinnerets. 

The above glands are comparatively large, and their ducts 
terminate in distinct tubular prominences. 

There are about 200 Piriform glands, all connected with the 
anterior spinnerets, where their ducts terminate in hair-like 
tubes. 

Finally, there are about 400 Aciniform glands, each 
posterior and each intermediate spinneret bearing the hair-like 
terminations of about a hundred ducts. 

Or thus, tabulating for one side only: 


GLANDS. | Ant, SPINNERET. | INTERMEDIATE. PosTERIOR. 
Ampullaceal . 1 on inner side 1 on inner side 
Aggregate 3 On inner side. 
Tubuliform 1 on outer side 2 on inner side, 
Piriform . A About 100 
Aciniform About 100 About 100. 


The question naturally arises as to the different functions 


32 CECIL WARBURTON. 


performed by glands apparently so distinct. Apstein attempts 
its solution by reasoning which is mainly indirect and, in my 
opinion, misleading. It occurred to me that the problem might 
be attacked in a more direct manner, and with this view the 
experiments to be now described were performed. 

A spider of this group usually trails a line from its spin- 
nerets while walking. With a little dexterity it can be quickly 
seized, and imprisoned in such a manner that the spinnerets 
from which the line is proceeding can be microscopically 
examined. 

This may be best effected by means ofa piece of wood about 
the size and shape of a microscope slide, with a narrow band of 
cloth attached by its end to one extremity. The cloth band 
is then held in front of the crawling animal, which may, with 
a little practice, be thus trapped between the cloth and the 
wood, so that the band passes beneath the cephalothorax, 
leaving the abdomen free for examination with the lately 
emitted line still attached. 

The fourth pair of legs must be kept from interfering with 
the experiment by pins suitably adjusted. The spinnerets will 
now be in their quiescent position, and the precise origin of the 
threads therefore invisible. If, however, it be gently drawn 
forwards, i.e. towards the animal’s head, certain facts with 
regard to it become at once clear. As, however, the phe- 
nomena differ at different times, we must take the various cases 
in succession. 

In the simplest case (fig. 9) one of the anterior spinnerets will 
be pulled forward with the thread, which will be easily seen to 
consist of a single line emanating from one large tube. 

More frequently (fig. 10) the line will be double issuing 
from similarly situated tubes on the inner sides of the two an- 
terior spinnerets. This is probably the most usual case, and I 
have drawn out from a spider many yards of such a double line 
of silk, its origin being all the time plainly visible. 

It is important to note that there is no adherence between 
the two lines, which remain perfectly distinct throughout their 
whole parallel course. 


SPINNING APPARATUS OF GEOMETRIC SPIDERS. 33 


The spider will probably tire of having its silk thus drawn 
out—a process which it can only influence indirectly. Were 
its hind legs free it would seize the thread and break it. It 
sometimes contrives to do this by a rapid movement of its spin- 
nerets, but occasionally it decides to strengthen the thread 
instead. The spinnerets are accordingly actively rubbed to- 
gether, and a little flocculent mass of silk appears upon the 
line, which is thereafter seen to consist of four strands, two of 
finer calibre having made their appearance between the former 
lines (fig. 11). To see their origin the anterior spinnerets must 
be kept forward by a gentle strain on the thread, and the pos- 
terior spinnerets thrust aside with a needle. The new lines 
may then be traced to the intermediate spinnerets, and proceed 
from large spinning tubes on the inner side. Again, the four 
lines remain distinct and non-adherent. 

Should the spider still resolve on strengthening the line a 
further rubbing together of the spinnerets occurs, and presently 
a large number of strands are seen to proceed from the nume- 
rous hair-like tubes on the anterior spinnerets (fig. 12). The 
four previous lines are still distinguishable by their greater 
thickness. 

If after drawing out several inches of this compound line it 
be slightly slacked, a puff of air separates the strands, showing 
that, though contiguous, they are not adherent. 

Lastly, upon rare occasions, the whole battery of tubes seems 
to be brought into play, the posterior spinnerets contributing 
their quota to the strengthening of the line. Thus the “ trail- 
ing line,” as I have called it, will be found at any moment 
to be constituted as indicated in one of the cases above 
described. 

It appears, therefore, that such a line usually consists of 
either two or four non-adherent threads emanating from what 
Apstein has shown to be the origin of the Ampullaceal 
glands, and that it may on occasion be strenghened by con- 
tributions from the Piriform and Acinate glands opening 
upon the anterior and posterior spinnerets respectively. 

It was next attempted to apply the same direct method to 

VOL, XXXI, PART IL—NEW SER. c 


34 CECIL WARBURTON. 


the observation of the animal when employed naturally in its 
various spinning operations. Here the difficulties experienced 
were considerable, but some results were obtained by the aid of 
a simple contrivance, consisting of a pair of compasses with the 
points fixed some two inches apart, and between them a narrow 
strip of cloth stretched. 

A flat piece of wood was held behind the spider while at 
work, and between this and the strip of cloth the creature 
was suddenly trapped, the points of the compasses, which pro- 
jected the eighth of an inch beyond the cloth, being buried in 
the wood on either side. 

Flies were now placed in the various webs, and the spiders 
seized in the act of binding them up in the usual manner. The 
fly is held and rotated by means of the jaws, palps, and ante- 
rior legs, while the fourth pair of legs draw up from the spin- 
ners the bands of silk which are to enclose it. These silken 
bands were found to be constituted as shown in figs. 12 or 13. 
There seems no doubt, therefore, that the Aciniform 
and Piriform glands are mainly used in performing this 
operation. 

The structure of the geometric snare was next investigated. 

This is a familiar object, and may be said to consist of— 

(1) a sort of frame or scaffolding, to which are attached 
the distal ends of 

(2) the radial lines ; 

(3) the spiral line, extending from the periphery to near 
the centre. 

(1) The thread of the framework was generally found to be 
composed as exhibited in fig. 11. When necessary the spider 
strengthened the line by repeating the journey, and laying it 
down a second time. 

(2) The same line, or that of fig. 10, was also employed in 
constructing the radii of the snare. 

Thus the framework and radii of the geometric web are 
supplied by the Ampullaceal glands. 

(3) The spiral line requires a more detailed description. 

A low power shows it to consist of bead-like viscid globules 


SPINNING APPARATUS OF GEOMETRIO SPIDERS. 35 


strung upon a thread with remarkable regularity, as shown in 
fig. 14 d. 

It was until a few years ago supposed that these globules 
were separately deposited by the spider, whereas a uniform 
coating of viscid matter is given to the thread in the first in- 
stance, and its subsequent subdivision into globules is an 
entirely physical phenomenon. Boys! well describes the spider’s 
action as follows: 

“The spider draws these webs slowly, and at the same time 
pours upon them a liquid, and, still further to obtain the effect 
of launching a liquid cylinder into space, he pulls it out like 
the string of a bow, and lets it go with a jerk.” 

That this separation into globules is really a secondary 
phenomenon I have shown by taking upon a slide a portion of 
such a spiral immediately upon its completion. It readily 
stains with hematoxylin, and on microscopic examination 
shows the various stages indicated in fig. 14. 

We have thus separately to consider the ground-line (Grund- 
faden, Apstein) and the viscid matter with which it is en- 
veloped. 

Apstein imagines the ground-line to be furnished by the 
Aciniform glands, and to be many-stranded. 

I have not yet succeeded in tracing it with certainty to its 
origin, but have established the following facts with regard 
to it: 

In the first place, it is not many-stranded, but double 
only. 

When engaged upon this line the creature is so absorbed as 
to allow of pretty close examination with a hand-lens. I have 
at such times noticed that the posterior spinnerets are partly 
open, and that the line is, at first, distinctly double, fusing, by 
virtue of its viscid envelope, where grasped by the leg which 
draws it forth. Moreover, on staining and teasing the spiral 
line, the ground thread readily shows its double nature (fig. 
15), but no amount of teasing breaks it up into further strands, 
as would surely be the case if such existed, for their separate 


1 “ Quartz Fibres,” by C. V. Boys, F.R.S., ‘ Nature,’ July 11, 1889. 


36 CECIL WARBURTON. 


existence as threads implies a degree of dryness inconsistent 
with complete fusion. 

As far as I have been able to trace these lines they have 
appeared to emanate from the intermediate spinnerets. They 
are much more elastic, however, than the radial lines, and can 
therefore hardly proceed from the Ampullaceal orifices. 

The only other paired orifices on the intermediate spinnerets 
are those of the Tubuliform glands. Now, an important 
function of these glands is undoubtedly, as Apstein remarks, 
the spinning of the egg cocoon, for they are always distended 
with yellow fluid in the female just before the deposition of 
ova, and comparatively inconspicuous after, while the cocoon 
consists of yellow silk. 

If, however, they also furnish the ground-threads, this would 
help to explain their presence in the male spider, which has not 
hitherto been very easy to understand. 

The objections to this view are, first, that cocoon silk is not 
especially elastic, and secondly, that I have not been able to 
find threads in the cocoon of the precise diameter of the ground- 
threads. 

In spiders of the species under consideration the following 
thread-diameters were found to be fairly uniform: 


Cocoon line d : ; : : ; 006 mm. 
Anterior Ampullaceal . 5 ‘ : : 003) 7 
Ground-line of spiral . : : ‘ : 0025 ,, 
Intermediate Ampullaceal . : : : 0016 ,, 


The imperfect view I obtained of the origin of the ground- 
thread led me to think that though it proceeded from the inter- 
mediate organs, it had some subsequent relation to the posterior 
spinnerets. 

It is possible, therefore, that Apstein is correct in supposing 
that the Aggregate glands, which debouch on the inner side 
of the posterior spinnerets, deposit the viscid matter above 
described. | 

The arguments hitherto adduced in support of this view are, 
first, the convenient arrangement of the Aggregate orifices for 
such a purpose, and secondly, the presence of these glands in 


SPINNING APPARATUS OF GEOMETRIC SPIDERS. 37 


such spiders—and such only—on whose threads the viscid 
matter has been observed. On dissecting out the various 
glands from a spider, isolating them on slides, and crushing 
them, I found that the contents of the Aggregate glands 
retained their viscidity the longest. Evidence was also sought 
from histological changes in the glands themselves before and 
after web-spinning, and though a much larger series of obser- 
vations would be necessary to afford trustworthy results, 
alterations similar to those known to occur in active serous 
glands seemed to be taking place (figs. 19 and 20). 

This would show that the Aggregate glands are used in 
spinning the web, in which case they must furnish the viscid 
matter, all the other structures being accounted for. 

The unsafe nature of such indirect evidence is, however, 
freely admitted, but it may be pointed out that the certainty 
which now exists with regard to some of the glands gives 
greater probability of the true function being allotted to the 
remainder. 

One other web structure remains to be briefly discussed. 
Foundation lines are attached to surrounding objects, and 
ordinary non-viscid lines are glued to one another by little 
patches of silk which we may call attachment discs (Haft- 
scheibe, Apstein). The spider rubs its anterior spinnerets 
against a surface, emitting silk from the Piriform glands, and 
upon walking away a line is drawn out from the spinnerets. 

I have been best able to study these structures in a small 
bottle in which a spider was obliging enough to deposit its 
eggs, fixing the cocoon in its place by a multitude of cross 
threads fixed to the sides of the bottle at their ends, and to one 
another where they intercrossed. Their appearance is given in 
figs. 16—18. It was this structure which led to the belief in 
the highly compound nature of the spider’s line. 


38 CECIL WARBURTON. 


Summary. 


1. Facts newly established.—A spider’s line does not 
consist of many strands fused or woven together, but ordinarily 
of two or four distinct threads. 

The framework and the radii of circular snares are supplied 
by the Ampullaceal glands. 

The Acinate and Piriform glands are those mainly em- 
ployed in binding up captured prey. 

The “trailing line” consists primarily of Ampullaceal 
threads, sometimes strengthened by contributions from the 
Acinate and Piriform glands. 

The ground-line of the spiral is double only, and the two 
strands are bound together merely by the viscid matter which 
envelops them. 

2. Corroborative of Apstein.—The “attachment discs” 
are furnished by the Piriform glands. 

The Tubuliform glands supply the silk for the egg-cocoon. 

The viscid matter of the spiral is probably the product of the 
Aggregate glands. 

Finally, the origin of the spiral ground-line is uncertain, 
but it may proceed from the Tubuliform orifices on the inter- 
mediate spinnerets. 


Morphological Laboratory, Cambridge. 


SPINNING APPARATUS OF GEOMETRIC SPIDERS. 39 


EXPLANATION OF PLATE V, 


Illustrating Mr. Cecil Warburton’s paper on “ The Spinning 
Apparatus of Geometric Spiders.” 


Fic. 1.—Profile of Epeira diademata, sp. spinnerets. 
Fic. 2.—Ventral aspect of the same species. 

Fig. 3.—Ampullaceal gland. 

Fic. 4.—Ageregate gland. 

Fie. 5.—Tubuliform gland. 

Fie. 6.—Piriform gland. 

Fic. 7.—Acinate gland. 


Fic. 8.—External spinning organs at rest. a. Anterior, p. Posterior 
spinnerets. ¢. Anterior tongue-like fold. z. Terminal fold of abdomen. 


Fics. 9—13 show the composition of the “trailing-line” under various 
circumstances. 7¢. Intermediate spinnerets. 


Fic. 14.—Stages in the formation of the viscid globules. d. Shows the 
final arrangement. 


Fic. 15.—Teased spiral line, showing that the ‘ ground-line”’ is double. 
Fie. 16.— Attachment disc” (Haftscheibe, Apstein). 
Fic. 17.—The same, more in profile. 


Fie. 18.—Attachment disc, gluing together irregular strands which held 
an egg-cocoon in position. 


Fic. 19.—Section (somewhat diagrammatic) of aggregate gland at rest. 


Fic. 20.—Ditto of aggregate gland when the spider had just constructed 
its web. (The right half only of Figs. 19 and 20 is shaded.) 


7" 


* 
4 


STRUCTURE, ETC., OF CERATA OF NUDIBRANCHS. Al 


On the Structure and Functions of the Cerata 
or Dorsal Papille in some Nudibranchiate 
Mollusca. 


By 


Ww, A. Herdman, D.Sc., F.L.S., 
Professor of Natural History in University College, Liverpool. 


With Plates Vi; VIL, VIII, EX, and X. 


Most of the Nudibranchiate Mollusca are provided with 
brightly coloured and sometimes elaborately branched projec- 
tions from the sides and dorsal surface of the body. These 
include— 

1. The Rhinophores, or dorsal tentacles. 

2. The true Branchiz. 

3. The Cerata, or dorsal papille. 

The rhinophores are a pair of tentacles placed near the 
anterior end of the body, on the dorsal surface of the head. 
They are undoubtedly sense-organs, and are supplied by large 
nerves arising from the cerebral ganglia. They are present in 
all the forms discussed in this paper. 

The branchiz, although they may possibly not be true 
ctenidia, are specialised organs of respiration. They are not 
present in all Nudibranchs. 

The cerata, which were the special subject of my investiga- 
tion, vary greatly in number, size, and arrangement in the 
different genera and species, and the characteristic appearance 
of the animals is in a great measure due to these structures. 
They are often termed dorsal papille, or branchial papille, or 
even branchie ; and they have been supposed by many z0o- 


42 WwW. A. HERDMAN. 


logists to be organs of respiration. They are not present in 
all Nudibranchs, but in many cases they are very large and 
conspicuous. ‘They may be present along with true branchie. 
I find the cerata in the genera which I have examined to be of 
two kinds : 

1. There are those which contain large diverticula of the 
liver, as in the case of the genera Holis and Doto. 

2. There are those which are essentially processes of the 
body-wall, and have no connection with the liver, as in the 
genera Tritonia, Ancula, and Dendronotus. 

The term “ cerata” may, as introduced by Lankester in his 
article ‘‘ Mollusca,” ! be employed for these processes in 
general, while those in the first category might be specially 
denoted as hepato-cerata, and those in the second as parieto- 
cerata. All the forms which I have examined are either 
distinctly hepato-cerata or are parieto-cerata. I have found 
no intermediate conditions. In regard to their morphological 
nature, if the fold of integument overhanging the foot in 
Doris is to be regarded not as a mantle edge, but as an epi- 
podial ridge (see Lankester, loc. cit., p. 655), then the smooth 
or tuberculated dorso-lateral ridges in the genera Goniodoris, 
Polycera, and Idalia, the larger row of lateral tubercles of 
Aigires punctilucens, the lateral clavate processes of Triopa 
claviger, the palisade-like cerata of Ancula, the branched 
parieto-cerata of Tritonia and Dendronotus, and probably also 
the hepato-cerata of Doto, Holis, Proctonotus, and all other 
forms, may be considered as epipodial papilla—outgrowths from 
a more or less distinct epipodial ridge. 

The six common British genera, Doris, Ancula, Tritonia, 
Dendronotus, Doto, and Eolis, show very different con- 
ditions of the cerata and other dorsal processes, and form an 
instructive series of types. The general anatomy of all these 
forms is well known, thanks chiefly to the labours of Alder 
and Hancock and of Rudolph Bergh, and many points in the 
detailed structure of particular organs have been worked out 
by Bergh, Vayssiére, Trinchese, and others; but the method 


1 * Ency. Brit.,’ ninth edition, vol. xvi, p. 655. 


STRUCTURE, ETC., OF CERATA OF NUDIBRANCHS. 43 


of serial sections, giving the exact histological relations of the 
different parts of the body, has apparently not up to now been 
made use of by any of these writers on the structure of the 
Nudibranchiata. 

My specimens have been collected in the Liverpool Bay 
district, either in the neighbourhood of the Biological Station 
on Puffin Island, or at Hilbre Island, in the estuary of the Dee. 
They were generally killed with Kleinenberg’s picric acid, 
hardened with graduated alcohols, stained in picro-carmine, 
embedded in paraffin, cut with the Cambridge “ rocking” 
microtome, and mounted in Canada balsam. Some were 
soaked in gum, cut in the freezing microtome, and examined 
in water, in glycerine, and in Farrant’s solution for comparison 
with the others. My laboratory assistant, Mr. J. A. Clubb, 
who is working along with me in the collection and identifi- 
cation of the Nudibranchs of the district for the Reports upon 
the Fauna of Liverpool Bay, has given me a great deal of 
assistance in preparing the specimens and cutting the sections. 


Doris. 


In Doris (Pl. VI, fig. 1) there is a pair of short stout 
laminated rhinophores on the head, and a clump of well- 
developed branchie near the posterior end of the dorsal 
surface of the body. There are no cerata or other dorsal 
processes. The branchiz are in the form of a number (usually 
6 to 12) of pinnate plumes arranged in a circle round the 
anus. In sections the branchiz have the structure shown in 
fig. 2. The branches are subdivided and the surface is very 
irregular. The epithelium varies from nearly squamous to 
columnar, and there are large blood-lacunz forming irregular 
spaces and passages and coming into close relation with the 
surface, being only separated from the ectoderm in some 
places by a very thin layer of structureless connective tissue. 


ANCULA. 


In Anculacristata (fig. 3) there are rhinophores, well- 
developed branchiz, and large but simple unbranched cerata. 


44, WwW. A. HERDMAN. 


The rhinophores are large, and are placed in the usual position 
on the head. Each of them has its upper half strongly 
laminated or marked with parallel transverse ridges, while 
near the base of each two simple tapering branches arise, the 
one directed horizontally forwards and the other rather out- 
wards to the side. There are three branchial plumes, which 
are placed in the centre of the dorsal surface. The largest one 
is median and anterior, and the other two form a pair placed a 
little further back : they are all much branched. 

The cerata form a series of five erect, rod-like processes 
along each side of the back. They extend from the centre 
nearly halfway to each end of the body, and thus form a 
protecting palisade along the middle third of the back, at 
each side of the branchiz. 

In sections through the front of the head (fig. 4) it is seen 
that the branches of the rhivophore are, like the cerata further 
back, prolongations upwards of the body-wall composed of 
ordinary mesodermal tissue containing only the usual small 
blood-lacune. A few sections behind (fig. 5) we come upon 
the rhinophore proper, showing the broad lateral lamine, 
while the stem contains a large bundle of nerve-fibres and, 
further up, a ganglionic mass of nerve-cells (fig. 5, g). 

Some way further back in the body the cerata and branchie 
are seen in section. Fig. 6 shows in the centre the basal 
part of the first or median branchia cut near its anterior end. 
It contains a large blood-cavity. On each side is seen one of 
the cerata, that on the right having had several undulations near 
its middle. These cerata are seen from this and neighbouring 
sections to be direct continuations upwards of the ectoderm 
and mesoderm of the body-wall, and to contain no special struc- 
tures beyond the epithelium and the connective and muscular 
tissues of the integument (fig. 6). There are many small blood- 
lacune in the mesoderm, but these are not more numerous nor 
larger than the corresponding spaces in the body-wall, and 
nothing approaching the structure of a branchia is seen. 

Pl. VII, fig. 7, shows a section further back where the 
median branchia is cut through longitudinally about its central 


STRUCTURE, ETC., OF CERATA OF NUDIBRANOHS. 45 


part, while the second pair of cerata are seen one on each 
side. This shows well the laminated structure of the branchia 
and the bundle of muscle-fibres branching through its interior. 
A few spaces are visible in the branchia, but they are com- 
paratively small, and it is only under a higher magnification 
that the numerous lacunz lying in the mesoderm close under 
the ectoderm, and containing blood-corpuscles (fig. 8), become 
visible. This is part of a vertical section; while fig. 9 shows 
part of a transverse section similarly magnified. These two 
figures show the deep infoldings of the surface of the branchia, 
and the former (fig. 8) exhibits well the change in the character 
of the ectoderm cells from place to place. The general arrange- 
ment and structure is the same as in the section of the 
branchia of Doris (fig. 2), and is very different from the struc- 
ture of the cerata when similarly sectionised and magnified 
(see fig. 10). So that, although in sections such as are repre- 
sented in figs. 7 and 11 the cerata and the branchiz sometimes 
overlap and become displaced, small pieces of the one are 
always distinguishable by their structure from those of the 
other. The cerata (see figs. 6, 7, 10, and 11) have the ecto- 
derm very thick, and the infolds are not nearly so_ deep or so 
close as those of the branchie. A layer of longitudinal muscle- 
fibres (fig. 10, m) lies under the ectoderm in the cerata, and 
there are only a few small lacune in the mesoderm. 

Fig. 11 represents a section further back, in which parts of 
all three branchiz and of two pairs of cerata are seen. The 
lateral branchie have their inner surfaces much more deeply 
folded than their outer surfaces, and this is especially the case 
near their bases. This is shown in fig. 12, a vertical sec- 
tion of the base of one of the lateral branchiz, where the left 
side shows the outer surface next to the cerata, while the right 
side is the inner surface nearest to the middle line of the body. 
Some of the deepest infolds of the ectoderm are seen to end in 
little crypts where the ectoderm cells become suddenly large 
and are arranged in a radiating manner around the end of the 
infold so as to form a spherical clump (fig. 12, gl). These 
are probably glandular. 


46 W. A. HERDMAN. 


TRITONIA. 


In Tritonia (e. g. Tritonia, or Candiella, plebeia) the 
body is long and low, nearly square in transverse section, and 
tapers rapidly to the posterior end (fig. 13). The rhinophores 
are large and complicated, having the base surrounded by a 
sheath and the terminal part divided up into a number of 
branches. There are no true branchie such as are present in 
Doris and in Ancula, but placed along each side of the dorsal 
surface is found a row of short branched cerata (fig. 13, c). 
These are seen in sections (figs. 14 and 15) to be merely 
processes of the body-wall containing no special structures and 
only a few small lacunz, such as are present under the integu- 
ment all over the body. In some transverse sections, where 
the sides of the body are much corrugated the irregular folds 
of the surface are almost as much branched as the cerata, and 
have very much the same appearance (see / in fig. 14). 

It is clear then (1) that true branchiz, such as those of 
Doris and Ancula, are not present in Tritonia plebeia; 
(2) that the cerata of the latter are merely processes of the 
body-wall like the cerata found along with branchie in Ancula; 
and (3) that although these cerata may become considerably 
branched (see fig. 15) they have not the structure of special 
respiratory organs. 


DENDRONOTUS. 


In Dendronotus arborescens there is practically the 
same condition asin Tritonia. Branchiz are absent, but the 
rhinophores are large and complicated, and the branched 
cerata arranged along the sides of the back are so greatly de- 
veloped as to form the most conspicuous part of the animal 
in the living condition (Pl. VIII, fig. 16). There are usually 
six pairs of these cerata, with occasional much smaller ones 
scattered between. In a specimen 4 cm. in length the largest 
pair of cerata may be 1 cm. in height, and have stems 3 mm. 
in diameter at the base. They branch repeatedly so as to form 
an arborescent structure. 


STRUCTURE, ETC., OF CERATA OF NUDIBRANCHS. 4.7 


The cerata of Dendronotus have been generally described 
as branchie, and have been universally supposed, until quite 
recently, to contain large digestive ceca or diverticula of the 
“liver.” I regard them, however, as being merely excessive 
developments of the small cerata found in Tritonia, and as 
having no special branchial function ; while last summer Mr. 
Clubb and I showed! that no digestive czeca penetrate into the 
cerata in Dendronotus. Such ceca were described and 
figured originally by Alder and Hancock,? and more recently 
by Dr. R. Bergh,® but these distinguished anatomists worked 
entirely, I believe, by means of fine dissections, and I can 
explain, I think, how it is that a deceptive appearance of 
hepatic diverticula is produced which has led to error when 
not corrected by the examination of serial sections. 

The so-called liver is a very large organ lying underneath 
(ventral to) the ovo-testis. It consists of a posterior and right 
and left anterior lobes, as correctly described by Bergh. It 
gives off a few diverticula directed dorsally, but these do not 
reach to the bases of the cerata, but end blindly in the body- 
wall. In the specimens examined by Mr. Clubb and myself 
last summer we found such prolongations going towards the 
rhinophores and the two first pairs of cerata, and sometimes, 
but less definitely, towards the smaller succeeding cerata, but 
in no case, either in dissections or in sections, were they found 
to reach the base either of rhinophores or of cerata. 

Dissections alone are apt to be misleading, as there are large 
blood sinuses (a) in the side walls of the body close to the liver 
and (4) extending up into the cerata, and these cavities join and 
open into the dorso-lateral veins close to where the hepatic 
diverticula terminate, so that it is easy to imagine a direct con- 
tinuity between the slender end of the diverticulum and the 
blood sinus and so proceed to trace the supposed hepatic ceca 
onwards into the cerata. In serial sections, however, the pro- 


1 ©Proe. Liverpool Biol. Soc.,’ vol. ili, p. 228, 1889. 

2 ¢ Ray Soc. Monograph,’ pt. ii. 

3 “ Bijdragen tot de Dierkunde,” ‘Natura Artis Magistra,’ Afl. xiii, viii, 
p. 25, Amsterdam, 1886. 


48 W. A. HERDMAN. 


longations of the liver can be followed with exactness until 
their terminations in the body-wall are found. In one case, 
for example, amongst our preparations the hepatic cecum 
going towards the left rhinophore can be traced forwards 
through sixty-six sections, gradually narrowing until it ends 
blindly, the last section passing through its anterior wall. At 
this point it has not nearly reached the base of the rhinophore. 

Dr. Bergh has figured! the cecal extremities of the hepatic 
diverticula in the terminal branches of the cerata as seen in 
transparent preparations, and I freely admit that such appear- 
ances are sometimes to be seen and that they look superficially 
very like granular, dark-coloured liver ceca; moreover, when 
one of the cerata is cut off near its base from a living Den- 
dronotus the cut surface sometimes (i. e. in darkly coloured 
individuals) shows an outer clearer zone, and then a dark 
chocolate-coloured ring which is very suggestive of the hepatic 
czeca, as seen in the cerataof some species of Holis. Sections, 
however, show that in both such cases, the terminal twigs and 
the freshly cut stumps of the cerata, the appearance is due to 
branching masses of pigment-cells lying in the solid mesoderm, 
always a little way in from the surface and sometimes more 
densely aggregated around the blood sinuses. I found that a 
specimen killed and hardened rapidly, soaked in syrup and 
gum, and cut at once in the freezing microtome without 
staining, showed these pigment-cells much better than did the 
specimens carefully hardened and stained and embedded in 
paraffin. PJ. VIII, fig. 19, shows the arrangement of the pig- 
ment in such a fresh section: it is of a rich reddish-brown 
colour. 

I believe, then, that the appearance of hepatic prolongations 
in the cerata, which have been described by various careful 
investigators, is due to the presence (1) of blood sinuses, and 
(2) of a good deal of dark pigment in the mesoderm, and that 
the hepatic ceca are not really prolonged into the cerata, 

Dr. Bergh has lately suggested to me in conversation that 
possibly my results might be due to the ceca being contractile, 


1 Loc. cit., pl. ii, figs, 21, 22. 


STRUCTURE, ETC., OF CERATA OF NUDIBRANCHS. 49 


and having been in some specimens retracted completely into 
the body ; but that cannot have been the case, because, in the 
first place, it is difficult to understand how a system of czeca 
extending up into the terminal twigs could be completely with- 
drawn from a densely branched structure like the cerata; and, 
in the second place, some of my sections were made from spe- 
cimens in which the cerata were suddenly cut off from the 
living animal with a pair of fine scissors, when fully expanded 
and healthy, in a dish of sea-water, and these showed the same 
structure when sectionised as did the other preserved speci- 
mens. I mention this, here, to show that this conceivable ex- 
planation of the absence of the cxca, which might occur to 
other readers, had been foreseen, and found not to be possible. 
The argument that as Dendronotus belongs to the group 
Kladohepatica it is very unlikely to be without hepatic ceca in 
its cerata is worthless, as Bergh has described an Eolid (Bor- 
nella excepta) which has absolutely no prolongations of the 
liver into the cerata.! 

The large cerata of Dendronotus are, then, as we would 
expect from our previous examination of the smaller similar 
cerata of Tritonia, prolongations upwards of the mesoderm 
and ectoderm of the body-wall, and contain no special structures, 
such as are found in the cerata of Holis and Doto. 

The upper part of fig. 17 shows a longitudinal section of one 
of the cerata, and fig. 18 is a drawing of a transverse section. 
The ectoderm-cells throughout are of moderate size, of low 
columnar form, and are not differentiated in any part. The 
mesoderm, which has the same structure as that of the cerata 
of Ancula and Tritonia, is penetrated by large, irregular 
spaces, containing blood-corpuscles. These may be called the 
ceratal sinuses ; they are near the centre of the mesoderm, and 
run in the main longitudinally (see fig. 17,c.s.); they occasionally 
branch, and they open into the numerous minute lacune which 
exist in the mesoderm here as elsewhere. Fig. 18 shows a 
transverse section where several branches of the ceratal sinus 
are present, In fig. 19 also several spaces containing blood- 

1 See ‘Report upon the “ Challenger ” Nudibranchiata,’ p. 41. 

VOL, XXXI, PART I.-—NEW SER, D 


50 Ww. A. HERDMAN. 


corpuscles are seen. Fig. 20 shows a small portion of the 
mesoderm more highly magnified, to show the network 
of connective tissue and the small blood-lacune in the 
meshes. 

At the bases of the cerata these large ceratal sinuses are 
continued into the body, and their communication can be 
traced in sections with the anterior and posterior dorso-lateral 
veins (fig. 17, d. J. v.), which open directly into the auricle. 
The junction between the ceratal sinus and the dorso-lateral 
vein is effected by means of a narrow transversely running 
branch, and from this point the ceratal sinus is continued 
ventrally through the mesoderm of the body-wall outside 
the “ liver,” and may be called the lateral sinus (fig. 17). It 
is with the upper part of these lateral sinuses that the prolonga- 
tions from the liver come in some places into close proximity, 
and so may have given rise in dissections to the appearance of 
a direct continuity between the liver and the blood-spaces in 
the cerata. . 

The cerata contain also bands of muscle-fibres, mostly lon- 
gitudinal in direction, nerves, pigmented connective tissue, 
forming branched masses and ramifying threads of a rich 
brownish colour, and finally masses of large distinctly nucleated 
cells, lying in meshes of fibrous connective tissue (see fig. 21). 
These occur chiefly in the smaller branches of the cerata, and 
are possibly mucus-secreting glands; they resemble the small 
groups of gland-cells seen under the ectoderm in the cerata of - 
some species of Holis (see fig. 837). The contrast in structure 
between transverse sections of the cerata of Dendronotus and 
of Holis is seen by comparing ‘figs. 18 and 34 or 35. 


Doro. 


In the genus Doto there are no true branchie; the rhino- 
phores are large with simple filiform distal ends, but having 
their’ bases surrounded by large funnel-shaped sheaths. The 
cerata form a row along each side of the back (fig. 22); they 
are very large and complicated, being swollen, tuberculated, 
usually brightly coloured, and forming the most conspicuous 


STRUCTURE, ETC., OF CERATA OF NUDIBRANCHS., 51 


part of the body. They contain large branched hepatic 
diverticula, and are therefore hepato-cerata. 

Transverse sections of Doto coronata show the relatively 
very large size of the hepato-cerata (fig. 23), and the manner 
in which they are occupied by numerous branches of the large 
hepatic ceca; ten or a dozen branches may often be found 
lying in one section. In fact, in this form, there is far more of 
the liver in the cerata than in the body proper. The median 
portion of the liver is reduced to a small tube flattened dorso- 
ventrally, which lies along the under surface of the large ovo- 
testis (see figs. 23 and 27, m. /.), and gives off at intervals lateral 
branches, which run up the sides of the ovo-testis (figs. 25 and 
26, h. c.’) to enter the cerata, and there expand into the large 
branched ceca. The difference, then, between this state of 
affairs, where the part of the liver in the body is little more 
than a duct leading from the hepato-cerata to the stomach, and 
that seen in Ancula, Tritonia, and Dendronotus, where 
the liver is wholly in the body, and the parieto-cerata are 
merely processes of the mesoderm and ectoderm of the integu- 
ment, is very great, and affords sufficient ground, I think, for 
the separation of the cerata into two categories. 


Eo.is. 


For my present purpose it is convenient to use the term 
Eolis in its older, wide sense, as employed for example by 
Alder and Hancock, and as including the modern genera 
Facelina, Flabellina, Coryphella, Galvina, &c. In 
these forms we have much the same condition as in Doto. 
There are no true branchie, rhinophores are present, and there 
are also large coloured hepato-cerata arranged along the back 
(Pl. X, fig. 29), and constituting the most conspicuous part of 
the animal. 

The hepatic diverticula in the cerata are either simple or not 
so much branched as in Doto, and are not cecal, but com- 
municate indirectly with the exterior at their apices. The 
hepato-cerata also contain at their apices cnidophorous sacs 


52 W. A. HERDMAN. 


which open at the upper end to the external world, and at the 
lower into the extremity of the hepatic diverticulum. 

This state of affairs was long ago pointed out by Alder and 
Hancock! as seen in transparent specimens, and it has more 
recently been demonstrated by Bergh in Phidiana Selence, 
Facelina Janii, Chlamylla borealis and Gonieolis 
typica; but the communication has often been denied or 
doubted, and Ray Lankester probably expressed the mental 
attitude of most zoologists towards the matter when he wrote 
in 1883 that the supposed communication of the hepatic ceca 
in the dorsal papille or cerata of some of the Ceratonota with 
the exterior by means of apertures in the apices of the papille 
‘requires confirmation.”* Last year Mr. Clubb and I de- 
scribed and figured® sections showing the exact manner in 
which the communication takes place in specimens of an Eolis 
from the Puffin Island Biological Station, and I now give 
some more detailed figures here (Pl. X, figs. 32, 36, and 37). 

The upper end of each of the hepato-cerata is occupied by a 
sac containing a number of large cells (the cnidocysts) filled 
_ with ecnida or thread-cells. This cnidophorous sac is evidently 
an invagination of the ectoderm, the cnidocysts being modified 
ectoderm cells (figs. 32, 36), and it communicates with the 
exterior by a small but perfectly distinct and clearly-defined 
aperture at its apex, through which the thread-cells are some- 
times found protruding (fig. 36). 

The size and shape of the cnidophorous sac varies in different 
species. Figs. 28a to 28c represent the upper ends of hepato- 
cerata from Facelina drummondi where the sac is greatly 
elongated, may become irregularly shaped, and overlap the 
upper end of the hepatic cecum (fig. 30). The ecnida are 
ovate or nearly spherical in shape (figs. 30, 38, and 39), with a 
small terminal projection, and the everted threads bear some 
large spines arranged in a spiral round the base (fig. 39), and 
smaller ones projecting alternately from opposite sides all the 

1 *Ray Soc. Monograph,’ part iii. 
® Article “ Mollusca,” ‘ Ency. Brit.,’ ninth edition, vol. xvi, p. 659. 
§ * Proe, Biol. Soc. Liverpool,’ vol. iii, p. 233, and pl. xii, 


STRUCTURE, ETC., OF CERATA OF NUDIBRANCHS. 53 


way along. In (?) Cuthona nana (figs. 36, 37) the sac is 
short and rounded, and the cnida are much smaller than in the 
last species, but still spherical in form; while in Galvina 
picta (figs. 31, 32) the sac is more elongated, the cnidocysts 
are very distinct (fig. 40), and the cnida are narrow rod-like 
bodies (fig. 33). 

The hepatic czeca occupying the greater part of the interior 
of the cerata (see figs. 34 and 35, which show transverse sections 
of two species) reach nearly or quite to the lower end of the 
cnidophorous sac, and communicate with it by means of a 
longer or shorter slender tube with thin walls strengthened 
by a few muscle-fibres. In Facelina drummondi (figs. 
28 and 30) the connecting-tube is very long, and may be bent 
upon itself. In the small species of Eolis shown in figs. 36 
and 37 (probably Cuthona nana) the cnidophorous sac is 
nearly spherical, and the connecting-tube is short and has a 
distinct muscular thickening, forming a sphincter around the 
small opening into the hepatic cecum. This condition suggests 
that possibly in all cases the communication between the 
hepatic cecum and the exterior through the cnidophorous sac 
may not be permanently open, but be kept closed when required 
by the contraction of the sphincter muscle. 


FUNCTIONS OF THE CERATA. 


In regard to the functions of these various kinds of cerata 
in the Nudibranchiata, in the first place I do not think that in 
any case they are specially branchial. In Ancula, as I have 
shown above, there are parieto-cerata existing along with true 
branchiz, and the two have a distinct structure, so that, 
although in sections pieces of the cerata and of the branchiz 
may become displaced, they can be distinguished by their 
structure from one another. Then in Tritonia and in 
Dendronotus I have shown that the parieto-cerata agree in 
structure with those of Ancula, and not with the true 
branchiz of Doris and Ancula. From a recent conversation 
with Dr. Bergh I learn that he regards the cerata as having a 
branchial function, and even in Ancula, where there are 


54 W. A. HERDMAN. 


large true branchiz present, he thinks that the cerata are 
supplementary respiratory organs. I am still of opinion, how- 
ever, that, considering the relatively large size of the branchize 
and the perfection of their adaptation to their function and 
the absence of any such adaptation in the cerata, the action 
of the latter in effecting respiration must be so feeble, com- 
pared with the action of the branchie, that it may be neglected. 

Dendronotus arborescens is the form in which it might 
be most readily supposed that the parieto-cerata have ac- 
quired, secondarily, a branchial function, but a close com- 
parison of sections shows that these processes do not contain 
more blood-cavities than the general body-wall, and have not 
even so many small lacune close to the surface as some parts 
of the dorsal and lateral integument. Hence, although they 
may by their extended surface aid somewhat in respira- 
tion, still they cannot be regarded as in any way specialised 
branchie. 

Then, again, in Eolis and Doto, although from their 
relatively very large size the hepato-cerata may be of some 
importance in respiration, it is merely as being an extension of 
the general integument, and not as being special respiratory 
organs. Nearly the whole of the space in the hepato-cerata 
in these two genera is occupied by the large hepatic czeca, and 
there are only a few small blood-lacunz to be seen scattered 
here and there in sections. Specimens of both Eolis and 
Doto continue to live after being deprived of most of their 
cerata; so, both from their constitution and as the result of 
experiment, it may be inferred that these structures cannot be 
of primary importance as respiratory organs. One function, 
of course, of the cerata in these genera is to contain the greater 
part of the liver; and no doubt this has led to an increased 
size and some modification of structure. In Eolis, finally, the 
apices of the hepato-cerata accommodate the cnidophorous 
sacs, which act, doubtless, as important organs of offence. 

But I believe that, in addition to these minor functions, the 
cerata of the Nudibranchiata are of primary importance in 
giving to the animals, by their varied shapes and colours, 


STRUCTURE, ETC., OF CERATA OF NUDIBRANCHS. 55 


appearances which are in some cases protective and in others 
conspicuous and warning ; and in this, it seems to me, we have 
an explanation of the extraordinary development and variety of 
these otherwise mysterious processes of the dorsal body-wall. 

For several years past I have been paying some attention to 
the colours of Nudibranchs and their variations in connection 
with their habits and natural surroundings. In October, 1888, 
I described! a peculiarly coloured Doris (Archidoris) tuber- 
culata which was especially well protected from observation, 
and since then I have found the same species repeatedly lying 
in hollows in the surface of large sponges, generally Hali- 
chondria panicea, and simulating the colours of its sur- 
roundings so closely as to be quite inconspicuous.” Giard has 
recently noticed this same point on the coast of Normandy, 
and has also recorded a few other cases of protective colouring 
amongst common Nudibranchs.” 

The view which I have given above in regard to the primary 
function of the cerata occurred to me early last summer, when 
observing some of the Nudibranchs in their natural conditions 

_on the shore at Puffin Island, and I have since brought the 
theory briefly before the notice of the Liverpool Biological 
Society and before Section D of the British Association at the 
recent Newcastle-on-Tyne meeting. Since then Mr. Garstang 
has independently arrived at practically the same conclusions 
in regard to the function of the cerata from his observa- 
tion of the colouring and habits of the Nudibranchs at 
Plymouth. 

I shall now give a few instances from my own observations 
in support of my views. 

Tritonia (or Candiella) plebeia is fairly abundant at 
Puffin Island and at Hilbre Island, near Liverpool, and is 
always found (so far as I have noticed) in these localities 

1 ¢ Proc. Biol. Soc. Liverpool,’ vol. iii, p. 13. 

2 IT see that Mr. Walter Garstang, in his recently published “ Report upon 
the Nudibranchs of Plymouth Sound,” has noticed this same instance of 
protective colouring. ‘Journ. M. B. A.,’ vol. i, No. 2, p. 174. 


3 * Bulletin Scientifique de la France et de la Belgique,’ t. xix, 1888, p. 492. 
4 See Abstract in forthcoming volume of Reports. 


56 Ww. A. HERDMAN. 


creeping over the surface of colonies of Aleyonium digi- 
tatum. The specimens of Tritonia plebeia are marked 
with many colours (none of them bright) including tints of 
yellow, brown, blue, grey, black, and opaque white ; and when 
examined in a vessel by themselves considerable differences 
between individuals are noticed, but when in their natural con- 
dition on the Aleyonium colony they are nearly all equally 
inconspicuous. The colonies of A]cyonium differ consider- 
ably amongst themselves in tint, some being whiter, others 
greyer, and others yellower than the rest. Different parts of 
the same colony also vary in appearance on account of the 
different states of expansion of the polypes, and on account of 
irregularities of the surface and of adhering sand and mud, so 
that the varieties of colouring found in Tritonia plebeia do 
not render it conspicuous, but are suited to the varying con- 
ditions of the Alcyonium colonies. The small branched 
cerata along the back of the Tritonia aid the protective re- 
semblance not only by contributing to the general colouring, 
but also by their similarity in appearance to the crown of 
tentacles of the partially expanded polypes. They are placed 
at just about the right distance apart, and have the necessary 
tufted appearance. 

Then, again, Doto coronata when isolated is a very con- 
spicuous and brightly coloured animal, but I find it at Hilbre 
Island invariably creeping on the under surfaces of ledges and 
stones on which are large colonies of the zoophyte Clava 
multicornis, and in that position the Doto is not readily seen. 
The gay appearance of this Nudibranch is mainly due to the 
large and brightly coloured cerata, and these agree so closely 
in their general effect with the upper ends of the zooids of 
Clava, covered with the numerous tentacles and the clusters of 
sporo-sacs, that when the Doto remains still it is hidden to a 
very remarkable extent. 

Dendronotus, again, with its large branched cerata and 

1 Mr. Garstang tells me, in a letter just received (October), that at Ply- 


mouth the specimens of Doto are not so highly coloured, and are found upon 
Calyptoblasts, Clava being rare there. 


STRUCTURE, ETC., OF CERATA OF NUDIBRANCHS. 57 


its rich reddish-brown and yellow markings, is a handsome and 
most conspicuous object, but I have frequently found it 
amongst masses of brown and yellow zoophytes (coarser forms 
such as Sertularia abietina and Hydrallmania falcata) 
and on purplish-red seaweeds, where it was very completely 
protected from observation and I did not for several seconds 
recognise what I was looking at.! 

Now, these are all cases where the colouring is protective, 
and I have no doubt there are many other similar instances to 
be found amongst the Nudibranchiata,? but the species of 
Eolis appear to belong to a different category. They are 
noted for the very brilliant hues of their cerata and they are 
always conspicuous, so far as I have noticed, even in their 
natural conditions. 

Then, again, the species of Holis are rarely found hiding in 
or on other animals; they are not shy, and they are active in 
their habits—altogether they seem rather to court observation 
than to shun it. When we remember that the species of 
Eolis are protected by the numerous stinging-cells in the 
cnidophorous sacs placed on the tips of all the cerata, and that 
they do not seem to be eaten by other animals, we have at 
once an explanation of their fearless habits and of their con- 
spicuous appearance. The brilliant colours are in this case of a 
warning nature for the purpose of rendering the animal pro- 
vided with the stinging cells noticeable and easily recognisable. 
It is, of course, important for the soft-bodied Nudibranch that 
it should be not only disagreeable to taste but also as con- 
spicuous as possible, in order that it may not run the risk of 
being tried by voracious animals. An experimental snap from 
a fish might cause the death of the Nudibranch even though it 
was immediately rejected as food. 

These, then, are the grounds upon which I base my view 
that the cerata from their structure cannot be important respira- 


1 Professor Giard finds it at Wimereux amongst red seaweeds of the genus 
Callithamnion. 

2 Such as the interesting cases of Hermea bifida and H. dendritica, 
described by Garstang, loc. cit., p. 191. 


58 W. A. HERDMAN. 


tory organs and that their chief function is by their varied 
shapes and colours to enable the animals to assume protective 
or warning appearances as may be found best suited to their 
surroundings and mode of life. 

It is still necessary for the satisfactory establishment of this 
theory that I should have some more definite experimental 
grounds for my opinion that such forms as Doto and Dendro- 
notus are edible, while Eolis is distasteful to (say) fishes, and 
I have lately arranged a series of experiments which will be 
conducted in the fish-tanks of the aquarium here, with the 
kind assistance of Mr. T. J. Moore, the curator of the Liver- 
pool Museum. We have just commenced observations, and 
have got satisfactory results so far with eight species of shore 
fishes, but at this season it is almost impossible in this neigh- 
bourhood to get Nudibranchs in any quantity. As soon as 
more material can be obtained the experiments will be resumed, 
and I shall give a detailed account of the results when suffi- 
cient evidence has been accumulated. 


Summary. 


1. In Doris there are true branchiz and no cerata. In 
Ancula both branchie and cerata are present. In Tritonia 
and Dendronotus there are cerata, but no true branchie. 
In Ancula, Tritonia, and Dendronotus the cerata, whether 
simple or branched, large or small, are merely processes of the 
body-wall (parieto-cerata) and contain no special organs or 
structures. 

2. In Doto and Eolis there are no true branchie. The 
cerata (hepato-cerata) are large, and contain extensive hepatic 
diverticula. 

3. In Eolis the hepato-cerata contain also cnidophorous 
sacs which communicate on the one hand with the distal end 
of the hepatic cecum, and on the other with the exterior at the 
apex of the ceras. 

4, Morphologically, all the forms of cerata are probably 
epipodial processes. 


STRUCTURE, ETC., OF CERATA OF NUDIBRANCHS. 59 


5. The large, elaborately branched parieto-cerata of Dendro- 
notus are merely a further development of the small tufted 
parieto-cerata of Tritonia, and although they may on account 
of their extended surfaces have secondarily acquired to a certain 
extent a respiratory function, they cannot be regarded as 
specialised branchie. 

6. The cerata, whether they are large branched parieto- 
cerata as in Dendronotus, or hepato-cerata containing the 
greater part of the liver as in Doto, or having cnidophorous 
sacs in addition as in Holis, are not of primary importance 
either in respiration or in digestion, but give to the animals, 
by their varied shapes and colours, appearances which are in 
some cases protective and mimetic, and in others conspicuous 
and warning, as may be found best suited to their individual 
surroundings and mode of life. 


EXPLANATION OF PLATES VI, VII, VIII, 1X, X, 


Illustrating Prof. W. A. Herdman’s paper “ On the Structure 
and Functions of the Cerata or Dorsal Papille in some 
Nudibranchiate Mollusca.” 


a, Artery. ap. Aperture of cnidophorous sac. 4. c. Blood-corpuscles. 
br., br., 6r?. Branchie. 6.8. Blood-space. ¢., c'., c®. Cerata. ex.s. Cnido- 
phorous sac. c. ¢. Connecting tube between hepatic caecum and cnidophorous 
sac. c. tis. Connective tissue. d./.v. Dorso-lateral vein. ec. Ectoderm. 
ec’. Invaginated ectoderm-cells (cnidocysts) which produce ecnida. ec’’. Young 
cnidocysts or cells of cnidophorous sac. f Foot. / g/l. Foot-glands. g. Gang- 
lion. gl. Gland-cells. 4. c. Hepatic cecum. 4. ce’. Narrow part of hepatic 
cecum in the body of Doto. 7. Junction of Cerata with body in Doto. 
k. Folds of the integument. 7. Liver. 7. m. Longitudinal muscles. m., 
Muscle-bands. m’. Longitudinal muscle-bands in body of Dendronotus. 
m.l, Median part of liver in body of Doto. mes. Mesodermal tissues. z. 


60 W. A. HERDMAN. 


Nerves. o.¢. Ovo-testis. yp. Pigment. 7. c. Renal cavity. 74. Rhinophore 
sph. Sphincter muscle. 7¢. m. Transverse muscles. ¢z. Oral tentacles. 


Diameters. 
S. | =Swilt’s lin-objoc. 2  ; ; 5 . magnifying about 45 
Seca oe ee : : ; : us »» 230 
Sepa cme FU ee ier ‘ ; ; bs » 93030 
Z. jz; = Zeiss’s 7; ,, (oil immersion), oc. 2 sf DUB 


5 if BS oc; 4. 33 950 to 1363 

Where not otherwise stated, the drawings were made from specimens 

hardened in Kleinenberg’s picric acid and graduated alcohols, stained in picro- 
carmine, embedded in paraflin, and cut with the ‘‘ rocking” microtome. 


PLATE VI. 


Fic. 1.—Outline of a Doris seen from the left side, showing the rhino- 
phores (r2.) and the branchie (d7.). About natural size. 


Fic. 2.—Part of a longitudinal section through the branchia of Doris 
(Acanthodoris) pilosa. 0.8. Large blood-space. S. 3. 
Fic. 3.—Outline of Ancula cristata seen from the right side, showing 


the rhinophores, the branchiz, and the cerata (¢.). x 3. 


Fic. 4.—Upper part of a transverse section through the head of Ancula 
(middle of odontophore), showing the tentacle-like branches of the rhinophores 
(rh.) cut in longitudinal section. The mesoderm contains only a few small 
blood-spaces (4. s.). 5. 1. 

Fic. 5.—Transverse section through the front of the body of Ancula, 
showing the rhinophores cut longitudinally. m. Muscles. x. Nerves. g. 
Ganglion. 8S. 1. 


Fic. 6.—Transverse section through Ancula at the anterior end of the 


median branchia (d7'.), showing the first pair of cerata (cl.), and the large 
blood-space in the branchia (4. s.). 8, 1. 


PLATE VII. 


Fic. 7.—Transverse section through Ancula in the middle of the median 
branchia, and first pair of cerata. S. 1. 


Fic. 8.—Small part of median branchia in longitudinal section, showing 
the blood-spaces in the mesoderm. S. 4. 

Fic. 9.—Small part of same branchia in tranverse section. 8. +. 

Fic. 10,—Part of the outer edge of one of the cerata shown in Fig. 7, near 
base. 58, 2. 


STRUCTURE, ETC., OF CERATA OF NUDIBRANCHS. 61 


Fic. 11.—Transverse section of Ancula through tie lateral paired branchis 
(é77.) _S. 1. 


Fig. 12.—Vertical section of part of base of lateral branchia. S. +. 


Fig. 13.—Tritonia (Candiella) plebeia, from right side. x 3. 

Fie. 14.—Outline of transverse section of Tritonia plebeia through 
rhinophores, showing processes of the body-wall (4.). 8. 1. 

Fic. 15.—Outline of transverse section of Tritonia plebeia through 
middle of body, showing cerata. S. 1. 


PLATE VIII. 


Fic. 16.—Dendronotus arborescens from the left side. Natural size. 


Fic. 17.—Transverse section of Dendronotus near the middle of the 
body, showing one of the second pair of cerata cut in longitudinal section 
(c.). The relative positions of the liver (/.), short hepatic cecum (A. ¢.), 
ceratal blood-sinus (c. s.), and dorso-lateral veins (d. J. v.) are shown. S. 1 
(reduced). 

Fic. 18.—Transverse section of one of the third cerata of Dendronotus, 
showing the solid mesoderm (mes.) containing muscles (m.) and ceratal blood- 
sinuses (¢.s.). 8.3. 

Fic. 19.—Transverse section near the base of one of the cerata cut off 
the living animal, rapidly hardened, cut in the freezing microtome, and ex- 
amined unstained in water, to show the distribution of the dark-brown 
pigment. S. 1. 


Fic. 20.—A small piece of the mesodermal tissue as seen in some sections 
of the cerata, to show the lacune containing blood-corpuscles. S. 2. 

Fic. 21.—A section near the top of one of the cerata, from a specimen 
killed with corrosive sublimate and glacial acetic acid, then hardened gradually 
with alcohols of different strengths, soaked in gum, cut with the freezing 
microtome, and stained with picro-carmine, to show the masses of gland-cells 


lying in meshes of connective tissue. 8, 1. 


PLATE IX, 


Fie. 22.—Outline of Doto coronata, from the left side. x 3, 


Fic. 23.—Transverse section of Doto coronata through the posterior 
part of the body, showing the large cerata and their contained hepatic cca 
(i ¢.). 8.1. 


Fic, 24,.—Part of a similar transverse section more highly magnified, showing 
the junction of one of the cerata with the body, §, 2, 

Fic, 25.—Part of another similar section, showing the continuation of the 
hepatic cecum (4. c!.) into the body alongside the ovo-testis (0. ¢.). S. 2, 


62 W. A. HERDMAN. 


Fic. 26.—Part of another similar section, showing the narrow continuation 
of the hepatic cecum (4. c!.) sinking into the body and moving to a more 
ventral position, so as to reach the median tube lying under the ovo-testis. 
8. 2. 

Fic. 27.—Part of another section, showing the median tubular part of the 
liver (m. 7.) lying in the body below the ovo-testis (0. ¢.). 5.2. 

Fic. 28, a, B, c—The extremities of three cerata of Eolis (Facelina) 
drummondi preserved in glycerine and then treated with potassic hydrate, 
and slightly squeezed to show the long recurved tube connecting the tip of 
the hepatic caecum with the cnidophorous sac. S. 1. 


PLATE X. 


Fic. 29.—Eolis (Galvina) picta from the left side. x 6. 

Fic. 30.—Part of the tip of one of the cerata of Eolis (Facelina) 
drummondi from Hilbre Island, preserved in glycerine and treated with 
potassic hydrate, showing the connecting tube (c. ¢.) between the cnidophorous 
sac (ez. s.) and the hepatic cecum (4. ¢.). sph. Sphincter muscle. ec. Hcto- 
derm, covered with fine cilia. 7. m. and ¢. m. Longitudinal and transverse 
muscle-bands. 5. 2. 

Fic. 31.—Transverse section near the tip of one of the cerata of Holis 
(Galvina) picta, showing the cnidophorous sac (cz. s.) containing large 
invaginated ectoderm-cells (ec’.), in which are placed the enida. 8. 3. 

Fic. 32.—Longitudinal section through the tip of one of the cerata of 
Eolis (Galvina) picta, showing the ectoderm (ec.) turning in at the 
terminal aperture (ap.) to form the large cells or cnidocysts (ec’.) in the 
cnidophorous sac (cz, s.). 8. 2. 

Fic. 33.—Three of the cnida of Eolis (Galvina) picta. Z. 34. 

Fic. 34.—Transverse section (with freezing microtome) of one of the 
cerata of Eolis (Acanthopsole) coronata from Puffin Island. ec. 
Ectoderm. mes. Connective tissue and muscle-fibres. 4. s. Blood-spaces in 
the mesoderm. 4. c. Hepatic cecum. S. 2. 


Fic. 35.—Transverse section of one of the cerata of Holis (Cuthona) 
nana (?), showing in addition to the hepatic caecum (2. c.) groups of large 
gland-cells (g/.) embedded in the mesodermal tissues. §&. . 

Fic. 36.—Longitudinal section through the tip of one of the cerata of 
Lolis (Cuthona) nana (?), showing the opening (ap.) of the enidophorous 
sac (cz. s.) to the exterior, the continuation inwards of the ectoderm-cells 
(ec.) to form the enidocysts (ec’.) which contain the enida, the sphincter 
muscle (sph.) round the opening of the cnidophorous sac into the hepatic 
cecum, and the clumps of gland-cells (g/.) lying in the mesodermal tissues, 
8.4, 


STRUCTURE, ETC., OF CERATA OF NUDIBRANCHS. 63 


Fig. 87.—A neighbouring section to the preceding one, showing the con- 
necting tube (ce. ¢.) between the cnidophorous sac and the hepatic czecum, cut 
open in the greater part of its length, being only crossed by a few of the 
fibres of the sphincter muscle at its lower end. The other parts are before. 8. 4. 


Fic. 38.—Group of enida of EHolis (Facelina) drummondi in various 
positions ; the upper four are unexploded, the lower one has the thread everted. 
Z. > 5- 

Fic. 39.—Cnida of Holis (Facelina) drummondi more highly mag- 
nified (x about 1360), showing the arrangement of the spines on the basal 


part of the thread. Z. 54, oc. 4, tube. 


Fic. 40.—One of the large cells or enidocysts (ec’.) from the cnidophorous 
sac of Holis (Galvina) picta, in which the cnida are formed, showing the 
nucleus and nucleolus and the numerous elongated cnida embedded in the 
protoplasm. Two younger cells (ec”.) are seen at the base. Z. 54. 


‘ er 
eee | pay UE ee ee 
. 
> Ld > 
A ee ee 
one @ r my (We On ree es bd Al eu ty 
. .. « 4 ¥ C a bre ‘(ake ae 
' Pe i ali aa ss eT? . ie 
. ia ceep hi vi ro) 
ial bal vrei pet Taos me | 
gas ; F - sat? cial Ww alta’ a 


= MA Ci Cet 
; ~ ; ; i Tom 
i , .* re & 
‘ 4 
=z 
' 
s . 
, 
be - 
F P 
. es - . 
. 
ci @ 
F 
- * @ 


. - 
ed - 
‘ 
or r } * 
= 
- 
P E 
, 
a 7 
‘ ne 
be Lg © 


OBSERVATIONS ON THE HISTOLOGY OF STRIPED MUSCLE. 65 


Further Observations on the Histology of 
Striped Muscle. 


By 


Cc. F. Marshall, M.B., M.Se., 
Late Platt Physiological Scholar in the Owens College. 


With Plate XI. 


In a recent paper published in this Journal! I gave the 
results of some researches on the histology of muscular fibre. 
The present paper is a record of my further investigations on 
this subject. These are somewhat incomplete owing to want 
of time to finish them satisfactorily, but several results have 
been obtained which I trust are worthy of publication. 

In order to render the questions dealt with more intelligible 
the summary of results arrived at in the former paper is repro- 
duced. 


Summary of Former Paper. 


1. In all muscles which have to perform rapid or frequent 
movements a certain portion of the muscle is differentiated 
to perform the function of contraction, and this portion takes 
on the form of a very regular and highly modified intra- 
cellular network. 

2. This network, by its regular arrangement, gives rise to 
certain optical effects which cause the peculiar appearances of 
striped muscle. 

1 «Observations on the Structure and Distribution of Striped and Un- 
striped Muscle in the Animal Kingdom, and a Theory of Muscular Contrac- 
tion,” ‘Quart. Journ. Mier. Sci.,? 1887. 

VOL. XXXI, PART I.—NEW SER. E 


66 O. F. MARSHALL. 


3. The contraction of the striped muscle-fibre is probably 
caused by the active contraction of the longitudinal fibrils 
of the intra-cellular network: the transverse networks appear 
to be passively elastic, and by their elastic rebound cause the 
muscle to rapidly resume its relaxed condition when the longi- 
tudinal fibrils have ceased to contract; they are possibly also 
paths for the nervous impulses. 

4. In some cases where muscle has been hitherto described 
as striped, but gives no appearance of the network on treat- 
ment with the gold or other methods, the apparent striation is 
due to optical effects caused by a corrugated outline to the 
fibre. 

5. In muscles which do not perform rapid movements, but 
where contraction is comparatively slow and peristaltic in 
nature, this peculiar network is not developed. In most if not 
all of the invertebrate unstriped muscle there does not appear 
to be an intra-cellular network present in any form; but in the 
vertebrate unstriped muscle a network is present in the form of 
longitudinal fibrils only ; this possibly represents a form of net- 
work intermediate between the typical irregular intra-cellular 
network of other cells and the highly modified network of 
striped muscle. 

6. The cardiac muscle-cells contain a network similar to 
that of ordinary striped muscle. 


Discussion of the Views of Recent Observers. 
Before commencing the subject-matter of this paper I propose 
to discuss the results arrived at by several recent observers 
concerning the structure of muscle-fibre, and also some of the 
criticisms which have lately appeared concerning the existence 
of an intra-cellular network in striped muscle. 

The most important paper to discuss is that of Rollett.? 
He considers the muscle-fibre to consist of longitudinal fibrille 
grouped together into “ muscle-columns,” the cross-sections of 
which correspond to Cohnheim’s areas. Filling up the spaces 
between the muscle-columns is the interfibrillar material or 

‘ «Arch. f. mikr, Anat.,’? 1888, pp. 233—265. 


OBSERVATIONS ON THE HISTOLOGY OF STRIPED MUSCLE. 67 


“sarcoplasma.”? Each muscle-column consists of alternating 
thick and thin segments ; in the centre of each thin segment is 
a dark granule. The thick segments of the muscle-columns 
correspond to “‘ Bowman’s sarcous elements,” and the dark 
granules in the centre of the thin segments correspond to 
“ Krause’s membrane” or the “ transverse network” (fig. 13 


=n 


Diagram A (high focus). Diagram B (low focus). 


The sarcoplasma is the part which is stained in gold prepara- 
tions. In short, Rollett regards the appearance of gold 
preparations as due to the staining of the sarcoplasma, and 
considers this to be a honeycomb of interfibrillar material, and 
not a true intra-cellular network. 

The latter portion of Rollett’s paper is devoted to an 
elaborate criticism of the network view of the structure held 
by Melland, Van Gehuchten, Carnoy, and myself. He states 
that what we describe as the fibrils of the network are only 
transverse and longitudinal sections of the walls of the honey- 
comb or sarcoplasma (‘‘ Was beide Autoren an diesen als Faden 
beschreiben sind nur Quer- oder Liingsschnitte der Winde des 
Wabenwerkes, welches das Sarcoplasma um die Muskelsaiulchen 
bildet,” p. 252). 

He describes the appearances seen in fresh muscle-fibre as 
follows :—At low focus (diagram B) the muscle-columns appear 
dark and in a line with the granules, the sarcoplasma appearing 
light. At high focus (diagram A) the sarcoplasma appears dark, 
the muscle-columns light, and two rows of granules appear in a 
line with the sarcoplasma and alternating with the muscle- 
columns. He states that the dark lines drawn by Melland 


68 O. F. MARSHALL. 


joining the dark granules are only parts of the optical longi- 
tudinal sections of the walls of the sarcoplasma and are only 
seen at high focus; and that they do not lie in a line with the 
dark granules, but alternate with them, the dark granules being 
only seen at low focus. The granules which appear at high 
focus he considers to be thickenings of the sarcoplasma. 

The above diagrams, modified from Rollett’s figures, will 
make these points clearer. Diagram B represents the appear- 
ance at low focus, and diagram A the appearance at high focus 
according to Rollett. In B the muscle-columns appear dark 
and the sarcoplasma light ; in A the sarcoplasma dark and the 
muscle columns light. 

In the same way he states that the rows of granules seen in 
gold preparations are thickenings of the sarcoplasma lying 
between the thin segments of the muscle-columns, and that the 
true row of granules which correspond to and are parts of the 
muscle-columns are only seen at low focus, and then do not lie 
in a line with the dark lines, but alternate with them. 

He describes the same appearance at low and high focus in 
hardened muscle, and states that Melland and Van Gehuchten 
place the granules alternating with the muscle-columns, whereas 
they are really in a line with them. 

In short, he concludes that a network does not exist, and 
that the appearances described by Melland, Van Gehuchten, 
and myself are due to errors in the interpretation of micro- 
scopic appearances and confusion of high and low focus (‘ Ich 
komme also zum Resultate dass ein Netzwerk im Sinne von 
Melland, Marshall, und Van Gehuchten und ein Enchyleme 
im Sinne des Letzteren in der quergestreiften Muskelfaser 
nicht existirt,” p. 262). 

In answer to this criticism I venture to make the following 
observations. 

1. If the appearance of the network in fresh muscle is due 
to the optical appearance of the “sarcoplasma” at high 
focus, there must always be a double row of granules—one on 
each side of Krause’s membrane. (I use this term for the sake 
of convenience, and take it to represent the row of dark 


OBSERVATIONS ON THE HISTOLOGY OF STRIPED MUSCLE, 69 


granules of Rollett’s muscle-columns, and the row of granules 
of the transverse network of Melland, &c.). 

2. If the granules seen at high focus in gold preparations 
are thickenings of the sarcoplasma stained by gold, there 
must again be a double row, one on each side of the true row 
of granules which are only seen on the low focus, and further, 
the former rows of granules must alternate with the latter. 
(Vide diagram A.) Although I have examined several hun- 
dred preparations of muscle-fibre, I have never seen two 
rows of granules in any gold or fresh preparation, nor have I 
seen any change in the position of the granules at high and 
low focus. The granules have always appeared to me to be 
ina line with, and connected with the longitudinal lines of, the 
network (or “ sarcoplasma”’) at both high and low focus. 

8. Rollett states that the “sarcoplasma” is a honeycomb, 
and that the appearance of the network is due to the optical 
sections of the walls of that honeycomb. If this is true, I do 
not understand why there is never any appearance of honey- 
comb structure in finely teased preparations, whereas isolated 
pieces of network are easily obtained. Moreover, one prepara- 
tion obtained by Melland is, I think, almost conclusive in 
favour of a network (fig. 15). 

4, According to Rollett, the “muscle-columns” are the 
essential parts of the fibre, and the “ sarcoplasma” is simply 
interfibrillar material ; we should therefore expect the latter 
to be least abundant in the most perfectly developed muscles. 
Now, in insects which possess the most powerful and most 
rapidly contracting muscles of all animals, the part stained by 
gold is more strongly marked than in other animals. This 
seems to point to the fact that it is the most essential part of 
the fibre and not interfibrillar material, and is therefore in 
favour of the network view. 

5. Again, in developing muscle-fibre I have shown that the 
network is present but only demonstrated with difficulty ; 
whereas if it were inert interfibrillar substance, one would expect 
it to be relatively more abundant in the embryonic fibre. 

6. If, as I haye attempted to show, the nerve-ending is 


70 CO. F. MARSHALL. 


connected with the part of the fibre stained by gold, this 
again points to the latter being the essential portion of the 
fibre ; while if Rollett’s view is correct, we should expect the 
nerve-ending to be connected with the muscle-columns. 

7. The apparent connection in some cases of the network 
with the intra-nuclear network of the muscle-corpuscles is in 
favour of the network view. 

8. Lastly, the network view places the muscle-cell on a basis 
of comparison with other cells having intra-cellular networks, 
whereas all other views of the structure are at variance with 
such a comparison. 

Dr: Klein! adopts Rollett’s view in the new edition of his 
text-book, and states that ‘“‘the reticulation described by 
Melland, Marshall, and others, is due to the coagulation of 
the sarcoplasma brought about by certain hardening reagents.” 

Dr. Michael Foster apparently holds the same view of the 
structure in the new edition of his book,’ for he states that the 
muscle-substance is composed of longitudinal fibrillze, embedded 
in interfibrillar substance which stains with gold, and hence 
appears as a network. He says, “ The interfibrillar substance 
is relatively to the fibrille more abundant in the muscles 
of some animals thanin those of others, being, for instance, 
very conspicuous in the muscles of insects, in which animals 
we should naturally expect the less differentiated material to 
be more plentiful than in the muscles of the more highly 
developed mammal.” 

Now, I think I am right in saying that the muscular system 
of insects is the most highly developed in the animal kingdom ; 
certainly the muscles are far more powerful in comparison with 
the size of the animal than in mammals, and among insects 
are found the most rapid movements. It therefore appears to 
me that the fact of the meshwork being more conspicuous in 
the muscles of insects is strongly in favour of its being the 
active part of the pike, and not of the nature of interfibrillar 
substance, 


1 «Blements of Histology,’ p. 76. 
2 *Text-book of Physiology,’ vol. i, p. 91. 


OBSERVATIONS ON THE HISTOLOGY OF STRIPED MUSCLE. 71 


I shall next give arésumé of the results of some recent 
observers who are in favour of the existence of a network. 

Van Gehuchten! has described a network in striped muscle 
similar in most respects to that described by Melland, myself, 
and others, but differing in some details. 

Carnoy? also adopts the network view, and remarks that 
“La cellule musculaire est une cellule ordinaire dont le réti- 
culum s’est regularisé, et l’enchyleme chargé de myosine.” 

Haswell® has recently published an important paper on this 
subject. His observations were made on the gizzard of 
various species of Syllis, where he claims to have found very 
primitive forms of striped muscle. He divides striped muscle 
into simple and compound types, the simple type showing 
only transverse striation, not due to network; the com- 
pound both transverse and longitudinal. These two types 
correspond to what I have termed true and false striation ; 
the characteristics of the compound or true striped muscle 
being as follows : 

1. Each fibre consists of a bundle of fibrils. 

2. Each fibre is composed of two alternating series of 
anisotropous and isotropous segments, the former of which 
are more easily stained than the latter. 

3. Running across the fibre in the middle of each iso- 
tropous segment is the transverse network, or Krause’s mem- 
brane. 

4. Between the fibrils run the strands of the longitudinal 
network. 

5. Each fibre is formed from a single cell, the nucleus of 
which divides and forms “a multinucleated protoplasmic 
body, by modification of whose protoplasm the muscle sub- 
stance and networks are formed.” 

He states that in these animals the elements of the fibre 
are on a larger scale than in Vertebrates and Arthropods, and 

1 «Etude sur la structure intime de la cellule musculaire Striée.” Extrait 
de la Revue ‘La Cellule,’ t. ii, 2 fascicul. Louvain, Gand et Liege, 1886. 


2 “La Biologie cellulaire,’ 1884. 
3 © Quart. Journ. Micr, Sci.,’? 1889. 


72 C. F. MARSHALL. 


are hence favorable for the study of the structure of mus- 
cular fibre, which ‘seems to lead to that view of the structure 
of compound striated fibres advocated by Retzius, Bremer, 
Melland, C. F. Marshall, and others; the only point of 
importance in which there seems reason for dissenting from 
that view being with reference to the relation of the trans- 
verse networks to the fibrils.” The chief differences between 
the network Haswell describes and that described by the 
above observers are these: (1) He states that the transverse 
networks are not only in the interspaces between the 
“ fibrils,” but partly also penetrate them. (2) The longitu- 
ninal networks, though mostly longitudinal, have oblique 
strands and anastomoses. (3) The transverse networks some- 
times appear as rows of spindle-shaped granular bodies, but in 
crushed specimens he says these are seen to consist of a close 
reticulum of delicate threads, the spindle bodies being con- 
densed parts of it. (4) The distance between the transverse 
networks is much greater. 

Haswell regards the “ fibrils” as the contractile part, and 
not the network. 

I shall refer to Haswell’s observations again in the subse- 
quent portions of this paper. 

A. B. Macallum! has published some interesting observa- 
tions dealing with striped muscle. He confirms the existence 
of the network of Retzius, Melland, and others, and considers 
it the contractile part of the fibre. He states that the muscle- 
nuclei are marked by furrows, sometimes transverse and some- 
times longitudinal, and that these are probably caused by pres- 
sure of the trabecule of the network. He also states that in 
some nuclei there is an intra-nuclear network similar to that 
of the fibre itself. 

It thus appears that the view of the existence of a true 
intra-cellular network in striped muscle-fibre has received 
much support, and has been confirmed by several recent 
observers ; and that the view held by Rollett and others that it 

1 “Qn the Nuclei of the Striated Muscle-fibre in Menobranchus,” ‘ Quart. 
Journ. Mier. Sci.,’ 1887. 


OBSERVATIONS ON THE HISTOLOGY OF STRIPED MUSCLE. 73 


is interfibrillar material has not sufficient evidence for it to be 
accepted. I shall therefore, in accordance with the results of 
Retzius, Bremer, Melland, Carnoy, Haswell, Macallum, and 
myself, assume that the former view represents the true 
structure of striped muscle. 

The present paper deals with the following points: 

1. The connection of the transverse networks with the 
muscle-corpuscles. 

2. The development of the network. 

3. The connection of the nerve-ending with the network. 


Connection of Network with Muscle-corpuscles. 


It was shown by Retzius! that the transverse portions of the 
muscle network were directly connected with the muscle- 
corpuscles. He states that the protoplasm of the muscle- 
corpuscle is produced into several processes from which finer 
processes arise forming the transverse networks. Retzius’ 
results were obtained by a modification of gold staining. 
The fresh muscle-fibre was first placed in a1 per cent. solution 
of formic acid for a few seconds, then in gold chloride +—4 
per cent. for twenty-five minutes, then in formic acid 1 per 
cent., and exposed to light for 10—20 hours. 

By a special method of staining I have been able to confirm 
Retzius’ results, and have made specimens showing the un- 
doubted connection of the transverse networks with the 
muscle-corpuscles. 

Method of Preparation.—The method of er I 
adopted is a modification of that employed by Mays for 
demonstrating nerve-endings in muscle. He uses the follow- 
ing solution : 


Arsenic acid 4 per cent... : : , 20 parts. 
Gold chloride a per cent ; : ‘ Ain 
Osmic acid 2 per cent. . ; ; ‘ S55 


This solution although it preserves the nerve-endings disinte- 
grates the muscie-fibre. This I found was due to the arsenic 


‘Zur Kenntniss der Quergestreifter Muskelfaser.? Biologische Unter- 
suchungen, 1881. 


74 C. F. MARSHALL. 


acid. I therefore tried various strengths of acetic and formic 
acid in place of the arsenic, and found the best combination 
was the following : 


Acetic acid 1 per cent. . ; : : 20 parts. 
Gold chloride 1 per cent. k : : Air 553 
Osmic acid 1 per cent. . : ; : Laan 


The muscle-fibre was placed in this solution for fifteen 
minutes, after previous immersion in acetic acid 1 per cent. 
for a few seconds ; then in acetic acid 1 per cent. in a warm 
chamber for one or two hours. 

1. Dytiscus.—Fresh muscle-fibre of Dytiscus stained by 
the above method shows the muscle-corpuscles in the form of 
one or more chains of nuclei in the substance of the fibre, the 
nuclei being surrounded by a small amount of undifferentiated 
protoplasm. The transverse networks are seen directly con- 
tinuous with the nuclei. This is well shown in fig. 1. Fig. 2 
shows a portion of fibre with two rows of muscle nuclei; the 
transverse networks are seen to be connected with both sets of 
nuclei in some places. Fig. 3 shows several isolated nuclei 
with the transverse processes attached to them. 

Transverse views of the network and nuclei are more diffi- 
cult to obtain. Fig. 4 shows a transverse view of an isolated 
nucleus, with part of the transverse network connected with it. 

2. Dragon-fly.—The muscle-corpuscles of the muscle- 
fibre of the dragon-fly are situated peripherally, i.e. just 
under the sarcolemma, contrary to the general rule in insects. 

In one preparation of this muscle I could trace the trans- 
verse networks into the muscle-corpuscles ; and, moreover, the 
networks appeared to be distinctly connected with the intra- 
nuclear networks of the muscle-corpuscles (fig. 5). 

3. Crayfish.—In a preparation of crayfish muscle pre- 
pared by Retzius’ method I could apparently trace the 
connection of the muscle network with that of the muscle- 
corpuscle. In this case it was somewhat difficult to tell 
whether the effect was not due to the network lying over the 
muscle-corpuscle ; but by careful focussing I think the con- 
nection could be made out (fig. 6). 


OBSERVATIONS OF THE HISTOLOGY OF STRIPED MUSCLE. 75 


These observations confirm Retzius’ results, viz. that the 
transverse portions of the muscle network are directly con- 
nected with the muscle-corpuscles ; and, furthermore, that the 

network is directly continuous with the intra-nuclear network 
of the corpuscles. 


Development of the Network. 


This I have studied in embryos of the trout and rat. 

Trout.—In some gold preparations of embryo trout, taken 
from the ova, I found developing muscle-fibres in an early 
stage. These consisted of a portion of undifferentiated proto- 
plasm containing the nucleus, and a portion already trans- 
versely striated. Under a comparatively high power the 
transversely striped portion showed darkly stained masses of 
an ellipsoidal shape arranged side by side, and causing the 
appearance of striation. Under a very high power (4 immer- 
sion) the network could be seen between these darkly stained 
masses, and in the same form in which it appears in the adult 
fibre. The dark masses appear to be some substance altered 
by the method of staining, and shrunken in the meshes of the 
network (figs. 7,8 and 9). No connection was seen between 
the network and the nucleus. 

In older fibres the network is more fully developed, but still 
no connection appears to exist between the nuclei and the 
network (fig. 10). 

Rat.—-In developing muscle-fibre from the embryo rat the 
fibres consist of an axial core of undifferentiated protoplasm 
containing the nuclei, and a peripheral part with developing 
network. Here, again, no connection was observed between 
the nuclei and the network. 

It thus appears that— 

1. The network appears at a very early stage. 

2. It develops in its permanent form, and is not produced by 
the transformation of an irregular network into the adult type. 

3. Each muscle-fibre is probably developed from a single 
cell, and is not formed by a coalescence of cells, either end to 


76 C. F. MARSHALL. 


end or laterally (as Calberla! states), because the fibres of 
trout muscle examined were evidently single cells, and had the 
network well developed. It is difficult to conceive that these 
become fitted together, either end to end or laterally, so that 
the network of one cell should exactly fit on to that of the 
next. 

4, The network develops centripetally, and commences at the 
part of the cell farthest away from the nucleus; moreover, it 
does not appear to become connected with the nuclei till the 
fibre is fully developed. 

Haswell,? from his observations on the muscle of the gizzard 
of Syllis, forms a view of the ontogeny of striped muscle which 
does not agree with that described above. 

In the same organ, in various species of Syllis, he finds in 
one case bundles of non-striped fibres ; in another, compound 
hollow striated fibres, consisting of bundles of fibrils similar to 
the above, and bound together by a single transverse network. 
In a third there are three transverse networks, and so on up to 
the fully developed type of striped muscle found in Vertebrates 
and Arthropods. He therefore regards each striped fibre as 
derived from a bundle of non-striped fibres. He thinks the 
transverse network probably the equivalent of a transverse 
line of nuclei of the unstriped fibres, which occupies a similar 
position. 

If this is correct each striped fibre must be a multicellular 
structure, and the network intercellular, and not intra-cellular, 
However, in another part of his paper he speaks of each fibre 
being formed from a single cell, the nucleus of which divides and 
forms “‘a multinucleated protoplasmic body, by modification 
of whose protoplasm the muscle-substance and networks are 
formed.” 


1 ¢ Arch. f. mik. Anat.,’ xi, 1875. 
? Loc. cit. 


OBSERVATIONS ON THE HISTOLOGY OF STRIPED MUSCLE, 77 


Connection of Nerve-ending with Muscle 
Network. 

Nerve-endings may be divided into two main types, the 
circumscribed and the diffused form. Between these types 
there are many intermediate forms, as shown by Kiihne in his 
classical work on ‘ Nerve-Endings.’! The nerve-endings in the 
muscle of the snake form a good example of the circumscribed 
variety. Here we have a localised end-plate, compact and non- 
ramifying. This is also the common type in mammalian muscle. 
In Dytiscus, on the other hand, we find the diffuse form, which 
branches and ramifies nearly to the end of the fibre. 

In each case the nerve-ending, when stained by any of the 
gold methods, is seen to consist of a slightly stained ramifying 
portion, with darkly stained irregular masses between the 
branches of the lighter stained portions. The lighter stained part 
appears to be continuous with the axis-cylinder of the nerve. 

In attempting to demonstrate a connection between the 
nerve-ending and the muscle-network there is great difficulty, 
because the methods which best show the nerve-endiugs usually 
disintegrate the muscle-fibre and destroy the network. The 
best method of obtaining the nerve-endings is that employed 
by Mays,’ and also, with modifications, by Kiihne,? viz. the mix- 
ture of arsenic acid, gold chloride, and osmic acid mentioned in 
the first part of this paper. But this method usually destroys 
the network, owing to the action of the arsenicacid. Another 
difficulty arises from the fact that the connection, if it exists, 
must take place underneath the nerve-ending, and hence in the 
normal position of the parts it could hardly be seen. 

Dytiscus.—In one preparation of Dytiscus muscle prepared 
by May’s method I succeeded in finding a portion of fibre with 
the network still intact (fig. 11). A portion of the ramifying 
nerve-ending is seen crossing the fibre transversely ; both the 

' “Untersuchungen iiber motorische Nervenendigung,” ‘Zeit. f. Biol.,’ 
Bad. xxiii. 

> «Zeit. f. Biologie,’ Bd. xx, 1884, 

3 Loe. cit. 


78 Cc. F. MARSHALL. 


fibre and nerve-ending are stretched out transversely more than 
in the normal state. The nerve-ending appears to have been 
broken off from the upper part of the network and tilted over to 
show its inner surface, which here appears to be connected with 
the longitudinal bars of the network. The upper border of the 
nerve-ending, in the position of the figure, I take to be the 
external surface next the sarcolemma, the lower border to be 
the internal surface ; the outline of the latter seems to me to 
point strongly to the fact that it is really connected with the 
longitudinal bars of the network. 

Crayfish—lIn preparations of crayfish muscle there are 
frequently found what I may, for waut of a better term, call 
“streaks” of slightly stained matter usually crossing the fibre 
transversely. In some preparations these are seen to be dis- 
tinctly continuous with the nerve-fibre going to the muscle, 
and are hence presumably portions of the nerve-ending. They 
appear to correspond to the lighter stained part of the nerve- 
endings as usually seen. 

In several specimens these streaks appeared to be connected 
with the longitudinal bars of the network, in the same way as 
in the specimen of Dytiscus muscle. Fig. 12 shows one of 
these streaks of nerve-ending connected with the network. 
The reason that these are so much stretched out appears 
to be on account of their connection with the network, which 
stretches them with it when it becomes stretched itself. In 
this figure the triangular deeply-stained bodies appear to 
represent the deeper stained part of the normal nerve-ending. 
They also appear to be continuous with the network at their 
apices. 

Although the above results are imperfect, and not so con- 
clusive as those discussed in the previous portions of this 
paper, nevertheless it appears to me that the nerve-ending is 
connected with the muscle-network, and apparently chiefly 
with the longitudinal fibrils of the network. 

A recent paper by Macallum! on the termination of nerves 
in the liver of Menobranchus has an important relation 

' «Quart. Journ. Mier. Sci.,? March, 1887. 


OBSERVATIONS ON THE HISTOLOGY OF STRIPED MUSCLE. 79 


to the above question. He states that he has traced the 
terminal fibrils of the nerves into direct connection with 
the intra-nuclear network of the hepatic cells, and the figures 
he gives seem to place this beyond doubt. This by analogy 
gives support to the view that the nerve-ending of muscle is 
connected with the network; and possibly this is the normal 
method of termination of nerves in connection with cells. 

Haswell’s! observations are also interesting in connection 
with this question. He describes a ganglion-cell at the end of 
each fibre, and occupying the axis of the fibre. This sends out 
numerous branching processes, which penetrate between the 
fibrils. He states that the core of the fibre is occupied by 
granular protoplasmic material, through which runs a network 
of fine threads, “‘ connected with the fine branches of the nerve- 
processes of the ganglion-cell, and with the network of the 
muscle substance.” Special branches of nerye-processes also 
ramify on the surface of the fibre, and probably enter into con- 
nection with the transverse networks. 

Hence both Macallum’s and Haswell’s observations lend 
support to the view that the network is connected with the 
nerve-ending. 


Conclusion. 


In reviewing the foregoing observations we are at once met 
by an apparent contradiction, viz. that in the young muscle- 
fibre the network does not appear to be connected with the 
nucleus, whereas the connection is definitely established 
(Retzius, Marshall) in the adult fibre. This at first sight is an 
absolute contradiction: for, on the one hand, it is difficult to 
conceive that processes originating as outgrowths from the 
nucleus could exactly hit off and fuse with the already formed 
network ; on the other hand, it is almost equally inconceivable 
that the fibres should grow into the nucleus. This apparent 
contradiction can, I think, be to a great extent explained. 
Firstly, we must bear in mind that histological differentiation 
proceeds during its development from without inwards, i.e. 

1 Loe. cit, 


80 C. F. MARSHALL. 


centripetally ; and that the special characters of the adult cell 
appear first and are most marked at the periphery, i. e. farthest 
from the nucleus. Examples of this rule are seen in the de- 
velopment of epidermic- cells, dentine, cartilage, and bone. In 
bone and dentine the processes are not supposed to grow from 
the cells, but to be formed by the lengthening of connecting 
strands by deposit of new matter. 

Secondly, the nerve-ending being on the surface of the fibre, 
and the network also appearing first on the surface, the con- 
nection between the two can be established from the first ; 
whereas if the network grew out from the nuclei it could only 
be connected with the nerve-ending at a much later period. 
Hence the fibre would be useless till it was far advanced in 
development, because for the network to be of any use it must 
from the first be connected with the nerve-ending. Fora new 
structure of any kind to be developed it must always have been 
of use from the first, either for its ultimate purpose or for 
some other. If the network grew out from the nuclei it 
could not be of use till it got to the nerve-ending, i.e. to the 
surface. 

In considering this point we may imagine that the cell 
divides into ‘‘ formed and unformed matter” (Beale), the 
formed matter being characteristic of the particular cell. In 
the case of the muscle-fibre the protoplasm divides into net- 
work and muscle-plasma all along, beginning at the periphery 
and gradually extending to the nucleus. We are here met by 
an apparent difficulty, for in the muscle-fibre the “formed 
matter”? which is characteristic of the cell is the muscle- 
plasma, and not the network (which is presumed to be the 
contractile part). But the special feature of striped muscle is 
not the fact that it contracts, but the mode in which this is 
brought about, i.e. the rapidity; and this will be a matter of 
nutrition which will depend on the muscle-plasma, There- 
fore the special feature of striped muscle is probably the mode 
of nutrition of its specially contractile part, i. e. quick repair. 
The active network bathed in such a fluid is placed in a good 
position for such rapidity. 


OBSERVATIONS ON THE HISTOLOGY OF STRIPED MUSCLE. 81 


Summary. 


1. The transverse portions of the network of the striped 
muscle-fibre are directly connected with the muscle-cor- 
puscles. 

2. The nerve-ending appears to be connected with the 
muscle-network, and chiefly with the longitudinal bars of the 
network. 

3. The development of the network takes place at a very 
early stage in the development of the fibre, and the network 
develops from the first in its permanent form. 

4. The network develops first at the surface, and grows cen- 
tripetally. It does not appear to be connected with the 
muscle-corpuscles till the fibre is fully developed. 

5. Each muscle-fibre appears to be developed from a single 
cell, and not by a coalescence of cells. 

The investigations described in this paper were carried on 
in the Physiological Laboratory of Owens College during the 
winter of 1887. My thanks are due to the Council of the College 
for a special grant to enable me to carry on the research. 

I must express my thanks to my brother Professor Milnes 
Marshall for many valuable suggestions in producing this 
paper, and for examining several of the preparations. I must 
also thank Professor Stirling for much assistance to me in my 
work. 


DESCRIPTION OF PLATE XI, 


Illustrating Mr. C. F. Marshall’s paper, ‘ Further Observa- 
tions on the Histology of Striped Muscle.” 


[The main details were drawn under the camera in all the figures. ] 


Fic. 1.—Muscle-fibre of Dytiscus, showing transverse networks connected 
with muscle-corpuscles. The longitudinal bars of the network are omitted 


VOL. XXXI, PART I.—NEW SER. F 


&2 C. F. MARSHALL. 


for the sake of distinctness, 1th obj. Acetic acid, osmic acid, and gold 
chloride. 

Fic. 2.—Portion of another fibre, with transverse networks connected with 
two rows of muscle-corpuscles. -;th immersion obj. Same method. 

Fic. 3.—Three isolated muscle-corpuscles of Dytiscus muscle, with trans- 
verse networks attached. jth immersion obj. Same method. 


Fic. 4.—Muscle-corpuscle of Dytiscus, with portion of transverse network 
connected with it. Transverse view. Gold preparation. ;+,th immersion obj. 


Fic. 5.—Muscle-fibre of dragon-fly, showing two muscle-corpuscles with 
their intra-nuclear network. Some of the transverse networks are seen to 
be connected with the intra-nuclear network of the upper corpuscle. 5th 
immersion obj. Acetic acid, osmic acid, and gold chloride. 


Fic. 6.—Muscle-corpuscle of crayfish, with portion of the muscle network 
apparently connected with its intra-nuclear network. 1th immersion obj. 


Fics. 7 and 8.—Developing muscle-fibres of trout, showing striation. 2. 
Nucleus. -{;th immersion obj. Acetic acid 2 per cent., a few seconds; gold 
chloride 1 per cent., fifteen minutes ; formic acid 25 per cent., thirty minutes 
in warm chamber. 


Fie. 9.—Portion of the striated part of one of the above fibres, showing 
the network and the darkly stained bodies in its meshes. »>;th immersion obj. 


Fic. 10.—More fully developed fibre of trout. 2. Nucleus. ,th im- 
mersion obj. 


Fig. 11.—Muscle-fibre of Dytiscus, with a portion of nerve-ending appa- 
rently connected with the longitudinal bars of the network. 4th imm. obj. 
Mays’ method. 


Fie. 12.—Muscle of crayfish, showing a “streak” of nerve-ending appa- 
rently connected with the longitudinal bars of the network. 35th imm. obj. 
Retzius’ method. 


Fies. 13 and 14.—Diagrams comparing the view of Rollett and others 
with the network. In Fig, 13 the structure, according to Rollett, is marked 
in full lines, and the network marked in dotted lines. In Fig. 14 the thick 
segments of Rollett’s muscle-columns are shown by dotted lines in the 
meshes of the network. x. Network. s. Muscle-columns. g. Granules. 


Fic. 15.—Portion of muscle-fibre of Dytiscus, showing network very 
plainly. One of the transverse networks is split off, and some of the longi- 
tudinal bars are shown broken off. (Copied from Melland, loc. cit., fig. 6.) 


[Note.—In figure 13 the lines connecting the thick parts of Rollett’s 
muscle-columns should be much thicker. They represent the thin segments 
of the columns. ] 


ON CHATOBRANCHUS. 85 


On Chetobranchus, a New Genus of 
Oligochetous Chetopoda. 


By 


Alfred Gibbs Bourne, D.Sc.Lond,, F.L.S., C.M.Z.S., 
Fellow of University College, London, and of the Madras University. 


With Plate XII. 


Habitat.—I discovered this very remarkable worm in the 
mud from a pond (Anglo-Indian “tank’’) in Madras town. 
The mud had been placed in a bottle and allowed to stand, so 
that the Naids for which I was searching might come to the 
top. On examination I found projecting from the surface of 
the mud, among numerous individuals of Nais and Dero, 
several specimens of this worm. These at once attracted my 
attention on account of the branchial processes, which could 
be seen with the naked eye. The mud is of a very finely 
divided character and is brown in colour, and contains very 
little animal and hardly any vegetable life, and very little or- 
ganic débris. I obtained a large quantity of it, and by 
the use of muslin sieves secured numerous specimens of 
Chetobranchus. 

The worm does not secrete any glutinous material, and so 
make itself a mud tube; but, as I have been enabled to ascer- 
tain by keeping them in a small aquarium nearly filled with 
the mud, makes for itself long tracks in the mud—“ burrows,” 
they might be called ; and each worm appears to reside in its 
own “burrow,” at times projecting from the surface of the 
mud into the water, and at other times withdrawing itself com- 
pletely into its “burrow.” The mud is of such a character 


84. ALFRED GIBBS BOURNE. 


that unless disturbed in some way the “ burrow” remains as a 
permanent structure. The mud can be dried in cakes, and 
when such a cake is broken across, the “ burrows” are as 
obvious as they are in a lump of earth in which earthworms 
have lived. 

Some of these mud-living Oligochzta secrete a glutinous 
substance, so that when they are removed and the loose par- 
ticles washed away a tube of mud remains, while others secrete 
no such substance; and when these latter are removed and 
the loose mud washed away the worm remains quite unpro- 
tected. Chetobranchus belongs to the latter category. 

External Characters—Branchial Processes—Sete. 
—The worms vary in size, but when stretched out and crawling 
on a slide an average sized individual is about 1} inches to 2 
inches in length, and about =, inch in breadth (fig. 9), and 
consists of about 130 segments. The anterior extremity is a 
little thinner than the rest of the worm, and the body wall 
in this region is slightly pigmented; the pigment is to a cer- 
tain extent arranged in transverse bands on the dorsal surface, 
each band corresponding to a segment. There is no proboscis, 
and the eye-spots are absent. 

The most remarkable and striking feature of the worm is the 
presence of dorso-laterally placed processes, of which there is 
a pair to each of the anterior segments, commencing with the 
second! segment. It is difficult to say how many pairs of these 
processes exist, as, after the ten to twelve most anteriorly 
placed pairs, they gradually diminish in size until they become 
mere warts on the surface of the worm, and in the posterior 
segments are entirely absent (fig. 1). I have counted about 
sixty to seventy pairs. In most of the individuals I have exa- 
mined the five or six most anteriorly placed processes are a 
little shorter than those immediately following, but there is no 


1 T have assumed that the first setigerous segment is the second segment 
of the body, the first segment of the body being the buccal segment with the 
prostomium. The nomenclature of the segments in earthworms is usually 
based upon this assumption, and it would be more convenient if a similar 
practice were always adopted with regard to other Oligocheta, 


ON CHAU TOBRANCHUS. 85 


regularity about this, and I am inclined to believe that these 
get more or less injured in the “ burrowing.” The length of 
the processes relatively to the size of the body may be judged 
of from an examination of fig. 1. 

These processes are obviously branchial in function. The 
structure of one of these branchial processes is shown in fig. 2. 
Each is virtually a hollow prolongation of the body wall; the 
coelomic corpuscles may occasionally be seen pushed out into 
it. The epidermis is bounded externally by a distinctly visible 
cuticle, through which project very fine cilia—so fine that they 
might easily escape notice but for the commotion which 
they create among particles of mud in the water. At the 
extremity are a few stiff processes, doubtless sensory in 
function. 

Into each of the longer processes (about the first fifty) runs 
a loop of the lateral vessel (see below, circulatory system). 

Entirely contained within each process are all, in the case of 
the more anterior, or some in that of the more posterior, pro- 
cesses of the setz belonging to the dorsal bundle; so that in 
the anterior portion (about thirty segments) of the worm no 
dorsal setz project freely outside the body wall, while in the 
region immediately following (about thirty segments) two at 
least of the dorsal setz do not freely project, while one seta 
of the dorsal bundle does so project, and in the remaining 
portion of the body all the setz project. 

There are no muscular structures in the branchial pro- 
cesses, which are kept fairly rigid, and are moved by the 
dorsal setze, and thus serve the worm as locomotor organs. 

The seta bundles are placed in four rows—two ventral rows 
and the two dorsal rows mentioned above. The dorsal seta 
bundles commence in the same segment as do the ventral seta 
bundles, i.e. the second body segment. Two kinds of sete 
occur in the dorsal bundles: the one kind is the straight capil- 
lary seta shown in fig. 5; they vary in length; in the anterior 
segments they are very long. The seta drawn in fig. 5, if in- 
tended for one of the longest, should have been drawn double 
the length it is, to be on the same scale as the sete drawn in 


86 ALFRED GIBBS BOURNE. 


figs. 6 and 7a. The relation of these long capillary sete to 
the branchial processes is described above. The other kind of 
seta occurring in the dosal bundle is always of the same length, 
and has a curved sickle-shaped free extremity (fig. 6). Such 
sete do not occur in the more anterior bundles, and in passing 
backwards one comes across intermediate conditions between a 
short straight capillary seta and the sickle-shaped form figured. 
As arule, there are two or three straight and two or three 
curved sete in each bundle. 

The ventral sete are all “ crotched-shaped ;” in the most 
anterior segments the free extremity has the shape drawn in 
fig. 76; but this, in going backwards, soon passes into the 
shape shown in fig. 7a. Each of these ventral sete has a little 
swelling placed rather nearer to the free extremity than to the 
root. ‘There are four to six sete in each ventral bundle. All 
the sete diminish rapidly in size as one approaches the poste- 
rior extremity, which presents therefore, as in Naids, the 
appearance of being a region of continued growth. 

Viscera.—The alimentary canal presents no special feature 
of interest ; there is no enlargement corresponding to the so- 
called gizzard of many Naids. : 

The celom is, as in most Oligocheta, incompletely divided 
into a series of chambers by diaphragms placed between the 
segments. 

The coelomic corpuscles are rounded (fig. 8), and like those 
of Naids. They contain numerous olive-green granules, which 
look like droplets of fatty matter. They may be seen passing 
from segment to segment, and at times into the branchial 
processes. 

The circulatory system consists of a dorsal and ventral 
vessel and a series of lateral vessels, a pair in each segment, 
which run from the dorsal to the ventral vessel, and in those 
segments provided with well-developed branchial processes 
loop out into the process (fig. 2). 

The walls of the dorsal vessel are much pigmented, as in 
many Oligochzta, while those of the ventral vessel, and, as a 
rule, the lateral loops, are unpigmented. 


ON CHATOBRANCHUS. 87 


The nephridia are not very clearly seen, but are undoubtedly 
present, a pair in each of the segments. In the middle region 
of the body they have exactly the same appearance as have the 
nephridia of Naids. 

Asexual Reproduction.—Although I have examined a 
very large number of individuals I have found a few specimens 
only in the act of asexual reproduction, and this process 
appears to be more like one of simple fission, as opposed to 
gemmation, than it is in Naids and Chetogaster. In these 
latter forms a new region, a “ budding zone,” is produced 
between two existing segments ; this divides into two portions: 
the anterior portion forms the new tail of the anterior daughter 
zooid, and the posterior portion forms a head for the posterior 
daughter zooid, while the previously existing segments! of the 
parent zooid undergo little or no change. 

In Chetobranchus I cannot find any “ budding zone.” In 
the specimen which is drawn in fig. 10, for instance, I counted 
over 200 segments. The most anterior sixty segments bear 
recognisable branchial processes, which, as usual, get smaller 
as one passes backwards ; then there are about thirty segments 
which bear no trace of processes; then about forty-five seg- 
ments on which branchial processes can be counted, the anterior 
ones almost as large as those usually found in the head region, 
and the posterior ones becoming so small as to be unrecognis- 
able; behind the last process-bearing segment I counted about 
sixty-five segments, the posterior ones very much crowded 
together, as though active growth were taking place in this 
region. 

This individual would doubtless soon have divided into two 
zooids, the posterior one of which must form a head, consisting, 
at any rate, of a buccal segment and a prostomium. The re- 
markable feature in this process is the new growth which occurs - 
in connection with so many segments of the parent zooid, viz. 
the development of the branchial processes in all those seg- 
ments which bear them, and which subsequently form part of 
the posterior daughter zooid. 

1 Semper, ‘Arb. Zool. Zoot. Institut, Wurzburg,’ Bd. iv, 1877-8. 


88 ALFRED GIBBS BOURNE. 


I have found several individuals preparing in this way for 
fission, but never found any trace of a “ budding zone.” At 
some other time of year I shall, I hope, be able to make further 
observations with regard to this process, and also with regard 
to the generative organs. I have hitherto found no trace of 
generative organs. This is unfortunate, as it leaves the 
systematic position of the worm still open to some doubt. 

Systematic Position.—This is without doubt the worm 
referred to by Semper as occurring along with Dero philip- 
pinensis.! In many details of its structure, as well as 
in the fact that it exhibits fissiparous reproduction, Cheto- 
branchus resembles the Naidomorpha, and represents, I be- 
lieve, a family closely allied to the Naidomorpha (Vejdov- 
sky) and the Cheetogastridz (Vejdovsky). The most remarkable 
feature in its structure is, of course, the possession of 
branchial processes, and these processes are themselves re- 
markable in completely surrounding the whole or a portion 
of the dorsal seta bundle. With the exception of the con- 
tained sete these branchial processes closely resemble in 
structure the long branchial processes found in the anal region 
of some species of the genus Dero. 

The only other Oligochete which possesses branchial pro- 
cesses is Alma nilotica; but, judging from Vejdovsky’s 
remarks anent this form, it is very different from Cheto- 
branchus. 

Generic Description.—Chetobranchus, g. n.—Capil- 
lary sete present. Each of the anterior segments, from 
the second segment backwards, bears a pair of dorso-laterally 
placed branchial processes, which entirely include some or all 
of the sete in the dorsal bundle. Dorsal setae commence in 
the same segment as do the ventral sete, i.e. the second 
segment. 

Chetobranchus Semperi, sp. u.—Fresh stagnant water, 
Madras town. 

I have dedicated the species to Professor Semper, as the 


' *Arb, Zool. Zoot. Institut, Wurzburg,’ Bd. iv. 


ON CHETOBRANCHUS. 89 


worm which he discovered in the Philippine Islands is un- 
doubtedly either the same or a closely allied species. 

It is impossible, in dealing with a single species such as 
this, to define the specific characters ; the absence of eye-spots, 
the character of the sete, the length of the branchial processes, 
the nature of the pigmentation, will probably serve, among 
other characters, to mark the species. 


DESCRIPTION OF PLATE XII, 


Illustrating Professor A. G. Bourne’s paper “On Cheto- 
branchus Semperi.” 


Fic. 1.—Lateral view of Chetobranchus; the branchial processes and 
dorsal and ventral sete of one side only are drawn. pr. Prostomium. m. 
Mouth. az. Anus. 


Fic. 2.—A single branchial process. s. Dorsal sete. a. Afferent blood- 
vessel. ¢. Efferent blood-vessel. 


Fic. 3.—A portion of the dorsal blood-vessel. 7. A pair of lateral vessels. 
Fic. 4.—A portion of the ventral blood-vessel. 7. A pair of lateral vessels. 
Fie. 5.—A capillary seta from a dorsal seta bundle. 


Fie. 6.—A seta from a dorsal seta bundle, with a sickle-shaped free 
extremity. 


Fic. 7.—a. A seta from a ventral seta bundle in the posterior region of 


the body. 4. The free extremity of a similar seta from the anterior region of 
the body. 


Tic. 8.—A ccelomic corpuscle. 
Fic. 9.—Dorsal view of Chetobranchus. Natural size. 


Fie. 10.—A “budding” individual of Chetobranchus. About twice 
the natural size. 


J Z 
Mitr DOW 


F Huth, Lith® Edint 


RANVIER S CONSTRICTIONS IN THE SPINAL CORD. 91 


The Presence of Ranvier’s Constrictions in the 
Spinal Cord of Vertebrates. 


By 


Dr. William Townsend Porter, 
of St. Louis. 


With Plate XII, dis. 


No question in the controversy over Ranvier’s constrictions 
is more interesting than that of distribution. Torneaux and 
Le Goff! declared in 1875 that constrictions are present in 
the spinal cord. Their discovery remained almost unnoticed 
until Schiefferdecker,? not aware that he had been anticipated 
by the French observers, offered silver pictures as proof that 
the constrictions are not confined to the peripheral nerves. 
The statements of these authors have been contested. Boll? 
wrote against Torneaux and Le Goff; and Kolliker,* seeking 
to test the results of Schiefferdecker, worked over the same 
ground, but arrived at opposite results,-and brought new argu- 
ments in favour of the old opinion. 

It is of importance to have this question examined by many 
observers, for the presence of constrictions in the spinal cord 


1 “Note sur les étranglements des tubes nerveux de la moelle épiniére,” 
‘ Journal de |’Anatomie,’ 1875. 

2 “Beitrage zur Kenntniss des Baues der Nervenfasern,” ‘ Arch. f. mikr, 
Anat.,’ Bd. xxx, 1887. 

3 “Ueber Zersetzungsbilder der markhaltigen Nervenfasern,”’ ‘Arch, f. 
Anat. u. Entwickelungsges.,’ 1877. 

4 ¢Handbuch des Gewebelehre des Menschen,’ Leipzig, 1889, Bd. i, p. 
154. 


92 WILLIAM TOWNSEND PORTER. 


must influence our conceptions of the form and the purpose 
of these structures in the peripheral fibres. 

In order to prove that constrictions are present in the 
central nervous system, it is necessary to demonstrate them in 
silver and in osmium preparations, in isolated fibres and in 
sections. 

Silver Preparations—teased.—Small pieces of the 
white sustance of the spinal cord treated with silver-osmium 
solution give these results: the colour ranges from pale grey 
to deep black, according to the size of the piece, the strength 
and quantity of the osmic acid, the duration of its action, the 
distance of the fibre from the surface of the piece, and the 
thickness of the myelin. Beneath the outer darker parts of 
the fragment used is a portion whose pale grey colour and soft 
consistence show that the osmic acid has affected it but slightly. 
It is here that the silver crosses are seen at their best, for the 
silver solution readily penetrates the entire piece, and its deep 
brown impregnations contrast strongly with their grey sur- 
roundings. The teasing should not be too thorough; the 
nerves are easily broken at the constricted points, and even in 
the most careful preparations many will be found with the 
silvered constriction torn in two. In the grey fibre-groups are 
numerous axis-cylinders cross-striped with Frommann’s lines, 
and sticking out in all directions from the edges of the group. 
Small granular silver masses, of a great variety of forms, may 
at first be taken for constriction stains; some of these are 
probably the remains of broken crosses, and others are pre- 
cipitations between the fibres. The medullary sheath is poorly 
marked, often quite unrecognisable, and again only indicated 
here and there by faint lines, more or less parallel to the axis- 
cylinder. Well-shaped crosses are hard to find, but crosses 
whose sharpness of outline is somewhat blurred by the silver 
granules lying about them are common. A sceptical observer 
will sooner or later be convinced by finding an unmistakable 
cross in a fibre whose medullary sheath is still recognisable, and 
whose isolated position guards against deception. The lines of 
Frommann have in my preparations been seen to best advan- 


RANVIER’S CONSTRICTIONS IN THE SPINAL CORD. 93 


tage in the frog and the guinea-pig; but the crosses are larger 
and clearer in the spinal cord of the ox. 

Osmium Preparations—teased.—Difficulties surround 
the demonstration of constrictions in teased preparations of 
central fibres treated with osmic acid alone. Only a small 
number of the fibres are properly “ fixed.” The absence of 
Schwann’s sheath makes isolation difficult, and from the same 
cause by far the greater number of the constrictions are torn 
by the needles. We have no reason to suppose that the rule 
according to which the interval between any two constrictions 
is approximately directly proportional to the thickness of the 
fibre is any less true of central than of peripheral fibres. It is 
the larger central fibres which are most easily isolated, and 
their constrictions, if this rule holds good, must be widely 
separated. Inthe animals used by me it seemed impossible to 
isolate pieces long enough to show two constrictions. Only 
when the locus minore resistentiz has escaped the teasing 
needles and the fibre is torn through the myelin can constric- 
tions be found. If the teasing is not too thoroughly done, 
small groups of fibres, lying for the most part parallel, will be 
secured. It is near the edges of these groups that the con- 
strictions are most likely to be found, because the outermost 
fibres protect the rest. In both rabbit and ox I have seen 
entirely isolated fibres with good constrictions. 

Osmic Acid Preparations; Sections.—Longitudinal 
sections of central fibres are best made from the spinal cord of 
the ox. The part used by me was taken from the neighbour- 
hood of the median fissure. For the detection of the con- 
strictions sections of the average thickness of the fibres are 
most suitable. The constrictions are not so easily seen when 
the nerves are cut in the plane of the axis-cylinder. Only the 
outer fibres of a section should be used, because they are 
usually found lying with more space between each fibre than 
is the case elsewhere, and are better “ fixed.” 

Many of the interruptions in the continuity of the nerves 
are of considerable width. They are so broad that one at once 
thinks of artificial separations or of places where fibres have 


94 WILLIAM TOWNSEND PORTER. 


been cut through as they rose above or sank below the plane 
of section. Examination with a high power shows that this 
is true of only a small number, and that the myelin at many 
of the other interruptions is plainly neither cut nor torn. The 
myelin has here the shape of a cylinder, ending in a truncated 
cone. When the fibre is seen a little obliquely, the observer 
looks into the end of this cone and sees that out from it comes 
a pale, ribbon-like or cylindrical structure, which runs a more 
or less wavy course until it enters another myelin sheath. It 
has the look of a constriction elongated by the process of pre- 
paration. A little search shows that the space which separates 
two medullary sheaths is often not greater than the breadth of 
the fibre, and in such cases appears large because the fibres 
are very broad. In these narrower interruptions the axis- 
cylinder is usually well shown. In a good section are inter- 
ruptions not wider than half the breadth of the fibre. On 
either side is the cone-shaped ending of the medullary sheath ; 
between the two, looking very small and colourless in contrast 
with the thick myelin, is the axis-cylinder—straight, of uni- 
form calibre, and often faintly striped in a longitudinal direc- 
tion. All the structures lie in parallel planes. The fibre is 
neither cut nor torn. It is a true constriction, and closely 
resembles, except that the sheath of Schwann is absent, the 
Key and Retzius! drawings of peripheral constrictions in 
man. 

With a strong immersion system it is possible to see that 
most of the constrictions are crossed about the middle by a 
very fine line, which seems to lie closely upon the axis-cylinder, 
and sometimes appears as an indistinct ring. From the 
margins of this ring I have sometimes seen a line of equal 
fineness, passing on either side towards the medullary sheath. 
These delicate structures are just within the limits of visi- 
bility. They correspond to the lines found in silvered central 
fibres, and are probably portions of the ensheathing neu- 
roglia. No sheath of Schwann is present. Long fusiform 


1 «Studien in der Anatomie des Nervensystems und des Bindegewebes,’ 
Stockholm, 1876, ii, Bd. i, plate vii, figs. 11, 18, 14, 15. 


RANVIER’S CONSTRICTIONS IN THE SPINAL CORD. 95 


nuclei sometimes lie against the myelin, but I have found no 
proof that these do not belong to the neuroglia. The myelin 
was not broken by the incisures of Schmidt and Lantermann. 

The value of these results depends of course on the value 
of the method by which they were obtained. Well-founded 
objections have been brought against nitrate of silver. This 
reagent may be reduced in all parts of the nerve-fibre, and 
even between the individual fibres! Combinations of various 
forms of impregnations give a great variety of pictures. In 
teased preparations, for reasons previously stated, the crosses 
are not often found isolated. These facts make deception 
easy, but do not lessen the force of the truth that unmistak- 
able crosses are proof of the presence of Ranvier’s constric- 
tions. It may be said that the constrictions found in osmium 
preparations were artificially produced. Naturally, the only 
evidence which can be offered here is the testimony of accu- 
rate drawings. 

The demonstration of the presence of constrictions in cen- 
tral fibres helps directly in the settlement of some vexed 
questions and is of indirect value with respect to many others. 

Ranvier® advanced the opinion in his original memoir that 
the myelin is impervious to crystalloids, and that nitrate of 
silver enters the nerve at the constriction. This, he said, goes 
to show that uutritive fluids take the same route; and his 
explanation is accepted by most physiologists. There is, as 
Boveri insisted, no reason for supposing that the nutrition of 
medullated central fibres is different from that of medullated 
peripheral fibres. Boveri,> not finding the constrictions in 
the spinal cord, rejected Ranvier’s hypothesis and substituted 
his own, which was that the constrictions served a purely 
mechanical end, permitting great freedom of motion, like a 


1 T can confirm Kolliker’s statement that stripings similar to Frommann’s 
lines occur in silver preparations of small blood-vessels. Boll (I. ¢., p. 310) 
saw cross-striping in elastic fibres treated with nitrate of silver. 

2 “Recherches sur l’Histologie et la Physiologie des Nerfs,’ ‘Arch. de 
Physiologie,’ iv, Mars, 1872, No. 2. 

3% “ Beitrage zur Kenntniss der Nervenfasern,” ‘Abhandl. der math. physik. 
classe d. k. Bayer. Akad. der Wissensch.,’ Bd. xv, 1886, 


96 WILLIAM TOWNSEND PORTER. 


jointed chain, and removing the danger of injury from the 
sudden bendings to which many peripheral nerves are ex- 
posed. This can be accepted as one of the uses of peri- 
pheral constrictions, but the objection to Ranvier’s explana- 
tion falls to the ground with the confirmation of Torneaux and 
Le Goff’s discovery, and the constrictions may be looked upon 
as a food-way to the axis-cylinder. 

Constrictions are present in the spinal cord; the sheath of 
Schwann is absent from the spinal cord; therefore the sheath 
of Schwann is not concerned in the formation of constrictions 
in central fibres, and is probably not an essential part of peri- 
pheral constrictions. We can safely reject the dictum of 
Hans Schultze :1 The sheath of Schwann is “ die formgebende 
Ursache der Ranvier’schen Markunterbrechungen,” an idea 
* prominent in many researches, and may look with suspicion on 
theories that find in Schwann’s sheath an explanation of the 
structure of the constrictions. : 

Further inferences are very tempting. The central nerve- 
fibres are epiblastic, and lie in an epiblastic framework 
(Geriist); no mesoblastic tissue ensheathes them (Gierke) ; 
in Palemon squillathe medullated nerves have no connective- 
tissue sheath ; therefore, the myelin of central fibres, like the 
axis-cylinder, is of epiblastic origin, and the medullary sheaths 
of peripheral nerves are probably formed from the axis- 
cylinder. If the myelin is formed from the protoplasm of 
the axis-cylinder, then the function of the sheath of Schwann 
is that of a simple connective-tissue envelope. The many 
theories of development and degeneration that affirm a generic 
relation between the sheath of Schwann and the medullary 
cylinder are incorrect; spinal nerves are developed and regene- 
rated through their cells of origin, and changes in the nuclei of 
Schwann’s sheath are not the cause of the growth or decay of 
the medullary substance. But the evidence in our possession 
does not warrant our going so far, for, although Gierke® de- 

1 * Axen-cylinder und Ganglionzelle,’ p. 27. 


2 i.e. Axis-cylinder and medullary sheath. 
3 « Die Stiitzsubstanz des central Nervensystems,’‘ Arch. f. mikr, Anat.,’ 


RANVIER’S CONSTRICTIONS IN THE SPINAL CORD. 97 


clares that no connective tissue surrounds the individual nerve- 
fibres in the spinal cord, yet it is by no means settled that the 
medullary sheath of central fibres is not mesoblastic. Blood- 
vessels enter the cord at a very early date in the life of the 
embryo,’ and it is not certain that the mesoblastic tissue 
which forms and surrounds them is limited, as Gierke thinks, 
to the blood-vessels themselves. Joseph’s? assertion that a con- 
tinuous network penetrates and binds together axis-cylinder 
and medullary sheath in the peripheral nerves—a fact of much 
importance if true—has been denied by Retzius,? who found 
the network only in the axis-cylinder. The discovery of 
Retzius* that Palemon squilla has medullated nerves which 
lack Schwann’s sheath and are provided with oval nuclei 
lying between the medullary sheath and the axis-cylinder has 
not yet been confirmed. There are finally many who still 
deny the correctness of the views of His* regarding the for- 
mation of the central nervous system. 

Such reflections are therefore merely suggestive, and are of 
use only as they emphasize the fact that a solution of many 
weighty problems is to be found through a sufficient explana- 
tion of the origin of myelin. 

I take this opportunity to gratefully acknowledge the kind- 
ness of Professor Flemming, under whose direction this work 
was done at the Anatomical Institute in Kiel. 

xxv, 1885, p. 533 :—* Alle faserigen Elemente der Stiitzsubstanze Fortsatze 
von Gliazellen sind; andere Faden, als elastische oder Bindegewebsfibrillen, 
sind durchaus zwischen den Nervenfasern nicht zu finden,” 

1 His found blood-vessels in the cord between the fourth and fifth week. 
See “Zur Geschichte des Menschlichen Riickenmarkes und der Nerven- 
wurzeln,” ‘Abhandl. der math. phys. Classe der Kgl. Sachs. Gesellsch. der 
Wissensch.,’ 1886, xiii, No. 6. 

2 “Ueber einige Bestandtheile der peripherischen markhaltigen Nerven- 
faser,” ‘Sitzungsber. d. Berlin, Akad.,’ Bd. ii, 1888, p. 1281. 

3 «Der Bau des Axencylinders der Nervenfasern,”’ ‘Verhandlung des 
Biol. Vereins in Stockholm,’ Bd. i, Jan., 1889, No. 4. 

4 “ Ueber myelinhaltige Nervenfasern bei Evertebraten,” ‘ Verhandl. d. Biol. 
Ver.,’ Stockholm, Bd. i, Dec., 1888, No. 3. 

° « Every nerve-fibre is a process from its own nerve-cell ; this is its genetic, 
nutritive, and functional centre.” His, ]. ¢. p. 513. 


VOL. XXXI, PART I.—NEW SER. G 


98 WILLIAM TOWNSEND PORTER, 


DESCRIPTION OF PLATE XII, dis, 


Illustrating Dr. William Townsend Porter’s paper, “ The 
Presence of Ranvier’s Constrictions in the Spinal Cord of 
Vertebrates.” 


Fics. 1, 2, 3, 4, 5.—White substance near median fissure. Ox. Osmic 
acid 2 per cent., nitrate of silver 1 per cent., each 1 part, two hours. Dilute 
solution of caustic potash, about five minutes. Teased in glycerine diluted 
about one third with distilled water. Leitz, ;, ocular 3, draw-tube out. 

Fig. 6.—Spinal Cord. Guinea-pig. Osmium-silver mixture as above, 
three hours. Teased in diluted glycerine. Leitz, 3, oc. 3, d. t. out. 

Fie. 7.—Medulla oblongata. Rabbit. Osmium-silver mixture, two hours. 
Picro-carmine, twenty-four hours. Teased in diluted glycerine. Leitz, 3, 
oc. 1. Axis-cylinder shows three silver cross-stripes at constriction; the 

‘lines between the ends of the medullary cylinders represent the light reflex. 

Fic. 8.—White substance near median fissure. Spinal cord, Ox. Osmic 
acid 2 per cent., twelve hours. Celloidine, cloves, balsam. Leitz, 3;, oc. 3, 
d. t. out. The section was cut the average thickness of the fibres. 


EXPERIMENTAL IMITATION OF PROTOPLASMIC MOVEMENT’. 99 


Professor Butschli’s Experimental Imitation of 
Protoplasmic Movement. 


Proressor Burscuur, of Heidelberg, has recently made 
some extremely interesting observations upon a substance 
which simulates ina remarkable way the appearance and move- 
ments of the protoplasm of an Ameeba, or of the plasmodium 
of Mycetozoa. He has been kind enough to send to me some 
oil in a suitable condition for use, with directions as to the 
exact details of the experiment. In my laboratory, by fol- 
lowing his directions, the movements described by him have 
been observed in a satisfactory manner. In order to obtain 
the best results some experience and care is requisite, and 
probably they cannot always be obtained by a single experi- 
ment. The subject is so interesting, and so fitted for further 
investigation by all who have leisure and a taste for the 
study of the vital phenomena of the Protozoa and of living 
protoplasm in general, that I think it will be of advantage to 
readers of this Journal to have Professor Biitschli’s directions, 
which he has permitted me to publish, placed in their hands. 

E. Ray Lanxesrer, 
March, 1890. 


HEIDELBERG, February lst, 1890. 


You have kindly asked me how I prepare the protoplasma- 
like drops which I have described. As you yourself feel 
greatly interested in this discovery, and presumably a like 
interest exists among other English biologists and micro- 
scopists, I hasten to satisfy your desire, and to explain some- 
what more fully the methods which I have described in a 
previous publication. 


100 PROFESSOR BUTSCHLI’S 


As you well know already, I use in the preparation of these 
globules—showing protoplasma-like streaming—ordinary olive 
oil. My first experiments were made with a small quantity of 
olive oil which had been standing for along time in my labora- 
tory in a small bottle. By some happy chance this oil had 
just the right properties which are necessary for the success of 
the experiment, for not every sort of olive oil is suitable. As 
far as my experience goes, it tends to show that the ordinary 
oil cannot be directly used, because it is too thin, or is perhaps 
deficient in other qualities on which the success of the experi- 
ment depends. In order, therefore, to prepare a suitable oil, I 
proceed in the following manner :—A medium-sized watch-glass 
or flat dish is filled with a thin layer of common olive oil, and 
is placed on a water-bath or in a small cupboard, such as are 
used for embedding in paraffine, at a temperature of about 
50° C. Under the influence of the higher temperature the oil 
gradually loses its yellow colour and becomes thicker. The 
great point now is to select the right moment at which the oil 
will have attained the proper degree of thickness and viscosity, 
as also the other properties which at present I am not able to 
define more exactly, but on which much of the success seems 
to depend. The exact moment can, however, only be found 
out by systematic trials. After the oil has been thickening for 
three or four days a trial should be made with a drop of 
it in the manner described below. Should the drop not 
become finely vesiculate, and exhibit little or no streaming, 
continue the heating process and experiment again on the 
following day. If the oil should have become too thick it will 
form good frothy drops, but will scarcely show any streaming. 
In this case mix it with a small quantity of ordinary olive oil, 
and thus render it more liquid. If it has become much too 
thick it will form a good froth, but the latter dissolves very 
rapidly in glycerine. 

You see thus that the process to obtain the suitable oil is 
somewhat slow, but I do not at present know of any other 
method by which the result can be arrived at more quickly 
and surely. 


EXPERIMENTAL IMITATION OF PROTOPLASMIC MOVEMENT, 101 


To prepare the vesiculate drops I proceed in the following 
way :—In a small agate mortar I grind a small quantity of 
pure dry carbonate of potash (K,CO;) to a fine powder. I 
then breathe on to the salt till it becomes slightly moist, and 
with a glass rod add to it a drop of oil, mixing the two con- 
stituents to a thickish paste. The success of the experiment 
depends, however, more upon the nature of the oil than upon 
the proportions of oil and salt in this mixture. Then witha 
glass rod or a needle I place a few drops of the paste, about 
the size of a pin’s head or smaller, on a cover-glass, the corners 
of which are supported by small pegs of soft paraffine. I then 
place on a slide a drop of water, and put the cover-glass over 
this in such a manner that the drops of the paste are immersed 
in the water, but are not much compressed, to which end 
the corners of the cover-glass have been supported by the 
paraffine. The preparation is then placed in a damp chamber, 
and remains there about twenty-four hours. The drops have 
now a milk-white and opaque appearance. The preparation 
is then well washed out with water by applying blotting- 
paper to one edge of the cover-glass, and supplying water 
at the other edge from a capillary tube. 

If the drops have turned out well, they will begin almost 
immediately after this to move about rapidly, and change their 
shape continuously. The water under the cover-glass must 
now be displaced by glycerine, diluted with an equal bulk of 
water, and the drops will then exhibit a vigorous streaming 
and forward movement, becoming gradually quite transparent. 
The amceboid movements are generally more distinct if the 
drops are somewhat compressed. If the drops do not show 
the streaming movement you may succeed in producing it by 
tapping the cover-glass slightly, by applying gentle pressure, 
or sometimes by breaking up the drops. For it seems as if 
at times incrustations were formed on the surface of the drops, 
which prevent or impede the streaming movement, and which 
can, in part at least, be removed by the above-mentioned 
manipulations. 

It is especially interesting to see how fast and beautifully the 


102 PROFESSOR BUTSCHLI’S 


drops creep to and fro in water, or in half-diluted glycerine, 
even when they are not compressed. The streaming movement, 
on the other hand, is better seen if the drops are somewhat 
compressed, which may be done by inserting under the cover- 
glass a piece of a broken cover-glass of medium thickness, and 
then removing the paraffine pegs. Then draw away the liquid 
until the necessary pressure is obtained. This streaming 
movement is best demonstrated twenty-four hours after the 
addition of the glycerine, as the drops will then be thoroughly 
cleared and transparent. Further, it is interesting to note 
that a progression of the drops takes place in the direction in 
which the streaming moves. 

As this forward movement is rather slow in compressed 
drops, it is necessary to use a micrometer ocular to satisfy one- 
self of the advance. 

Unfortunately the oils which I have prepared since my first 
experiments do not move and stream so well or so rapidly as 
those I employed then. The movement and streaming show 
themselves much more markedly and distinctly if they are 
examined on a warmed stage at a temperature of 50°C. If 
you should be in a position at your demonstrations to conduct 
the experiment at this temperature, the phenomena will cer- 
tainly be much more evident. 

From the preceding description you will see that it will 
be necessary, to obtain good results, to gradually get hold of 
the methods, and you must not doubt the correctness of the 
phenomena which I have described if the first trials do not 
give the desired results. 

At all events, you will have at first to make some experi- 
ments so as to obtain an insight into the conditions and sort 
of phenomena, but I do not doubt that you will succeed in 
observing the appearances and in demonstrating them to others, 
though perhaps in not so vigorous a degree as I might 
desire. 

I have lately made some trials to render olive oil 
suitable for these experiments by heating it more rapidly. 
Although at present I have no entirely reliable results, it 


EXPERIMENTAL IMITATION OF PROTOPLASMIC MOVEMENT. 103 


seems to me that by heating ordinary olive oil to 80°—90° C. 
for twelve or twenty-four hours, a suitable medium may be 
obtained. 

Finally, I would like to remark that I am the last person to 
defend the view that these drops, exhibiting protoplasma-like 
movements, are directly comparable to protoplasm. Composed 
as they are of oil, their substance is entirely different from 
protoplasm. ‘They may be, however, compared with the latter, 
in my opinion, firstly with regard to their structure, and 
secondly with regard to their movements. But as the latter 
depend on the former, we may assume that the amceboid 
movement of protoplasm itself depends on a corresponding 
physical constitution. 

These drops, too, resemble organisms inasmuch as they 
continue for days to exhibit movements, due to internal 
causes, which depend on their chemical and physical structure. 
I do not believe that up to this time any substance has 
been artificially prepared which in these two points, viz. 
structure and movement, has so much resemblance to the most 
simple form of life as have these vesiculate drops. I hope, 
therefore, that my discovery will be a first step towards 
approaching the problem of life from the chemico-physical 
side, and towards passing from vague and general hypotheses 
of molecular constitution to the surer ground of concrete 
conceptions of a physical and chemical nature. 

It is, however, a special satisfaction to me to hear that in 
your country, which has given rise to so many aad so cele- 
brated men in biological science, my investigations are followed 
with interest and sympathy. 

With friendly greetings, 
Yours sincerely, 
O. ButscH11. 


4 


7’ di 


a e =a 
‘ Pcs is 
res 
\ A" c 
. \ 4 = i 
Se . a 8 


i | - : , ; itt 
j 
vib 
~ i 4 , 
“ * rf 
*\e : 
it vies. — Pest ‘ ) Jo Tie 
: ® ® - Lee 
(sits 
- = é P ‘2 
ee: } the eer Oana t 7. 
(or). af TLOSMsRUS (AeNee ts Ores an? eng 


ae ae 2 ayes: bas lit ePr ol hea ~ aie 
bers be OE Ot BO nove PRUSL ae + , ‘. les 


sluts ae te, Uke Most oD Nie is oo fe hoor a 
whe ; ’ be _ r - ‘~ ‘i 

Ai OOK 4 a é Pali E : ‘ wy ‘ 
: ‘i te na ** ig ‘ iS Merraptie? 


, ; ; - 
s7TOGE ie 


2 


A 8) i ast ® - 
5.1) if ae 

j jim i ; , ‘ i Ae e ‘ 

> oo, eet 
; P Tye bos ae Ls ; i | wigs? 7 Mid4t Wr ene 
 ¢ = « 
. as 
a P 3 ’ i - « Sot iti tes “Tae ~) 
* - 


voll Pate, siutittaatal >it 2) Oi cmaene ae Co 


ait Lo dai 
bab 


aA 


wy tie OL the 


THE EMBRYOLOGY OF A SCORPION. 105 


The Embryology of a Scorpion (Euscorpius 
italicus). 


By 


Malcolm Laurie, B.Sc., 
Falconer Fellow of Edinburgh University. 


With Plates XIJI—XVIII. 


Srtyce 1870 there has been no detailed work on the de- 
velopment of the Scorpion. As it seemed likely that with 
modern methods of section-cutting and the great advance 
which has been made of late years in the field of embryology, 
a renewed examination might yield interesting results, I 
have, at Professor Lankester’s suggestion, examined and cut 
sections of a large number of embryos of Euscorpius italicus 
preserved for him by the Zoological Station at Naples. I have 
also examined a number of embryos of Scorpio (Buthus) 
fulvipes preserved and sent over from Madras by Professor 
Bourne. These, however, chiefly owing to the small amount 
of food-yolk, show such a great difference from E. italicus 
in their mode of development that it seems better to postpone 
the description of them to a future paper. 

The Scorpion is interesting not only as being the lowest, and, 
as far as we know, the oldest type of air-breathing Arachnid, 
but also as being exceptional among Arthropods in that the 
whole development takes place within the body of the female— 
in the ovarian tubes. The only other instances of this with 
which I am acquainted are Phrynus, which is also viviparous, 

VOL, XXXI, PART II,—NEW SER. H 


106 MALCOLM LAURIE. 


and Spherogyna ventricosa, one of the Acarina in which 
the young are born sexually mature. 

I may fitly here express my thanks to Professor Ray Lan- 
kester not only for the suggestion that I should work at this 
interesting subject, and for the generous way in which he has 
provided me with material, but even more for his continual 
and invaluable assistance and advice while the work has been 
in progress. 


HistoricaAL INTRODUCTION. 


Johannes Muller! gave a short description, with five or six 
figures, of the development of Buthus. Owing to its brevity 
and the absence of any attempt to ascertain the internal 
arrangement, his paper is of little value except from an histori- 
cal point of view. 

Duvernoy” gives also only a few figures of Buthus and 
of another form, probably Euscorpius. He describes at some 
length a cord (baguette) which he says passes from the appendix 
of the follicle in Buthus to the mouth of the embryo, and which 
Miller had compared to an umbilical cord. I hope to be able 
in a future paper to give a detailed account of this and other 
curious points in the development of Buthus. The chief value 
of Duvernoy’s paper was that he reconciled the contradictory 
descriptions of the ovary which had been given by Miiller and 
Rathke.? While doing this he makes a rather serious mistake 
in describing the ovum of Buthus as occupying the whele of 
the diverticulum of the ovarian tube, instead of only a small 
space at the top. 

The next writer on this subject is Léon Dufour,* who gives 

1 Joh. Miiller, “ Beit. z. Anat. des Skorpions,’”’ * Meckel’s Arch. f. Anat. 
u. Phys.,’ Bd. xiii, 1828. 

* Duvernoy, “ Fragments sur les organes de la génération de divers Ani- 
maux,”’ ‘Mém. de l’Acad. des Sci. de l'Institut,’ t. xxiii. 

3 Rathke, “Zur Morphologie,” ‘ Reisbemerkungen aus Taurien,’ 1837, 
Riga, 4to. 

4 Dufour, “ Hist. Anat. et Phys. de Scorpions,” Mém. Presentés a |’Acad. 
des Sciences,’ t. xiv, 1856. 


THE EMBRYOLOGY OF A SCORPION. 107 


an elaborate description with numerous figures of the anatomy 
of the adult. His description of the embryo is, however, very 
brief and his figures unsatisfactory. 

Elias Metschnikoff! is the only writer who has treated of the 
development of the Scorpion with any degree of fulness. He 
gives a detailed account of the whole development, and his 
paper, which deals chiefly with the surface views and optical 
sections, contains a large amount of accurate and laborious 
observation. It is the classic on this subject, and up to 1886 
no attempt was made to add to it or supersede it. 

In 1886 Kowalevsky and Schulgin® published a short account 
of the development of Androctonus ornatus. Unfortu- 
nately their paper has no figures, which detracts much from its 
value. I find reason to differ from them on a few points, but 
it is quite possible that this may be due to our having worked 
on different genera. 

The only other paper on this subject which I am acquainted 
with is by G. H. Parker,? who treats at some length of the 
development of the central and lateral eyes. I had worked at 
this point before the appearance of his paper, and on the whole 
agree with his conclusions. These are briefly that the lateral 
eyes are monostichous, being formed from the hypodermis 
without invagination. The median eyes, on the other hand, 
are formed by invagination, and are therefore three-layered, 
all the layers being derived from the hypodermis. The retina 
is the second layer, the third being reduced to a post-retinal 
membrane. The material at Mr. Parker’s disposal did not 
enable him to go back to the commencement of the formation 
of the central eyes and their connection with the cerebral 
invaginations. 


1 Metschnikoff, ‘‘ Embryologie des Skorpions,” ‘ Zeit. f. wiss. Zool.,’ 1870. 

2 Kowalevsky and Schulgin, ‘‘ Entwickelungsgeschichte des Skorpions,”’ 
‘ Biol. Centralblatt,’ Bd. vi, 1886. 

3 Parker, “The Eyes in Scorpions,” ‘Bull. Mus. Comp. Zool., Harvard,’ 
vol. xiii. 


108 MALCOLM LAURIE. 


THe Ovary AND Ovarian Eee. 


The ovary consists, as is well known, of three longitudinal 
tubes connected by transverse anastomoses, so as to form eight 
quadrilateral meshes. The oviducts arise from the lateral 
angles of the two foremost meshes and run forward to open 
together on the genital operculum. The ovary appears to be 
embedded in the liver, the chief mass of which lies dorsal to 
it; this is not really the case, for, though lobes of the liver 
pass through the meshes of the ovarian network they do not 
unite on its ventral side. Both the longitudinal and transverse 
tubes bear ova, which project from their outer surface as 
oval bodies each attached by a short pedicle and measuring 
when ripe about 1:2 mm. in length and *83 mm. in breadth. 
Ova in all stages of development are present on the ovarian 
tubes at the same time, and there are in addition the corpora 
lutea (v. p. 111). 

The microscopic structure of the ovarian tubes is shown in 
Pl. XIII, fig. 2. They are there seen to be formed of two 
layers surrounding an irregular lumen. ‘The outer layer, o. ., 
which is purely skeletal in function, consists of irregularly 
polygonal cells, with circular nuclei and strongly marked cell 
outlines. The contents of these cells are highly refractive. 
Towards the inside of this layer the cells become flattened 
so as to form a distinct, cellular, limiting layer. The inner 
layer, which surrounds the lumen of the tube, is formed of 
very long and thin columnar cells, with oval nuclei and clear, 
faintly granular protoplasm. The nucleiare for the most part 
confined to a central zone, leaving a large part of the outer 
and a smaller part of the inner ends of the cells clear. It is 
from this inner layer of cells that the ova and their follicles 
are formed ; it is, in fact, the germinal epithelium. 

The first sign of the formation of an ovum is that one of 
the cells of the inner layer of the ovarian tube begins to 
increase in size (fig. 1, ov.). It contains finely granular 
protoplasm, a large and distinct oval nucleus, and a darkly 


THE EMBRYOLOGY OF A SCORPION. 109 


staining nucleolus. There is at first no sign of its presence 
on the outside of the ovarian tube. As it increases in size, 
however, it pushes its way, at the head of a column of cells, 
towards the outside. The outer layer of the ovarian tube 
becomes very thin, but remains as a membrane containing 
few, if any, nuclei (fig. 2, fol’.). By the time the ovum is 
about ‘04 mm. in length (fig. 2) it has passed completely 
through the outer layer and is visible as a small protuberance 
on thesurface of the ovary. It remains connected to the inner 
layer of the tube by a column of cells which is somewhat 
expanded over the base of the ovum. The nucleus has not 
increased in size in proportion to the growth of the cell. 

The nuclei of the cells of the column which connects the 
ovum to the inner layer of the ovarian tube next arrange 
themselves so as to leave a clear space down the centre of the 
column (Pl. XIII, fig. 3, mz.). They also grow round the ovum 
so as to form a follicle (fig. 3, fol.) one cell thick. The 
cells of this follicle rapidly become flattened and their nuclei 
become smaller. The cells which remain clustered at the base 
of the ovum (fig. 3, ger’.) on the other hand increase in size, 
and shortly after the stage represented in fig. 3, which is a 
drawing of an ovum of about +1 mm. in diameter, they begin 
to secrete the yolk of which the greater part of the ripe egg is 
composed. The outer layer of the ovarian tube can still 
be traced as a thin and apparently structureless membrane 
(fig. 3, fol’.) surrounding the egg outside the cellular follicle. 
The nucleus has increased in size and is now a distinct oval 
body with a round, granular nucleolus. 

In fig. 4. is shown a longitudinal section of an egg of about 
‘4 mm. in length and ‘28 mm.in breadth. A considerable 
quantity of yolk is now present in the form of spheres 
ranging in size from mere granules up to as much as ‘025 mm. 
These spheres are clear, homogeneous, sharply defined bodies 
showing no internal structure except that there is, in the 
larger ones, a collection of granules at one point near 
the outside. Round the nucleus the yolk-spheres are small, 
and round the margin of the egg the protoplasm is coarsely 


110 MALCOLM LAURIE. 


granular, the rest of the space being filled up with the larger 
spheres. 

The nucleus (fig. 4, 2.), which retains its central position, 
is large (‘05 mm.) but indistinct in outline and is probably 
breaking down as I have been unable to find any trace of 
it in eggs larger than that here figured. The nucleolus 
(fig. 4, v’.), which is situated towards one side of the nucleus, is 
also large, staining darkly with carmine and showing a very 
distinct circular outline. It often contains one large, clear, 
circular vesicle and a number of smaller ones. 

The whole egg is surrounded by a distinct, rather thick 
vitelline membrane (fig. 4, v. m.). No trace of pores or any 
other structure was made out. Outside the vitelline membrane 
the egg is surrounded, except at the base, by the follicle in 
which the two layers (fol. and fol’.).of the ovarian tube can 
still be traced. The cells of the inner layer of the follicle are 
now flattened and small. The large yolk-forming cells at the 
base of the egg (ger”.) have increased in size and arranged 
themselves in a circle the centre of which is occupied by a 
prolongation of the ovarian tube (mz.). ‘The egg is only 
separated from this prolongation of the lumen by the vitelline 
membrane. The spermatozoa are thus enabled to reach and 
fertilize the egg while it is still in its follicle. 

Pl. XIII, fig. 5, shows the base of a ripe egg attached to the 
ovarian tube. The pedicle has become shortened and its lumen 
has increased very much in size. The yolk-forming cells have 
degenerated, their flattened nuclei (ger’.) being, however, still 
distinguishable, and the follicle has become much thinner owing 
to the growth of the egg. The egg itself is a mass of tightly 
compressed yolk-spheres, among which I have in vain sought 
for the nucleus. It is probable, however, that the nucleus and 
the greater part of the protoplasm migrate to the base of the 
ege as segmentation commences there. 

The yolk (Pl. XIII, fig. 6) consists of spheres, ranging up to 
‘2 mm. in diameter. They are not homogeneous, but contain 
spherical or prismatic bodies, which stain darkly with borax 
carmine, These bodies are very large in the smaller yolk- 


THE EMBRYOLOGY OF A SCORPION. 11] 


spheres, which contain one, two, or more of them. In the 
larger spheres they are much more numerous and much 
smaller. Many of the spheres show round holes as if the 
darkly staining bodies had dropped out. It may be, however, 
that these cavities contained a fatty or oily substance, which 
has been dissolved out in the course of embedding and 
mounting. 

The only structures remaining to be described in connection 
with the ovary are the corpora lutea mentioned above (fig. 7). 
These are irregularly shaped bodies of about ‘12 mm. in 
diameter, showing a slight tendency to radiate structure, and 
containing a considerable number of nuclei, which are scattered 
about without any definite arrangement. They project from 
the surface of the ovarian tubes, and are evidently the collapsed 
remains of the follicles after the egg has passed out. I was 
confirmed in my idea that these were corpora lutea by their 
resemblance to the structures described by v. Siebold! in the 
ovary of Apus. They differ from these latter, however, in not 
containing fluid. 


First Periop.—Formation of Blastoderm. 


The egg is fertilized in the follicle, from which it does not 
begin to pass out until the end of this period. It then passes 
into the ovarian tube in which it undergoes the rest of its 
development, the young when born being exactly like the 
parent in form. Kowalevsky and Schulgin’ state that the egg 
in Androctonus is not fertilized until it has entirely left the 
follicle, and passed into the ovarian tube, or, as he calls it, 
uterus. I can hardly believe this to be the case, but it is quite 
possible that it leaves the follicle at an earlier stage in 
Androctonus than in Euscorpius. 

Stage A.—I have not, unfortunately, been able to observe 
the processes of fertilization and the formation of the first 
segmentation-spheres. I should think it probable that the 

1 vy, Siebold, ‘Beitrage zur Parthenogenesis der Arthropoden,’ Leipzig, 


187 1s.px 19) 
= Loe, cit., p. 526, 


112 MALCOLM LAURIE. 


greater part of the protoplasm with the nucleus collects at the 
base of the egg. The youngest stage in my possession is shown 
in surface viewin Pl. XIV, fig. 8, and in sectioninfig.9. The blas- 
toderm forms a circular patch about ‘2 mm. in diameter, lying on 
the surface of the yolk at the end of the egg nearest to the micro- 
pyle, and consists of about twenty large cells, those in the centre 
measuring about ‘(03 mm. in diameter. In section (Pl. XIV, 
fig. 9) it is seen to be a single layer, the cells of which are 
about ‘023 mm. thick in the centre. Round the margin the 
cells are wedge-shaped so that the blastoderm les flush with 
the surface of the yolk. The cell-contents are coarsely granu- 
lar, rather more so towards the lower side. The nuclei are 
large, round and granular with distinct outlines. 

The yolk-spheres under the blastoderm appear to be breaking 
down. The blastoderm and yolk are closely surrounded by 
the structureless vitelline membrane (v. m.). This stage seems 
to be a little younger than that figured in Metschnikoff’s paper 
in Pl. XW, fig. 6. 

Stage B.—In the next stage (Pl. XIV, fig. 10) the blastoderm 
is somewhat larger, measuring ‘*23 mm. in diameter. The 
blastoderm is now almost twice as thick (‘045 mm.). Some of 
the cells are columnar, and occupy the whole depth of the 
blastoderm, but the majority have divided in a plane parallel to 
the surface, so that it is in places two or even three cells deep. 
The nuclei vary in shape, those in the columnar cells being 
oval. 

Stage C.—In the next stage (Pl. XIV, fig. 11) the blastoderm, 
now ‘3 mm. in diameter, is formed of an irregular mass of cells 
showing as yet no trace of arrangement into layers. The cells 
are comparatively small with well-marked outlines and large 
nuclei. Round the margin of the blastoderm the cells form a 
single layer on the surface of the yolk, but in the centre the 
blastoderm is five or six cells thick, and the cells push their 
way in between the yolk-spheres to which some of the cells 
attach themselves. These cells, which attach themselves to 
yolk-spheres, lose their definite outline and take, as far as I 
have been able to ascertain, no part in the further growth of 


THE EMBRYOLOGY OF A SCORPION. rs 


the embryo. There is no doubt that these yolk-cells are 
derived from the blastoderm in this and the next stages, and 
do not arise in the yolk by any process of free cell-formation. 
Kowalevsky is also of this opinion. The yolk in the Scorpion’s 
egg shows no sign of segmenting as does that of the Spider. The 
yolk of the Spider’s egg seems! to represent the hypoblast, and 
takes an active part in the building up of the embryo; that of 
the Scorpion, on the other hand, remains throughout develop- 
ment an inert mass of food-material. This fundamental 
difference in the segmentation makes any comparison of the 
early stages of these two groups impossible, and would 
seem to point to an independent origin for their abundance of 
food-material. If the segmentation in Scorpions is a modifi- 
cation of the centrolecithal type, as would seem probable from 
the modes of segmentation in other groups of the Arachnida, 
it is a very extreme one, and almost all trace of its origin has 
been lost. 


Srconp Prertop.—-Formation of the Three Layers and the Em- 
bryonic Membranes. 


Stage D.—It is difficult to get good sections at this stage as 
the blastoderm is often humped up at the end of the egg and 
compressed by the ovarian tube into which it is beginning to 
pass. In one, and only one, series of sections I have seen what 
appeared to be a longitudinal groove in the blastoderm. This 
primitive groove is figured by Metschnikoff (PI. XVII, figs. 2 
and 3), but he may have been misled by the edges of the serous 
membrane which is growing up and might easily give the 
appearance of a groove in surface view. If the primitive groove 
exists, which I am inclined to doubt, as the appearance in my 
sections may have been due to shrinking, it is a very temporary 
structure. Towards the posterior end of the blastoderm the 
cells are proliferating and forming what I shall call the primi- 
tive thickening. From this primitive thickening is formed the 
mass of hypoblast which is found later on in the tail-segment. 


1 Locy, “ Observations on the Development of Agelina nevia,” ‘ Bull. 
Mus,, Harvard,’ vol. xii, 


114 MALCOLM LAURIE. 


It would seem to represent a modified invagination, and is com- 
parable to the primitive streak in the chick. I was at first in- 
clined to call this the primitive cumulus, but considering the 
fundamental differences between Scorpions and Spiders, and also 
the-fact that, while Balfour! places what he calls the primitive 
cumulus at the posterior end of the embyro, Locy? gives the 
same name to a thickening at the anterior end, it seemed better 
to avoid a term which might suggest erroneous homologies. 

A layer of cells (fig. 12, pr. hy.) is seen to be forming under 
the rest of the blastoderm, though not yet extending to its 
edges. This is well marked in the next stage, and forms the 
greater part of the primitive hypoblast or hypomesoblast. It 
would seem to be simply split off from the epiblast. I have 
seen no appearance of a “‘ down-sinking ” of cells to form the 
hypoblast, such as is described by Kowalevsky and Schulgin ;$ 
but, without the help of figures, it is not easy to be certain 
of their exact meaning. Whether this “down-sinking” is 
supposed to take place over the whole blastoderm or only 
at the primitive thickening is not clear from their descrip- 
tion. 

Round the edges of the blastoderm a single layer of large 
cells (fig. 12, s. m’.) is seen to be spreading a little way over the 
surface of the yolk. These peripheral cells, which are at present 
continuous with the epiblast, form later on the continuation of 
the serous membrane. ‘This serous membrane, or outer layer 
of the amnion, is seen growing up as a single layer of cells 
from the edges of the blastoderm (Pl. XIV, fig. 12, s. m.). 
It spreads over the surface of the blastoderm from all sides, 
and its edges ultimately meet and fuse in the middle line. At 
this stage the edges have not yet come together, and the cells 
of the layer are still small and similar in appearance to those 
of the rest of the blastoderm. 

The yolk is broken down to a considerable extent, and the 


1 Balfour, “Notes on the Development of the Araneina,” ‘ Quart. Journ. 
Mier. Sci.,’ vol. xx, 1880. 

2 Locy, loc. cit. 

3 Loc. cit., p. 526. 


THE EMBRYOLOGY OF A SCORPION. 115 


cells in it (fig. 12, y. ¢.) are numerous. Their nuclei are very 
large and granular, and of irregular shapes. The cell-outlines 
have entirely vanished, the cells being swollen up by an enor- 
mous quantity of yolk-stuff. According to Kowalevsky and 
Schulgin these cells are capable of amceboid movements. Cells 
continue to be added from the under surface of the blastoderm 
to those already in the yolk up to the end of thisstage. Their 
function—of breaking down the yolk—is carried on at a later 
period by the hypoblast. 

Stage E.—In the next stage the blastoderm (Pl. XIV, 
fig. 13) has assumed an oval form, the thickened part or 
ventral plate measuring *35 mm. in length and :25 mm. in 
breadth, though the peripheral cells extend some way beyond 
this. I have not been able, either in surface view or section, to 
find any trace of the primitive groove, and imagine that, if ever 
present, it has filled up. The primitive thickening (fig. 14, pr. t.) 
is better developed than in the last stage, and the single layer 
of primitive hypoblast (figs. 14 and 15, pr. hy.) is now quite 
definite and extends a little way beyond the thick part of the 
blastoderm, and forms a layer (hy’.) of cells under the peri- 
pheral cells. These last (s. m’.) extend a good deal further than 
in the last stage. The serous membrane (s. m.) is now com- 
pleted over the surface of the ventral plate. 

Stage F.—In the next stage the embryo, of which fig. 16 
shows a longitudinal section, consists of two somites—those 
which will afterwards bear the chelicerze and chelee—in addition 
to the head- and tail-segments. The head- and tail-segments 
are large, and a third somite is beginning to be formed from 
the tail. The first somite is smaller than the second, and not 
as yet very distinctly marked off from the head. It does not 
become fully separated from the head until a much later stage 
(eight somites). Except for this curious delay in the forma- 
tion of the first, all the somites are formed and separated in 
regular succession from the tail-segment. 

The epiblast has undergone little change since the last stage, 
except that it is somewhat thinner between the somites than in 
them. It is beginning to grow up at the edges over the surface 


116 MALCOLM LAURIE. 


of the ventral plate as a single layer of flat cells to form the 
inner embryonic membrane—the amnion proper (fig. 16, am.). 
This amnion never loses its connection with the epiblast as 
the serous membrane has now done, but remains attached 
to its edges and only extends round the egg as the epiblast _ 
extends. 

The most important change in this stage is the formation of 
the mesoblast (mes.). This layer is formed under the whole 
ventral plate by a multiplication of the cells of the primitive 
hypoblast, from which it is in places not yet distinguishable. 
The mesoblast extends across the whole ventral plate from 
side to side, and is much thicker in the somites than between 
them. 

The serous membrane (s. m.) has, as mentioned above, now 
lost all connection with the blastoderm, and is continued round 
about two thirds of the egg by the “ peripheral cells,’’ which 
are now beginning to separate from the egg and form a definite 
membrane. The cells of the serous membrane are becoming 
large and flat. 

The hypoblast extends a little way beyond the ventral plate, 
forming a single layer of cells (Ay.) in the periphery of the 
yolk immediately under the serous membrane. 

By the time the embryo has reached a stage with three somites 
completely formed (Pl. XIV, fig. 17) most of the changes which 
were going on in the last stage are completed. The amnion has 
entirely closed over the embryo (fig. 18, am.), though its cells 
have not yet attained their characteristic form. The mesoblast 
(mes.) is entirely separated from the hypoblast, and remains 
henceforth a distinct and independent layer. The hypoblast 
(hy.) is now a single layer, extending under the whole ventral 
plate, except in the tail-segment, where it consists of a spheri- 
cal mass. This hypoblastic mass in the tail-segment is the 
direct product of the primitive thickening. The hypoblast 
extends somewhat further round the egg than the other 
layers, as is diagrammatically shown in fig. 19. 

The description given above of the mode of formation of the 
serous membrane and amnion differs very considerably from 


THE EMBRYOLOGY OF A SCORPION. 117 


that of Kowalevsky and Schulgin. They describe it as a fold, the 
outer layer of which forms the serous membrane while the 
inner forms the amnion. This is probably the more primitive 
mode of origin for these structures, and the mode described 
above for E. italicus is probably derived from it either by a 
hastening of the formation of the serous membrane or a retarda- 
tion of that of the amnion. Iam unable to confirm their state- 
ment that mesoderm cells are present between the two layers. 


THirp Prrtop.—Up to the Formation of Nine Somites. 


This period covers the rest of the time before the appendages 
begin to form. The egg has by this time entirely passed into 
the ovarian tube. It has also increased considerably in size, 
but I am unable to say whether this is due in any degree to 
absorption of fluid or whether it is entirely due to internal 
changes. 

Stage G.—In the first stage belonging to this period which 
I have examined (Pl. XV, fig. 20) the embryo consists of 
nine somites. The first of these—that which will bear the 
cheliceree, is much smaller than the others, and is seen in section 
to be not yet fully separated from the head. The second 
somite, which will bear the chele, is larger than those following 
it. The next four are the ambulatory, and the seventh will 
bear the genital operculum. A slight groove (nz. g.) runs 
down the middle line of the body; this is chiefly due to 
the mesoblast having divided into two longitudinal bands 
(figs. 21 and 22, mes.). 

The epiblast is moderately thick in the somites, and is 
beginning to grow as a single layer round the rest of the egg 
(fig. 21, ep’.), carrying the amnion with it. By this stage it 
has extended almost as far ashas the hypoblast. The cells in the 
middle line show a more definite arrangement than the rest of 
the epiblast. This is preparatory to the formation of the 
neural groove. The cells of the amnion (am.) have developed 
their characteristic nuclei—spindle-shaped in section—and 
form a well-marked thin membrane lying close over the embryo. 

The mesoblast (figs. 21 and 22, mes.) shows most impor- 


118 MALCOLM LAURIE. 


tant changes. As mentioned above, it has now separated into 
two longitudinal bands. This separation does not extend into 
the tail-segment (fig. 23, mes.), where the mesoblast remains 
as a solid mass of cells somewhat thinner in the middle line. 
The coelomic spaces are now formed by a splitting of the meso- 
blast in the somites. They are best seen in the posterior 
somites (fig. 21, cw.), where the mesoblast is thin and forms 
only a single layer on each side of the celomicspace. Further 
forward (fig. 22) the mesoblast is thicker and the celomic 
space is not so well marked. 

The hypoblast has undergone very little change. It is still 
visible in the tail as a solid mass (fig. 23, hy.m.), and spreads 
under the ventral plate and a little way beyond its margin as 
a single layer (figs. 21—23, hy.). The cells of this single layer 
have large oval nuclei which stain less darkly than those of 
the epi- and meso-blast. These nuclei are somewhat widely 
separated from each other, and the cells seem to contain a 
considerable amount of food-stuff. 

The serous membrane (figs. 21—23, s. m.) is by this time quite 
separate from the egg all round. - It has attained its final 
structure, the nuclei being enormously large (‘05 mm.), flat, 
and at a considerable distance from each other. As far as my 
observations go I can confirm Blochmann’s statement? that the 
nuclei of the serous membrane divide directly without forming 
any karyokinetic figures. As the serous membrane plays a 
purely passive part in the future development it will not be 
necessary to refer to it again. 

Stage H.—In the next stage (Pl. XV, fig. 24), which is 
the last before the formation of the appendages, the embryo 
consists of nine somites. The first is very much smaller than 
the others, while on the second, which is the largest, a trace 
of the appendages is just visible. The first six somites are 
clearly distinguished from those further back, owing to their 
sloping backwards and outwards, while the posterior ones are 
at right angles to the axis of the embryo. 


1 “Ueber direkte Kerntheilung in der Embryonalhiille der Skorpione,”’ 
‘Morph. Jahrb.,’ vol. x. 


THE EMBRYOLOGY OF A SCORPION. 119 


A distinct groove, the neural groove (n. g.), runs down the 
middle line and extends some distance into the head-segment. 
It is due to a thinning of the epiblast in the middle line 
(figs. 25 and 26, n.g.). The ventral nervous system is 
formed by a thickening of the epiblast along each side of 
this groove. 

The epiblast now spreads as a single layer beyond the 
hypoblast (ep’.) and extends over nearly half the yolk, carrying 
the amnion with it. This is diagrammatically shown in fig. 27. 
In the head-segment (fig. 25) the epiblast is irregularly 
grooved and thickened. This is the commencement of the 
formation of the cerebral ganglion. In the thoracic somites 
(fig. 26) the epiblast is very thick and solid at the corners (ap.) 
where the appendages are about to appear. It is also some- 
what solid just at each side of the neural groove (n. th.). 
This is the commencement of the thickening which will form 
the ventral nervous system. 

The mesoblast is a thin layer in the head-segment (fig. 24, 
mes.), but shows the ccelomic space (c@.) distinctly. This 
development of a head ccoelom does not, of course, as Balfour 
has pointed out, necessarily indicate that the head-segment is 
equivalent to a body somite. In the body somites (fig. 
26) the mesoblast is pretty thick and the ceelomic space is 
almost entirely closed up. The mesoblast does not extend 
across the middle line or beyond the limits of the ventral 
plate. 

The hypoblast (figs. 25, 26, hy.) shows no change from the 
last stage but remains as a single layer, except in the tail- 
segment, where the hypoblastic mass is distinctly visible. 

As the next stage shows the commencement of a large 
number of new structures, the ventral nervous system, the 
appendages, &c., it seems advisable to give a short summary 
of what has taken place so far. 


First Period. 


(1) The blastoderm commences as a single saucer-shaped 
layer of cells at one end of the egg (Stage A). 


120 MALCOLM LAURIE. 
(2) These multiply and form a thick mass (Stages B, C). 


Second Period. 

(3) The serous membrane grows up from the edges of the 
blastoderm over its surface as a single layer of cells, and is 
continued round the yolk by the peripheral cells (Stages D—F). 

(4) The hypo-mesoblast is formed partly as a single layer of 
cells split off from the under surface of the blastoderm and 
partly, at the tail end, as a thick mass, the primitive thickening, 
which probably represents an invagination. Before and up 
to this stage cells pass from the blastoderm into the yolk 
(Stage D). 

(5) The mesoblast is formed as a layer several cells thick, 
extending right across the blastoderm. The hypoblast remains, 
after the formation of the mesoblast, as a single layer, except 
in the region of the primitive thickening, where it is a spherical 
mass (Stage E). 

(6) The amnion is formed as a single layer of cells growing 
up from the edges of the epiblast, with which it retains its 
connection. ‘The serous membrane has by this time lost all 
connection with the blastoderm, and spreads round the greater 
part of the yolk (Stage F). 

The embryo by this time consists of three somites and the 
large head- and tail-segments. The somites are formed from 
the tail in regular succession. 


Third Period. 

(7) The mesoblast divides into two longitudinal bands, and 
celomic spaces are formed in the somites and in the head 
(Stage G). 

(8) The epiblast and amnion begin to spread round the egg 
beyond the limits of the ventral plate (Stage G). 

(9) The neural groove is formed by a thinning of the epiblast 
in the middle line (Stage H). 

(10) The epiblast in the head-segment begins to thicken to 
form the cerebral nervous system (Stage H). 


THE EMBRYOLOGY OF A SCORPION. 121 


Fourtu Preriop.—From the Formation of the Appendages to the 
Hatching of the Embryo. 


Stage I.—The first stage of this third period shows—as 
mentioned above—the commencement of some of the most im- 
portant structures. The embryo, of which a surface view is 
given in Pl. XV, fig. 28, now consists of twelve somites in 
addition to the head- and tail-segments. These somites are 
no longer separate thickenings as in the last stage, but have 
grown close up to one another, and are marked off by narrow 
grooves. ‘The epiblast extends as a single layer all round the 
egg. The longitudinal neural groove is well marked and 
extends the whole length of the body with the exception of the 
tail-segment. 

The first six somites bear appendages, i.e. the chelicere, 
chelz, and four pairs of walking legs. These appendages are 
simple outgrowths, and are, with the exception of the first two 
pairs, of approximately equal size. The chelicerz are much 
smaller, and the chele somewhat larger than the other appen- 
dages. The appendages are an outpushing of the epiblast and 
the outer layer of mesoblast or somatopleure (Pl. XVI, fig. 31). 
They are hollow, the spaces being prolongations of the coelomic 
pouches. There is at this stage no sign of appendages on the 
somites behind those bearing the walking legs. 

The embryo has a strong dorsal flexure so that the cephalic 
segment curves round the end of the egg. This is best seen in 
longitudinal section (Pl. XVI, fig. 29). The anterior margin 
of the cephalic segment is deeply cleft in the middle line, the 
segment being thus divided into two lobes. The lobes are in 
much the same state as in the last stage, and show no signs of 
the cerebral invagination from which a greater part of the brain 
is formed. In the middle line, and a very short way behind 
the bottom of the cleft, is a circular raised area with a pit in 
its centre (Pl. XV, fig. 28, st.). This pit is the stomodeum. 
It is seen in section in Pl. XVI, fig. 29, and is a simple inpush- 
ing of the epiblast. 

VOL. XXXI, PART II.—NEW SER. I 


122 MALCOLM LAURIE. 


The ventral nervous system consists of a pair of thickened 
bands of epiblast running the whole length of the body on each 
side of the neural groove (Pl. XV, fig. 28). The bands are 
cut up into blocks by the grooves which separate the somites. 
The epiblast is not evenly thickened, but the nuclei are ar- 
ranged so as to present a wavy outline. This is characteristic 
of the formation of nerve-tissue in this animal, and was well 
seen in the cerebral lobes in the last stage (Pl. XV, fig. 25). 
The small ganglia of the cheliceral somite are well seen at 
this stage (fig. 28, g. I). 

The tail-segment, from which the six caudal somites have yet 
to be formed, has begun to be pushed out (Pl. XVI, fig. 29). 
The epiblast in this region is very thick, and the cavity of the 
outpushing is lined by a thick layer of hypoblast, which is the 
*“hypoblastic mass ” of earlier stages (fig. 29, hy. m.). 

Besides this mass in the tail-segment the hypoblast extends 
as a single layer round the whole egg (Pl. XVI, fig. 29, hy.). 
Along the ventral side the cells of this layer are close together, 
but towards the sides and back they become more scattered, 
and are to a great extent involved in the yolk. It is from the 
mass in the tail-segment that the mesenteron is chiefly formed. 
The hypoblast along the ventral surface also takes some part in 
its formation, but that round the sides and back is not involved, 
though it aids in the formation of the great digestive gland or 
liver. 

The mesoblastic bands (P]. XVI, fig. 31) are not yet united 
across the middle line. The celomic spaces (Pl. XVI, figs. 30 
and 31) are well marked and quite separate for each segment. 
Those in the first six somites are prolonged into the appendages. 
The somatopleure is several cells thick ; the splanchnopleure, on 
the contrary, consists of a single layer of cells. The mesoblast 
in the cephalic segment is thinner than in the body somites, 
and the ccelomic space is narrower. 

Stage K (Pl. XVI, fig. 32).—The thoracic appendages have 
increased very much in size, and the cheliceree and chele are 
both bifurcated at the extremity. A section through the base 
of one of the ambulatory appendages (Pl. XVI, fig. 33) shows 


THE EMBRYOLOGY OF A SCORPION. 123 


a well-developed process extending inwards towards the middle 
line. This is undoubtedly the sternocoxal process, which is 
present on the second, third, and fourth appendages of the 
adult. Lankester! characterises the presence of this process 
as a very important point of resemblance between the thoracic 
appendages of Limulus and Scorpio. It is therefore interest- 
ing to find it at this early stage present on all four pairs of 
ambulatory appendages. A series of sections through the base 
of the fifth appendage, i.e. third ambulatory (Pl. XVI, fig. 34, 
a—h), shows the first stage of another structure characteristic 
of Limulus and the Arachnids—the coxal gland. This consists 
of a simple tube opening to the exterior at the base of the fifth 
appendage (fig. 34 @), and running forwards through the meso- 
blast to open in fig. 34 @ into the ceelomic space. There can be 
no doubt that it is a nephridium. Gulland’s researches® on 
the coxal gland in the young Limulus point to the same con- 
clusion. I have been unable to find traces of nephridia in 
any other somites, unless, indeed, the genital tubes are partly 
nephridial. The six abdominal segments also bear appendages 
(Pi. XVI, figs. 32 and 35). These appear on surface view much 
more prominent than they really are owing to their white 
colour, which is due to the greater thickness of cells. In 
section (Pl. XVI, fig. 35) they are seen to project very slightly, 
and to be formed by a thickening of the epiblast and somato- 
pleure, but with no definite outpushing such as there is in the 
thoracic appendages. The first pair of these appendages—the 
genital opercula—is very small, and concealed by the last pair 
of walking legs. The other five pairs—the pectines and four 
pairs of lung-books—are all of approximately equal size and 
structure. Ihave been unable to find the smallest trace of 
appendages on the somites behind these, i.e. somites 13—17, 
and do not believe they exist. 

The cephalic segment is not so deeply cleft as in the last 
stage, and the mouth has shifted posteriorly sc that now it lies 
between the bases of the chelicere. In the centre of each 


1 *Timulus an Arachnid,’ p. 20. 
2 «Quart. Journ. Mier. Sci.,’ vol. xxv. 


124 MALCOLM LAURIE. 


cephalic lobe is seen a dark spot (fig. 32, ce. in.). These spots 
are the cerebral invaginations. They begin in a somewhat 
earlier stage (Pl. XVI, fig. 36) as a pair of small inpushings. 
These extend rapidly backwards and meet in the middle line, 
their two lumens becoming continuous. This is seen in Pl. 
XVI, fig. 37 a—p, in which four transverse sections through 
this region are figured. Owing tothe strong cephalic flexure in 
this stage the stomodzum (s¢.) is also shown in section. The 
cells, both at the sides of the cephalic lobes and throughout the 
greater part of the invaginations, are rapidly increasing in 
number to form the cerebral ganglia. Those in the centre of 
the cerebral lobes remain as a thin layer, and take no part in 
the brain formation. The cells also on the dorsal side in the 
middle, where the two invaginations have united (Pl. XVI, fig. 
37 D, oc.), are more closely packed than the others, and take no 
part in the formation of the brain. They are the beginning of 
the retinal layer of the central eyes. 

The ventral nervous system is in much the same condition 
histologically as it was in the last stage. The commencement 
of its separation from the hypodermis can, however, be seen 
(Pl. XVI, fig. 35) where the hypodermis is growing over it 
from each side as a thin layer. 

The tail segment is now divided into six somites, and extends 
forward along the ventral surface of the body, reaching, at this 
stage, to the third abdominal somite. The epiblast is thickened 
on the ventral surface to form the nervous system. This is 
not shown in fig. 35, as the section passes between two thicken- 
ings. The cavity of the tail is occupied by a tubular extension 
of the hypoblast (fig. 35, Ay.) surrounded by mesoblast. There 
is as yet no trace of the proctodeum. 

The cclomic spaces in the thoracic somites have not 
developed much. ‘Those in the abdominal somites, however 
(Pl. XVI, fig. 35, cw.), have extended enormously, and now 
reach round almost one third of the egg. The mesoblast, ex- 
cept in abdominal appendages, consists of two single layers of 
cells. In the tail the coelomic spaces are not yet formed. 

Stage L.—The embryo, of which fig. 88 (Pl. XVII) shows a 


THE EMBRYOLOGY OF A SCORPION. 125 


surface view, has by this time made considerable progress in 
several important points. . The thoracic appendages are slightly 
segmented (Pl. XVII, fig. 39, ap.), though this is not apparent 
in a surface view. The chelicere have moved in towards the 
middle line, and the mouth is now concealed between their 
bases. The chele are very large, and have their pincers well deve- 
loped. The coxal gland, which opens at the base of the fifth 
pair of appendages, is no longer a straight tube, but has become 
bent on itself, so that a section through it (Pl. XVII, fig. 39, 
cox.) shows the tube cut in three places. It can still, however, 
be traced through a series of sections as a simple tube opening 
into the celom. The abdominal appendages have undergone 
great changes. The genital opercula are still simple thickenings 
of the epi- and meso-blast, but the pectines (Pl. XVII, fig. 
40) have become folded in a direction parallel to the long axis 
of the body, i.e, transverse to their own axis. The most im- 
portant change is, however, that of the four following abdo- 
minal appendages. These (Pl. XVII, fig. 41) are pushed in 
so as to form shallow cup-shaped cavities. The inpushing is on 
the posterior part of the appendage, and is directed slightly 
forwards. This is the commencement of the formation of the 
lung-book. 

The cephalic segment, which is shown in Pl. XVII, fig. 38, 
extended in the same plane as the ventral surface of the 
embryo, is no longer so distinctly bilobed as in the last stage. 
The cerebral invaginations (Pl. XVII, figs. 42, @ and 4, and 43) 
are much shallower, and have entirely joined together, so that 
there is now only a single inpushing. This lies justin front of 
the chelicere (Pl. XVII, fig. 38). The brain is being formed 
from the sides of the inpushing, and shows a very characteristic 
structure. ‘The mass of cells is more or less grouped round 
small circular clear spaces (fig. 43), which give to this part of 
the brain the appearance of being composed of a number of 
small vesicles. I have not succeeded in tracing the development 
of the nerve-fibres, which occupy the centre of the cerebral 
ganglion (fig. 43). This central portion appears at this stage 
perfectly transparent and empty. 


126 MALCOLM LAURIE. 


The retina of the central eyes is still a thickening of the 
dorsal layer of the cerebral invagination (Pl. XVII, fig. 43, 
rin.). It is visible in surface view (fig. 38, oc.) as a white spot 
on the margin of the invagination. The hypodermis imme- 
diately outside it is somewhat thickened, and will in this 
region form the vitreous layer (fig. 45, vit.). 

The ventral nervous system is now completely separated from 
the hypodermis (figs. 40 and 41, ”.c¢.). The cells are beginning 
to congregate together to form the ganglia, though the nerve- 
cord between the ganglia is still largely cellular. Nerves are 
seen growing out from the ganglia as thick cords of cells 
(fig. 40). The ganglia contain a clear space in their centre 
which later is occupied by a mass of fibres. 

The tail (Pl. XVII, fig. 38) has now attained its full 
number of segments but the sting is not yet formed. The 
gut extends up almost the whole length of the tail. There is 
no sign yet of the formation of the proctodeum. The hypo- 
blast in the rest of the body remains as a scattered layer of 
cells. 

The mesoblast has now grown round the body as a double 
layer, with the coelomic space between. In the middle line of 
the back, where the right and left folds of mesoblast meet, 
there is a somewhat irregular thickening in which both soma- 
topleure and splanchnopleure seem to be involved. From this 
thickened band, which extends from close behind the brain to 
the beginning of the tail, the heart is formed. On the ventral 
side in the thoracic region the mesoblast of the outer layer is 
broken up into long strings of cells—the muscles—so that the 
coelomic space can no longer be very definitely made out. 

The stomodeum reaches as far as the back of the cerebral 
ganglion. This is the limit of its growth, and it remains a 
closed tube until, at a much later stage, the gut has grown 
forward and united with it. 

Stage M.—The embryo (Pl. XVII, figs. 44 and 45) does 
not show very much change in surface view. The thoracic 
appendages are longer and distinctly segmented. They overlap 
across the middle line and conceal the pectines. The chelicere 


THE EMBRYOLOGY OF A SCORPION. 127 


are further forward in relation to the mouth, which can now 
be seen lying between the bases of the chele. 

The genital opercula begin to grow out from the body wall 
and the genital duct begins to be formed. This last (Pl. XVII, 
fig. 46) is developed in the mesoblast as a tubular portion of 
the coelom, but does not open to the exterior up to the time of 
hatching. It may be nephridial in its nature, but this very 
late formation of the external aperture is not very favorable 
to such an hypothesis. The pectines are separated at their 
outer ends from the body wall. The inpushings for the lung- 
books are much deeper, and the cavity, which extends forwards 
from the opening, is divided up by lamelle which grow down 
from its upper end (Pl. XVII, fig. 47). It is in close relation 
to a space in the mesoblast which contains blood-corpuscles. 

The cephalic segment (Pl. XVII, fig. 45) is now rapidly 
approximating to its finalshape. The cerebral ganglion, which 
is seen from the surface as a four-lobed white mass (fig. 45, ce.), 
has now lost all connection with the epiblast. The invagina- 
tion remains, but its sides no longer give rise to nerve-tissue 
(Pl. XVII, figs. 48 and 49). The thickening for the central 
eye (figs. 48 and 49, rin.) is more largely developed, and 
pigment is deposited in the ends of the cells furthest from 
the invagination. The eye is plainly visible as a double 
black spot on the surface. The upper edge of the invagina- 
tion is growing down to close its orifice. The hypodermis 
lying immediately above it is clearly marked off from the 
rest as the vitreous layer (fig. 49, vit.). A considerable space 
still separates the retina from the vitreous layer. 

The lateral eyes now appear for the first time as black spots on 
what Lankester terms the “optic area,” 7. e. the front margin 
of the head (Pl. XVII, fig. 45, oc.). Their development, as 
Parker! has shown, is strikingly different from that of the 
central eyes. Hach eye, and in this species there are at first 
three, is formed (fig. 50) by a slightly cup-shaped thickening 
of the hypodermis. The nuclei of this thickened portion 
become larger, and pigment soon begins to be deposited at the 

1 Loe. cit. 


128 MALCOLM LAURIE. 


outer ends of the cells. Fig. 50 @ shows a somewhat later 
stage, in which the cupping of the hypodermis has become 
flattened out. There is no invagination of any sort, and the 
eyes are, as Lankester and Bourne! described them, mono- 
stichous. The ventral nervous system has not undergone much 
development. It has sunk somewhat deeper and is separated 
from the hypodermis by the mesoblast. 

The tail has now developed its terminal segment—the sting. 
The cavity of this last is partly occupied by the paired poison 
gland, apparently formed by inpushing of the hypodermis (PI. 
XVIII, fig. 51, p. g/.). Hach mass is connected to the super- 
ficial hypodermis by a short duct. 

The gut extends down the whole length of the tail, and the 
proctodeum is present in the form of a solid mass of hypodermis 
cells blocking up its end (Pl. XVIII, fig. 51, proct.). The gut 
has also begun to grow forward (Pl. XVIII, figs.52 and 53). In 
the last abdominal segment it is a complete tube surrounded 
by a thin layer of mesoblast (fig. 52, ant.). It gives rise to 
two tubular outgrowths from its dorsal side, which are the 
Malpighian tubes (fig. 52, mlph.). These run first towards the 
back and then bend forward. ‘There can be no doubt as to 
their hypoblastic origin in this form, as the proctodzeum is not 
yet formed. They have been already shown to be outgrowths 
of the mesenteron in some Spiders by Loman,’ and also in ter- 
restrial Amphipoda by Spencer.* Further forward (Pl. XVIII, 
fig. 58, int.) the gut is simply asemi-cylindrical layer of hypoblast 
supported by a string of mesoblast and open to the yolk on its 
dorsal side. In the thorax it has not yet begun to form. 

The mesoblast is broken up into strings and bands. The 
ceelom is still pretty distinct in the abdominal region (Pl. X VIII, 
fig. 53, cw.), and the heart is a large thin-walled tube apparently 
connected with both somatopleure and splanchnopleure. As 
mentioned above, the genital tube is formed in the somato- 
pleure in the seventh somite and is a portion of the celom. 


1 ¢Quart. Journ. Mier. Sci.,’ vol. xxiii. 
2 ¢ Tijdschrift der nederl. Dierk. Vereen,’ i. 
3 © Quart. Journ. Mier, Sci,,’ vol. xxv. 


THE EMBRYOLOGY OF A SCORPION. 129 


Stage M.—The changes from the last stage up to the time 
of hatching are not very numerous, though very important. 
The body attains a structure almost exactly lke that of the 
adult, the appendages being segmented and the whole animal 
covered by a thin, structureless, highly refracting cuticle. The 
coxal gland still opens by a small aperture to the exterior at 
the base of the fifth appendage (Pl. XVIII, fig. 54). This 
aperture, which is lined for a short distance by the cuticle, 
leads to a straight duct (fig. 54) lined by cubical cells with 
round nuclei, which closely resemble the cells of the gland. 
The gland itself is distinguishable into medullary and cortical 
portions as described by Professor Lankester! in the adult. The 
tubules have distinct lumens surrounded by acubical epithelium. 
The gland and its duct are surrounded by a thin capsule of flat 
mesoblast cells. 

The genital tubes have pushed their way some distance 
between the lobes of the liver, but they are not yet connected 
by transverse tubes nor do they open to the exterior. The two 
layers of which the tube is composed in the adult (v. supra, 
p. 108) are not yet distinguishable. The pectines approximate 
very closely to their adult structure. 

The ninth, tenth, eleventh, and twelfth appendages are also 
very similar to those in the adult (Pl. XVIII, fig. 55). The 
number of lamelle is not so great, but their structure is very 
well shown. Lach lamella is covered by a thin cuticle, and its 
cavity is in direct communication with a blood-sinus (fig. 55, 
bl. s.). The cells which form the lamelle are very large, espe- 
cially towards the base of the appendage. Towards the apex 
they become smaller, and finally pass into a mass of different 
cells from which more lamelle are formed as the animal grows. 
The spaces between the lamellee (fig. 55, a. c’.) are narrower and 
in communication with the exterior through the stigma. 

The head is now completely formed, the mouth having 
shifted so as to lie behind the chelicere. The invagination 
which forms the central eyes has closed up. <A stage immedi- 
ately after its closure is shown in Pl. XVIII, fig. 56. Here the 

1 * Proc, Roy. Soc,,’ vol, xxxiv, 1882-83, 


130 MALCOLM LAURIE. 


lips of the invaginations have come together, but not fused. 
The posterior layer of the invagination is visible as a thin layer 
of cells (fig. 56, rtn'.), separated from the retina by a narrow 
space. The vitreous layer (vit.) is distinctly marked as a 
thickening of the hypodermis on the top of the head, the nuclei 
in that region being elongated, but there is still a small space 
separating it from the retina. By the time the embryo is 
hatched the eye (Pl. XVIII, fig. 57) has lost all connection 
with the hypodermis at the point where it was invaginated. 
The cells are long and deeply pigmented round their margins. 
The pigment is not equally abundant throughout the whole 
length of the cell, but five alternately more and less deeply 
pigmented zones can be distinguished (fig. 58). The base of 
the cells is most deeply pigmented, and their superficial ends 
come next. The nuclei of the retinal cells lie in zone 4, 
but I have been unable to find any trace of either rhabdomes 
or phaospheres. I have also not been able to trace any migra- 
tion of mesoderm cells among the retinal cells. The posterior 
layer of the invagination can with difficulty be made out in 
depigmented sections owing to the flatness of its nuclei, and it 
is absolutely undistinguishable in sections from which the pig- 
ment has not been removed. This posterior layer forms the 
post-retinal membrane of the adult eye. The optic nerve is 
beginning to grow out from the cerebral ganglion, but has not 
yet come into connection with the eye. The hypodermis, im- 
mediately in front of the eye, is formed of a single layer of 
large transparent cells with faintly staining oval nuclei 
(Pl. XVIII, figs. 57 and 58, vit.). This vitreous layer is 
covered by a thin cuticle exactly like that which covers the 
rest of the surface of the body. The only sign of the formation 
of the lens is a slight cupping of the vitreous layer at one point 
(fig. 58). The hollow formed here is, however, not as yet filled 
up by any cuticular substance, but the cuticle passes straight 
over it. Round the area, where the lens will form the hypodermis 
is deeply pigmented. The cells are much smaller than those 
of the vitreous layer, and their nuclei are irregular in shape. 
The cells of the lateral eyes (Pl. XVIII, fig. 59) are about 


THE EMBRYOLOGY OF A SCORPION. 131 


the same size as those of the median ones. The pigment is not, 
however, arranged in definite zones, though it is more abundant 
at the base and at the outer ends of the cells than in the middle. 
There is a small third lateral eye present, and the hypodermis 
around and—the lateral eves being so to to say on the edge of 
the head—below the eyes is all pigmented. I have been 
unable to find the nerve to these eyes and think it is probably 
not yet formed. The cuticle over the lateral eyes is not 
thickened to form the lens, and I have seen no sign of the 
peculiar mode of lens-formation described by Parker,' i. e. the 
ends of the perineural cells pushing in front of the retina. It 
is, of course, possible that this does not take place till later. 

The brain and ventral cord have almost attained their 
adult structure. In the nerve-cord there is a string of 
cells in the middle line (Pl. XVIII, fig. 60) dorsal to the 
cords proper, which seems to represent the centre of the neural 
groove. 

The tail is exactly similar to that of the adult, and is carried 
in the same way curved over the back. The poison-glands are 
fully formed and surrounded by muscles, but do not occupy so 
much of the terminal segment asin the adult. The proctodeeum 
is lined by flat cells and has pushed its way almost to the 
base of the tail. The mesenteron is fully formed only in the 
hind segments of the body (Pl. XVIII, fig. 60). From the 
end of the stomodzeum to where the last hepatic ceca 
join it the intestine (fig. 61) is surrounded by a definite 
cylindrical layer of mesoblast which is continuous with that 
surrounding the lobes of the liver, but the hypoblast cells 
lining this cylinder (fig. 61, Ay.) are not yet definitely arranged. 
The nuclei are scattered about in groups for the most part 
near the outside, and the cells are drawn out into irregular 
more or less pyramidal form, the apex of the pyramid pointing 
towards the centre of the tube. There is no definite lumen, 
the space between the cells being filled up by small yolk-spheres 
(fig. 61, yk.). 

The liver-follicles are much the same in structure as the 

1 Loe. cit., p. 199. 


182 MALCOLM LAURIE. 


intestine. They contain, however, a rather larger proportion of 
yolk. The scattered layer of hypoblast cells, which in the 
preceding stages surrounded the yolk, takes a large part in their 
formation. They open into the intestine in pairs by wide 
ducts. 

The Malpighian tubes (Pl. XVIII, fig. 60, m/ph.) have not 
undergone much development. They reach well forward in the 
body, and open into the intestine in the first caudal segment. 

It is evident from the structure of the intestine that the 
young scorpion does not need food for some time after hatch- 
ing. The large amount of yolk which still exists must last it 
for some weeks, or most probably till the next spring. If this 
is the case embryonic life practically lasts twelve months as the 
eggs are fertilized in May. 

The outer layer of the mesoblast has now for the most part 
formed itself into muscles. The inner layer is very much com- 
plicated, being folded in so as to surround the gut and the 
lobes of the liver. The spaces between the lobes of the liver, 
which are undoubtedly the true ccelom, are filled up by a net- 
work of trabecular tissue (Pl. XVIII, fig. 60). The heart, 
pericardium, and blood-vessels are fully formed and contain a 
considerable number of large nucleated corpuscles. 


Summary of the Changes during the Fourth Period. 

(1) The thoracic appendages begin as simple outpushings of 
the body wall containing a portion of the coelom (Stage I). 

They rapidly increase in length and the chelicerz and chelz 
become bifid at their extremities. Sternocoxal processes are 
present on the third to sixth appendages (Stage K). 

The chelicerze, which were at first behind the mouth, gradu- 
ally move forward relatively to it till they come to lie in front 
of it (Stage L). 

(2) The coxal gland begins as a simple tube opening to the 
outside at the base of the fifth pair of appendages, and opening 
at the other end into the celom (Stage K). The tube soon 
becomes coiled, but the external opening persists until after 
hatching. It is undoubtely a nephridium. 


THE EMBRYOLOGY OF A SCORPION. t33 


(3) The abdominal appendages appear as thickenings of the 
epi- and meso-blast on the seventh to twelfth somites (Stage K). 
The first pair (genital opercula) does not develop further till 
a late stage (L). 

The second pair (pectines) form a number of short longi- 
tudinal ridges on the surface of the abdomen (Stage K). They 
then separate from the body, the separation beginning at their 
outer ends (Stage M). 

The third to sixth pairs (gill-books) begin to be pushed in 
(Stage L). The inpushing becomes deeper, and begins to be 
divided up (Stage M), and by the time the embryo is hatched 
they have attained their adult condition in every respect except 
size and number of lamelle. 

(4) The cerebral ganglion and central eyes begin as a pair of 
invaginations on the cephalic lobes. These invaginations 
meet in the middle (Stage K). The cerebral ganglion is formed 
from the sides of the invaginations, which rapidly become 
shallower and unite so as to open in the middle line. The 
dorsal surface of the invagination becomes thickened to form 
the retina of the central eyes (Stage L). 

The brain becomes entirely separate from the hypodermis, 
the invagination remaining to form the eyes(Stage M). The 
invagination closes up and its lumen disappears. ‘The cells of 
its lower layer form the post-retinal membrane. Those of the 
upper layer form the retina, and come in contact with the hy- 
podermis on the top of the head, which is thickened in this 
region to form the vitreous layer. The retinal cells become 
deeply pigmented (Stage N). 

(5) The lateral eyes form as cup-shaped thickenings of the 
hypodermis in the “ opticarea,”’ the cells of which become 
pigmented. There is no invagination, and they consist of a 
single layer (Stage M). 

(6) The ventral nervous system forms as a pair of thickened 
segmented bands, one on each side of the neural groove 
(Stage I). The nerve-cords sink down, a thin layer of hypo- 
dermis growing over them. ‘There is at this time (Stage K) 
a distinct postoral pair of ganglia for the chelicere. The cells 


134 MALCOLM LAURIE. 


become aggregated to form ganglia, and the cheliceral ganglia 
become fused with the cerebrum. 

(7) The tail grows out, lying along the ventral surface 
of the abdomen. The poison-gland in its terminal segment is 
formed by a pair of invaginations of the epiblast. 

(8) The hypoblast consists of an irregular layer under the 
whole embryo and a solid mass at the tail end (Stage I). As 
the tail grows the hypoblast grows into it as a tube reaching 
down to the last somite (Stage K). 

The hypoblast forms the gut in the abdominal portion of the 
body, growing forward in a sling of mesoblast at first as a flat 
layer, which soon becomes bent round into a cylinder. The 
Malpighian tubes are formed as outgrowths from the mesenteron 
in the first post-abdominal somite (Stage M). 

The gut does not reach forward to the stomodeum till 
shortly before hatching, and at this period the portion of 
it into which the liver-follicles open is not fully formed 
(Stage N). 

(9) The stomodeum is formed early. It lies at first in front 
of the chelicere (Stage 1), but soon shifts its position and 
comes to lie behind them. It extends inwards as far as the 
back of the brain. 

(10) The proctodeum is formed much later than the stomo- 
deum. It is at first a solid plug of cells (Stage M). As it 
increases in size it appears to replace the hypoblast in the 
last four somites. 

(11) The mesoblast consists at first of a pair of segmented 
bands with a separate coelomic space in each somite, and also 
one in the cephalic segment (Stage 1). The ceelomic spaces 
soon unite, and the mesoblast bands join across the ventral 
surface. Somewhat later they extend round—the ccelomic 
space extending with them—and unite in the middle line on 
the dorsal surface (Stage L). From the thickened band where 
they have united on the dorsal surface the heart is formed. A 
portion of the ceelom in the seventh segment becomes separated 
off to form the genital tubes (Stage M). These do not open to 
the exterior. The outer layer of the mesoblast forms chiefly the 


THE EMBRYOLOGY OF A SCORPION. 135 


muscles of the body. The inner layer becomes folded so as fo 
surround the liver and intestine, and the cceelomic space becomes 
partly filled up by trabecular tissue. 


Conclusion. 


The development cf this Scorpion, of which I have tried to 
give an outline above, is interesting in many points. It does 
not agree closely with any other Arachnid type as yet described, 
and I have for the present given up all attempts at comparison. 

The development of the central and lateral eyes entirely bears 
out Lankester and Bourne’s description of their structure. It 
is true that the central eyes are three-layered, but as the retina 
is the second layer from the surface—the third layer forming 
only apost-retinal membrane—they may be called diplostichous. 
The account given above of their development agrees in all 
essential respects with that of Parker, but, having a larger 
supply of embryos, I have been able to trace the earlier stages 
and the connection of the eyes with the cerebral invagination. 
Their mode of origin resembles very closely Locy’s! description 
of the development of the eyes in Agelena nevia, the chief 
difference being that in Agelena the optic invaginations appear 
to have no connection with the formation of the brain. Locy 
does not, however, give a detailed description of the formation 
of the latter. 

The description given above of the development of the lateral 
eyes also agrees pretty closely with that of Parker. In these, 
as in the central eyes, Lankester and Bourne’s conclusions are 
confirmed, and Patten’s*® conclusions as to what the structure 
of the eyes must be in order to fit in with his theories are 
shown to be without foundation. The lateral eyes are mono- 
stichous, being simply somewhat specialised hypodermis cells. 

The mode of formation of the ventral nervous system is 
exceptional among Invertebrates, resembling rather that of 

1 «Bull. Mus. Comp. Zool., Harvard,’ vol. xii, p. 85. 


2 “yes of Molluscs and Arthropods,” ‘Mitth. Zool. Stat. Naples,’ 
Bd. vi. 


136 MALCOLM LAURIE. 


Chordata. The nerve-cord instead of peeling off from the 
superficial layer of epiblast sinks down bodily, and is covered 
by a layer of epiblast which grows over it from each side. 

The development of the coxal gland leaves, I think, no room 
to doubt that it isa nephridium. That of the genital tubes is 
less conclusive, but I should think it probable that they are 
also, at least in part, nephridial. 

The gill-books are undoubtedly appendages comparable to 
the abdominal appendages of Limulus. Whether they are 
really invaginated, i. e. whether the edge of each lamella in the 
Limulus appendage corresponds to the bottom of the fold 
between the lamellz in the Scorpion’s gill-book, or whether the 
whole appendage has become sunk in a hollow in the abdominal 
surface without being invaginated, it is difficult to say. Un- 
doubtedly, the surface now exposed to the air has always been 
the external surface, but that would be the case with either of 
the above modes of derivation. Although the second alterna- 
tive has the advantage that it is easy to see how the change 
could take place gradually, I am inclined to think the first is 
probably the true way in which they have arisen. One argu- 
ment in its favour is that if the second alternative were correct 
one would expect the gill-book to commence as a distinct out- 
growth, which would become sunk ina pit. Now, there is no 
such outgrowth in the formation of the gill-book. The first 
thing to appear on the thickened portion of the epiblast, from 
which the gill-book is formed, is a pit (Pl. XVII, fig. 41). The 
lamellz do not begin to form till a later stage. Again, the 
abdominal appendages of Limulus are directed towards the tail 
as one would expect abdominal appendages to be. Now, if the 
appendage had sunk in without invagination, one would expect 
it to be still directed towards the tail unless there were some 
very good reason for its having changed its direction. If, on 
the contrary, it had become invaginated it would naturally be 
directed in the opposite direction towards the head, and this is 
what we find in the Scorpion. The inpushing is from the 
beginning towards the head, and the aperture opens towards 
the tail (Pl. XVII, fig. 47). I think it is quite conceivable 


THE EMBRYOLOGY OF A SCORPION. 137 


that the changed conditions of development, due to terrestrial 
life, and the consequent pressure on the embryo, may have pro- 
duced this change. A detailed account of the development of 
these appendages in Limulus may throw more light on the 
matter, but, unfortunately, though many authors have attacked 
the problem, a complete and satisfactory account of the develop- 
ment of Limulus is not yet in existence. 


EXPLANATION OF PLATES XIII, XIV, XV, XVI, 
Mb GoOav 


Illustrating Mr. Malcolm Laurie’s paper on “The Em- 
_ bryology of a Scorpion (Euscorpius italicus).” 


Abbreviations. 


a. c. Air-cavity in gill-book. ac’. Air-spaces between the lamelle of gill- 
book. «ab. ap. Abdominal appendage. am. Amnion. am. c. Amniotic cavity. 
ap. Appendage. 0/. Blastoderm. 4/7. s. Blood-space. 47. c. Blood-corpuscle. 
cau. Caudal segment. ce. Cerebral ganglion. ce. iz. Cerebral invagination. 
ceph. Cephalic segment. ce. Celom. cow. Coxal gland. cor. d. Duct of 
coxal gland. ep. Hpiblast. ep’. Extension of epiblast beyond ventral plate. 
fol. Follicle. fol’, Outer non-cellular layer of follicle. g. I. Ganglion of 
cheliceral somite. ger. Germinal epithelium or inner layer of ovarian tube. 
ger’. Yolk-forming cells derived from germinal epithelium. ge. ¢. Genital 
tube. At. Heart. Ay. Hypoblast. %y’. Extension of hypoblast beyond 
ventral plate. Ay. m. Mass of hypoblast in caudal segment. z¢. Intestine. 
i. Gastric gland. mes. Mesoblast. mé#. Prolongation of ovarian tube to egg. 
mlph. Malphigian tubes. 2. g. Neural groove. z. z'. Nucleus, nucleolus. 
w.c. Nerve-cord. 2. gl. Nerve-ganglion. 2. ¢h. Neural thickening. oe. 
Central eye. oc’. Lateral eye. o. 7. Outer layer of ovarian tube. ov. Ovum. 
p. gt. Poison-gland. pr. hy. Primitive hypoblast (hypomesoblast). proct. 
Proctodeum. pr. ¢. Primitive thickening. tz. Retina of central eye. 
rtn'. Third layer of central eye, post-retinal membrane. s. m. Serous mem- 
brane. ss. m'. “ Peripheral cells.” som. mes. Somatic mesoblast. spd. mes. 
Splanchnic mesoblast. s¢. Stomodeum. ste. p. Sternocoxal process.  s¢g. 


VOL. XXXI, PART 11.—NEW SER. K 


138 MALCOLM LAURIE. 


Stigmata. ¢e. Telson. ¢r. mes. Trabecular mesoblast occupying celom. vit. 
Vitreous layer of centraleye. y.c. Cells in yolk. ys. Yolk. The somites 
are numbered I, 11, 111, &c. 


PLATE XIII. 


Fic. 1.—Transverse section of ovarian tube, showing the two layers; one 
cell of the inner layer enlarging to form an ovum. xX 222. 

Fic. 2.—Transverse section of ovum and ovarian tube. The egg has now 
pusbed its way through the outer layer, and appears as a small protuberance 
on the ovarian tube. The follicle is beginning to form from the cells of the 
inner layer, which have accompanied the ovum. x 229. 

Fig. 3.—Longitudinal section of ovum of ‘1 mm. diameter, showing the 
two-layered follicle and the yolk-forming cells (ger’.). x 222. 

Vic. 4.—Longitudinal section of egg of ‘4 mm. in length, showing yolk- 
spheres, indefinite nucleus, and strongly marked nucleolus. The egg is sur- 
rounded by a vitelline membrane. The rest asin Fig.3. x 229, 

Fic. 5.—Section through the base of a ripe egg. x 212, 

Fic. 6.—Yolk-spheres from ripe egg, showing the darkly stained spherical 
and prismatic bodies and the clear spaces. xX +12. 

Fie. 7.—Section through a corpus luteum and part of ovarian tube. 
x 220, 


PLATE XIV. 


Fic. 8.—Surface view of one-layered blastoderm. x 32. 

Fic. 9.—Section through one-layered blastoderm, same stage as Vig. 8. 
x 28°, 

Fic. 10.—Section through blastoderm later than Fig. 9, showing the cells 
multiplying to form a mass at one pole of the egg. x 229 

Fic, 11.—Section through more advanced blastoderm. x 222. 

Fic. 12.—Transverse section through blastoderm at time of formation of 
primitive hypoblast and serous membrane. The yolk and yolk-cells are drawn 
in detail in this figure to show the breaking down of the former. x 222, 

Tic. 13.—Surface view of blastoderm now becoming oval. x 2°. 

fre. 14.—Transverse section through posterior end of embryo figured in 
Fig. 13, showing serous membrane, primitive thickening, primitive hypoblast, 
and “ peripheral cells.” x 229. 

Fic. 15,—Transverse section through anterior part of embryo about the 
same stage. X 22°, 

Fic. 16.—Longitudinal section through an embryo a little younger than 
Fig. 17, showing two somites with a third forming. The mesoblast is 
forming from the primitive hypoblast, the amnion is growing up from the 


THE EMBRYOLOGY OF A SCORPION. 139 


edges of the hypoblast, and the primitive thickening is well seen in the caudal 
segment. xX 142, 

Fic. 17.—Surface view of embryo, with three somites fully formed. 
x 30, 

Fic. 18.—Transverse section through posterior end of Fig. 17, showing 
hypoblastic mass, mesoblast, &&. x 112, 

Fic. 19.—Diagrammatic representation of the relative extension of the 
various layers in an embryo of the stage of Fig. 17. 


PLATE XV. 


Fic. 20.—Surface view of an embryo with seven somites, drawn as if 
flattened out. a—d and c—d are the planes of the sections figured in Figs. 
21 and 23. x 3°. 

Fic. 21.—Transverse section through one of the posterior somites of an 
embryo with seven somites (e—d in Fig, 20), showing the three layers, 
epiblast thinning in centre, and mesoblast thin; amnion, serous membrane, 
and ccelomic spaces. X 219. 

fic. 22.—Transverse section through one of the anterior somites of Fig. 
20. x 242. 

Fic. 23.—Transverse section through tail-segment (c—d) of Fig. 20, 
showing the undivided mesoblast and the hypoblastic mass. x +22, 

Fic. 24.—Surface view of embryo of nine somites, drawn as if extended. 
x 3e, 

Fic. 25.—Transverse section through head-segment of Fig. 24, showing 
epiblast thickening to form cerebral nervous system and spreading (ep’.), with 
the amnion beyond the ventral plate, neural groove, thin mesoblast, with small 
cceelomic space and hypoblast. x 142. 

Fic. 26.—Transverse section through one of the anterior somites of Fig. 
24, showing the epiblast very solid where the appendage will develop (ap.) 
and form the neural thickening (z. ¢#.) at each side of the neural groove. 
Mesoblast thick, and ccelom not very evident. x 242. 

Fic. 27.—Diagrammatic representation of the relative extension of the 
various layers in an embryo of the stage of Fig. 24. 

Fic. 28.—Surface view of embryo at Stage I (ten somites) extended in a 
plane, showing appendages, cheliceral ganglion, stomodeum, &. x 22, 


PLATE XVI. 


Fic. 29.—Longitudinal section of Stage I in the middle line, showing 
dorsal flexure of the embryo, commencement of tail outgrowth, stomodxum, 
&. x 22. 


140 MALCOLM LAURIE. 


Fie. 30.—Longitudinal section to one side of the middle line, showing the 
appendages and ceelomic spaces. X ~°. 

Fic. 31.—Transverse section through the third somite of Stage I, showing 
the formation of the appendage, the neural thickening, &. x -%©. 

Fic. 32.—Surface view of embryo at Stage K, showing the cerebral invagi- 
nations, abdominal appendages, tail, &. x 4°. 

Fic. 33.—Transverse section through the base of a thoracic appendage, 
showing the sternocoxal process. 

Fic. 34, a—h.—Series of sections through base of fifth appendage, showing 
the coxal gland. 

Fic. 35.—Transverse section through one of the abdominal appendages and 
the tail, showing the appendage, the neural thickening beginning to separate 
from the epiblast, the gut forming in the tail, &. x 7,2 

Fic. 36.—Transverse section through the cephalic segment of a somewhat 
earlier embryo, showing the beginning of the cerebral invagination. x +2. 

Fic. 37, a—p.—Sections through the head of an embryo of Stage K, show- 
ing the cerebral-optic invaginations. x +42. 


PLATE XVII. 


Fie 38.—Surface view of an embryo of Stage L extended in a plane, 
showing the cerebral invagination, the central eyes, &. x 3°-. 

Fic. 39.—Transverse section through the base of the fifth appendage, 
showing the coxal gland. x 45. 

Fic. 40.—Section through the pectines. x 5. 

Fig. 41.—Section through an abdominal appendage, showing the inpushing 
to form the gill-book. x -%%. 

Fie. 42, a, 6.—Transverse sections through the cerebro-optic invaginations. 
x $2, 

Fic. 43.—Longitudinal section through the cerebro-optic invaginations, 
showing the formation of the brain and the central eye. x 2345. 

Fic. 44.—Surface view of an embryo of Stage M, from the ventral surface. 
im So, 

Fic. 45.—Surface view of the dorsal side of the head of the same embryo, 
showing the central and lateraleyes. x 3°. 

Fig. 46.—Section through the seventh somite, showing the formation of 
the genital tube from part of the celom. x 4°. 

Fie. 47.~Longitudinal secion through a gill-book, showing the commence- 
ment of the formation of the lamelle. x 14+, 

Fic. 48.—Longitudinal section through the head end, showing the stomo- 
deum and the cerebro-optic invagination from which the brain is now entirely 
separate. Xx +*. 

Fic. 49.—Transverse section through the same region. x 4 


-: 


THE EMBRYOLOGY OF A SCORPION. 141 


Fie. 50.—Longitudinal section through the lateral eye, showing its forma- 
tion by a thickening of the hypodermis. x 152, 

Fie. 50¢.—Section through a somewhat older lateral eye in which the 
inpushing of the hypodermis has disappeared. x 222. 


PLATE XVIII. 


Fie. 51.—Longitudinal section through the tail end of Stage M, showing 
the poison-gland, proctodseum, intestine, &ec. Xx 32. 

Fic. 52.—Transverse section through the posterior end of the body, showing 
the intestine, with the Malpighian tubes, the heart, &e. x 52. 

Fic. 53.—Transverse section a little further foward than Fig. 52, showing 
the intestine, which has not yet closed intoa tube. x $2. 

Fic. 54.—Section through the coxal gland of a newly-batched scorpion, 
showing the opening to the exterior, &. x Z®. 

Fie. 55.—Longitudinal section through a gill-book of a newly-hatched 
scorpion, x 589, | 

Fic. 56.—Longitudinal section through the central eye of an embryo a short 
time before hatching, showing the closure of the cerebro-optic invagination 
and the three layers of the eye. x 2. 

Fic. 57.—Longitudinal somewhat oblique section through the central eyes 
of a newly-hatched scorpion. 222, 

Fie. 58.—A few cells of the same eye more highly magnified, and showing 
the inpushing in the vitreous layer. 

Fias. 58, a, 6, c.—Transverse sections through the same eye at different 
levels. 

Fic. 59.—Section through the lateral eyes of a newly-hatched scorpion. 
x 222, 

Fic. 60.—Transverse section through the posterior part of the body of a 
newly-hatched scorpion, showing the fully-formed intestine, the Malpighian 
tubes, the nerve-cord, and the trabecular tissue filling up the ccelomic space. 
x 222, 

Fie. 61.—Transverse section through the intestine further forward, where 
it is not yet properly formed, showing the irregular hypoblast cells and yolk- 
spheres. xX 222, 


ee Oe oe Cee eon 


ctr) * 
e 
Mh : 
Sai) iv fy 
aryl) a a! 
: E ET d io? 
“ 
} dvi WaltA, A 
AA) bY Ln 2 s r; 
ie oe y acer! ; 
' 
- 
Fd é i ¥ 
wl Lr 32s 
‘ ° + , 
ahi 
‘ 1 Fd 
H ; ‘ f r i. 
neat . 
3 off gi | ; 
- 4 x 
FY) A 
: “ A ue 
‘ af at 


pa 


ens DP 4 
n 


’ ee 
| . ; iy} avy j 4 ® 4 are Se —— r, 
ot, a - os i e i ig? ‘ =~ ' ‘ ’ © be j coin ih 
any z 7 ae ry f : | 7 


pir ee té f he s 
2 
™? . 
‘ . © wl ail 
m - ' ie aa iy ‘ j fy ib 
: : é U _ . 
— 4 ’ j t » ) ‘ iw wad ect emiay rcs 
Ru Oe = Ao lett. S } bul . " 
phalite ‘eb. ian tad at ghee abe as ae 
1) tes ae Gis por . : 


‘ TH ; “ais 3 bu [ Fs hina 
= © 


P 
j i Z 14 
aTiGe pity ie ee ipa yray ae 


' iy 
aie) oh Pe. a Gy i‘, of ’ 


MORPHOLOGY OF COMPOUND EYES OF ARTHROPODS. 143 


On the Morphology of the Compound Eyes of 
Arthropods. 


By 


S. Watase, 
Fellow of the Johns Hopkins University. 


With Plate XIX, 


Preratory Note.—The following extract and accompanying plate, from a 
memoir recently published in the ‘Studies from the Biological Laboratory, 
Johns Hopkins University,’ vol. iv, are here reproduced because it seems to 
me that they form a very important contribution to a subject which has 
been largely discussed in the pages of this journal, whilst they may not be 
readily accessible to European morphologists. In Mr. Watase’s original 
paper a description and figures of the structure of the lateral eyes of 
Limulus, and of the compound eyes of other Arthropods, are given; but the 
diagrams in the plate now reproduced are sufficient to indicate the author’s 
conclusions. 

The distinctive feature about Mr. Watase’s views is that he does not, as I 
did in my paper published conjointly with A. G. Bourne, on the “ Eyes of 
Scorpio and Limulus,” ‘ Quart. Journ. Mier. Sci.,’ vol. xxiii, endeavour 
to derive the compound eye of Arthropods from a segregation of such dip- 
lostichous eyes as the central eyes of Arachnida; but, leaving these entirely 
aside, derives the commoner type of Arthropod compound eye from the 
monostichous lateral eyes of Limulus. 

This seems to me to be a very happy suggestion. At the same time, I 
regret that the author has not, apparently, accepted the statements made by 
Bourne and myself as to the simple monostichous structure of the lateral eyes 
of Scorpions—nor investigated their structure for himself. Had he done so, 
he would have been able to assign even a clearer and simpler starting point 
for the “ ommatidium ” of the compound Arthropod eye than that afforded by 
the lateral eyes of Limulus, where the anomalous central ganglion-cell dis- 
covered by him, presents us with a complication, ‘There is, I believe, no 


144. S. WATASE. 


longer any doubt that a simple monostichous structure characterises the 
lateral eyes of Scorpions as originally shown by Bourne and myself. It is 
perhaps well to note that Grenacher was the first to describe the charac- 
teristic monostichous structure of the lateral eye of Limulus, on which Mr. 
Watase’s theory is based. 

Interesting matters for investigation and speculation are opened up by Mr. 
Watase’s views. Such matters are the relation of the diplostichous mono- 
meniscous central eyes of the Arachnida to compound or polymeniscous om- 
matidial eyes of Arthropoda generally. And especially interesting, it seems 
to me, would be the attempt to account for that incipient segregation of the 
retinal cells into groups united by a five-fluted rhabdom which we find in 
both the lateral and the central eyes of Scorpio. 


March, 1890. EK. Ray LANKESTER. 


Balfour! has given a sketch of the possible evolution of a 
visual organ. He starts with a simple organism in which a 
spot on the surface of the body may become spontaneously 
pigmented and therefore become specially sensitive to light. 
The cuticular covering of the body may become thickened at 
this spot and act as an apparatus for condensing the light upon 
the pigmented spot lying beneath it. He further expresses 
his view elsewhere,’ that the lens-like dioptric apparatus of 
the eye, formed either as a thickening of the cuticle or as a 
mass of cells, was at first formed simply to concentrate the 
light on the sensitive spot; the power to throw an image of 
external objects on the perceptive part of the eye was acquired 
gradually afterward. 

The part which is played by pigment in the physiology 
of vision is considered a most obscure problem. I quote 
the following, clearly put forward by Foster,® as a physiological 
aspect of the question bearing upon the discussion at issue : 


1 BF, M. Balfour, “ Address to the Department of Anatomy and Physiology, 
British Association, 1880,” ‘ Nature,’ vol. xxii, p. 417, 1880. 

2 «Comparative Embryology,’ vol. ii, chap. xvi, “Organs of Vision,” p. 
470. 

3 M. Foster, ‘A Text Book of Physiology,’ 4th edition, book iii, chap. ii, 
“ Sight: the Photochemistry of Retina,” pp. 515, 516, 


MORPHOLOGY OF COMPOUND EYES OF ARTHROPODS. 145 


“But in order that light may produce chemical effects (upon 
protoplasm), it must be absorbed; it must be spent in doing 
the chemical work. Accordingly, the first step towards the 
formation of an organ of vision is the differentiation of a 
portion of protoplasm into a pigment at once capable of 
absorbing light and sensitive to light—i.e. undergoing decom- 
position upon exposure to light. An organism, a portion of 
whose protoplasm had thus become differentiated into such a 
pigment, would be able to react towards light. The light 
falling on the organism would be in part absorbed by the pig- 
ment, and the rays thus absorbed would produce a chemical 
action and set free chemical substances which before were not 
present. We have only to suppose that the chemical substances 
are of such a nature as to act as a stimulus to the protoplasm 
of other parts of the organism (and we have manifold evidence 
of the exquisite sensitiveness of protoplasm in general to 
chemical stimuli), in order to see how rays of light falling on 
the organism might excite movements in it or modify move- 
ments which were being carried on, or might otherwise affect 
the organism in whole or part. Such considerations as the 
foregoing may be applied to even the complex organ of vision 
of the higher animals. If we suppose that the actual termina- 
tions of the optic nerve are surrounded by substances sensitive 
to light, then it becomes easy to imagine how light, falling on 
these sensitive substances, should set free chemical bodies 
possessed of the property of acting as stimuli to the actual 
nerve-endings, and thus give rise to visual impulses in the 
optic fibres.” 

Lubbock! advances essentially the same idea as Balfour’s in 
his recent work on the subject, illustrated with some lucid 
diagrams. “In the simple forms,” he says, ‘‘the whole surface 
is more or less sensitive. Suppose, however, some solid and 
opaque particles of pigment deposited in certain cells of the 
the skin. Their opacity would arrest and absorb the light, 
thus increasing its effect, while their solidity would enhance 


1 “Qn the Senses, Instinct, and Intelligence of Animals,’ Inter. Sci, 
Series, vol. Ixix, 1888, 


146 S. WATASE. 


the effect of external stimulus. A further step might be a 
depression in the skin at this point, which would serve some- 
what to protect these differentiated and more sensitive cells, 
while the deeper this depression the greater would be the 
protection.” 

That such steps of gradual development of visual organs 
have actually taken place in some forms is quite probable. In 
Arthropods, it seems to me worthy of remark that the omma- 
tidium of the lateral eye of Limulus makes the nearest 
approach to this primitive condition. It is nothing more and 
nothing less than a depression in the skin, with the thickened 
chitinous cuticle fitting in the open cavity, and acting as a lens 
to condense the light. The cells which form the sensory part 
of the structure are modified ectodermic cells, and, like the 
rest of the ectodermic cells lying on the surface of the body, 
secrete the chitinous cuticle on a part of their surface. 
Assuming the surface where the chitin is secreted to be the 
exterior, we may describe the ommatidium of the lateral eye 
of Limulus as a group of modified ectodermic cells aggregated 
around and beneath the funnel-shaped depression in the 
skin. A glance at the diagrams (Pl. XIX, figs. 1 and 2) 
will show this clearly. The sensory cells of the ommatidium 
come into direct contact with the conical lens, which is the 
thickened part of the general cuticle; or, to express this in 
the phraseology of Lankester and Bourne, the ommatidium in 
the lateral eye of Limulus is “epistatic.’ The cornea 
and crystalline cone as such have no separate existence in 
this stage. 

Suppose such an ommatidium to become duplicated until a 
considerable number be formed, as we may safely imagine to 
have been the case, from the general tendency in the perfection 
of a visual organ. What will be the result? The first effect 
of such an increase in the number of ommatidia in a given 
area will be the lengthening of each unit in the direction of 
the ommatidial axis, and the cells (Pl. XIX, fig. 3, V.) 
which were situated directly on the outside of the retinule will 
travel over and above the sensory portion (R?. and G., fig. 3). 


MORPHOLOGY OF COMPOUND EYES OF ARTHROPODS. 147 


The distal ends of such (V.) which were thus pushed over will 
meet one another in the median or the “ optic axis” of the 
ommatidium; further, they will continue to secrete chitin 
(fig. 3, c.c.) from their original chitin-secreting surfaces which 
are now median and axial. The chitin thus secreted will 
have an independent existence from the cornea, thus 
forming the rudiment of the crystalline cone, and the cells 
themselves will form the vitrelle (figs. 3, 4, V., &c.). Finally 
as the deepening still further goes on, the corneal lens (C.) and 
crystalline cone (c.c.) will be entirely separated, thus producing 
a condition somewhat similar to that which obtains in Serolis 
(figs. 5 and 6). 

From this point onward, the three chitinous structures 
cornea, crystalline cone, and rhabdomere, undergo a 
different development in different Arthropods. In some the 
crystalline cone assumes a transparent semi-liquid state, while 
the whole cell becomes extremely elongated, forming the 
crystalline cone of certain crustacea (Pl. XIX, fig. 7); it 
may form a hard chitinous ball as in Serolis (Pl. XIX, 
fig. 6); or a cuneiform chitinous structure, as in Talo- 
chestia (Pl. XIX, fig. 9); or, finally, the whole cell may 
remain as a Clear, transparent body, as in several insects, form- 
ing Grenacher’s “ aconous type”’ of the compound eye. 

The forms assumed by the rhabdomeres in different Arthro- 
pods are equally diverse. The rhabdomere may exist as a 
plain cuticular covering over the non-pigmented part of the 
retinula, as in Limulus or in Serolis; it may become 
extremely elongated and narrow as in Musca or in Calli- 
nectes (Pl. XIX, fig. 8, Rd.); it may become transversely 
folded asin Cambarus (Pl. XIX, fig. 7, Rd.) ; these transverse 
folds may become still finer, showing the chitinous serrature 
along the axial edge of the retinula, as in Penezus and 
Homarus; or this transverse serrature may become ex- 
tremely fine and regular, as in Squilla. . 

The cornea undergoes equally diverse modifications accord- 
ing as it is purely protective, or partly protective and partly 
dioptric in function. The range of variation is shown by the 


148 S. WATASE. 


degrees of curvatures and by the varieties of its thickness. In 
several of the Decapod Crustacea which I have examined, as 
Peneus, Cambarus, Homarus, Callinectes, Gebia, 
&c., the curvatures of the individual cornea on both surfaces 
are very slight; it is biconvex in an extremely small degree. 
In Talorchestia both surfaces of the corneaare parallel. In 
Serolis, four species of which I have studied, all having well- 
developed compound eyes, there exists a considerable difference 
in different species in the nature of the cornea. In some the 
curvature on the proximal surface is very strong, and the whole 
structure is quite thick; while in others the cornea is rather thin, 
and a slight development of curvatures exists. This is interest- 
ing, showing that even within the group of nearly allied species 
there are considerable differences in this respect. 

This fact is easy to understand when we remember the func- 
tional property of the cornea and the crystalline cone. As has 
been noticed already, the crystalline cone is always dioptric in 
function, while the cornea may be partly protective and partly 
dioptric, or wholly protective. When the cornea becomes 
partly dioptric, as in Serolis and in several other Arthropods, 
the dioptric function in an individual ommatidium comes to 
be performed by two structures, the crystalline cone and the 
corneal lens. When the two structures act together for the 
same end at the same time, it is easy to see how a certain 
trivial peculiarity of the one may induce a correlative modi- 
fication of the other, and how a slight specific peculiarity may 
appear exaggerated in the thickness or in the degree of 
curvature of the corneal lenses in different species. 

After so much has been said in regard to the unity of 
structure of the ommatidium in different Arthropods, one 
important point awaits our consideration, viz. the homology 
and fate of the central ganglion-cell found in the ommatidium 
of Limulus. Unless a great many forms of ommatidia in 
different Arthropods be compared, a discussion on this point 
appears to be unprofitable. The consideration which follows 
is therefore a purely provisional one. 

There can be no question that the central ganglion-cell is an 


MORPHOLOGY OF COMPOUND EYES OF ARTHROPODS. 149 


important factor in the ommatidium of Limulus, nor can we 
doubt the existence of a fundamental homology between the 
retinule of Limulus and those of all the other Arthropods 
which I have examined. With the exception of a few pro- 
blematical bodies, such as the “hyaline cells” of Serolis, 
there are no structures in the ommatidia of most Arthropods 
which correspond to the central ganglion-cell of Limulus, in 
spite of the existence of a fundamental homology in the other 
elements of the ommatidium. 

What has become of the central ganglionic element of the 
ommatidium? Was it lost in the course of the phylogenetic 
history of a more complex ommatidium? Or is it reasonable 
to suppose that some ommatidia came into existence without it 
from the beginning? Or, if it were lost at all, is there any 
evidence which makes this supposition probable ? 

The colourless ganglionic cell and the pigmented rod- 
bearing cells which surround the former I consider as the two 
primitive morphological factors in the unit of the sensory part 
of the Arthropod retina, somewhat in the same way as the 
circle of rods with a cone in the centre are the two essential 
factors in the neuro-epithelial Jayer of the human retina. 
In the absence of enough comparative data in Arthropods at 
present we have to dwell largely on the analogy suggested in 
the other groups of animals. Whatever be the views as to the 
fundamental homology of the ommatidium of Limulus toa 
structural unit of the sensory part of the human retina, a 
superficial resemblance of the one to the other is certainly very 
strong. The structural resemblance is paralleled by a physio- 
logical one. The place where the light acts in the visual end- 
organ of Arthropods and of man may alike be considered as 
consisting of a number of definite groups of cells, each group 
being a morphological and a physiological unit; or, in other 
words, the sensory part of the retinz in both cases consists of 
a mosaic of several sensitive spots. The image formed on such 
a surface is therefore a mosaic one, whether in an Arthropod 
or in a Vertebrate. 

Fundamental as this arrangement appears to be in the 


150 S. WATASE. 


human retina, these two factors are liable to variation in their 
relative distribution in different Vertebrates. In fact, the 
variation takes place between the two extremes where the rods 
alone exist on one hand and where the cones alone constitute 
the essential part of the retina on the other. Thus, according 
to Schultze, “ either form of percipient element (rod and cone) 
may be represented by the other” in the Vertebrate. This 
range of variability in the distribution of the cones and rods 
occurs even in a single group of Vertebrates, as in mammalia, 
showing that the variation in the distribution of the essential 
factors, even within a tolerably well circumscribed group of 
animals, is sometimes quite extensive. The group of Arthro- 
pods is a heterogeneous one, and I see no a priori objection to 
believing in the existence of a phenomenon analogous to what 
we find in Vertebrates, viz. that the two percipient elements 
represented by the central and the peripheral cells in the 
ommatidium of Limulus may be differently represented in 
different Arthropods. 

There is no doubt whatever that the retinula cells are 
homologous throughout the Arthropods. In fact, in most 
Arthropods which I have examined no other elements but the 
retinule have any connection with the optic nerve-fibres, and 
they often undergo an enormous development and acquire 
most complicated structures, as in Homarus or in Peneus, 
giving rise to the much discussed “ spindle.” 

But what has become of the central element which is so 
conspicuous in the ommatidium of Limulus, if the retinule 
in all Arthropods are homologous? I believe the central cell 
is fully functional, judging from its position and from its 
veritable connection with optic nerve-fibre in Limulus. 
What in other Arthropods strongly reminds one of this cell is 
the “hyaline cell”? at the bottom of the ommatidial pit in 
Serolis and, according to Beddard, also in the Cymo- 
thoidz. One important difference, however, exists between 
the “hyaline cell” of the Isopods and the central cell in the 
ommatidium of Limulus, viz. that, while in the latter the 
cell is connected with the optic nerve, the “ hyaline cell” in 


MORPHOLOGY OF COMPOUND EYES OF ARTHROPODS. 151 


the former has no connection with the central nervous system 
whatever. Hence the “hyaline cell” cannot be sensory, even 
if it be homologous with the central cell of Limulus, which 
it resembles in its general appearance and in its position. 
The number of “hyaline cells” in Serolis is always two, 
while its supposed homologue in Limulus is, as a general 
rule, only one. This fact does not offer any objection to my 
view of their homology when we bear in mind that other 
elements in different ommatidia, as vitrelle and retinule, 
show a wide range of variation so far as their numbers are 
concerned, and yet they can be considered as perfectly 
homologous. 

A further embryological and comparative knowledge in 
regard to the ‘hyaline cell”? in Isopods is necessary for the 
determination of its exact homology. Meanwhile I would 
observe that if the central and the peripheral cells which we 
see in the ommatidium of Limulus may be taken as the two 
essential factors of the sensory element of the typical Arthropod 
retina, the case of Serolis may be taken as a loss of balance 
in the relative development of these two factors, the central 
cells having lost their sensory function and remaining as a sort 
of supporting mechanism. We can imagine this change in the 
function of the central cell as carried still further, and with 
the excessive development of the peripheral elements, the 
retinule, the central element may finally have disap- 
peared. 

All this is, however, a mere suggestion, and my interpreta- 
tion of the nature of the Arthropod ommatidium in general 
does not lose its force even if this section of my views in regard 
to the fate and homology of the central cell or cells be proved 
untenable. It is quite possible that the ommatidia in which 
there is no element corresponding to the central cell of 
Limulus may have originated without it from the beginning. 
It seems, however, more natural to suppose that such an 
ommatidium had it originally and lost it later, observing that 
the simplest form of ommatidium possesses it in its fully 
functional, sensory form. 


152 S. WATASE. 


Finally, we have to consider the nature of the compound eye 
as a whole as presented in various types of Arthropods. 

That a certain structure in the body of an animal may 
repeat itself and give rise to a secondary aggregate, or to 
a compound organ, is a well-known fact; the repetition of 
similarly constructed uriniferous tubules forms the essential 
part of a Vertebrate kidney, or the similar repetition of gill- 
filaments forms the respiratory organ of a Lamellibranch. 
Sundry other examples of this nature might be given, but the 
above two will suffice. Tracing, as I have attempted to do, 
the most complicated ommatidium into a simple, open, ecto- 
dermic pit, there is to my mind no difficulty in believing that 
the compound eye of the Arthropod is one of the most astonish- 
ing examples of the formation of an organ by the vegetative 
repetition of the similar structure. Thus, according to 
Lubbock, there are about 4000 facets in the compound eye of 
the house-fly (Musca), each facet corresponding to a single 
tubular invagination of the skin, the ommatidium. There are 
4000 independent invaginations in the area in the head of the 
fly occupied by the compound eye; in the gad-fly (Aistrus), 
7000; in the goat-moth (Cossus), 11,000; in the death’s- 
head moth (Sphinx atropos), 12,000; in a _ butterfly 
(Papillio), 17,000; in a dragon-fly (Aischna), 20,000; in a 
small beetle (Mordella), as many as 25,000. On the other 
hand, the number of ommatidia seems to have reached its 
minimum in certain Copepods, as in Coryczus, where the 
whole visual organ seems to be represented by a single colossal 
ommatidium. 

Certain forms of Collembola! seem to have a very small 
number of ommatidia; thus in Templetonia only one 
ommatidium exists on each side of the head; Orchesella 
has six on each side of the head; Tomocerus, Ipsoma, 


1 Lubbock, ‘Monograph of the Collembola and Thysanura,’ the Ray 
Society, 1873, p. 57, pls. lv and lvi. Lubbock uses the term “ ocellus” to 
designate a single element of the eye, which I here called an ommatidium, 
If the structure of this “ ocellus” differs from the ommatidium of other 
Arthropods, it has, of course, nothing to do with the discussion at issue. 


MORPHOLOGY OF COMPOUND EYES OF ARTHROPODS. 153 


have seven; Degeeria, Lepidocyrtus, Smynthurus, and 
Papirius, eight. In the ants we observe a similar gradation 
in the number of ommatidia. 

What reasons can we assign for this enormous multiplica- 
tion of similarly constructed parts? What advantage follows 
from this arrangement? If the view of the nature of the 
compound eye which is put forward in the preceding pages be 
a true one, Miiller’s celebrated theory of mosaic vision is the 
only one that can account for the enormous multiplication of 
the similarly formed pits in the skin. The subject has been so 
fully discussed by Lubbock that I need not enter into details 
here. “ According to his (Johannes Miiller’s) view, those rays 
of light only which pass directly through the crystalline cones, 
or are reflected from their sides, reach the corresponding 
nerve-fibre. The others fall on and are absorbed by the 
pigment which separates the different facets. Hence each 
cone receives light only from a very small portion of the field 
of vision, and the rays so received are collected into one spot 
of light. The larger and more convex, therefore, is 
the eye, the wider will be its field of vision; while 
the smaller and more numerous are the facets, the 
more distinct will the vision be. In fact, the picture 
perceived by the insects will be mosaic, in which the number 
of points will correspond with the number of facets.”?! The 
whole explanation of the problem seems to me to be contained 
in the passage above cited; and no further comment will be 
necessary more than a statement that the increase in the 
number of ommatidia is a decided advantage to their possessor. 
An eye like that of Limulus might bya slight change be 
converted into one of a more protuberant nature so as to 
command a wider field of vision, as we see in some species of 
Serolis or in some Trilobites; a slight change again might 
produce a protuberant ocular area mounted on an ophthalamic 
stalk, and accompanied by the accessory apparatus of vision, 
such as the socket for protection or the set of muscles to 
move the eye-stalk in different directions so as to command a 

Lubbock, ‘Senses, Instinct, and Intelligence,’ p. 163. 

VOL. XXXI, PART II],—NEW SER. L 


154 S. WATASE. 


still wider field of vision. In this connection I may refer to a 
series of diagrams (Pl. XIX, figs. 10-17). The black heavy 
layer represents the ectoderm, and the region in which the 
ectoderm is thrown into folds the area of the compound eye. 
The yellow-coloured layer outside represents the chitin, and 
the dotted line beneath the ectoderm the basement mem- 
brane. 

In Limulus (fig. 10) the ectoderm is thrown into a series 
of shallow folds, which, when viewed from above, would be a 
group of shallow pits in the skin. Each pit is an ommatidium. 
In Serolis (fig. 11) the invagination of the skin is a little deeper 
than that of Limulus, and the whole ocular area is more pro- 
minent. Fig. 12 represents the condition of the ectodermal 
folding in Notonecta, and fig. 13 thatof Agrionlarva. Fig. 
14 represents the eye of Branchipus, only a part of the stalk 
being shown in the figure. Fig. 15 represents the eye of 
Cambarus; fig. 16 that of Penezus; and fig. 17 that of 
Lucifer. 

It must not be understood that the number of folds given 
in the diagrams have anything to do with the actual number of 
ommatidia that may exist in the actual specimens; no more 
than a morphological expression of the eye in a simplest 
possible form was intended. If one suppose a single invagina- 
tion of the skin, say of fig. 15, to be divided into three strata 
and the cells in the bottom stratum to send out nerve-fibres, 
those in the middle to form the crystalline cone and those in 
the outermost to form the cornea (fig. 7), the interpretation of 
the diagram will be complete. 

According to this view the compound eyes of Arthropods, 
either in the sessile or in the stalked form, are nothing more 
than a collection of ectoderm pits whose outer open ends face 
towards the sources of light, and whose inner ends are con- 
nected with the central nervous system by the optic nerve- 
fibres. The cells forming the walls of the pit arrange them- 
selves into three strata, in most cases accompanied by three 
regional functional differentiations. Grenacher’s classification 
of the compound eyes of insects into “‘ acone,” “ pseudocone,” 


MORPHOLOGY OF COMPOUND EYES OF ARTHROPODS. 165 


and ‘‘eucone” types refer to the condition of the cells and 
their products in the middle stratum—the vitrelle. 

Morphologically, then, the compound eye of an Arthropod 
is strictly single-layered, although, as is evident, the pre- 
sent conception is entirely different from the monostichous 
theory maintained by some recent writers. From Limulus 
to Squilla we have a series of forms showing all degrees of 
modification in the general structure of the eye as well as the 
structure of its individual elements, and there is not here a 
single form which invalidates the view maintained in the pre- 
sent paper. Moreover, this view has the advantage of greatly 
simplifying our conception of these structures, reducing, as it 
does, all of them to one primitive structure, a depression in 
skin, in which several organs of ectodermal nature, often of a 
very complicated type, find their common morphological origin. 
And when thus the nature of the unit is reduced into a simple 
invagination of the skin, the formation of the compound eye 
appears to be but another instance of the well-known method 
in the formation of a morphological organ, namely, the vege- 
tative repetition of a similar structure. 


SUMMARY. 


In studying the structure of the ommatidium of the com- 
pound eye of Serolis it has been found that it may be 
reduced to a simple ectodermic invagination of the skin. 
Extending my researches over several other Arthropods, of 
which Talorchestia, Cambarus, Homarus, and Calli- 
nectes were mentioned in the preceding pages, the same inter- 
pretation of the ommatidium may be applied without excep- 
tion. This view of the ommatidium finds its strongest support 
in the fact that in Limulus the ommatidium is an open pit 
of the skin. 

By supposing that the ommatidial pit of Limulus became 
deeper, and that this was accompanied by modifications in the 
structure and arrangement of the component cells, we can 
show the probability of our first supposition that the omma- 


156 S. WATASE. 


tidium of the compound eye of an Arthropod is an independent 
invagination of the skin. If this view is correct, the unit of 
the compound eye of an Arthropod is not, after all, so com- 
plex a structure as has been supposed by some; and the 
enormous increase in the number of ommatidia in a given area 
of the skin which results in the formation of the compound 
finds its parallel in the well-known method of the formation 


of the morphological organs, viz. the duplication of a simple 
unit. 


DESCRIPTION OF PLATE XIX, 


Illustrating Mr. Watase’s paper on “ The Morphology of the 
Compound Eyes of Arthropods.” 


Fics, 1—5.—Diagrams showing the probable evolution of the three- 
layered ommatidium from the single-layered surface depression in the skin 
by the gradual subsidence of the neuro-epithelial elements, Aé. and G., Fig. 
1. In Fig. 2 the ommatidium of Limulus is represented, which is con- 
sidered a step further advanced from the condition shown in Fig. 1. The 
distal end of the retinula (2z.) instead of being pointed toward the exterior 
as in Fig. 1, in Limulus points towards the median axis of the omma- 
tidium. The chitinous substance being still secreted on the outside, a distinct 
body of chitin beneath the lens-cone (C.) is formed, the rhabdom (Ré.). In 
Fig. 3 this deepening is supposed to have gone still further, resulting in the 
formation of another independent chitinous body, the crystalline cone (c. c.). 
In Fig. 4 this deepening is considered to have advanced still further, the 
crystalline cone (c. ¢.) being entirely separated from the corneal lens (C) by a 
distinct stratum of cell, the corneagen (cg.). In Fig. 5 an ommatidium with 
three strata of cells, each secreting chitinous substance on the part of their 
surface, is formed. These three strata of cells are known as the corneagen 
(cg.), the vitrella (V.), and the retinula (Rt.). Three chitinous bodies 
secreted by each group of cells above mentioned are the cornea (C.), the 
crystalline cone (¢.c.), and the rhabdomere (Ré.), respectively. 


Fic, 6.—Serolis. Diagram of the ommatidium of Serolis. The general 
arrangement of cells in this is not very different from that shown in Fig. 5 
The place of ganglion-cell in Fig. 5, G., is taken by a pair (of which only 
one is shown in the diagram) of transparent “ hyaline cells” (4.). 


MORPHOLOGY OF COMPOUND EYES OF ARTHROPODS. 157 


Fic. 7.—Cambarus. This is introduced in comparison with the hypo- 
thetical ommatidia. 

Fic. 8.—Callinectes. 

Fig. 9.—Talorchestia. 


In the last three forms no element corresponding to the central ganglion 
cell of Limulus nor to the “hyaline cell” of Serolis can be found. The 
sensory element of the ommatidium is represented by the retinule (2R¢.) only. 


Fie. 10.—Limulus. Diagram of the compound eye of Limulus, the 
black, heavy line representing the ectoderm, and each depression in this layer 
corresponding to an ommatidium. 


Fic. 11.—Serolis. In the same way as the above, the eye of Serolis 
may be represented by a series of folds. 

Fic. 12.—Notonecta. 

Fie. 13.—Agrion (larva). 

Fie. 14.—Branchipus. 

Fie. 15.—Cambarus. 

Fie. 16.—Peneus. 

Fic. 17.—Lucifer. 


> 7 
7 7 Po - 
e451 aed : u Let Vee 1 oe a 


5 s = 
i U - p ~ 

rT nT) : " ‘es Fee Ay 

if off Dibeagt : Pi 
‘J while + P a 
4h4 i vo‘ [ap 2 


wet ae 

Lo bt eae 

. valine Ae 
tonsa tae 
tes cht we east a) 
toa tnt ie 
j ork lee ce 


STRUCTURE OF EARTHWORM OF GENUS DIACHMTA. 159 


On the Structure of a Species of Earthworm 
belonging to the Genus Diacheta. 


By 


Frank E. Beddard, M.A., 
Prosector to the Zoological Society of London. 


With Plate XX, 


I RECEIVED some time since, through the kindness of 
Mr. Windle, a number of examples of earthworms from the 
Bermudas. Some of these belonged to a species of Lum- 
bricus, while others seemed to be referable to the genus 
Urocheta; under this generic name I described a year ago 
in ‘Nature’ the remarkable characteristics of the sete of the 
hinder end of the body of the worm. 

1 believe now that the earthworms belong to Mr. Benham’s 
genus Diacheta (5), though probably representing a new 
species of that genus, which is at present only known by Mr. 
Benham’s account of the anatomy of Diacheta Thomasii. 

I cannot, however, be certain about this, as Mr. Benham’s 
description of this species is not complete, any more than is 
the description of the Bermuda form contained in the following 
pages. 

The worms measure about four inches in length, are com- 
paratively stout, particularly in the anterior region, comprising 
about the first ten segments, which is smooth and swollen. 

So far as I can see, there is no prostomium, but the mouth 
opens terminally, as in D. Thomasii! and in Urocheta. 

1 T have been as careful as possible in ascertaining this, because I find 


that I have made a mistake in stating that Thamnodrilus has a terminal 
mouth and no prostomium ; Mr. Benham suggested the possibility of an error 


160 FRANK E. BEDDARD. 


There is some little difficulty in counting the anterior seg- 
ments, for the reason that the first two or three appear to be 
retractile, as they are stated to be in Urocheta. 

Moreover, the first two are very narrow (see Pl. XX, fig. 1), 
and although there is a distinct furrow separating them,! their 
total diameter is less than that of the 111rd segment. From 
this point onwards, the only difficulty in mapping the segments 
accurately is caused by the fact that a number of the 
anterior segments of the body are entirely un- 
provided with sete. In three specimens, which were par- 
ticularly examined with regard to this point, the first five 
segments were entirely without setz, so that the first seti- 
gerous segment is the 6th body-segment. 

So far as I am aware, this is the only instance among the 
Oligocheta of so large a number of segments whose setz have 
disappeared ; in certain forms, such as Nais, the sete of the 
first few anterior segments are considerably reduced in number 
by the disappearance of the dorsal bundles, while in Cheto- 
gaster there is apparently a number of segments without any 
sete intercalated. between the first and second setigerous 
segments. 


in my description, having discovered a long prostomium in an example of a 
worm which seems to be identical with my Thamnodrilus Gulielmi (2); 
on re-examining my specimens I find that a prostomium is present in that 
species, and that I actually figured if (2, woodcut, fig. 2, p. 157) in its 
retracted condition. The ornamented sete are, moreover, not confined to the 
clitellum, but are found all over the body, though the ridges upon them are 
very much less marked, and quite escaped my attention. It is therefore 
evident that my Thamnodrilus is identical with Rhinodrilus. It is 
possible that Anteus does not generically differ from either of these forms. 

1 Tam not at all prepared to state positively that there is really this divi- 
sion. In longitudinal sections the supposed division looked of no more 
importance than the division between the annuli of the succeeding segments ; 
the brain is as usual placed between two segments, or rather near to the 
posterior boundary of the segment containing it. If this is reckoned the 
third, as in other earthworms, then the supposed two segments will be 
really only one; but then we shall have the apparently anomalous position of 
the testes and vas deferens funnels in the 1xth and xth, instead of in 
the xth and xith segments. 


STRUCTURE OF EARTHWORM OF GENUS DIACHATA. 161 


The sete of this Diacheta are remarkable for three 
different reasons. 

In the first place, they are irregularly disposed, and the 
irregularity commences, as in D. Thomasii, from the very 
first, and continues to the end of the body ;' to the end of the 
xith segment the sete are grouped in two lots on either side of 
the median ventral line, as shown in fig. 1, but their 
arrangement does not exactly agree in any two consecutive 
segments; further back there are four sete separated from 
each other by considerable intervals, and two closely approxi- 
mated on each side to form a pair; towards the tail end 
the paired arrangement is again lost, and the irregular dis- 
position of the sete is returned to. Fig. 8 illustrates this 
arrangement. 

Another remarkable fact about the setz of this species of 
Diacheta is that in the first three (setigerous) segments, at 
any rate, they are not situated along one line running 
parallel with the intersegmental furrow; they have a 
curiously alternate arrangement as seen in the figure (fig. 1), 
some being situated further forwards, others further back in the 
segment to which they belong ; they form, in fact, almost 
a double row. It might, perhaps, be supposed that this 
appearance is due to the simple fact that I have confounded 
two segments and regarded them as one; and this supposition 
is strengthened by the fact that in two out of the three seg- 
ments a faint groove divided the segment into two. Never- 
theless, I regard this groove as the division between two 
annuli; it is frequently the case among earthworms that the 
anterior segments, when large, are marked by one or more 
annulations; moreover, in Segments vi and vit the sete, 
although showing the curious arrangement referred to, are only 
eight in number. The 1xth segment, on the other hand, 
in one of the three specimens, was certainly fur- 
nished with nine set; indeed, I actually counted eleven; 
but it is possible that two of these, being placed each very close 


1 The irregularity in the arrangement of the sete is much like what 
Schmarda (11) has described in his genus Pontoscolex. 


162 FRANK E. BEDDARD. 


beside another seta, may represent only setz ready to replace 
the others; but as they were perfectly mature, and appeared to 
protrude from the body, it is at least equally possible that this 
segment was furnished with two pairs of sete in addition to 
seven scattered sete. I did not observe anywhere else more 
than eight setz in the segment, though their arrangement in 
two ill-defined rings was very noticeable on several of the 
anterior segments. In Pericheta Houlleti I have occa- 
sionally seen a similar dislocation of the ring of setz, but in no 
other forms, though I have not yet specially investigated the 
point. These various facts with regard to the sete have some 
bearing upon, though of course I do not pretend that they 
entirely explain, an important question in Annelid mor- 
phology. 

One of the most remarkable facts in the structure of this 
group is the varying position of the different organs in 
different genera and families. For example, in most earth- 
worms the testes are in x, x1, and the ovaries in x11; in 
Nais, on the other hand, the testes and ovaries are respec- 
tively developed in the vth and vith segments. Are we to con- 
sider in such cases that the testes of one form are perfectly 
homologous with the testes of another genus which are placed 
in a different segment? or are we rather to regard them as 
‘serially homologous,’ believing that any and every segment 
can develop testes, which are actually developed here in one 
case and there in another? A third possible alternative is to 
consider that Nais and Lumbricus have been evolved along 
different lines from a common ancestor, in which there were at 
least six pairs of testes, occuping Segments v, VI, VII, VIII, Ix, 
x, x1. If either of these views were true, we might expect to 
meet with abnormal specimens with a larger number of testes; 
but this is not the case—such individuals have not been met 
with.! Connected with this view as to the reason for the 


1 It is true that Lumbricus is often provided with what appear to be an - 
additional rudimentary pair of ovaries in Segment xu. But this is hardly a 
case in point, inasmuch as the normal presence of the ovariesin Phreoryctes | 
and EKudrilus seems to indidate, as do other considerations, that two ovaries 


STRUCTURE OF HARTHWORM OF GENUS DIACHATA. 168 


varying position of important organs, and apparently confirm- 
ing it, is the observed fact that the growth of the worm—the 
addition of new segments—takes places at the tail end. It is 
common to find worms in which the last segment, or the last 
two or three, have no sete at all, or fewer than the normal 
number, while the internal organs show a corresponding em- 
bryonic condition. In spite, however, of these facts, which 
are undoubtedly true so far as they go, another mode of for- 
mation of new segments occurs in earthworms. Fritz 
Miiller (6), in describing the anatomy of Urocheta core- 
thrura (termed by him Lumbricus corethrurus), called 
attention to the invariable presence, at a constant distance from 
the clitellum, of a group of segments evidently newly formed, 
for the reason that they were unprovided with sete. Here, 
therefore, is an example of an intercalary growth of seg- 
ments; but as it takes place in the least differentiated part of 
the body, it is perhaps to be regarded as being essentially 
different from the process of formation of new segments at the 
caudal end. 

The structure of Diacheta appears to me to be 
suggestive in the light of the hypothesis of an in- 
tercalary growth of segments at the anterior end. 

It has been mentioned that the first three segments are very 
much narrower than those which follow, and they can be 
apparently retracted into the buccal cavity ; it may be, there- 
fore, that they are in course of disappearance—the initial steps, 
i.e. the disappearance of the setz andthe reduction in size, 
having already occurred. The opposite interpretation is of 
course possible. But whatever may be the value of these facts, 
the arrangement of the setz on the vith, vith and virrth seg- 
ments (or vth, vith and vith) seems to be inexplicable except on 
the hypothesis of the intercalary production of new segments. 

Throughout the body the sete, although they are so far 
irregular in their arrangement that they do not correspond in 
position in successive segments, are nevertheless regular, in that 


corresponding to the two testes on each side may be the original condition in 
the Oligocheeta. 


164. FRANK E. BEDDARD. 


they are disposed along a line encircling the segment; in 
two individuals (the only ones examined from this point of 
view) this regularity in the first three setigerous segments 
was lost. The sete of each segment were placed in 
two quite distinct lines, separated by a slight 
furrow. This abnormal state of affairs may perhaps seem 
capable of being explained by the simple theory that the 
supposed “ segment” was in reality two segments. But in 
this case there should be sixteen setz, whereas I counted them 
very carefully and only found a total of eight, except in one of 
the three segments of one individual; here there were eleven 
setz, distributed fairly equally between the two annuli of the 
segment. On first noting these facts (which, as they occurred in 
three individuals selected at random, must surely be fairly con- 
stant) it appeared possible that there was in this earthworm 
just a trace left of a primitive condition, in which the sete, 
like the nephridial pores, were scattered irregularly. There 
is, however, no evidence that the setz have, like the nephridia, 
been derived from some unsegmented aucestor. If this be the 
case, there is no reason to suppose that the setze were not dis- 
posed in a perfectly metameric fashion from the first. 

It seems, therefore, that these facts, if they have any mean- 
ing, point to the conclusion that the scattering of the setz is 
a preliminary to the formation of two segments out of one; it 
does not seem likely that the reverse interpretation—viz. the 
nearly complete fusion of two segments—is the right one. If 
this be allowed, we then have a means of under- 
standing the varying position of certain important 
organs in allied genera. 

The third point in which the sete of Diacheta are 
remarkable is their very unusual degree of specialisation. 
Examples of the various set are illustrated in figs. 3—7, all 
carefully drawn to scale by the help of the camera lucida. 

On the first few segments of the body the setz are all of the 
form indicated in fig. 6: there is a long shaft ornamented 
with raised arches, like the sete of Rhinodrilus, and the 
clitellar setee of Urocheta; the distal end of the setz is 


STRUCTURE OF EARTHWORM OF GENUS DIACH®TA. 165 


sharply curved; these sete are found on about the first 
ten segments; after this they still occur, being sometimes 
(fig. 6a) of very large size, but are mingled with sete (fig. 5) 
of the usual lumbricid pattern, and of which no special 
description is necessary ; these sete gradually come to be the 
only only ones present, and towards the end of the body their 
free extremities are much more hooked. Benham has re- 
marked (5, p. 90) that in Diacheta Thomasii “the extre- 
mity is more strongly curved than in the ordinary setz of 
Lumbricus.” 

At the tail end the setze become enormously enlarged (figs. 
3, 4); they are so large that their amber-yellow colour can 
be distinctly recognised by the unarmed eye, and their strongly 
recurved apices catch in the skin when the worm is held by the 
tail, and produce the impression of some sticky substance. 
In a recent number of ‘ Nature’ I have referred to the pro- 
bable use of these setz, viz. to give the worm a stronger hold 
upon the ground when it is lying outside its burrow with only 
the tail concealed. These large hooked sete, although all are 
larger than the setz of the segments in the middle of the body, 
vary much in size, as will be seen from an inspection of 
figs. 8,4. On the whole, they increase in size towards the 
end of the body. The last few segments have fewer sete ; 
their numbers in the last few segments of one individual were 
as follows—8, 7, 7, 7. 

Though most earthworms are furnished with two kinds of 
setee, I am not acquainted with any case which is quite com- 
parable to the present. It is remarkable to find that the 
sculptured sete of the anterior segments become less nume- 
rous upon the clitellum than they are in the segments in front 
of it, for the reason that the similar sete of Urocheta and 
Rhinodrilus, two genera evidently allied to Diacheta, 
are either confined to or better developed upon the clitellum. 

The clitellum occupies a large number of segments, per- 
haps as many as in D. Thomasii (5), but I am not able to 
state the exact number. 


So much, then, for external characters, the description of 


166 FRANK E. BEDDARD. 


which has occupied an unusual amount of space; this worm, 
however, possesses a more interesting exterior than is gene- 
rally the case. 


INTERNAL ANATOMY. 


It appears from Mr. Benham’s account of the internal seg- 
mentation of Diacheta that there is some difference between 
his species and mine in the segments occupied by the various 
organs. There are six strong intersegmental septa (fig. 9), the 
first bounding the vith segment ‘posteriorly, and the last inter- 
vening between the xith and x11th segments ; the position of 
these is the same as in D. Thomasii, but there is one more. 

The spermatheca are situated a segment further back, 
the first (of the three pairs) being in Segment vu, behind 
instead of in front of the first thickened intersegmental 
septum. Benham mentions the presence of a single pair of 
long ‘‘tongue-shaped” sperm-sacs attached to the anterior 
wall of Segment xur. I find two pairs of sperm-sacs, both 
small, and confined to the x1th and x11th segments respec- 
tively. The vas deferens funnels are in Segments x, x1; 
only one pair of these organs are mentioned by Benham, 
those occupying the xith segment. The testes are in the 
same segments, and occupy the usual position, i.e. they are 
attached to the front of the segment opposite to the funnels 
of the vasa deferentia. 


§ Integument. 


There are two points in the structure of the body wall which 
require notice. The first point is the absence of those peculiar 
structures in the epidermis which have been regarded as 
possibly abortive setz; it is important to mention their 
absence in Diacheta, for the reason that they are present in 
Urocheeta, which isin many other respects so closely allied to 
Diacheta. 

In all the segments of the body, commencing with the vith 
(the first setigerous segment), the circular muscular layer is 


STRUCTURE OF EARTHWORM OF GENUS DIACHATA. 167 


interrupted in the middle line, as shown in the figure (fig. 15) ; 
a number of muscular fibres of different appearance from those 
out of which the circular muscular layer is formed; they are 
of less diameter, and not composed of a bundle of closely 
united fibrils. These few fibres are shut off in a compart- 
ment which runs completely round the segment, and which is 
partially divided up by a fine network in the meshes of which 
lie the individual fibres. This peculiar specialisation was to be 
seen in the anterior segments of the body commencing with 
the vith ; it is possibly this which gives rise to the appearance 
of annulation already referred to as seen in the anterior 
segments. 


§ Nephridia. 

The nephridia do not seem to differ much from those of 
D. Thomasii. ‘The first six segments are occupied by a large 
pair of nephridia, which differ much in minute structure from 
the “mucous glands” of Urocheta, although agreeing with 
them in being very much larger than the following nephridia; 
the apertures of these nephridia appear to be upon the rvth seg- 
ment, and all the following segments are provided with a pair. 
The external apertures are placed very far forward in the seg- 
ments, quite inthe intersegmental furrow; they are furnished 
with the remarkable cup-like structure (fig. 14) at their 
external aperture, which has hitherto been only found in 
Urocheta; it seems to be clear that the structures are 
muscular, and perform the office of a sphincter; the muscular 
fibres are chiefly disposed in a radial direction, but outside 
is a thin layer of circularly disposed fibres, which, though 
fewer, are thicker than the radial fibres, and may be of equal 
strength. 


§ Alimentary Tract. 


There are two noticeable peculiarities about the alimentary 
canal of this earthworm: the first is the presence of a large 
thin-walled crop of equal diameter with the gizzard, into 
which it opens immediately; the gizzard itself is situated in 


168 FRANK E. BEDDARD. 


the vith ring, just in front of the first of the specially thickened 
intersegmental septa. The second peculiarity concerns the 
calciferous glands, or rather their equivalents; for, like Mr. 
Benham, I have been unable to find any calciferous glands 
like those of Urocheta. The csophagus is a narrow tube 
with greatly vascularised walls, there being apparently a con- 
tinuous blood sinus below the lining epithelium. Its inner 
wall is raised into numerous irregularly arranged folds ; in the 
xuith segment it suddenly increases in diameter, and exhibits 
a remarkable structure, which is illustrated in Pl. XX, figs. 9g/., 
10ca., 11. The folds which are distinctive of the cesophagus 
become greatly increased in depth, and at the same time 
regularly arranged ; each fold consists of two layers of epithe- 
lium, enclosing in the space between them a blood channel. 
The structure of this part of the gut, which occupies three 
segments, is not unlike that of the calciferous glands in 
Acanthodrilus (Beddard 8); its epithelium is not ciliated. 


§ Vascular System. 

The vascular system of this worm is distinguished by the 
enormous size of two pairs of “ hearts,’ which unite the dorsal 
and ventral vessels in the xth and xith segments. Both in 
dissection and longitudinal sections these vessels were seen to 
be gorged with blood. The interior of each vessel was dis- 
tended with a coagulated mass of blood, which, however, did 
not consist of a uniform yellow-coloured clot; but, as shown 
in the accompanying figure (Pl. XX, fig. 18), contained 
scattered through the blood certain curious structures, con- 
cerning the nature of which I feel rather doubtful. The two 
pairs of hearts opened each by a relatively very narrow opening 
into the two longitudinal trunks. The orifice of communica- 
tion was in each case guarded by a well-marked valve, which 
consists of a mass of elongated cells (see Pl. XX, figs. 12, 13), 
evidently a special growth of cells which line the blood-vessel 
throughout. It is interesting to note the valve between the 
heart and the dorsal vessel (fig. 13) projected into the former, 
while in the case of the communication between the heart and 


STRUCTURE OF EARTHWORM OF GENUS DIACHHTA. 169 


the ventral vessel (fig. 12) the valve projected into the ventral 
vessel, thus showing the course of the blood to be from the 
dorsal vessels through the hearts into the ventral vessel, as 
in Lumbricus. 


§ Nervous System. 

The position of the cerebral ganglia, as already mentioned, 
is somewhat anomalous. They lie close to the posterior 
boundary of a segment which, if the oviducts are to open on to 
the xivth, and the other organs of the reproductive system to 
be normally placed, must be regarded as the rvth, i.e. a seg- 
ment further back than is usual. The cerebral ganglia give 
off two intertwined bundles of fibres to the pharynx, which 
represent the stomatogastric system, usually developed in 
earthworms. There are not many data regarding the minute 
structure of this stomatogastric system. In Megascolides 
Spencer describes and figures (12) this system, and expressly 
notes the absence of ganglionic corpuscles. In longitudinal 
sections of Diacheta, the presence of numerous ganglion- 
cells in the branches forming the stomatogastric system was 
quite obvious. 


§ Testes, Sperm-sacs, and Vasa Deferentia. 


In Diacheta Thomasii there are a pair of extraordinarily 
long sperm-sacs attached in front to the septum separating 
Segments x1, x11, and extending back through more than 
twenty segments. Similar sperm-sacs have been described in 
Urocheta (10), and, though much smaller, by myself in 
Typheus (4). I find, however, in Diacheta two pairs of 
these organs in the x1th and x11th segments (fig. 9,vs), not at all 
large, though containing abundant developing spermatozoa. It 
is possible that the single pair of sacs described by Benham may 
be the result of a fusion between two sacs on each side; but 
the matter requires further study, and, in the meantime, 
Benham only discovered one pair of vas deferens funnels. 
Two pairs of these organs were quite obvious in longitudinal 
sections, and, corresponding to them, two pairs of testes in 

VOL, XXXI, PART II,—NEW SER, M 


170 FRANK E. BEDDARD. 


the Diacheta investigated by myself. There is nothing in 
the structure of these orgaus that calls for particular remark. 


§ Ovaries and Oviducts. 


There is a single pair of ovaries in Segment x111; the ovi- 
ducts open into the same segment and on to the exterior on 
Segment xiv, on a line with and between the ventral sete ; 
aperture is distinct. 


§ Spermathece. 


There are three pairs of these organs in Segments vit, 
VIiL, 1x. 

The first spermatheca, although lying in Segment vu, opens 
on to the exterior in Segment vi; its duct perforates the 
thickened intersegmental septum separating these two seg- 
ments, and opens on to the exterior distinctly in front of the 
intersegmental groove. This fact appears to me to be of 
some little importance, and for the following reasons. Ben- 
ham has mentioned that in D. Thomasii the sperma- 
theca are in Segments VI, VII, VIII, 1.e. a segment farther 
forward than in the present form; but the apertures are, 
according to Benham, placed posteriorly in each segment ; 
accordingly, the present species, though differing in many par- 
ticulars from D. Thomasii, and, among others, in the fact 
that the spermatheca are in vii, viII, 1x, instead of v1, v1, v1, 
agrees in the important fact that the external aperture has 
remained in the same place, uninfluenced by the slight alte- 
ration of the segmentation; a very slight shifting in the 
attachment of the intersegmental septum, such as occurs in 
many worms—for example, in Hormogaster (Rosa, 9, 11)— 
would place the spermatheca entirely in the vith segment. 

I have myself pointed out an analogous change in the posi- 
tion of the spermatheca of Allolobophora complanata. 
In this case, therefore, it is clear that the spermathece of 
Diacheta are homologous in both species, and that their posi- 
tion is only apparently and not really changed. 


STRUCTURE OF EARTHWORM OF GENUS DIACH@TA. 171 


It is clear from the above very brief account of the organi- 
zation of Diacheta that, as Mr. Benham pointed out, it is 
closely allied to Urocheta. 

But the two species of Diachzta differ from each other in 
most of their points of agreement with Urocheta. In both 
species there is no prostomium, and the set alternate in posi- 
tion from segment to segment as they doin Urocheta, though 
the alternation begins from the very first in Diacheta, and 
not until later in Urochezta; there are five strong septa 
commencing behind Segment vi, and three pairs of simple 
spermathece. Diacheta Thomasii agrees with Urocheta 
in having only a single pair of testes, sperm-sacs, and 
vasa deferentia. 

Diacheta Windlei, as I desire to name my species, 
agrees with Urocheta in that the clitellum commences at 
Segment xv, and in the mass of muscles which surrounds the 
aperture of the nephridia. 

In the following table the principal points of resemblance 
and difference between the three forms are shown :— 


172 FRANK E. BEDDARD. 


DIACHETA DIACHETA 
UrocHzta. THOMASII. WINDLEI. 
Setz . {Irregular in distri-|[rregular in distri-|First five segments. 
bution after Seg-| bution from the| without sete, irre- 
ment x; f-shaped,| first. _f-shaped,| gular in distribution 
with bifid extre-| with pointed ex-| on remaining seg- 
mity. Some of cli-| tremity. Clitellar) ments. Sete highly 
tellar sete orna-| sete not different?) specialised, there 
mented with ridges being three forms : 
(1) simple “shaped, 
(2) ornamented sete 
| as in Urocheta, 
(3) large hooked 
sete. 
Epidermic Present Absent Absent. 
glands be- 
tween setzx 
Prostomium ./Absent Absent Absent. 
Clitellum xv (xvi)—xxm 0 (xx—xxxmI xv—? 
& pore »|KIX/XX XXII ? 
Atria. . Absent Absent Absent. 
Destes.. . One pair in XI One pair in x1? =| Two pairs in x, XI. 


Vasa deferen- One pair in x1 
tia funnels 
Sperm-sacs 


One pair in xI Two pairs in xX, XI. 

. One elongated pair|One elongated pair|Two pairs (small) in 
extending from x1] extending from XI] Xt, XII. 
to XIV to XXXVIII 

Simple sacs without|Simple sacs without/Simple sacs without 
diverticula in vut,| diverticula in v1,| diverticula in vu, 
VIII, 1X (opening| VII, VIII (opening VIII, Ix (opening at 
at anterior border} at posterior border| posterior border of 
of segment) of segment) Segments VI, VII, 

VIII). 
. Anterior pair form-|Anterior pair form-|Anterior pair form- 


Spermathece 


Nephridia . 


ing a branched] ing a “mu-| ing a “mucous 
“mucous gland”| cous gland,’ not| gland,” not branch- 
opening on Seg-| branched, opening| ed, opening on Seg- 
ment u. First) on Segment 1.) ment iv. Ordinary 
pair of ordinary} Ordinary nephri-| nephridiacommence' 
nephridia in Seg-| dia commence in| in following seg- 
ment iv. External} Segment tv. No| ment. Orifices 
opening surround-| sphincter. guarded by asphine- 
ed by a “ sphinc- ter. 
ter ” 
Posterior Present Absent Absent. 
glands con- 
nected with 
nephridia 
Alimentary |Three pairs of cal-/No  calciferous {No calciferous gland, 
tract ciferous glands | glands but a portion of in- 


testine(in Segments 
xlI—xIv) with a 
similar structure. 


STRUCTURE OF EARTHWORM OF GENUS DIACHATA. 173 


It may perhaps be considered that the points of difference 
enumerated in the above Table are sufficient to distinguish 
generically all three forms; but I defer the discussion of this 
matter until something is known about the structure of other 
allied forms, which may occur in the West Indies. It is to 
be noted, however, that the species, which for the present I 
refer to the genus Diacheta, although much specialized in 
the shape of the setz, and in the loss of the setz of the ante- 
rior segments, connects in some ways (e.g. in the structure 
of the generative organs) the genus Urocheta with allied 
forms, such as Geoscolex. 


List oF MEMOIRS REFERRED TO. 


1. Bepparp, F. E.—‘‘ On the Tail Bristles of a West Indian Earthworm,” 
‘Nature,’ vol, xxxix, p. 15. 
2. Bepparp, F. E.—‘‘ On the Structure of a New Genus of Lumbricide 
(Thamnodrilus Gulielmi),” ‘Proc. Zool. Soc.,’ 1887, p. 154. 
8. Bepparp, F, E.—‘ On the Specific Characters and Structure of Certain 
New Zealand Earthworms,” ‘ Proc. Zool. Soc.,’ 1885, p. 810. 
4, Bepparp, F. E.—‘“‘ Note on some Earthworms from India,” ‘ Ann. and 
Mag. Nat. Hist.,’ ser. 5, vol. xii, p. 213. 
5. Brennam, W. B.—‘‘ Studies on Earthworms,” No. 2, ‘ Quart. Journ. 
Mier. Sci.,’ vol. xxvii, N. S., p. 77. 
6. Miuter, Fritz.—‘On Description of a New Species of Earthworm, 
Lumbricus corethrurus,” ‘Ann. and Mag Nat. Hist.,’ ser. 2, 
Vol. xx, p; lo: 
7. Perrier, H.—“ Recherches pour servir alhistoire des Lombriciens 
terrestres,” ‘Nouv. Arch. d. Mus.,’ t. viii (1872). 
8. Perrier, H.—‘ Organisation des Urocheta,” ‘Arch. d. Zool. Exp.,’ 
t. ni (1874), p. 331. 
9. Rosa, D.— Sulla Struttura dello Hormogaster Radii,” ‘Mem. R. 
Accad. Sci. Torino,’ t. xxxix, ser. 2, p. 3. 
10. Rosa, D.—‘‘ Lombrichi raccolti nell’ isola Nias, &c.,” ‘Ann. Mus. civ. 
Genoa,’ ser. 2a, vol. vii. 
11. Scumarpa.— Neue wirbellose Thiere,’ Bd. i, pt. ii. 
12. Spencer, W. B.—“The Anatomy of Megascolides australis,” 
‘Trans. Roy. Soc. Victoria,’ vol. i, pt. i, 


174 FRANK E. BEDDARD. 


DESCRIPTION OF PLATE XX, 


Illustrating Mr. Frank E. Beddard’s paper “ On the Structure 
of a Species of Earthworm belonging to the Genus 
Diacheta.” 


Diacheta Windlei. 


Fic. 1.—Ventral view of anterior segments, to illustrate arrangement of 
sete and their disappearance from the first five segments. 

Fic. 2.—Front view of first segments, to show absence of prostomium. 

Figs. 8 and 4.—Hooked sete of posterior segments. 

Fic. 5.—Sete of middle segment. 

Fic. 6.—Sete of anterior segments. a. Seta of larger size. 

Fic. 7.—Extremity of one of these sete, more highly magnified to show 
ridges. 

Fic. 8.—Tail end of body. s. Large hooked sete. m. Muscles moving 
them. JZ. Last segment. z. Intersegmental groove. 

Fic. 9.—Semi-diagrammatic longitudinal view. c. Brain. xph. Aperture 
of anterior nephridia. g. Gizzard. sp. Spermathece pore. 7¢. Testes. vs. 
Sperm-sacs. g/. Calciferous gland. v. d.f. Funnels of vasa deferentia. 
ov. Ovary. od. Oviduct. The segments are numbered. 

Fic. 10.—Longitudinal section through calciferous gland. @. @sophagus 
s. Septa. ca. Gland. 

Fic. 11.—A portion of ditto more highly magnified. yp. Peritoneal cover- 
ing. J, 7. Muscular layers. 4/. Blood-space. ep. Epithelium. 

Fie. 12.—Section through point of opening of heart into ventral vessel. 
h. Heart. 0. Ventral vessel. v. Valve. 

Fie. 13.—Similar section through opening of heart into dorsal vessel. 
h. Heart. 0%. Dorsal vessel. v. Valve. a. Fibrous matter contained in 
blood. 

Fic. 14.—Sphincter (sp/.) at aperture of nephridium (xp.). 

Fie. 15.—Section through body-wall in middle of Segment vir or vir. ep. 
Epidermis. ¢7. Transverse muscles, /. Longitudinal muscles. z. A tract of 
transversely running fibres enclosed in a separate fibrous sheath. 


HEKATEROBRANCHUS SHRUBSOLII. L75 


Hekaterobranchus Shrubsolii. 
A New Genus! and Species of the Family 
Spionide. 


By 


Florence Buchanan, 
Student of University College. 


With Plates XXI and XXII. 


THis worm was found at Sheppey by the members of the 
University College Biological Society during an expedition 
made therein July, 1889. It appears to have been already 
known to naturalists living at Sheppey, but no one had tried 
to identify it. Not being able to find any published account 
of it, I believe it to be as yet undescribed, and have therefore, 
at Professor Lankester’s kind suggestion, undertaken the exa- 
mination and description of it.? 

Occurrence.—The worm was always found associated with 
Haplobranchus (described by Dr. Bourne in the ‘ Quart. Journ. 
Micr. Sci.,’ 1883), and occurs therefore in soft mud at the 
bottom of gullies, usually overlain by an inch or so of water. It 
is not so tenacious of life as Haplobranchus, and is hence not 
always to be found in mud containing Haplobranchus. Its 


1 See, however, note at the end of this paper. 

2 I have been greatly helped in my investigations by the kindness of Mr. 
Shrubsole, of Sheerness, who has sent me up from time to time, as I required 
it, fresh material. I will take this opportunity of thanking him for the kind 
way in which he has allowed me to encroach upon his time and patience ; for 
collecting and searching through mud to see that a particular animal, and that 
a very minute one, is present in it is no very easy nor interesting task. 


176 FLORENCE BUCHANAN. 


other associates are Nais littoralis, Hemitubifex (Clitel- 
lio) ater, nematodes, and planarians. It is, however, more of a 
marine form than its associates, since, after heavy rain at low 
water, itis, Mr. Shrubsole informs me, seldom to be found, while 
the other forms of life may be still abundant. When present it 
can, as a rule, be recognised readily by its nematode-like move- 
ments and red colour, and the four tentacles waving on its head. 

It is usually from about 6 to 10 mm. in length, the size 
varying according to the number of segments. It is, therefore, 
slightly larger than the Haplobranchus. It forms loosely 
coherent tubes by gathering up particles of mud round it, but 
inhabits each only for a very short time. It is more frequently 
to be found moving about in the mud. 

Anatomy.—The number of segments varies. I have 
never counted more than forty-eight, and the greater number of 
specimens examined had between thirty and forty. The body 
is divided into regions which, as in other members of the 
family, are not so distinctly marked off from one another as in 
most sedentary annelids. 

Cephalic Region.—The Ist or head-segment has a 
well-developed prostomium, on which are two well-marked 
pairs of eye-spots, one pair more dorsal and median than the 
other. In two out of the many specimens examined there 
were eight eye-spots, not, however, arranged as four pairs, 
but scattered and at very unequal distances apart. In another 
specimen there were five eye-spots, three on one side and two 
onthe other. It is not unusual for the number of eyes to vary 
individually in marine annelids; it is, indeed, usual for the 
number to be greater in the larva than in the adult; and it would 
therefore seem that the eight-eyed condition is to be explained 
rather as a retention of a larval feature, than as due to the divi- 
sion of the four eyes normally found in the adult. 

Behind the eye-spots, at the base of the prostomium, between 
it and the body of the Ist segment, are the cephalic ten- 
tacles, each containing a single contractile blindly-ending 
vessel (P]. XXI, figs. 1 and 2,7¢.). They are richly ciliated 
all round, the cilia not being confined, as in most other 


HEKATHROBRANCHUS SHRUBSOLII. E77 


members of the family, to a single longitudinal groove. The 
tentacles have an annulate appearance, due to slight surface 
ridges on which are the cilia, and to greenish-yellow streaks 
crossing the tentacles here and there. Between the ridges are 
short, stiff, tactile hairs. The contractile vessel (figs. 2 and 12) 
lies freely in the cavity of the tentacle which is part of the 
celom, and in which, in transparent specimens, ccelomic cor- 
puscles can be seen. The tentacles are situated more laterally 
than in most members of the family: they are placed on either 
side of the mouth, and slightly above it. When the animal is at 
rest they are bent forwards in search of food, and infusorians 
may be seen carried down by their cilia to the mouth. When 
the animal is moving and tosses its head, the tentacles stand 
up more or less vertically ; or, when it is moving in a definite 
direction, they are bent back over the dorsal surface, reaching 
back usually to the 3rd or 4th segment. 

Behind these tentacles, which, for want of a better name, 
I have merely called “ cephalic,” and dorsad of them, situated 
on the body of the Ist segment, is a pair of organs with the 
characteristic structure of Spio branchiz, although a great deal 
larger than these usually are (figs. l and 2, dr.). They are 
about half as long again as the “ cephalic ”’ tentacles, and of a 
reddish-orange colour, due to the presence of an ascending and 
descending blood-vessel, forming together a simple loop in each. 
They are ciliated, but the cilia are shorter than they are on the 
“cephalic ” tentacles, and they do not appear to be ciliated quite 
allround. The vessels, not being contractile, are not readily 
seen except in section (fig. 10). They run close to the epidermis, 
projecting into the cavity of the branchia which is a prolonga- 
tion of the celom. The one vessel is rather larger in calibre 
than the other. Jike the “cephalic” tentacles, they may 
either be carried erect, or bent back over the dorsal surface. 
Their length, also, varies much individually. Usually when 
bent back they would cover the first five segments ; sometimes, 
however, they reach back over more than eight. At the base 
of each branchia are two or three short capillary chete (fig. 1, 
and fig. 12, ntp!.) 


178 FLORENCE BUCHANAN. 


On the same segment (the lst), placed ventro-laterally, 
almost vertically below but a little behind each branchia 
(fig. 1), is another group of three or four rather longer capil- 
lary chzetze, behind and below each of which is a membranous 
lobe—the ventral ‘cirrus’ of most authors, the neuropodial 
“ lamina” of others. 

The body of the first segment reaches further forward on 
the ventral than on the dorsal surface, and is there folded, 
forming a kind of ventral collar (figs. 1 and 3, ». coll.). 
This fold can be traced up laterally to the base of the branchie, 
which appear to be attached to it. 

Thoracic and Abdominal Regions.—On all the other 
segments of the body there are, as on the first, two groups of 
cheetze on each side, but the cheete are longer (when of the 
same kind) and more numerous than on the Ist segment. 
Their number varies in different individuals and in different 
segments of the same individual, but with no constancy. 
Five, six, and seven are usual numbers, but sometimes there 
are aS many as nine ina group. Seeing that they may so very 
easily be knocked off, and that new ones may always be form- 
ing, not much importance can be attached to their exact 
number in different segments and in different individuals. 
In the dorsal groups throughout the whole length of the body 
the cheete are capillary only (fig. 4, a.). In the ventral groups 
they are so also in the anterior region of the body; but from 
the 8th segment onwards there are, as well as these, also 
hooked or crotchet cheete. We may, therefore, consider the 
thoracic region to extend as far as the 7th segment (in- 
clusive), and the abdominal region to beginin the 8th. There 
are at first two or three crotchets to about four or three 
capillary cheete. More posteriorly there are usually about 
five hooked chete to two capillary ones. LKach crotchet (fig. 
4, c.) is three-toothed at the extremity, the one tooth being 
larger and more prominent than the other two, so that in 
some views it alone is to be seen clearly (fig. 4, c.'). The 
hooked extremity is surrounded by a membrane. 


1 The name “cirrus” would imply a homology with the cirrus of the 


HEKATEROBRANOHUS SHRUBSOLII. 179 


The chete all arise from sacs, each group from one sac, 
firmly implanted in the body-wall and projecting into the body- 
cavity, though the ventral sacs do not project so far as the 
dorsal. The wall of each sac is supplied by muscles, by means 
of which the sac can be moved in and out asa whole. Springing 
separately from the body-wall behind both dorsal and ventral 
chetze-bundles slightly dorsad, of the dorsal ones and ventrad 
of the ventral ones, are the membranous lobes known either as 
“cirri” or ‘‘ parapodial lamine.”! They can readily be seen 
in the first few segments, but then gradually grow smaller, 
and are not found in the posterior region of the body. It is 
difficult to determine exactly in which segment they cease to 
exist, and whether this is constant in all, since, to see them 
clearly, the animal must be living and moving, and it is then 
not easy to count them. When the animal is killed the lobes 
become, by the position taken up by the worm, very difficult 
to see and be certain of. The dorsal ones are much closer to 
one another, i. e. nearer to the median line, anteriorly than 
posteriorly, and in the 2nd segment (the first one in which they 
exist) they are so close together that they seem to form, or form 
part of, a collar, which is therefore dorsal, and quite distinct 
from the ventral collar of the 1st segment (figs. 1 and 12, d. coll.). 
There are very minute stiff hairs on all these lobes, resembling 
cilia, but without their movement. Such hairs are also found 
elsewhere on the cuticle of the body-wall. 

Internal Anatomy.—The bod y-wal] consists of—(1) An 
outer epidermic layer of cells,in parts more than one layer 
thick, with a fine cuticle (figs. 5, 6, 7, and 8, epid.). The 
epidermis is thicker on the ventral surface than elsewhere. 
parapodium of an Errant annelid, e.g. Nereis or Phyllodoce ; and it is difficult 
to say whether this homology exists without first deciding, by the comparison 
of a large number of forms, to which families of the Errantia the Spionide are 
most nearly allied, taking the Spionidee to be, as they probably are, the living 
representatives (though probably degenerate) of the most primitive of the 
Sendentaria. A cirrus may vary so much both in form and position that we 
can see no reason why these membranous lobes in the Spionide should not 


represent cirri; but this, of course, does not alone in the least prove them 
to be true cirri. 


180 FLORENCE BUCHANAN. 


Here and there in the epidermis, and occupying its whole 
thickness, are a few large coarsely granular cells with well- 
marked large nuclei and nucleoli, probably opening to the ex- 
terior, and secreting the material by which the animal holds 
the fragments together which compose its temporary tube. 
(2) The circular muscular layer (c. m.) is only very slightly 
developed, and can scarcely be seen except in longitudinal 
sections. It is best developed in the ventral region just 
over the nerve-cord (where there are no longitudinal ones), 
and can there be seen in transverse sections. (3) The 
longitudinal muscular layer, on the other hand, is very well 
developed, running in three bands, one dorsal and two ventral 
(figs. 5, 6,7, and 8, d./. m. and v./.m.). Although a single band, 
the dorsal one is much more feebly developed in the median 
line than on either side. (4) Below this again is a delicate 
layer of ceelomic epithelium, forming the outer wall of ccelom, 
and only to be distinguished by a few nuclei scattered here 
and there on the extremities of the muscle-fibres (figs. 5, 6, 7, 
and 8, c. ep.). 

Coming from and dividing the dorsal longitudinal muscles 
on either side, and stretching vertically downwards to be 
attached close to the thickened portion of the epidermis of 
the ventral surface on either side, are, in the anterior region 
of the body, i.e. from the 2nd to the 6th segments, very 
distinct dorso-ventral muscles (fig. 5, d.v.m.), dividing 
the cavity of each of these segments more or less completely 
into three longitudinal chambers. 

Besides these there are in every segment muscles going 
from the ventral epidermic thickening on each side to the two 
setal sacs (s. s.m.), but these appear to be rather continuations 
of the circular than of the longitudinal layer. Both these and 
the dorso-ventral muscles are covered by a delicate layer of 
celomic epithelium. 

Alimentary Canal.—The mouth is not terminal, but is 
overlapped by the prostomium (fig. 1, m.). The two “‘cephalic” 
tentacles, as already mentioned, arise just above it on either 
side. The pharynx extends through the first two segments 


HEKATEROBRANCHUS SHRUBSOLII. 181 


(fig. 2). Its anterior part is evertible and richly ciliated. It 
is almost always extruded at once when the animal is first 
compressed by acover slip (fig. 3, 8. ph.). The pharynx narrows 
in the posterior part of the 2nd segment to form the 
cesophagus, which is continued through the next few segments. 
The canal then gets much wider, and begins to be constricted 
intersegmentally by the septa. The segment in which this 
change from cesophagus to intestine takes place varies with 
the size of the individual. Posteriorly, i.e. in the posterior 
third or fourth of the body (again varying according to the 
size of the individual), it narrows again, and this part espe- 
cially is exceedingly contractile. The anus is terminal (fig. 
2). From it cilia can be seen moving upwards towards the 
mouth, indicating thereby some anal respiration. In some 
specimens, but not in all, ciliated ridges could be seen in the 
intestine just in front of the anus (fig. 11), probably the same 
thing as the richly ciliated swelling found in the larve of allied 
forms. ‘The alimentary canal is lined throughout by columnar 
epithelium, consisting of cells one layer deep, ciliated in the 
pharynx and esophagus, and also in the hinder unconstricted 
part, of the intestine, but apparently not in the anterior con- 
stricted part, which occupies the greater length of the body. 
This epithelium is much folded in the anterior region, especially 
in the pharynx (fig. 5, mt. ep.), not so much in the third 
and fourth segments, but again in the hinder wsophageal 
region. It is not folded, and the lumen of the canal is wide, 
in the anterior intestinal region (fig. 6); afterwards it again 
becomes folded to some extent (fig. 7). The cells forming the 
folds are longer and narrower than the others (figs. 5 and 7), 
but their nuclei, as in the other cells, are situated close to the 
peripheral wall, all the nuclei together forming a very regular 
circular layer. 

Outside the epithelial layer is a very thin circular muscular 
layer, best seen in the anterior region of the body (fig. 5, c. m”.), 
but not seen at all distinctly posteriorly, though its presence 
would seem to be indicated by the muscular contractions of the 
whole alimentary canal. There are no longitudinal muscles to 


182 FLORENCE BUCHANAN. 


be seen, and directly outside the circular muscle layer comes the 
celomic epithelium. Neither of these-last two layers takes any 
share in the folds. 

In one specimen which I had, which was evidently a young 
form with only about twenty segments, the alimentary canal 
was wide, and constricted intersegmentally in all the anterior 
segments of the body as far back as the 10th. Between the 10th 
and 11th segments was a deep permanent constriction, the canal 
continuing very narrow throughout the rest of the length of the 
body. This would seem to imply that the pharyngeal and 
cesophageal region of the alimentary canal developed late.!. In 
this specimen there was green pigment all down the sides of 
the alimentary canal, not, as far as I could see, enclosed in any 
way. There were also no thoracic nephridia. 

Like so many other Chetopods, this one has almost con- 
stantly present parasitic monocystes in its intestine, and these 
are often very numerous. They are broad at one extremity 
(apparently the anterior), and usually pointed at the other 
(fig. 13). The cortical substance forms a clear zone, wider at 
the anterior extremity. The medullary substance is coarsely 
granular, and in it, reaching to the posterior extremity, is usually 
a long narrow vacuole (vac.), which may sometimes be found 
bursting. Sometimes they have no vacuole, and such I at 
first mistook for eggs, until finding that they were in the 
alimentary canal and not in the ceelom. They may be seen 
moving backwards and forwards with the intestine, apparently 
incapable, while in the body at least, of any independent motion 
of their own. The nucleus (z.) is spherical and well marked, 
containing a nucleolus. 

Vascular System.*—There is a contractile dorsal vessel 

1 But it may be that the cesophagus is developed, but resembles the part 
of the intestine following it in being intersegmentally constricted, since this 
appears to be the case in the larva of what is probably a Spio or Nerine 
described by Leuckart in the ‘ Arch. f. Naturg.,’ 21st Jahresg., 1855, p. 63, &e., 
and pl. ii, fig. 1. Here, however, I did not observe anything marking off the 
two regions of the alimentary canal from one another, as Leuckart describes in 


his larva. 
2 The whole arrangement of the vascular system is not easy to determine, 


HEKATEROBRANCHUS SHRUBSOLII. 183 


(figs. 2 and 12, d.v.) in the anterior region of the body, con- 
tinued forwards into the prostomium. Just before it reaches 
the prostomium two vessels are given off, one to each branchia 
(d. br. v.). These run up the inner sides of the branchiz, and 
return by vessels on the outer side (v. 67. v.), which meet in 
the median line on the ventral surface in the posterior part of 
the first segment to form the ventral vessel (v. v.).1 Before they 
meet each appears to give off or be joined by the single con- 
tractile vessel going to the “cephalic” tentacle (¢.v.). The 
ventral vessel runs throughout the whole length of the body 
(figs. 2, 5, 6, 7, 8, v. v.), passes in the anal segment into a 
sinus (figs. 2, 7, 8, sz.) surrounding the intestine, and lying 
just outside the epithelium, probably between it and the cir- 
cular muscular layer; or it may be that the circular muscular 
layer is really absent in this region, and that the sinus lying 
between the intestinal and the ccelomic epithelium of the ali- 
mentary canal has some contractile power of its own, as it 
has in other sedentary annelids, e. g. Spirographis,? where, 
however, muscular fibres are present as well. This sinus com- 
pletely surrounds the intestine in the whole of its posterior 
non-constricted part, and is at first continued over part of the 
constricted part; then, however, a nucleated mass appears 
inside it on the median dorsal line of the wall of the intestine, 
and forms a longitudinal upstanding ridge. Part of the sinus 
closes in round this ridge, and becomes nipped off from the rest 
of the sinus (figs. 2 and 6, d.s. v.), and so is continued forwards 


and what is given in the text is only what appears to me—after the examination 
of numerous living specimens and series of sections—to be its probable 
distribution. When living the animal is too opaque, when dead the vessels 
are seldom in the same state of contraction or expansion in two individuals. 
The vascular system in the highest animals even is subject to individual 
variation, and it may be that there are really slight individual variations in its 
arrangement in worms, and in this amongst others. 

1 The direction in which the blood flows in the branchie cannot be deter- 
mined, as the vessels cannot be seen in the living. It probably may flow in 
either direction, from the ventral to the dorsal at one time and from the 
dorsal to the ventral at another. 

2 Claparede, 1873, ‘ La Structure des Aunélides sédentaires.’ 


184 FLORENCE BUCHANAN. 


on the intestine, the ridge inside it being separated from the 
intestinal epithelium by a very fine layer of coelomic epithelium 
only. Some series of sections would seem at first sight to 
show that the ridge was in its posterior part directly continuous 
with the intestinal epithelium ; but a more careful examina- 
tion leads rather to the conclusion that it is formed by the 
tucking-in of the ccelomic epithelium which lies outside the 
sinus on either side. It lies, however, especially posteriorly, 
exceedingly close to the intestinal wall. Its significance 
(whether physiological or morphological) is as difficult to deter- 
mine as that of the so-called “* Herzkorper” or “ cardiac body” 
of certain other Polycheets,! which it in all probability repre- 
sents. No lumen is to be seen in it here throughout its course. 

In the cesophageal region the nipped-off upper part of the 
sinus enclosing the longitudinal ridge (fig. 2) leaves the walls 
of the alimentary canal, and becomes the contractile dorsal 
vessel which runs upwards until it comes to lie just beneath 
the thin part of the body-wall in the dorsal median line, 7. e. 
where the longitudinal muscle layer is only very feebly deve- 
loped (fig. 5,d.v.). It is here surrounded by a well-developed 
circular muscular layer (¢. m’.) to which its contractile power is 
due. The walls of all the other vessels and of the sinus appear to 
consist only of coelomic epithelium. It is difficult to say what 
happens to the rest of the sinus (which is continued throughout 
the intestinal region) when the dorsal vessel finally leaves the 
wall of the alimentary canal in the cesophageal region. It 
certainly is not continued as a sinus, but whether it forms vessels 
or not is a difficult point to determine, since there are other very 
much coiled vessels in each segment of the cesophageal region. 
These coiled transverse or dorso-ventral vessels seem to me to 
connect the dorsal and ventral vessels (as shown diagramma- 
tically in fig. 12), but it may be that they connect the ventral 
not with the dorsal, but with lateral vessels which are continua- 
tions forwards of the sinus, lying, for some part of their course 
at least, close to the dorsal vessel. ‘The dorso-ventral vessels 

1 See Cunningham, ‘Some Points in the Anatomy of Polycheeta,” this 
Journal, vol. xxviii, 1887. 


HEKATEROBRANCHUS SHRUBSOLII. 185 


all lie freely in the ceelom. They are represented in the first 
segment by the vessels going to the branchiz. In the posterior 
region of the body, z. e. where there is the sinus, it is difficult to 
say whether transverse vessels are present or not. In some 
series of sections vessels may be seen here and there leaving 
the upper part of the remaining sinus where the dorsal vessel 
is just nipped off. More posteriorly, where the sinus is con- 
tinuous all round the intestine, vessels may sometimes be seen 
running from the ventral vessel (fig. 8, v.v.). These do not 
appear to occur regularly in every segment, and they cannot be 
seen. at all in some series of sections which show the other 
parts of the vascular system clearly. But it is difficult to say 
whether they are really not present, or whether they are merely 
contracted, and therefore not seen, or not recognised as blood- 
vessels. We should not, therefore, be justified in concluding 
that the sinus represents them, although this would seem not 
unlikely in the most posterior region where the sinus is com- 
plete. In other Polychets where there is a sinus (e. g. Scali- 
bregnia, Trophonria, Eumenia) transverse vessels running to it 
from the ventral vessel are long and well marked.' 

The blood flows from behind forwards in the sinus and 
dorsal vessel, from in front backwards in the ventral vessel. 
It is probably aérated both at the anus and in the branchiz on 
the head-segment, and also to some extent in the ‘‘ cephalic” 
tentacles. It would be interesting to note whether all forms 
that have a sinus round the intestine have also other indications 
of an anal respiration. The blood is red, coloured probably 
by hemoglobin. It contains, as far as I have seen, no cor- 
puscles. 

Celom and Nephridia.—The celom is partially divided 
into separate cavities by the septa, which are thin muscular 
partitions between the segments coated with ccelomic epithelium 
on either side. They move backwards and forwards with the 
intestine. There is a dorsal and ventral mesentery supporting 
the intestine (fig. 7), and thus dividing the ccelom longitudi- 

1 See A. Wiren, ‘ Beitrage ziir Anat. u. Hist. d. Anneliden. Konigl. Sv. 
Vet. Akademiens Handlingar,’ Bd. xxii, No. 1. 

VOL. XXX1, PART II],—NEW SER, N 


186 FLORENCE BUCHANAN. 


nally into two halves. Besides this there are in the anterior 
region the three longitudinal chambers separated from one 
another by the dorso-ventral muscles. 

There are nephridia of two kinds. In the anterior (tho- 
racic) region of the body there are at once seen in the living 
(fig. 1) two green tubes, one on either side of the alimentary 
canal. On further examination each is seen to be bent on 
itself, and cilia may be seen moving in it, especially well seen 
at the bend of the tube which is in the posterior part of the 
6th segment. As far as I can make out from examination of 
the living and from sections, the opening to the exterior is 
between the second and third ventral bristle bundles, in the 
hinder part of the 2nd segment. By analogy we should expect 
the internal opening to be in the septum dividing the 1st from 
the 2nd segment. Whether this is so or not I am unable to 
say ; I can trace the lumen of the internal limb in longitudinal 
sections up into the 2nd segment to the level of the second 
pair of bristle bundles, but it is difficult to trace further. It may 
be that the septum is temporarily bent back, so as to lie partly 
within the 2nd segment. In transverse sections the internal 
limb (fig. 5, meph. ¢.) is not at all easy to see and to trace, since 
it lies almost in the dorso-ventral muscles or is obscured by 
them. The external limb (fig. 5, neph.e.) lies below the 
internal one on either side of the ventral vessel, with it in the 
middle one of the three longitudinal cavities shut off by the 
dorso-ventral muscles. Both limbs consist of simple drain-pipe 
cells. ‘These nephridia are probably excretory in function.’ 

1 Such thoracic nephridia in other sedentary annelids have been called 
“tubiparous glands” by Claparéde and others; but it is more probable, as 
has been pointed out by Cosmovici, Soulier, and Brunotte (as quoted by 
Meyer in the ‘Zool. Mith. v. Neapel’ for 1888), that it is the unicellular 
glands of the epidermis, not the thoracic nephridia, which secrete the material 
for fixing together the particles of mud or sand of which the tube is formed, 
since worms from which the thoracic region of the body has been entirely 
removed can still form tubes, and since the tube does not begin to be formed 
until after the development of the unicellular glands. In favour of this view 
is the fact that, in forms most nearly allied to the one we are here considering, 
which are more tubicolous in habit, there are not these modified thoracic 
nephridia. 


HEKATEROBRANCHUS SHRUBSONII. 187 


In the young specimens above mentioned (p. 182) these 
tubes were not to be seen, showing probably that they also 
develop late with the cesophagus. 

The second kind of nephridium is found in the abdominal 
region of the body only of those individuals in which the 
gonads are developed, a single pair in each segment in which 
there are gonads. In such individuals they may be seen very 
distinctly in transverse section (fig. 8, xeph.). They are very 
short, simple, uncoiled, ciliated tubes (fig. 14). But here, 
again, it is very difficult to say with absolute certainty whether 
they lead through a septum from one segment to the next, or 
whether they lie wholly in a segment. 

In individuals in which there are no genital products 
present they are, if represented at all, at any rate functionless, 
and with no lumen. They serve, therefore, as genital ducts. 

Genital Organs.— The sexes appear to be distinct, 
though I am not sure that I have seen any specimens with 
ova. As is usual in marine annelids, the generative products 
develop only at certain seasons of the year, and at other times 
the males and females are indistinguishable. In living speci- 
mens which I examined in the summer I thought I saw eggs, 
i.e. I saw bodies resembling eggs, but forget whether I dis- 
tinctly saw them in the ceelom, or only inferred them to be 
there. They may, therefore, have been only parasites. Un- 
fortunately, thinking that I was sure to get plenty more with 
eggs, I did not preserve or cut sections of any of them. 

The sperm-mother cells are oval or spherical, with well- 
marked nuclei which may be seen dividing. Masses of them 
may be seen in the ripe male individual on either side of the 
intestine just above the nephridium, and attached to the 
ccelomic epithelium surrounding the sinus of the intestine in 
all the hinder abdominal segments (fig. 8, ¢es¢.). Spermatozoa 
with long tails may also be seen. Together they occupy 
almost the whole cavity of the coelom in the region where they 
are developed. In the one ripe male individual of which I 
was able to cut sections, which was a specimen with thirty- 
five segments altogether, the gonads (and consequently the 


188 FLORENCE BUCHANAN. 


nephridia) were present in the posterior twenty-two segments, 
i.e. from the fourteenth to thirty-fifth inclusive. 

Nervous System.—There is a supra-cesophageal ganglion 
nearly filling the prostomium (fig. 2, gzg.), and probably sup- 
plying the much-thickened epidermis of the anterior region of 
the prostomium, the eyes, and the anterior pair of tentacles.! 
From this a commissure goes down on either side to join the ven- 
tral nerve-chain, which runs throughout the whole length of 
the body as a double cord in the much-thickened epidermis of the 
ventral surface (figs. 5, 6, 7, and 8, 2. c.). The two cords are 
distinct from one another, although very close together. There 
are no ganglionic swellings on them. Very minute giant-fibres 
(‘neural canals,” “ fibres tubulaires,’ ‘ neurochords”) may 
be made out by careful staining in each cord on its dorsal and 
inner side. In sections stained with hematoxylin each 
appeared as a hollow tube, containing a shrunken homoge- 
neous mass inside (figs. 5, 6, and 7). In other sections, stained 
with borax-carmine (fig. 8), the giant-fibres were more difficult 
to distinguish from the rest of the nerve-cord, the homogeneous 
mass not having shrunk away from its sheath. <A similar posi- 
tion for these structures has been noted in the anterior region 
of Nerine foliosa, Sars, and in Scolecolepis vulgaris, 
Johnst., amongst the Spionide. In Prionospio there are also 
two neural canals, but these are inferior in position.? In forms 
belonging to other families the same position of the giant-fibre 
with regard to the nerve-cord is found, e. g.? Arenicola (Telethu- 
side), Trophonia (Chlorhemide), Sabellaria (Hermellide). 

Affinities.—Hekaterobranchus I take to belong to the 
family Spionide on account of (1) the single pair of tentacles 
containing a single blindly-ending vessel; (2) the branchie, 
each containing an afferent and efferent vessel not connected 
with one another by capillaries; (3) the very superficial posi- 

1 Jacobi, ‘ Polydoren d. Kieler Bucht,’ 1883, p. 23. 

* M'‘Intosh, “On the Structure of the Body-wall in the Spionide,” ‘ Proc. 
Roy. Soc. Edin.,’ vol. ix, pp. 1283—129. 


3 See Cunningham, “Some Points in the Anatomy of the Polycheta,” 
this Journal, vol. xxviii. 


HEKATEROBRANCHUS SHRUBSOLII. 189 


tion of the nerve-cord, the distinctness of the two cords from 
one another, and the absence of ganglionic swellings. 

Tt differs from all other genera of the family Spionidz in the 
possession of only one pair of dorsal branchiz.! These are on 
the first or head-segment. They are found in this position as 
well as on the following segments in some species of some of 
the other genera (e.g. Spio fuliginosus,? Scolecolepis 
vulgaris, and Scolecolepis cirrata®). The single pair 
of branchiz of Hekaterobranchus are much larger and more 
developed than are any of the numerous branchie of other 
Spionide. This reduction in number of the branchiz, their 
increase in size, and their position on the head together seem 
to indicate that the worm once led a more sedentary life than 
itnow does. Other facts leading to the same conclusion are— 
the presence of a ventral collar, the single pair of modified 
thoracic nephridia, the reduction of the parapodia, and, 
as in all other sedentary annelids, the possession of crotchet 
cheetee. 

It is true that most genera of the family Spionidz, with 
a much greater number of dorsal branchie, inhabit now 
much more distinct tubes than Hekaterobranchus. It 
would appear, therefore, either that they have not yet dege- 
nerated so far as Hekaterobranchus from the ancestral Spio, 
or that Hekaterobranchus has developed other modes of respira- 
tion which other Spionids have not, e.g. the anal respiration 
and that of the cephalic tentacles. For although in other 
genera the cephalic tentacles do serve as respiratory organs, 
they do so to a very slight extent only: they are ciliated only 
on one side, and they serve mainly as prehensile and tactile 
organs; whereas here they are ciliated all over, and probably 


1 See, however, note at the end of this paper. 

2 Claparéde, ‘Ann. Chet. du G. de Naples,’ 1868, part ii, pl. xxiii, fig. 1. 
In the text (p. 63) Claparéde says the branchiz begin on the second segment ; 
but this is evidently a slip, as, in his definition of the species (p. 62), he says, 
rightly enough, they begin on the first setigerous segment; and here, as in 
most, if not all the Spionide, the first segment is setigerous. 

* Malmgren, ‘ Ann. Polych.,’ 1867, pl. x, figs, 544 and a’, 


190 FLORENCE BUCHANAN. 


serve to a much greater extent for respiration, though retain- 
ing their other functions as well. 

The ventral collar, usually taken as a characteristic of the 
family Serpulide, has not been actually mentioned as present 
in the Spionide. Judging, however, from a very incomplete and 
imperfect figure given with the Report on Annelids by Webster 
and Benedict in the ‘Commissioner’s Report of Fish and Fish- 
eries for the United States’ in the year 1881, there would appear 
to be a collar of the same sort in Streblospio Benedicti.} 

Large modified thoracic nephridia are found in several 
families of sedentary annelids, e.g. the Terebellidze, Her- 
mellide, Serpulide, and Cirratulide, sometimes a single 
pair only, sometimes two or even three pairs, and sometimes, 
as in the Serpulide, one pair with a common opening to the 
exterior. They have not been described in other genera of 
the family Spionide. 

Thesinus round the intestine hasalso not,as far as lam aware, 
been described in other Spionids, but a vascular plexus occurs 
round the intestine in some forms, e.g. Nerine cirratulus? 
A sinus is found in the Serpulidz, Chetopteride, Ariciide, 
Terebellide, and in many of the Cirratulide amongst others, 
but cannot be regarded as of much classificatory importance. 

The dorsal collar of the 2nd segment is not found in other 
Spionids,’ nor in other families; but in some Spionids, e.g. 
S piophanes Verrilli (described in the same paper by Webster 
and Benedict), there is a membranous ciliated dorsal ridge con- 
necting the bases of the opposite so-called “cirri”? on every 
segment from the 6th backwards, and this may be something 
of the same kind. 

We see, therefore, that Hekaterobranchus has many charac- 
ters in common with the Serpulidz,? and I think there is good 
reason to regard it as the degenerate descendant of a form 


1 See note at the end of this paper. 

2 Claparéde, ‘ Annélides sédentaires,’ 1873. 

3 It also has many characters in common with the Cirratulidee which oth 
Spionids have not. This family is notably closely allied to the Spionide, 
being probably an earlier and more primitive offshoot than the Serpulide. 


HEKATEROBRANCHUS SHRUBSOLII. 191 


from which the ancestors of the two families (Spionide and 
Serpulidz) have been derived. It would appear to be nearest 
to the tribe Amphicoridz of the family Serpulidee. 

In order to grant that Hekaterobranchus does thus connect 
the two families we must follow Meyer in his recent and 
very interesting paper on the homologies of the branchiz of the 
Serpulidze (‘ Mitth. Zool. St. v. Neapel,’ vol. viii, 1888), and 
grant that the cephalic branchiz of the Serpulide are develop- 
ments of the cephalic tentacles of the Spionide. The branchiz 
of the Serpulids have been clearly shown to be prostomial 
organs, both in development and in innervation. The ten- 
tacles of the Spionids are, according to Leuchart,! and Leuck- 
art and Pagenstecher,? prostomial in origin. According to 
Jacobi (‘ Polydoren der kieler Bucht, Inaugural Dissertation 
zur Erlangung der Doctorwiirde, 1883, pl. ii, fig. 27, p. 23) 
they are innervated from the prostomial ganglion. So far we 
should have no difficulty in deriving the complex branchiz 
of Serpulids from the simple tentacles of Spionids. Were 
the simple remaining pair of dorsal branchie of Hekatero- 
branchus to disappear, still more work would be thrown on 
the cephalic tentacles; and an organ with these important 
functions (respiration, prehension, and tactile sensibility) local- 
ised in it would be subject to great variation, and would 
consequently develop rapidly. Thus we should find these 
organs first multiplying, but remaining simple as in Haplo- 
branchus and Manayunkia,? then each becoming more com- 
plex and giving off secondary rays, as in Fabricia and Amphi- 


1¢Arch. f. Naturg.,’ 21st yr., p. 63, &., pl. i. Leuckart says they 
develop on either side of the ‘‘ Kopfhécker ” (? = prostomial crest), between 
it and the enlarged ‘“‘upper lip.” This “upper lip,” he says, marks the 
boundary between the pro- and peri-stomium ; and where the tentacles join it 
there is a group of long cilia on either side, probably representing the remains 
of the cephalotroch of the larva. 

2 §Miiller’s Archiv,’ 1858, pp. 610—613, pl. xxiii. Here it is said that 
the middle ciliated band of the larva separates the body into two halves, from 
the anterior of which the prostomium with the tentacles is formed, and from 
the posterior of which all the body-segments are formed. 

3 Leidy, ‘Proc. Acad, Nat. Sci. Philadelphia,’ pp. 204—212, pl. ix. 


192 FLORENCE BUCHANAN. 


glena; and, finally, both primary and secondary rays multi- 
plying greatly until we get the complex condition of other 
Serpulids (tribes Sabellide, Serpulide proper, and Eriogra- 
phide) ; and in all these forms, even in the most elaborate, 
the same structures can be traced. 

In Haplobranchus and Manayunkia there is a single pair 
of tentacles, with the same single blindly-ending contractile 
vessel running up them as in Hekaterobranchus and other 
Spionids.!. The other tentacles, which I would regard as 
multiplications of this one, have not as yet the contractile 
vessel developed in them; but their cavity is, as I have been 
able to ascertain from sections of Haplobranchus, in continuity 
with that of the tentacle containing the blood-vessel some way 
above the base. 

In Amphiglena and Fabricia all the branchial rays, and 
not a single pair only, have the contractile vessel in them. 
That they also still retain their tactile function is shown 
by the fact that each ray is non-ciliated, but provided with 
short, stiff, tactile hairs at its apex. In all the other Serpulids 
there is the same single contractile vessel ending blindly in 
each secondary ray.” 


1 Bourne calls these “palps” in Haplobranchus, but says their homology 
is difficult to determine. They are certainly, as confirmed by sections, ventral 
in position. lLeidy says that they are dorsal in position in Manayunkia, but 
apparently this has not been confirmed by sections. If there is this difference 
in position in what would at first sight (cf. the figures of Bourne and Leidy) 
appear so very evidently to be the same thing, it seems to me that if would 
go far towards showing that all the tentacles on the head of either form are 
developments (multiplications) of one, it being indifferent in which one the 
original characteristic contractile vessel develops. 

2 It will be seen from the above that if I accept Meyer’s premises, I do 
not agree with him in his conclusions with regard to the relationships of the 
Serpulide inter se. That is to say, I do not regard the tribe Amphicoride, 
to which Haplobranchus, Manayunkia, Fabricia, and Amphiglena belong, as 
degenerate from higher existing tribes, but rather as primitive; i.e. I regard 
these forms as the descendants (degenerate undoubtedly in many ways) of a 
form more primitive than the ancestors of any of the other existing tribes. 
No doubt, as Meyer remarks, they once led a much more sedentary life than 
they now do; but it does not therefore necessarily follow that they are 


HEKATEROBRANOCHUS SHRUBSOLII. 193 


Granting, therefore, that the cephalic tentacles of the 
Spionidee are prostomial, there would seem to be little or no 
doubt of their homology with the branchie of the Serpulide. 
But may we grant this? The transverse section of Nerine 
given by Claparéde in his ‘Structure des Annélides séden- 
taires,’ 1873, pl. xv, fig. 1, would seem directly to contradict 
Jacobi’s figure and explanation already referred to. Sete are 
never found on the prostomium ; yet, according to Claparéde, 
there are sete at the base of the cephalic tentacles in Nerine 
cirratulus.! Also, according to Claparéde (‘ Beobactungen it. 
Anat. u. Entwickelungsgesch. wirbelloser Thiere,’ 1863, 
p- 71), and Claparéde and Mecznikow (‘ Zeitsch. f. w. Zool.’, 
vol. xix, pp. 172 and 177), they develop not from the pro- 
stomium but from the peristomium. Claparéde evidently never 
thought of these tentacles as being anything but peristomial, 
and a good many of the figures in his ‘ Annélides Cheetopods 
du Golf de Naples (1868) ’ would seem to point to the fact. 
He very seldom, however, gives here a ventral or lateral view, 
and it is therefore very difficult to determine what is truly 
prostomial. From dorsal views only one is very apt to mistake 
the crest which is developed on the prostomium, but also some- 
times on one or two of the segments as well, for the pro- 
stomium. Whether Claparéde would have come to different 
conclusions had he had the question before him is of course 
impossible for us to say, but he is, as a rule, such an accurate 
observer that, without further examination of the same forms 
that he describes, we are not, I think, justified in concluding, 
as Meyer does, that this was a “ Beobachtungsfehler.” The 
point can only be decided by a renewed study of the develop- 
ment, and by further observations on the living in as many 
genera as are obtainable,” since in spirit specimens, even when 


degenerate from ancestors of now existing forms. Meyer’s arguments are, 
in my opinion, insufficient to prove this. 

1 Unfortunately, in the only specimens of Nerine cirratulus I was able 
to procure for examination the tentacles had fallen off, and I was unable to 
confirm Claparéde’s observation. 

2 The forms described in the ‘ United States Fishery Reports’ would no 
doubt prove of great interest if properly figured, but unfortunately the writers 


194 FLORENCE BUCHANAN. 


the tentacles remain attached, it is very difficult to make out 
their point of attachment, though sections of well-preserved 
specimens would also be of value. Meyer appears to think 
that there is sufficient evidence of their prostomial nature, and 
he goes on to show that both they and the branchiz of Serpulids 
probably represent the palps, not the so-called “ prostomial 
tentacles” of the Errantia (e.g. Nereis and Polynoé).1 He 
considers (pp. 614, 615) that there is a good deal of evidence 
that the tentacles of Spionids originate ventrally on either 
side of and slightly in front of the mouth, and only later move 
upwards more on to the dorsal surface. He regards the small 
quite anterior tentacles of Polydora antennata and others as 
representing the “‘ prostomial tentacles ” proper of the Errantia. 
He also quotes in support of his view Pruvot’s observations 
on the nervous system of annelids, which led that observer to 
conclude that the branchie of Serpulids probably represent 
the palps of the Errantia.”_ Again, he shows that in the adults 
of some of the Spionide (e.g. Polydora antennata), 
Cheetopteride (e.g. Telepsarus costarum and Phylloche- 
topserus), and Cirratulide (e.g. Heterocirrus frontifilis) 
the tentacles are much more ventrally placed than in others; and 
in some cases, as in Heterocirrus, they are situated quite ventrally 
in front of the mouth. In this view,1f the prostomial nature of the 
tentacles may be granted, I entirely agree withhim. He might 
also have given instances amongst the Errantia in which the palps 
are much higher up on the lateral surface than usual, e. g. Stau- 
rocephalus.® It may also be of some significance, though perhaps 


of the report have taken care to figure most of the forms without their charac- 
teristic tentacles, or else to figure the tentacles with some very limited 
portion of the head only. 

1 For definitions of the different tentacles on the head of polychet worms 
see Bourne’s paper on “ Haplobranchus ” above referred to, footnote to p. 169. 
2 ¢ Arch. de Zool. Expérimentale,’ 2nd series, vol. ili, 1885, pp. 314, 322. 

3 Claparéde, ‘Ann. Chet. du G. de Naples,’ part i, pl. vii, fig. 2a. Com- 
pare also Bourne’s description of the palps in the Polynoina (‘ Trans. Linnean 
Soc.,’ 1883, 2nd series, ‘‘ Zoology,” vol. ii, part vii, p. 351). He says, “ The 
palps differ from all other tentacular structures in being muscular along their 
whole length. They are capable of great elongation and contraction. ..... 


HEKATEROBRANCHUS SHRUBSOLII. 195 


not much, that the ciliated groove of the tentacles of most 
Spionids is used in conveying food down to the mouth, the 
cilia being continuous with those of the mouth opening, which 
would make it the more probable that the tentacles were once 
outgrowths of the side of the mouth. According to this view, 
the name “ palp”’ given to the most ventral pair of cephalic 
tentacles of Haplobranchus by Dr. Bourne might be retained, 
but extended so as to include the other tentacles as well. Dr. 
Bourne in his paper regards these other tentacles as peristomial, 
from a certain superficial resemblance they bore to the peristo- 
mial tentacles of Nereis, and he tries to show that the cephalic 
branchiz of the Serpulide are peristomial; but, as Meyer has 
shown, there is no evidence of fusion of ganglia in the Serpulidze 
as there is in the Nereide. When Dr. Bourne wrote his paper, 
however, he was unaware of the existence of Manayunkia, which 
was first described in the same year (1883), or he might have 
been led to different conclusions with regard to what he has 
called ‘ peristomial tentacles.” ! 


They originate [sections of Polynoé (Harmothoé) areolata] just where 
the prostomium joins the peristomial and buccal somites, although they 
appear to have more connection with the prostomium than with the other 
somites. Their nerve-supply appears to come from the supra-csophageal 
ganglion.” Such a description would need but little modification to serve for 
the tentacles of some of the Spionide (e.g. Polydoraantennata) and 
Cirratulide (e.g. Heterocirrus frontifilis). Bourne does not mention a 
contractile blindly-ending vessel in them, but there is this in Staurocephalus. 
The chief difference is the absence of cilia on the palps of Polynoé. 

1 As regards the rest of Meyer’s view, I should like to point out that in 
Hekaterobranchus the lateral parts of the collar are evidently not formed 
by the ventral cirri, since these are present quite independently as well 
on the Ist segment. The collar appears to be a mere folding of the 
ventral surface ; if anything it could only be the dorsal cirrus which assists in 
forming it. I have especially looked for the ‘“‘ Wimper-organe ” on the pro- 
stomium, which Meyer mentions (p. 639) as a feature which we might expect 
to find in the Spionide, but have been unable to find any trace of them. I 
may also mention in reply to Meyer’s suggestion (vol. vii, p. 723) that there 
are probably thoracic nephridia with a single external duct in Haplobranchus, 
that in my sections of Haplobranchus I have been unable to find any trace of 
such an arrangement ; andI even doubt whether what Bourne marks “gl” in 
his figure of Haplobranchus are to be regarded as nephridia at all. There is 


196 FLORENCE BUCHANAN. 


SysteMATIC DeEscRIPTION. 
Family Spionide. 
Hekaterobranchus,! gen. n. 
Spio quadricornis, Lam., ‘Anim.s. Vert.,’ vol. v, p. 319.? 


A single pair of dorsal branchiz, situated on the Ist seg- 
ment, and very large. 


no trace whatever of a lumen in them, and consequently none of external or 
internal openings. In Haplobranchus there are the same pigmented organs 
of unknown function on the bases of the branchisx, in some specimens at 
least, as are described and figured by Mecznikow in Fabricia (‘ Zeit. f. wiss. 
Zool.,’ vol. xv, p. 331, and pl. xxiv, fig. 8). Can these represent the pro- 
stomial ciliated pits from which Meyer considers (vol. viii, pp. 629—634) the 
most anterior portion of the common median duct to the exterior of the 
modified thoracic nephridia of Serpulids is to be derived? They open 
separately to the exterior on either side beneath the collar. Mecznikow does 
not say whether the common aperture to the exterior of the pair of nephridia 
he describes in the 2nd segment is on the prostomium or not. 

I will also mention that there is a sinus round the intestine in Haplo- 
branchus, as Dr. Bourne suggests there may be. In the anterior region I 
cannot in my sections see the dorsal vessel he mentions. The alimentary 
canal nearly touches the body-wall dorsally. The cclom, however, in this 
region is divided into distinct longitudinal cavities, four, or more anteriorly, 
two, on each side. 

My sections would also seem to show that Haplobranchus has an evertible 
pharynx, which, when inverted, reaches back into the 2nd segment. But 
further investigation is needed. 

'“Exatepoc = each (singly) of two. The name is meant to imply that two 
kinds of branchial organs are present, and that there is one single pair of 
each kind. 

* I have identified this form with the Spio quadricornis of Lamarck, 
because it with its four horn-like tentacles (tentacles proper and branchiz) is 
exceedingly suggestive of the name—much more so than the animal—Spio 
crenaticornis figured by Montague, to which Lamarck refers. (Lamarck 
refers toitas “ Diplotus hyalina” because Montague’s figures are wrongly 
numbered, but according to the text the figures 6 and 7 of pl. xiv (‘ Trans. 
Linnean Soc.,’ xi) should be marked as Spio crenaticornis.) It is ex- 
tremely probable that Lamarck saw Hekaterobranchus, and gave the name 
“Spio quadricornis” to it; and that he wrongly identified it with Spio 
crenaticornis, Montague, which was probably also not a Spio at all, 
but a Leucodore. 


HEKATEROBRANCHUS SHRUBSOLII. 197 


Cephalic tentacles not grooved, but ciliated all over. 

Prostomium well developed, bearing four eyes. 

Ist segment prolonged forwards on the ventral surface to 
form a collar. 

Pharynx evertible and richly ciliated. 

A single pair of thoracic nephridia, opening to the exterior 
in the 2nd segment, reaching back into the 6th segment, and 
thence bending forwards again. 

Giant-fibres minute, one in each nerve-cord near the upper 
and inner surface. 

Dorsal ‘‘ cirri” forming a sort of collar in the 2nd segment. 

H. Shrubsolii, sp. n. 

The following characters are probably of specific value. 

Ventral crotchet chet begin in 8th segment, accompanied 
by a few capillary cheetz. 

Shape of chete (fig. 4). 

Sinus round intestine posteriorly. . 

Intra-vascular ridge in dorsal vessel. 

Thoracic nephridia green. 

Habitat.—The mouth of the Thames.! 


PostscRIPT. 


Wuitst the foregoing paper was in the press I succeeded 
in obtaining Webster’s original description of the genus Stre- 
blospio, which is in the 39th Aunual Report of the Trustees 
of the New York State Museum of Natural History. Had I 
seen the full description of this genus earlier, I should have 
been loth to make a new geuus for a form which may well be 
included in Webster’s genus. But having seen a figure of the 
head of Streblospio, and this not even suggesting to me 
the identity of the genus I was studying with the one figured, 
I contented myself with merely mentioning (p. 190) what 
appeared to be one point of resemblance between the two 


1 Mr. Shrubsole informs me that he has found both this form and Haplo- 
branchus as far up as Gravesend on the south side of the river. 


198 FLORENCE BUCHANAN. 


forms, and did not concern myself with hunting for the fuller 
description before sending my paper to the press. 

From the description which I have now seen, it appears to 
me highly probable that Hekaterobranchus will have to be 
regarded, for the present at least, in virtue of its single pair of 
branchie placed on the head, its ventral collar, and its dorsal 
collar (equivalent to the so-called dorsal ‘‘ pouch ” of Webster), 
as a new species of the genus Streblospio, rather than as a 
distinct genus. There are, however, differences between the two 
forms, which might rankas generic differences were there a large 
number of forms with their common characteristics known. 
For instance, the “ conical median papilla or cirrus” on the 
anterior margin of the first segment, the short conical dorsal 
cirri of the posterior segments succeeding and replacing the 
plate-like lobes of the anterior, the fact that the “ pro- 
boscis is incomplete above ” (though, as I have failed to grasp 
what this means, I cannot attach much importance to it), the 
absence of thoracic nephridia (unless the dark green colour 
observed in the first eight segments of a few specimens was due 
to their presence), are all features in Streblospio which might 
prevent the association of Hekaterobranchus with it as one 
genus. How much, or how little, importance is to be attached 
to these differences is difficult to say without comparing the two 
forms, or at least proper representations of them, side by side. 
The drawings published by Mr. Webster of his Streblospio are 
so fragmentary and rough that I cannot undertake to form a 
final opinion on the subject by their aid alone. 


HEKATEROBRANCHUS SHRUBSOLII. 199 


EXPLANATION OF PLATES XXI and XXII, 


Illustrating Miss Florence Buchanaun’s paper on “ Hekatero- 
brauchus Shrubsolii, a new genus and species of the 
family Spionide.” 


Fic. 1.—Lateral surface view of the whole animal. pro. Prostomium. m. 
Mouth. ¢. Cephalic tentacles, with cilia and tactile hairs. 47. Branchie. 
v.c. Ventral collar of Ist segment. d.c. Dorsal collar formed by the two 
“cirri” of the 2nd segment. a. Anus. The ventral crotchet chet are 
seen beginning in the 8th segment. 

Fic. 14.—The same, natural size. 

Fic. 2.—Semi-diagrammatic view of the left side of the animal from the 
inside, as would be seen were the animal cut in two by a longitudinal, vertical, 
nearly median section. The tentacle (¢.) and branchie (d7.) are supposed to 
have been cut separately. The digestive, vascular, nervous, and excretory 
systems are shown. mm. Mouth. pd. Pharynx. «@s. Gisophagus. iu. In- 
testine (large, intersegmentally constricted part). iz¢’. Intestine (narrow, 
non-constricted part). «Anus. d.v. Dorsal vessel. ¢.v. Vessel of 
tentacle (its junction with the branchial vessel, which meets the ventral vessel, 
is seen in Fig. 12). d. dr. v. Branchial vessel in connection with the dorsal 
vessel. v. dr. v. Branchial vessel in connection with ventral vessel.  v. v. 
Ventral vessel. siz. Sinus. d.s. v. Dorsal vessel nipped off from the rest of 
the sinus, but lying close to the wall of the intestine, containing the intra- 
vascular ridge. z.c. Nerve-cord. gg. Prostomial ganglion. xzeph.e. Ex- 
ternal limb of thoracic nephridium, opening to the exterior in the 2nd 
segment. zeph. i. Internal limb of thoracic nephridium. c@/. Colom. sept. 
Septa. Longitudinal muscles would really be seen in such a view underlying 
the epidermis, but are omitted for the sake of clearness. 

Fic. 3.—Ventral view of anterior extremity. A. Showing ventral collar 
(v. coll.), and prostomium lying beneath it. ‘The collar is pushed forwards, 
and so covers the mouth. 3B. The pharynx (pd.) everted, hiding the pro- 
stomium. 

Fie. 4.—Chete. A. Notopodial capillary cheta. B. Neuropodial capillary 
cheta. C. Neuropodial crotchet (side view). C’. The same (ventral view). 

Fic. 5.—Transverse section through the thoracic region of a specimen 
in which no gonads were developed (stained with hematoxylin). Zpid. 
Epidermis. c.m. Circular muscle layer of body-wall. c¢. m'. Circular muscle 
layer of dorsal vessel (d. v.). c. m?. Circular muscle layer of alimentary 
canal. d./, m. Dorsal longitudinal muscles of body-wall. . 2. m. Ventral 
longitudinal muscles of body-wall. c. ep. Ccelomic epithelium, ca/. Colom. 
int. ep. Intestinal epithelium. /p. Notopodial chet. xrp. Neuropodial 


200 FLORENCE BUCHANAN. 


chete. cn. Notopodial “cirrus.” env. Neuropodial cirrus. (The right 
side of the section is behind the left.) s. s. m. Muscles going to setal sacs. 
d.v. m. Dorso-ventral muscles. g/. Gland-cell in epidermis. mxeph.e. Ex- 
ternal limb of thoracic nephridium. ep. i. Internal limb of thoracic nephri- 
dium. vase. r. Intra-vascular ridge in d. v. dorsal vessel. v.v. Ventral vessel. 
d.v.v. Parts of dorso-ventral vessel. x. ce. Nerve-cord. g./. Giant-fibre. 

Fic. 6.—Transverse section of the same worm through the anterior ab- 
dominal region. The section is taken just in front of a septum, and therefore 
the intestine does not occupy so much room in section as it otherwise would. 
The dorsal vessel (d. s. v.) is separated from the intestinal sinus; the intra- 
vascular ridge (vasc. 7.) is very close to but distinct from the wall of the 
intestine. Other letters as in Fig. 5. 

Fic. 7.—Transverse section of the same speeimen through the posterior 
abdominal region, where the intestine is not constricted intersegmentally. 
sin. Sinus, complete all round the intestine. mes. Dorsal mesentery. mes’. 
Ventral mesentery. Other letters as in Fig. 5. 

Fic. 8.—Transverse section through another specimen, in which the gonads 
are developed (stained with borax-carmine). (est. Testes. xeph. Nephridia, 
seen opening to the exterior on the left-hand side of the figure in eat. A 
parasite (par.) is seen in the intestine. The section is taken farther back 
than that represented in Fig. 6, and consequently the sinus shows no trace 
of the dorsal vessel. From the ventral vessel (v. v.) a branch is seen passing 
off on the right-hand side. Other letters as in Fig. 5. 

Fic. 9.—Transverse section of cephalic tentacle. epid. Ciliated epidermis. 
ce. m. Circular muscle layer. 7. m. Longitudinal muscle layer. c. ep. Coelomic 
epithelium. c@/. Coelom (containing ccelomic corpuscles in the living). ¢. v. 
Vessel of tentacle. 

Fic. 10.—Transverse section of branchia. d. dr. v. Branchial vessel in 
connection with dorsal vessel. v. dr. v. Branchial vessel in connection with 
ventral vessel. cel. Colom. Other letters as in Fig. 9. 

Fic. 11.—Dorsal view of the posterior extremity of the body of a speci- 
men in which the ciliated ridges could be seen in the intestine (¢zé.), just in 
front of the anus (a.). 

Fic, 12.—Semi-diagrammatic dorsal view of the head to show the probable 
arrangement of vessels. Pro. Prostomium. wzép'. Notopodial chete of Ist 
segment. rp’. Neuropodial chete of 1st segment (with “cirrus”). d. ¢. 
Dorsal collar (notopodial “ cirri”) of 2nd segment. dd. v. Dorsal vessel. 
d. br. v. Branchial vessel in connection with dorsal vessel. v. dr. v. Branchial 
vessel in connection with ventral vessel. v.v. Ventral vessel. 7. v. Vessel 
of cephalic tentacle. d. v. v. Coiled dorso-ventral vessel. 

Fic. 13.—Parasite from intestine (Monocystis hekaterobranchii). 
vac. Vacuole. x. Nucleus. 

Fie, 14.—Abdominal nephridium, eat, Its external opening. 


AN ATTEMPT TO CLASSIFY EARTHWORMS. 201 


An Attempt to Classify Earthworms. 
By 


W. B. Benham, D.Sc., 


Assistant to the Jodrell Professor of Zoology, in University College, 
London. 


Introduction. 
I. Nomenclature. IV. Table of generic characters. 
II. Outlines of the classification. V. Index to genera. 
III. Characters of the families and | VI. Phylogeny. 
genera. VII. Diagrams. 


Tue hard and fast line between Earthworms and Fresh- 
water worms, as indicated by Claparéde’s terms! “ Terricolze ” 
and “ Limicole,” is gradually becoming less distinct; and as 
new knowledge in both these groups is acquired we are led to 
recognise that the plan adopted by Vejdovsky is the more 
natural one: although I believe that it is possible to arrange 
the various families into which he divides the Oligocheeta into 
certain groups. 

The anatomical characters which served Claparéde as points 
of distinction between his two groups were drawn from the 
knowledge of Lumbricus alone. They were (1) the posses- 
sion of two blood-vessels below the intestine, the subintestinal 
and subneural vessels ; (2) the presence of nephridia in the 
genital somites; (3) the position of the clitellum behind the 
male apertures ; and (4) the presence of a plexus of capillary 
blood-vessels on the nephridia. The investigations of Per- 

1 Claparéde, “Rech. sur les Oligochétes,” ‘Mém. de la Soc. de phys. et 
hist. nat. de Genéve,’ t. xvi, 1862. 

VOL, XXXI, PART I1.—NEW SER. ) 


202 W. B. BENHAM. 


rier,! and later of Beddard, Horst, Rosa, myself, and others, 
have shown that, of these four characters, only the last can be 
retained ; and at the same time numerous points of agreement 
between “ Limicole ” and “ Terricole ” have been brought to 
light, especially by the help of microscopic research. For 
example, Beddard has pointed out the similarity between the 
genital organs of the earthworm Moniligaster and the 
water-worm Stylaria. 

From his observations of the various specimens of earth- 
worms in the Paris Museum, Perrier suggested a subdivision 
of the ‘Terricole”’ into four groups: (1) Anteclitellian ; 
(2) Intraclitellian ; (8) Postclitellian ; and (4) Aclitellian. 

Beddard and Horst have already shown that the second and 
third of these names cannot be always applied to all species of 
the same genus, one species being Intraclitellian, others 
Postclitellian ; and A. G. Bourne® has described a species of 
Moniligaster, for which the fourth group was formed, in 
which a elitellum is present. 

Vejdovsky,® in his beautiful monograph on the Oligocheta, 
divides the members of the group into seventeen families, each 
of equal value. These are— — 


Family 1. Aphanoneura. 

. Naidomorpha. 

. Chetogastride. 

. Discodrilide. 

. Enchytreeide. 

. Tubificide. 

. Phreoryctide. 

. Lumbriculide. 

. Pontodrilide. 
10. Criodrilide. 
11. Lumbricide. 
12. Eudrilidee. 

1 Perrier, “ Rech. pour servir 4 V’hist. des Lombriciens terrestres,” * Nouv. 
arch. du Mus. d'hist. nat. de Paris,’ t. vili, 1872. 


2 A. G. Bourne, ‘ Proc. Zool. Soc.,’ 1886. 
5 Vejdovsky, ‘Systeme und Morphologie der Oligochaeten,’ Prag., 1884. 


CMO MOND OK & WO 


AN ATTEMPT TO CLASSIFY EARTHWORMS. 203 


13. Acanthodrilide. 

14. Perichetide. 

15. Plutellide. 

16. Pleurochetide. 
?17. Moniligastride. 


He gives the characters of the first eleven of these families, 
but does not characterise the remaining six. As will be seen 
below, I shall split up some of his families, and unite others 
into one family. 

Rosa! does this to some extent ; but I have not taken quite 
the same lines as he has. He retains Claparéde’s “Terri- 
cole ” for the members of the last nine families, and groups 
these into six families ; and regards “ Terricole ” as of equal 
value to any of the first eight of Vejdovsky’s families. 

Vaillant, in the recently published volume ‘ Annéles’ 
(‘ Suite a Buffon’), divides the Oligocheta into two groups : 

(1) Naidina, including Vejdovsky’s first three families ; 

(2) Lumbricina, including the remainder with the excep- 
tion of Discodrilide. 

But, as he only deals with the genera and species of earth- 
worms known up to and including Perrier’s memoir in 1872, 
his subdivisions are not of that value that we should expect in 
such a work. 


I. NoMENCLATURE OF CERTAIN ORGANS. 


Before proceeding to the classification which I have to sug- 
gest, I will make a few remarks on the words and terms 
employed therein. 

The sete ina large number of worms are arranged, as in 
the common earthworm, in “twos.”” These are nearly always 
spoken of as ‘‘ pairs” by writers on the subject ; but this word 
seems to me to be ill-chosen: by “ pair” we usually under- 
stand a right and a left organ of a bilaterally symmetrical 
animal. I suggest, therefore, the word “couple” to denote the 


’ Rosa, “* Nuova Classificazione dei Terricoli,” Boll. d. Mus, Zool. ed Anat. 
Comp.,’ tom. ili, 1882. 


204 WwW. B. BENHAM. 


two sete placed close to one another; and in place of the 
terms “dorsal” and “ventral” sete I shall employ the 
words “outer” and “inner” in describing the couples in 
reference to their position relative to the ventral mid-line of 
the body. When the eight sete become wider and wider 
apart it sometimes becomes advisable to speak of such sete as — 
being “ separated.” 

The four setz on each side may be termed, with Perrier,! 
1, 2, 3, 4; the first being most ventrally placed, the fourth 
most dorsally. 

The prostomium is frequently a lobe nearly as wide as 
the first somite of the body, and sometimes separated from it 
by a transverse groove; but very frequently grooves extend 
back into this first somite, one starting from each side of the 
prostomium, so that the prostomium appears as a narrow or 
broad lobe embedded in, or ‘‘ dovetailed into,” the first somite. 
Sometimes these lateral grooves stop after traversing the first 
somite for only a short distance; the prostomium is then only 
“ partially dovetailed ’’ into the first somite. If, as in Lum- 
bricus agricola, Hoffmeister, the lateral grooves reach 
the intersegmental groove dividing the first from the second 
somite, the prostomium is said to be completely “dovetailed.” 

The first somite, that surrounding the mouth, is frequently 
called the “ buccal segment or somite;” but, following the 
nomenclature used in works on Polychetes, I have adopted 
the word “peristomium” or “ peristomial somite”? in the 
following classification. 

Most of the writers of the present day regard this as the 
first somite, and this is the view I shall take. It may be 
noted that Vaillant? retains Dugés’ enumeration, and speaks 
of the first setigerous somite as the “ first’ somite. 

In referring to the position of an aperture between two 
somites I use the form ‘ x/x1,” for example, meaning that such 
an aperture lies between Somites x and x1. 

' Perrier, “Sur un nouv. Gen. des Lombriciens terrestres,’ ‘Arch. f. 


Zool. Exp. et Gen.,’ t. ii, 1873. 
? Vaillant, “ Annéles,” ‘ Suite 4 Buffon,’ 1889. 


AN ATTEMPT TO CLASSIFY EARTHWORMS. 205 


The glandular modification of the epidermis indicated by the 
word “clitellum” is in many cases equally developed all 
round the body, as in Pericheta; this is indicated by using 
the word “cingulum,” or speaking of the clitellum as 
“complete.” In Lumbricus and many others the glan- 
dular modification does not extend across the ventral sur- 
face: this is a true ‘“‘clitellum” in the strict sense of the 
word, or an “incomplete clitellum” using the word in its 
wider sense. 

Intermediate conditions are sometimes met with, where, as 
in Acanthodrilus, the clitellum is “ complete” in the an- 
terior part of its extent, and “incomplete”? over the last 
two or more somites. The lower edge of the incomplete 
clitellum is sometimes, as in Lumbricus and Rhinodri- 
lus, further modified, presenting the appearance of a linear 
band, or group of glands over more or fewer somites. To 
these the name “tubercula pubertatis” has been applied 
by Hisen. 

The external openings of the sperm-ducts are the ‘‘ male or 
spermiducal pores;” those of the oviducts, “ oviducal 
pores;” those of the spermathece, the ‘“ spermathecal 
pores.” The first of these is frequently placed on a more or less 
prominent papilla, and in many worms other papille, median or 
paired, are present in their neighbourhood. These “ copu- 
latory papille’’ are probably of value for the diagnosis of 
species; but it is only recently that exact observations 
have been recorded as to the number and position of these 
papillz in different worms, and as they appear to be fully 
developed only at the breeding season they are not of 
value for absolute identification. Of this nature are the 
depressed reddish papille in Somite xxvi in the common 
earthworm. 

The setz in certain somites are not unfrequently modified 
for the purpose of copulation: those in more or less imme- 
diate relation to the male pores are known as “penial 
setz ” (Lankester) ; whilst those which are found, for instance, 
in Acanthodrilus layardus and other species, in connec- 


206 WwW. B. BENHAM. 


tion with the spermathece, have been distinguished by Horst 
as “copulatory sete.” These two terms are, to all intents 
and purposes, identical: so that I shall distinguish penial sete 
in the neighbourhood of the prostates, or spermiducal pores 
as “male copulatory sete,’ or ‘ posterior penial sete,” and 
those in neighbourhood of spermathece, as “female copu- 
latory,” or ‘‘ anterior penial sete.” 

Certain terms in connection with the internal reproduc- 
tive apparatus may require definition. The testes are 
usually present to the number of two pairs, and these are nearly 
always placed in Somites x and x1. The difficulty in counting 
the somites, or the delicacy and fragility of the septa, some- 
times render it difficult to be certain of the true position of 
the testes, especially as in some worms the septa do not 
correspond exactly in position to the true limits of the 
somites; hence it may be that some, at any rate, of the excep- 
tions are apparent rather than real. The ovaries likewise 
are nearly universally placed in the Somite x11; the oviducts 
open externally on the Somite xiv.- 

The terms vesiculz seminales, and seminal reservoirs, are con- 
veniently replaced by “sperm-sacs” (i.e. the “testes” of the 
older authors). Similarly the rather clumsy term ‘“ recepta- 
culum ovorum,” which has been observed in several worms, 
may be replaced by ‘“ ovisac;” a word at the same time 
simpler than the original term, and also in agreement with 
“* sperm-sac,”’ both in nomenclature and in function. 

I shall use the word ‘“ sperm-duct” for the more old- 
fashioned ‘ vas deferens;’’ however, I shall retain “ ciliated 
rosette” as a more convenient and shorter term than “ funnel 
of the sperm-duct.” 

In many genera there is, attached to the sperm-duct near 
its external pore, a glandular diverticulum, which may be lobed 
as in Pericheta, or tubular and coiled asin Pontodrilus, 
andothers. Most writers have referred to these structures as 
“ prostates ;”’ but Beddard has recently, in two papers (‘ Zool. 
Anz.,’ No. 268, 1887, and ‘ Quart. Journ. Micr. Sci.,’ xxix, 
p. 117), sought to establish an homology between these pro- 


AN ATTEMPT TO CLASSIFY EARTHWORMS. 207 


states and the “ atrium,” or enlarged part of the sperm-duct of 
Tubifex, into which the cement gland opens; and in his various 
papers on earthworm anatomy uses “ atrium”? to signify the 
prostate of authors, and by “ prostate ” refers to the peritoneal 
(? glandular) covering of the terminal portion or “ atrium” of 
Moniligaster and Stylaria. 

I will not, here, enter into a discussion of this subject; I 
shall, however, reject ‘‘atrium,” and retain the older word 
“prostate” for these glands, which either pour their 
secretion into the sperm-duct, or open, independently, to the 
exterior. 

This prostate is a hollow structure, the wall of which is 
formed of club-shaped cells similar to those found in the 
clitellum: sometimes, a layer of columnar epithelial cells 
intervenes between the clitellar cells and the lumen; some- 
times the columnar cells are absent. A layer of muscles 
usually surrounds the clitellar cells; or the muscular layer 
may be confined to the proximal region of the prostate. 

In many cases, as in Pericheta, Pontodrilus, Eudrilus, 
&c., the prostate opens into the sperm-duct, so that we only 
have one “male pore.”” In other genera, as in Acantho- 
drilus, Deinodrilus, &c., the prostates open to the exterior 
independently of the sperm-ducts: in these cases we must dis- 
tinguish “spermiducal pores” from “ prostate- pores.” 

The word “spermatheca”’ is retained, in preference to the 
* receptacula seminis ”’ of other authors. 

The glandular structures met with in Lumbricus, Peri- 
cheta, Brachydrilus, and others, and usually called “ cap- 
sulogenous glands,” are misnamed. As far as we know they 
have nothing to do with the formation of the capsule or 
cocoon, which is formed by the hardening of the secre- 
tion of the gland-cells of the clitellum; but they give rise to 
the albuminous fluid found in the cocoon, in which the ova 
and spermatozoa are deposited, and which serves as nourish- 
ment for the developing embryos. Vejdovsky suggests the 
word “albumen-glands” for these structures, a term which 
I retain. 


208 W. B. BENHAM. 


The excretory system has quite recently had a new light 
thrown upon it by the researches of Beddard! and of Baldwin 
Spencer.” 

Perrier, in 1872, noticed that the nephridia were replaced in 
Pericheta by small tufts of tubules ; and a similar arrange- 
ment is found in many other worms. 

Beddard® described a species of Acanthodrilus with eight 
“nephridia” per somite; these he more recently* discovered 
were not really large tubules, but eight groups of small tubules 
opening to the exterior by numerous apertures. Again, more 
recently he has shown us that in Pericheta armata there is 
a network of delicate tubules in addition to a pair of large 
nephridia in each somite; and this network of delicate tubules 
is provided with many external apertures, sometimes without 
internal funnels, rarely with them. 

Baldwin Spencer, in his detailed description of Megasco- 
lides australis, demonstrates the transition between the 
large nephridia in the posterior part of the body and the deli- 
cate small tubules which line the inner surface of the body- 
wall in the anterior somites. Anteriorly there is a network 
of these delicate tubules, which have no internal funnels, but 
numerous external openings, the network being continuous 
somite to somite. About the middle of the body this net- 
work becomes confined to a band of small tubules round the 
somite, of which one tubule is larger than the rest. Further 
back this tubule increases in size, and ultimately attains a size 
almost equal to that of a nephridium in Lumbricus; more- 
over, this tube communicates with the ccelom by means of a 
funnel, but still retains a connection with the network of 
smaller tubules. Both Beddard and Spencer favour the idea 
of the development of a “ nephridium” from such a network 
of tubules. 


1 Beddard, ‘ Quart. Journ. Mier. Sci.,’ vol. xxix, and see literature therein. 

2 Baldwin Spencer, “The Anatomy of Megascolides australis,” ‘ Trans. 
Roy. Soc, Victoria,’ vol. i, 1888. 

3 Beddard, ‘ Proc. Roy. Soc.,’ 1885. 

4 Beddard, ‘ Quart. Journ. Mier. Sci.,’ vol. xxviii. 


AN ATTEMPT TO CLASSIFY EARTHWORMS. 209 


The possession by earthworms either (1) of such a network 
or *‘ tuft” of small tubules (a “ plecto-nephric ” condition), or 
(2) of a pair (rarely two pairs) of large “ nephridia” (a 
“ mega-nephric ”’ condition), suggested to me the possibility of 
dividing the earthworms into two groups, according to con- 
ditions of the excretory system. This has formed the basis of 
the classification here put forward ; but we are instantly met 
by the fact that some species of certain genera comprised in 
the first set-—i. e. those with a network of small tubules—have 
apparently large “ nephridia;”’ but is this really the case? 

Both in Acanthodrilus and in Pericheta species 
have been originally described as having large nephridia ; but 
a renewed and more careful microscopical examination has 
proved either that this large nephridium is accompanied by a 
network of tubules as in P. armata (Beddard), or that 
the supposed “nephridium” really consists of a mass 
or tuft of small tubules as in Ac. multiporus. So 
that when Fletcher states that several species of Crypto- 
drilus, a genus which is usually “ plecto-nephric,” have three 
pairs of large “nephridia”’ per somite, I think we are justified 
in assuming that these will turn out to belong to one or other 
of the above categories.1 I may add that Fletcher’s descrip- 
tions of numerous species are unaccompanied by figures, 
except in a few instances of external characters. 

The genus Perionyx, formed by Perrier for the reception 
of a worm very similar to Pericheta, has always been closely 
associated with this latter genus, and I feel considerable hesi- 
tation in removing it from association with Pericheta; but 
it differs from the latter in possessing a pair of large nephridia 
in each somite unaccompanied by smaller tubules, as well as in 
some other small details. This I have ascertained by the 
examination of sections, and by mounting a portion of the 
body-wall, with nephridia, complete. 

1 Beddard has recently (‘ Quart. Journ. Micr. Sci.,’ xxx, Feb.), described 
these new species of Acanthodrilus, in which only large nephridia are 
mentioned. Here again, I think, we may suspect that a network of small 
tubules is present in addition. 


210 W. B. BENHAM. 


The nephridium here has all the appearance of that in the 
common earthworm. 

For this reason I have separated it from its former allies ; 
and although, at present, we know but few points of apparent 
difference between Perionyx and Pericheta, yet it may 
be possible to separate some of the species of the latter genus 
and place them in the former genus. 

Assuming with Beddard and Baldwin Spencer that the 
excretory network is more primitive than paired nephridia, 
and regarding the perichetous condition as secondary, it 
is not impossible to conceive this condition making its 
appearance both in worms which still retain the network, 
and in those which have acquired the large nephridia; 
or perhaps Perionyx is a descendant from forms which 
had become perichetous whilst still retaining the network, 
but which have lost this latter character and retained the 
former. 

The histological structure of the nephridia still remains to 
be worked out, although, thanks to Beddard and Spencer, we 
have a fair knowledge of the details in the case of the small 
tubules of the network in Acanthodrilus and Megasco- 
lides. These, like the larger nephridia, are made up of a 
series of “‘drain-pipe” cells—no doubt ciliated in some part of 
the tubule—which form the mass of the tubules; but this 
intracellular lumen becomes converted into an intercellular 
lumen near its external opening, and the wall is here frequently 
provided with muscle-fibres. 

Of the exact arrangement of the convolutions of the more 
or less elongated tube of the nephridia of other forms we 
have little or no information. Gegenbaur’s well-known figure 
of the nephridium of Lumbricus still remains the only 
accurate drawing of such a nephridium. ‘The recent drawing 
given by Goehlich (‘ Schneider’s Beitrage,’ Bd. 11) is not quite 
accurate; and he is mistaken in thinking that the cilia are 
continuous from the funnel to the muscular duct. This is 
not the case; certain regions of the duct are provided with 
cilia, and others are deprived of them. 


AN ATTEMPT TO CLASSIFY EARTHWORMS. 211 


Goehlich’s drawing of the funnel of the nephridium of Lum- 
bricus is wrong ; and in the various genera we find nephridia 
with more and with less complicated convolutions of the tubule ; 
for instance, Microcheta and Moniligaster or Rhino- 
drilus. The character of the muscular region or “ duct” 
also varies. 

In many cases it is, as in Lumbricus, a mere continua- 
tion of the tube; in other cases the muscular region is, 
in proportion, very much larger, and the tube does not 
enter it at its extremity, but at some point along its 
side, so that the muscular region is produced as a 
“cecum,” or bladder; and to some extent the families are 
in part characterised by possessing either a simple “ duct ” 
or a “ cecum.” 

The family Eudrilide, for instance, all possess simple 
nephridia ; the tube is comparatively short, and the duct simple 
and in many casesill-marked. In the accompanying diagrams 
I have inserted the nephridia—so far as I have been able to 
obtain information on the point—in order to show, first, the 
presence or absence of a “‘ cecum ;” and secondly, the somite 
in which the series commence. 

In the family Rhinodrilide all the genera have cecal 
prolongations of the nephridial duct, more or less marked. 
And frequently this cecum is less developed in the most 
anterior than in the greater number of nephridia: for example, 
Urobenus and Microcheta; in the latter genus the 
muscular portion attains the greatest relative size, and the 
coiling of the tubule the greatest complexity to be found 
amongst earthworms. In Rhinodrilus the more anterior 
nephridia have a simple duct, those more posteriorly possess a 
cecum. 

Plutellus is noticeable for the alternation, from somite to 
somite of the position of the nephridiopore, which is placed 
in front of the second or the fourth seta on each side, counting 
from the most ventral seta. 

Perionyx saltans, A. G. B., presents a somewhat 
similar condition, and according to Fletcher, certain species 


212 WwW. B. BENHAM. 


of Cryptodrilus ;! but see above as to whether large nephridia 
are really present in this genus unaccompanied by a network. 

Brachydrilus possesses four nephridia per somite, all 
alike. 

In a large number of genera the anterior nephridia— 
both in those retaining a network and those with large 
nephridia—are more or less modified. For instance, Beddard 
was the first to show for Ac. dissimilis that a group of 
tubules on each side of the pharynx is connected, by 
means of a strong duct, with the buccal cavity ; the same is the 
case with Dichogaster and Digaster—all genera in which the 
nephridia are in the form of a network elsewhere in the body. 

In Megascolides, Baldwin Spencer has described and 
figured the presence of numerous nephridial tubules around 
the pharynx, which open separately into the cavity of the 
alimentary tract. 

In other cases, e.g. Ty pheeus and Deinodrilus, the tubules 
of the network are much more abundant in the first two or 
three somites, but do not communicate with the cavity of the 
pharynx. 

We are, therefore, entitled to consider that these anterior 
nephridia are used by the worm for some other purpose in addi- 
tion to excretion: they are probably used for softening or other- 
wise acting on the food, either when the everted buccal region 
has seized the food, or previously to this. How the external 
aperture of a group of tubules has shifted from its position on 
the body-wall to the pharyngeal wall, and how at the same 
time, in some cases, the numerous apertures have united into 
a duct, we do notknow. We can only form conjectures on the 
subject. The epiblast is known to grow in at the blastopore, 
so as to form the lining of the pharynx; and the shifting of 
the nephridiopores may perhaps be connected with this 
invagination. 

I have used in my diagnoses of the genera the term 
“‘pepto-nephridium®*” to indicate this modification of the 

1 Beddard describes a similar alternation in Ac. rose and Ac, dissimilis, 

* Tlerrw=I soften. 


AN ATTEMPT TO CLASSIFY EARTHWORMS 2138 


anterior nephridia for the purposes of alimentation, both for 
cases where they open into the digestive canal and where they 
merely open to the exterior, i. e. intra-buccal pepto-nephridia, 
and extra-buccal pepto-nephridia. 

This modification of the anterior nephridia is also found in 
some of the families in which large nephridia have replaced 
the network. In the Geoscolecide, this is the case: and 
although in Geoscolex the first nephridium is not very 
greatly modified, it will be seen to be slightly different from 
the following ones; and in fact the coiled tubule is thicker, 
and is distinctly glandular in appearance. The following 
ones are somewhat similar. The point of entrance of the 
tubule into the duct gradually shifts towards the pore, so that 
the cecum becomes more and more marked. But in the 
other two genera of the family, Diacheta and Urocheta, 
the first pair of nephridia are very different from the following 
ones, both in regard to their size and complexity, and are 
“‘ pepto-nephridia.” From a glance at the diagrams it will 
be seen that Diacheeta presents an intermediate stage between 
the more simple condition in Geoscolex and the more com- 
plicated in Urocheta. 

Rhinodrilus and Microcheta also present variations 
in their nephridia, not so markedly as in the last family, and 
not so pronounced in Microcheta as in Rhinodrilus, 
where the first pair of nephridia are much larger than the 
following ones, and lie underneath the pharynx. 

In connection with the nephridia it is worthy of note 
that in Criodrilus and Pontodrilus they are entirely 
aborted in the first dozen or more somites. Is this con- 
nected with their aquatic habits? Is the absence of anterior 
nephridia analogous to the absence of salivary glands in 
fishes ? 

The position of the nephridiopore is not in all cases charac- 
teristic of genera, though this is usually the case. 


The Alimentary Canal. 
The regions that I have distinguished are—(1) buccal 


214. w. B. BENHAM. 


region, (2) pharynx, (3) cesophagus, (4) gizzard, (5) tubular 
intestine, (6) sacculated intestine. 

The buccal region is always present, and is bounded pos- 
teriorly by the circumpharyngeal nerve-collar: this region is 
thinner-walled than the pharynx, and is eversible. 

The pharynx occupies some two to four of the following 
somites: it is probable that the buccal region occupies always 
the first and second somites,! and the pharynx the third to fifth ; 
but as there are no septa in this region the pharynx fre- 
quently appears more extensive than this: the thick muscular 
wall is confined nearly entirely to the dorsal surface. 

The following region, up to the gizzard, is the cesophagus. 
As will be readily seen by a glance at the diagrams, this region 
is extremely variable in extent, according to the position of 
the gizzard, which may lie in Somite v or in Somite xvi, or, 
as in Moniligaster, still further back. 

The presence of two or more gizzards is by no means un- 
common, and this leads to a repetition of the cesophagus. 
How far the position of the gizzard is a generic characteristic 
itis impossible to say: descriptions of the alimentary tract are 
in most cases very brief, and it is well known that the gizzard 
rarely lies opposite to the somite to which it belongs; the 
septa are very frequently pushed backwards, so that the 
septum bounding a somite may, in the middle of the 
body, come to lie at the level of a somite some little way 
behind: in addition to this, the septa in the region of 
the gizzard are not unusually very thin, and easily broken; 
and it thus comes about that, whereas the positions of the 
various parts of the reproductive apparatus are carefully 
noted, the real situation of the gizzard has sometimes been 
less accurately observed. From the peculiar constancy in 
position of the parts of the generative organs, and the varia- 
bility in the position of the gizzard, I believe it will repay 
future observers to turn their attention more particularly to 

1 In the diagrams I have represented the buccal region as occupying the 


first somite only : so little positive information is available on this point, that 
I have not attempted to mark out the limits of the region accurately. 


AN ATTEMPT TO CLASSIFY EARTHWORMS. 215 


the latter organ. For instance, most species of Pericheta 
have the gizzard occupying three somites, viz. VIII, IX, xX. 
Some species have been credited with a gizzard in Somite v 
or vi. I believe that the former is the typical position for 
Pericheta, and that a further examination of the forms 
with a forward position of the gizzard will lead to a separa- 
tion of these species. 

The absence of a gizzard in Criodrilus and Pontodrilus, 
in addition to the absence of the anterior nephridia, is a point 
worthy of note, both negative characters being, no doubt, con- 
nected with their habitat and characters of food. 

The fixity of the ovary and testes in the thirteenth and 
tenth somites respectively—or, rather, their nearly constant 
position—gives us a fixed point or centre from which to count 
the variations in position of other organs. And I am greatly 
tempted, with Rosa, to regard the ovary as always in the 
thirteenth: notwithstanding the apparent exceptions—M icro- 
cheta and Brachydrilus, where they appear to be in 
Somite x11, the error may be due to fusion of two of the 
anterior somites to form a single ‘“‘ peristomium.”! However 
this may be, the majority of earthworms possess an ovary in 
Somite x111—this may be taken as a fixed point, and we may 
compare the position of other structures in regard to their 
greater or less distance from this point: thus the gizzard 
lies so many somites in front, or behind, in the various 
genera. This fact seems to indicate that the gizzard of 
Eudrilusin Somite vi is not homologous or homogenetic with 
the gizzard in Lumbricus in Somites xvi and xvii1; but 
that a similar modification of the wall of the gut has occurred 
in different somites in different worms. 

The region following the gizzard, and before the typhlosole 
commences, is in Lumbricus very short, occupying only a 


1 This is certainly the case in Microcheta beddardi, where small sete 
can be detected in the apparent peristomium: and as this somite never carries 
sete, we have here one somite occupying the position of two morphological 
somites. In M.rappi I can detect no sete on the peristomium ; the fusion 
is here complete. 


216 W. B. BENHAM. 


couple of somites; but in those cases where the gizzard is 
far forwards there is a greater extent of intestine in which 
the typhlosole is absent, and in which the walls are not 
distinctly sacculated or nipped by the septa. This non- 
typhlosolar region is the “tubular intestine;” and on it, 
instead of in the cesophagus, are situated the calciferous or 
other diverticula of the gut in genera with an anteriorly 
situated gizzard. 

The typhlosolar region appears to begin, in the majority of 
cases, somewhere about Somites xiv to xv1; but we have very 
scanty material for generalising on this point. Many of the 
diagrams of the canal are correct only so far as the position of 
the calciferous glands go. In the majority of cases authors 
confine themselves to the position of gizzard and glands, and 
rarely state where the typhlosolar or sacculated region com- 
mences. 

Only a few cases is this typhlosole absent, e.g. Ponto- 
drilus, Microscolex. It is sometimes a mere thin, com- 
pressed fold; or it is cylindrical, and nearly fills the cavity of 
the intestine. Rhinodrilus is peculiar—at any rate, no record 
of this has been published—in having its typhlosole, which is 
a mere fold, attached along a spiral line running round the 
intestine, instead of hanging down from the dorsal mid-line. 

Ina few cases—Pericheta, Urobenus, Hormogaster— 
the hinder part of the “tubular’’ region is distinctly pouched ; 
not merely nipped by the successive septa, but with thick 
walls, giving rise to a number of paired pouches, the walls of 
which are probably more vascular than elsewhere. This 
thrusts the commencement of the sacculated or typhlosolar 
region backwards ; and the point of union is marked by a pair 
(rarely more pairs insome sp. Pericheta) of blind cylindrical 
outgrowths or “ intestinal ceca.” These lie in Somite xxi in 
Hormogaster, in Somite xxvi in Pericheta and Uro- 
benus. I may mention once more that some worms, referred 
to Pericheta, are deprived of these ceca; and I believe this 
negative character goes hand in hand with a forward position 
of the gizzard. 


AN ATTEMPT TO CLASSIFY EARTHWORMS. 217 


A more usual variety of diverticulum of the canal is that 
found in Lumbricus, and known as “ calciferous glands,” 
or “glandes de Morren.” MHere there are three apparent 
pouches on each side of the cesophagus, two pairs lying in 
Somite x1, and one pair in Somite x11. But of these, only the 
first pair actually communicates with the gut; the other two 
pairs are not pouches, but thickenings of the cesophageal wall, 
which is here hollowed out by a number of horizontal, antero- 
posteriorly directed cavities, which end blindly behind, and 
open into the first “ pouch” infront. The horizontal lamelle 
separating the chambers from one another contain blood- 
sinuses, and are lined by large cells which secrete CaCO, ; this 
escapes from the cells, or more probably the cells themselves 
break away, and find their way by means of the anterior pouch 
into the cesophagus. 

We have little or no detail as to the “calciferous glands ” 
in other worms; in some cases we do not even know whether 
they produce lime ; but throughout I speak of them as “ calci- 
ferous glands.” ‘They are very frequently absent, and when 
present are very variable in number and position. 

In some cases “salivary glands” are said to be found 
amongst the muscles of the pharyngeal wall. 


II. OuTLINES OF THE CLASSIFICATION. 


The class OticocHz#zta may be divided into two sub-classes, 
according to the presence or absence of asexual repro- 
duction. 

Sub-class I. NarpomorpHa. 
Order 1. Naidina. 
Families 1. Aphanoneura. 
2. Naidide. 
3. Chetogastride. 
{And the genus Ctenodrilus. | 
Small worms of relatively few somites ; blood uncoloured ; male 
genital pores in front of Somite vm, or in this somite. 
Asexual as well as sexual reproduction occurs. The 
VOL. XXXI, PART I].—NEW SER, P 


218 W. B. BENHAM. 


anterior few somites of the body are frequently different from 
the following somites, e.g. in Naids and in Chetogaster 
the sete have a different arrangement or form in the anterior 
somites (“cephalization” of Lankester) ; or the prostomium 
is pigmented and ciliated, as in Alolosoma and Cteno- 
drilus. Eye-spots are frequently present. 


Sub-class Il. LumpricomorpHa. 


Reproduction only by sexual process ; no “ cephalization ;” ! 
somites behind the peristomium all similar; and sete are 
similar throughout the body, except in special regions, e.g. on 
clitellum. 

Male genital pores behind Somite vit. 

No eye-spots (? Helodrilus, Hoffmeister). 

The various families included in this sub-class cannot really 
be separated by any very marked anatomical characters; but 
they may be divided roughly into— 

Order 1. Microdrili (Lumbricomorpha minora) ; 
2. Megadrili (Lumbricomorpha majora); 


which correspond to ‘ water-worms” and “ earthworms ”’ re- 
spectively. ‘The only constant difference between these two 
groups is the absence in Order 1 of a capillary network of 
blood-vessels on the nephridium, and the presence of such 
blood-vessels on the nephridium in Order 2; and this is very 
likely due to the difference in size, and to the character of the 
medium in which the members of the two groups live. 

Other characters which are usual to Order 1, and rarely 
present in Order 2, are as follows: 

Small size, and thin, transparent body-wall. 

Prostomium not separated from the peristomium by a 
groove. 

Sete always in four groups per somite, and usually more 
than two in each group: frequently the sete are of two sorts 


‘ In a paper ona new species of Diacheta, a proof of which Professor 
Lankester has very kindly allowed me to see, Mr. Beddard states that no sete 
are present on the first five somites. 


Po 


AN ATTEMPT TO CLASSIFY EARTHWORMS. 219 


in each somite, and may be capillary, uncinate, forked, or 
simple. The clitellum is always developed round the male 
pores, and generally occupies only two somites. There are no 
nephridia in the genital somites. There is no true gizzard, 
no typhlosole, no subneural vessel. [These last four negative 
characters, however, hold for some “ earthworms.” | 

The characters nearly constant in Order 2, in which 
“earthworms” differ from water- worms, are— 

Large size, varying from two inches (or less) to six feet; 
thick, pigmented, and opaque body-wall, though the pigment 
may be absent and the wall more or less transparent on the 
ventral surface. 

Prostomium separated by a groove from the peristomium. 

Sete frequently not arranged in groups; when they are so 
arranged there are never more than two sete in a group 
(? Echinodrilus, Vaillant). These sete are nearly always 
simple, or the modification when present takes a different 
direction from that in Order 1. 

The clitellum varies in position with regard to the male 
pores, and always occupies more than two somites (? certain 
species of Perichexta). 

A gizzard is nearly always present, except in such cases as 
Criodrilus, Pontodrilus, Microscolex, and Photo- 
drilus, where the character of the food renders it useless.! 

With the exception of a few genera, nephridia occur in all 
somites after the third or fourth, including the genital 
somites. 

Order 1 includes Vejdovsky’s “families” Discodrilide, 
Enchytreide, Tubificide, Phreoryctide, and Lum- 
briculide. 

Order 2 contains the remainder of his families, and these 
I will now proceed to group as follows: 


Branch I. Plectonephrica.? 
Excretory system in the form of numerous delicate tubules 


* In Pontodrilus there isa modification of the gut-wall which probably 
represents the gizzard, 
2 TlAex7n = a net. 


220 W. B. BENHAM. 


in each somite, uniting to form a network, with more or less 
numerous external apertures: a large “nephridium,”’ with 
celomic funnel, may be present in addition to these tubules. 
a. Sete, eight per somite (rarely twelve), either in couples 
or separate. 
1. Spermiducal pores on Somite xvii or xviI1; one pair 
of prostates in the same somite. 


Family I. TypHaip2. 


Genera :— 

. Typheus. 

. Megascolides. 
. Cryptodrilus. 
. Didymogaster. 
. Perissogaster. 
. Dichogaster. 

. Digaster. 


O OF B® CO WO 


~I 


2. Spermiducal pores on Somite xvi11; two pairs of pro- 
states in Somites xvii and x1x. 


Family Il. AcantuopRiLipa. 


Genera :— 
8. Acanthodrilus. 
9. Trigaster. 
10. Deinodrilus. 


b. Sete more than twelve, usually twenty to eighty per 
somite; arranged in a continuous or discontinuous 
circle. 


Family III. Pericuztips. 
Genus 11. Pericheta (including Megascolex). 
Branch II. Meganephrica.! 


Excretory network absent; replaced by a pair (rarely two 
pairs) of large nephridia in each somite. 


1 Meyac = large. 


AN ATTEMPT TO CLASSIFY EARTHWORMS. 221 


A. Prostates present. 

a. Spermiducal pores intersegmental, and placed far for- 
wards ; sperm-duct traversing only one somite; clitel- 
lum on Somites x to xIII. 


Family IV. Montuicastrip#. 
Genus 12. Moniligaster. 


6. Spermiducal pores on Somite xvi or xvii1; clitellum 
occupying all or any of the Somites x11 to xvitt. 
1. Eight setz per somite, in couples or separate. 


Family V. Euprivip2z. 


Genera :— 
13. Eudrilus. 
14. Teleudrilus. 
15. Pontodrilus. 
16. Photodrilus. 
17. Microscolex. 
18. Rhododrilus. 
19. Plutellus. 


2. More than eight sete, usually thirty or more, per somite, 
arranged in a ring. 


Family VI. Pertonycipaz. 
Genus 20. Perionyx. 


B. No prostates present. 
1. Spermiducal pores behind Somite xviii, within the area 
covered by the clitellum. 
a. One pair of sperm-sacs occupying several somites ; 
eight sete, separate or alternate in some part at least 
of the body. 


Family VII. Groscotecip2. 


Genera :— 
21. Geoscolex. 
22. Urocheta. 
23. Diacheta. 


222 W. B. BENHAM. 


6b. Two or more pairs of sperm-sacs; setz in couples, 
and exhibiting no tendency to alternate. 


Family VIII. Rutnopritipz. 
Genera :— 
24. Rhinodrilus. 
25. Microcheta. 
26. Urobenus. 
27. Hormogaster. 
28. Brachydrilus. 


2. Spermiducal pores in front of Somite xvii1, anterior to 
the clitellum. 


Family IX. Lumsricip2. 
Genera :— 
29. Lumbricus. 
30. Allolobophora. 
31. Criodrilus. 
32. Allurus. 


TII. CHARACTERS OF THE FAMILIES AND GENERA IN EXTENSO. 
Branch J. PLEcTONEPHRICA. 


The excretory system is in the form of a number of delicate 
tubules in each somite, more or less united to form a network, 
and having numerous external apertures in each somite. Added 
to this, there may be a pair of larger “ nephridia,” each pos- 
sessing one external aperture and a funnel communicating 
with the celom. 

a. The sete are never more than eight (with the exception of 
one genus) per somite; and these may be arranged in couples, or 
are more or less separated from one another. 

1. Spermiducal pores in Somite xvii or xvii1; only one pair 
of “prostates,” which lies in the same somite as these pores. 


Family I. Typuaipa, mihi (= part Eudrilide, Rosa). 


The clitellum, which is more feebly developed on the 
ventral than on the dorsal surface, begins on Somite x11 or 


AN ATTEMPT TO CLASSIFY EARTHWORMS. 223 


xiv, and includes from five to ten of the following somites. 
““Copulatory” papille are usually present in the neigh- 
bourhood of the spermiducal pores. The “prostates” are 
tubular and convoluted, or are lobed structures. 


Genus 1. Typuaus, Beddard, 1883. 


Sete in couples, all on the ventral surface. 

Clitellum occupies Somites xiv to xvi. 

The male pores are on Somite xvii, in line with the inner 
couple of sete; sacs with penial sete present. The “ pro- 
states’ are convoluted tubes. 

Sperm-sacs.—A single pair in Somite x, which may occupy 
more than one somite. 

One pair of testes and ciliated rosettes in Somite x, en- 
closed in a median portion of the sperm-sac. 

[Prostomium! nearly as broad as the peristomium, and not 
dovetailed into it. 

A single pair of spermathece in Somite vir; the aperture 
is placed in the anterior part of this somite. 

Dorsal pores present. 

A single pair of calciferous glands in Somite x11. The giz- 
zard occupies Somites v1, vil; peculiar intestinal glands farther 
back. | 

Species 1. T. orientalis, F. E. B., 1884; India. 
2. T. gammii, F. E. B,, 1888; India. 


See Beddard, ‘Ann. Mag. Nat. Hist.,’ 5th ser., vol. xii, 1883 ; 
and ‘ Quart. Journ. Micr. Sci.,’ vol. xxix, 1888. 


Genus 2. Mreascotipes, McCoy, 1878 (= Notoscolex, 
Fletcher, 1886). 
Setze in couples, all on the ventral surface. 
Clitellum occupies Somites x11 to xxr (or xxmit), and 
though feebly developed, extends across the ventral surface. 
The male pores are on Somite xvii, on slight papille. 
No penial sete. 


1 The characters placed in square brackets are less easily observed, or are 
more subject to specific variation. 


224 W. B. BENHAM. 


S perm-sacs.—Four pairs in Somites x1, x11, x11, xiv (in 
one species a less number has been described). 

Prostates tubular, and very greatly coiled. 

[Prostomium broad, not dovetailed into peristomium. 
Dorsal pores are present. 

Testes and ciliated rosettes in Somites x, x1, free; the two 
sperm-ducts on each side are separate throughout their course, 
only uniting at their junction with the prostate. 

Spermathecze, with appendices, two pairs in Somites viir 
and 1x (or four pairs). 

The nephridia are in form of network anteriorly, continuous 
from somite to somite, one tubule of which gradually enlarges 
till posteriorly there is a pair of large nephridia, together with 
a network of small tubules in each somite. 

Gizzard in Somite v or v1; no cesophageal glands ; no typhlo- 
sole. 

Intestine dilated in Somites x11 to xvim. 

Numerous intra-buccal pepto-nephridia are present. 

The septa of anterior segments greatly thickened. ] 


Species 1. M. australis, McCoy, 1878; Australia. 
: camdenensis, Fletcher, 1886; Australia. 
. grandis, Fletcher, 1886; Australia. 
tasmanicus, Fletcher, 1887 ; Australia. 
. tuberculatus, Fletcher, 1887 ; Australia. 
. mawarre, Fletcher, 1887; Australia. 
pygmeus, Fletcher, 1888; Australia(? genus). 
.illawarre, Fletcher, 1888; Australia. 

See Baldwin Spencer, ‘ Trans. Roy. Soc. Victoria,’ vol. i, 
1888 ; and Fletcher, ‘ Proc. Linn. Soc. N.S.W..,’ vols. i and iii. 


COUN So St Ee 02, 29 
Bee ee 2 


Genus 3. Crypropritvs, Fletcher, 1886. 
Sete separate. 
Clitellum on Somites x111 to xvi (or less), complete ven- 
trally, 
Male pores on Somite xvii1, not on papille; no penial 
sete. 


AN ATTEMPT TO CLASSIFY EARTHWORMS. 225 


Prostates lobate, and sometimes extending beyond their 
proper somite. 

Sperm-sacs.—Two pairs in Somites 1x and x11. 

[Prostomium small; dorsal pores present. 

Testes and ciliated rosettes in Somites x and x1, free. 

Spermathece with appendices, usually two pairs in Somites 
viii and 1x (rarely more). 

Gizzard in Somite v or vil. 

“ Salivary glands” are present, but do not appear to open 
into the pharynx. 

Intestinal calciferous glands occur in Somites 1x or x 
to XIII. | 


Species 1. C. rusticus, Fletcher, 1886; Australia. 

2. C. saccarius, Fletcher, 1886; Australia. 

3. C. fletcheri, F. E. B., 1887; Australia. 

4, C. rubens, Fletcher, 1887; Australia. 

5. C. mediterreus, Fletcher, 1887; Australia. 

6. C. fastigatus, Fletcher, 1888 ; Australia. 

7. C. unicus, Fletcher, 1888; Australia. 

8. C. singularis, Fletcher, 1888; Australia. 

9. C. canaliculatus, Fletcher, 1888; Australia. 
10. C. manifestus, Fletcher, 1888; Australia. 
11. C. mediocris, Fletcher, 1888; Australia. 
12. C. tenuis, Fletcher, 1888; Australia. 

13. C. illawarre, Fletcher, 1888; Australia. 

14. C. mudgeanus, Fletcher, 1888; Australia. 
15. C. sloanei, Fletcher, 1888; Australia. 

16. C. oxleyensis, Fletcher, 1888; Australia. 
17. C. purpureus, Michaelsen, 1889; Australia. 


See Fletcher, ‘Proc. Linn. Soc., N.S.W.,’ vols. i, ii, iv; 
Beddard, ‘ Proc. Zool. Soc.,’ 1887; Michaelsen, ‘Jahrb. d. ham- 
burgischen wiss. Anstalten,’ vi, 1889. 


Genus 4. DipymocastsEr, Fletcher, 1886. 
Sete separate, nearly equidistant. 
Clitellum feebly developed ; occupies Somites xiv to xvi11. 
Male pores on Somite xvui1, on papille. 


226 Ww. B. BENHAM. 


Prostates flattened, equally bilobed. 

Sperm-sacs.—Two pairs in Somites 1x, x11. 

[Prostomium small; dorsal pores present. 

Spermathece, three pairs, greatly elongated, in Somites vir, 
VIII, 1X; their apertures in Ix, x, x1; not intersegmental in 
position. 

Two gizzards in Somites vi, vi. 

No accessory intestinal diverticula. 

The intestine is dilated and very vascular in Somites x to 
xvi; the following region of the intestine is stated to be 
coiled like a corkscrew. 

Anterior septa greatly thickened. 

Dorsal vessel doubled in each somite. | 

Species 1. D. sylvaticus, Fletcher; Australia. 

See Fletcher, ‘ Proc, Linn. Soc. N.S.W.,’ 2nd ser., vol. i, 

1886. 


Genus 5. Perissocaster, Fletcher, 1887. 


Sete not in couples, but all close together on the ventral 
surface. 
Clitellum on Somites (xm) xiv to xvit1, complete ven- 
trally except on last somite. 
Male pores slit-like in Somite xvimr (?); penial set 
present, 
Prostates unequally bilobed. 
Sperm-sacs.—Two or four pairs in Somites Ix, xX, XI, XII. 
Testes and ciliated rosettes free, in Somites x, XI. 
Three gizzards in Somites v, VI, vit. 
[Prostomium wide ; no dorsal pores. 
Spermathecze with appendices, two pairs, in Somites viii ; 
Ix; apertures anterior, intersegmental. 
Intestine with (?) calciferous glands in Somites 1x to xiv; 
salivary glands present around pharynx. | 
Species 1. P. excavatus, Fletcher, 1887; Australia. 
2. P. nemoralis, Fletcher, 1888 ; Australia. 
3. P. queenslandica, Fletcher, 1888; Australia. 


AN ATTEMPT TO CLASSIFY EARTHWORMS. 227 


Genus 6. Dicnoaaster, Beddard, 1888. 

Sete in couples; ventral sete absent in Somites xvi, 
XVIII, XIX. 

Clitellum on Somites x1 to xx; more feebly developed 
ventrally. 

Spermiducal pores on Somite xvi1; no penial sete. 

Two pairs of prostate-pores on Somites xviri, xix. 

Sperm-sacs.—Three pairs in Somites x, x1, x11, connected 
across the middle line both above and below the intestine. 
Testes and ciliated rosettes in Somites x and x1, enclosed in 
sperm-sacs, 

Prostates tubular, slightly coiled, in Somite xvi. 

Additional club-shaped prostates, two pairs in Somites xv111 
and xix, without connection with sperm-ducts. 

Two gizzards, each occupying two Somites, vit, vu, and 
EX, X 

[Dorsal pores are present. Only one pair of spermathece 
in Somite vi11 ; aperture near ventral mid-line. 

Calciferous glands in xv, xvi, xvu1.] 


Species 1. D. damonis, F. E. B., 1888; Fiji. 
See Beddard, ‘ Quart. Journ. Micr. Sci.,’ xxix. 


Genus 7. Dicaster, E. P., 1872. 

The setz in four couples. 

The clitellum on Somites xiv to xv, complete ventrally, 
though not so well marked as on the dorsal surface. 

The male pores on Somite xvii (xvu, Perrier), 

Penial sete present. 

Sperm-sacs in Somites rx, x1 (or x, x1, Perrier). 

Prostates lobulated. 

Two gizzards in Somites v and vi. 

Tn Somites v and vii nephridia in groups, forming “ pepto- 
nephridia,” the duct of which opens into the pharynx. 

Nephridia tufted; though apparently, according to 
Fletcher, large ones are present in addition. 

Dorsal pores are present. 


228 W. B. BENHAM. 


Species 1. D. lumbricoides, E. P., 1872; New South 
Wales. 

2. D. armifera, Fletcher, 1886; New South 
Wales. 

3. D. perrieri, Fletcher, 1888; New South Wales. 


See Perrier, ‘ Nouv. Arch. du Mus. d’Hist. Nat. Paris,’ 
vill, 1872; Fletcher, ‘ Proc. Linn. Soc. N.S.W.,’ 2nd ser., 
i and ili. 

Remarks on the Typheide. 


The beautiful monograph on Megascolides australis by 
Professor Baldwin Spencer gives the details of the anatomy gene- 
rally found in this group. Many of the genera are very closely 
allied, no doubt. Fletcher’s descriptions of the various species 
are only verbal, and unaccompanied, except in his first paper, by 
any figures. For instance, the description given by him of Peris- 
sogaster and of Didymogaster rather lead me to think that 
the two genera should be included in one genus. In the 
former the prostates are said to be unequally lobed, i.e. the 
posterior lobe occupies two somites; in the.latter the two 
lobes are equal. In Perissogaster, however, no dorsal pores 
are present; in Didymogaster they are present. In Didy- 
mogaster the spermathecal apertures are not intersegmental 
as is the case in Perissogaster and other germs. This 
seems to be the only point of real difference; for, as I 
have remarked, the number of gizzards cannot of itself 
furnish a generic character. Dichogaster differs from any 
other genera in possessing three pairs of prostates; the first 
pair being in connection with the sperm-ducts, the two hinder 
pairs being independent of them. Typhzus, again, is a well- 
marked genus, in possessing a single pair of sperm-sacs and 
testes, 

In some species of Cryptodrilus, Fletcher and Beddard 
describe large nephridia in all somites, and nephridiopores 
which have an alternate or irregular arrangement with regard 
to sete. It is very probable that a network will be found, in 
all cases, in addition to these large nephridia, 


AN ATTEMPT TO CLASSIFY EARTHWORMS. 229 


In one species, Cr. purpureus, there are five median sper- 
matheceze, instead of two pairs usual to the genus. 


2. Spermiducal pores on Somite xv111; two pairs of tubular, 
more or less convoluted prostates, in Somites xvii and xix, and 
opening to the exterior on these somites. 


Family II. Acanthodrilide, Rosa (= partly L. postclitelliens, 
E. P.= partly Acanthodrilide, Claus, Vejdovsky). 


Setz, eight or twelve per somite, in couples or separate. 

Clitellum occupies x11 or xiv to x1x, or fewer somites, or 
extends to Somite x; either complete ventrally throughout 
or only anteriorly. 

Spermiducal pores, one pair on Somite xvii1, in line 
with the inner sete. 

Prostate-pores on Somites xvir and xix, in same line as 
spermiducal pores. 


Genus 8. AcantHopritvs, E. P., 1872 (= Mandane, 
Kinberg, 1866). 

Sete eight, in couples or separate. 

Clitellum usually occupies Somites x11 to x1x; some- 
times one additional somite at each end, sometimes fewer ; 
usually complete in anterior somites, but leaving a non-glan- 
dular ventral area on Somites xvii, XVIII. 

S perm-sacs.—Two pairs in Somites x1, XII. 

Prostates usually accompanied by sacs containing special 
*‘penial ” sete. 

[Prostomium more or less dovetailed into the peristomium, 
Dorsal pores are present. 

Testes and ciliated rosettes in Somites x and x1. 

Spermathecz with appendices, and sometimes accompanied 
by sacs containing “‘ anterior penial ” setz ; apertures in a line 
with the inner couple of sete. 

Gizzard usually single, occupying two somites; rarely two 
separate gizzards in vi and VIII. 

Pepto-nephridia are frequently present in the anterior 
somites. Calciferous glands present. 

Dorsal vessel sometimes double. 


230 


W. B. BENHAM. 


Some species have been described as possessing large 
nephridia without a network of small tubules (see above). ] 


Species 1 


2. 


“IO 


9. 


10. 
lel 
12. 


13. 
14. 
15. 
1G; 


lee 
18. 
19. 
20. 
ke 
22. 


A. obtusus, E. P., 1872; New Caledonia. 
A. ungulatus, E. P., 1872; New Caledonia. 


3. A. verticillatus, EH. P., 1872; Madagascar. 
4. 
5. A. capensis, F. KE. B., 1884; Cape of Good 


A. kerguelenensis, KE. R. L., 1879; Kerguelen. 


Hope. 


. A. buttikoferi, Horst, 1884; Liberia. 
. A. schlegelii, Horst, 1884; Liberia. 
. A. layardi, F. E. B., 1886; New Caledonia. 


(Horst believes this species to be identical with 
A. ungulatus.) 

A. nove-zelandie, F. E. B., 1885; New 
Zealand. 

A. multiporus, F. E. B., 1885 ; New Zealand. 

A. dissimilis, F. E. B., 1885; New Zealand. 

A. neglectus, F. E. B., 1887; New Zealand 
(?= dissimilis). 

A. annectens, F. E. B., 1888; New Zealand. 

A. beddardi, Horst, 1888 ; Liberia. 

A. australis, Michaelsen, 1889; Australia. 

A. georgianus, Michaelsen, 1889; South 
Georgia. 

. littoralis, Kinberg, 1866; Magellan. 

. scioanus, Rosa, 1888 ; Africa. 

. bovei, Rosa, 1889; Magellan. 

. antarcticus, F. E. B., 1889; New Zealand. 

rose, F. E. B., 1889; New Zealand. 

. dalei, F. E. B., 1890; Falkland Isles. 


> b> > > > b> 


For descriptions of this genus and its species see Perrier, 
‘Nouv. Arch. du Mus. d’Hist. Nat. de Paris, vii, 1872; Bed- 
dard, ‘Proc. Zool. Soc.,? 1885—1887 ; ‘ Quart. Journ. Micr. 
Sci.,’? xxviii, xxix, and xxx; Horst, ‘Notes from Leyden 
Museum,’ ix, x; Michaelsen, ‘ Jahrb. d. hamburgischen wiss. 
Anstalten,’ vi, 1889; Rosa, ‘Ann. d. Mus. Civico d. Stor. 
Nat. di Genova,’ ser. 2, vi, 1888, and vil, 1889. 


AN ATTEMPT TO CLASSIFY EARTHWORMS. 251 


Genus 9. Tricaster, Benham, 1886 (= Benhamia, 
Michaelsen, 1889). 

Setze in four couples, all on the ventral surface ; individual 
sete of each couple close together. 

Clitellum occupies Somites xiv to xx; complete ventrally 
only on the first few somites. 

Spermiducal pores in xvul, and prostate-pores in 
XVII and xIx, in a large pit or fossa occupying the middle of 
the ventral surface of Somites xvit to xx, the margins of 
which are formed by two papille. 

Sperm-sacs not observed. 

Prostates asin Acanthodrilus. No penial sete. 

[Prostomium not dovetailed into peristomium. No dorsal 
pores are present. 

Spermathecz simple pear-shaped sacs without appendices, 
opening close to mid-line on ventral surface. 

Three gizzards in Somites vil, vit1, and 1x. No calciferous 
glands. Anterior masses of nephridial tubules in Somites tv, 
V, VI, grouped to form ‘‘ pepto-nephridia.”’] 

Species 1. T. lankesteri, W. B. B., 1886; St. Thomas, 

West Indies. 
2. T. rosea, Michaelsen, 1889; West Africa. 

See Benham, ‘ Quart. Journ. Micr. Sci.,’ xxvii; Michaelsen, 
‘ Jahrb. d. hamburgischen wiss. Anstalten,’ vi, 1889. 


Genus 10. De1nopritus, Beddard, 1888. 

Setz twelve per somite, nearly equidistant. 

Clitellum complete ventrally ; occupies only Somites xiv 
to XVI. 

Spermiducal pores, prostate-pores, sperm-sacs, &c., 
as in Acanthodrilus. 

[Prostomium dovetailed into peristomium. 

Spermathecz with three small globular appendices, two pairs 
in Somites viii and 1x. 

A single gizzard occupies Somites vi, vu. No calciferous 
glands. 


252 W. B. BENHAM. 


The dorsal vessel is double throughout its length, and is 
enclosed in a special ceelomic tube. ] 


Species 1. D. benhami, F. E. B.; New Zealand. 
See Beddard, ‘ Quart. Journ. Mier. Sci.,’ xxix. 


DovustFrut GENUS. 


Neodrilus monocystis, F. HE. B., New Zealand. 
Founded on a single specimen, and differs from Acantho- 
drilus in possessing a single pair of prostates and a single 
pair of spermathece. It appears to me very doubtful whether 
this should be considered as a new genus, or whether the 
characters are merely some peculiar variations of Acautho- 
drilus. 


Remarks on the Acanthodrilide. 

The genus was originally characterised by the presence 
of two pairs of male pores; it is only recently that Beddard 
has shown that these pores belong to the prostates, and that the 
sperm-ducts open by a pair of pores on the eighteenth somite. 
The chief points of difference between Acanthodrilus and 
Trigaster lie in the fact that the male pores and atriopores 
in the latter genus are in a pit (in my original description 1 
placed the atriopores in xvi, xvi11; I believe that this statement 
is wrong, and that the prostate-pores and spermiducal pores are 
placed as in Acanthodrilus), and in the absence of penial or 
copulatory setee and the presence of three gizzards. When the 
genus was formed, the only worm with more than one gizzard 
(except Moniligaster) was Digaster. That the existence 
of three gizzards is not generic is now established by the 
formation of Michaelsen of a species, T. rosea, with only 
two gizzards. 

Three species of Acanthodrilus are known with two 
gizzards—A. buttikoferi and A. beddardi of Horst; 
and A. scioanus, Rosa. 

Horst also figures the prostate-pores in A. schlegelii as 
situated in a fossa. 

But the great extent of the clitellum in Trigaster, to- 


Bisa or. 


AN ATTEMPT TO CLASSIFY EARTHWORMS. 233 


gether with the position of the spermathecal pores close to the 
ventral mid-line, and the general appearance of the worm, 
warrant the retention of the genus. I may mention here that 
frequently a mere description of the position of pores and 
organs, unaccompanied by figures, might lead to the associa- 
tion of two worms, an examination of which would leave a very 
different impression as to their relation. 

The genus Deinodrilus is sufficiently interesting and pecu- 
liar in the possession of twelve sete per somite; but this interest 
is greatly enhanced on comparison of the internal organs with 
those of Acanthodrilus on the one hand and of Pericheta 
on the other. 

Some species of Acanthodrilus have large nephridia, the 
power of which alternate in position; but no statement is made 
as to whether these nephridia are accompanied by a network : 
I believe we may expect this to be the case. Many species 
have the dorsal vessel double to a greater or less extent. 

b. Setee more than twelve (usually many more) in most of 
the somites, arranged in a ring, which is continuous all 
round, or interrupted dorsally and ventrally. 


Family III. Perichetide, Claus (= partly L. postclitel- 
liens, E. P.= Perichetidz + Pleurochetide, Vejdovsky). 


Clitellum completely surrounding the body, obliterating 
entirely the intersegmental grooves, and extending over all or 
some of the Somites x11I—xvu. 

Spermiducal apertures on Somite xvii, on the ventral 
surface. 

Oviducal apertures close together on Somite xiv. 


Genus 11. Prrtcnmta, Schmarda, 1861 (includes Mega- 
scolex, Templeton, 1844; Pleurocheta, Beddard, 1883 ; 
and many of Kinberg’s genera). 

Setz from twenty to eighty, or even 100 per somite, on a 
ridge (at least in spirit specimens), either in a continuous ring 
or interrupted by a greater or less gap in the dorsal or ventral 

VOL. XXXI, PART 11.—NEW SER. Q 


234. WwW. B. BENHAM. 


mid-line, or both. Sete usually small and of equal size, and 
generally equidistant, though in some species more or fewer 
of the more ventral ones are larger than the rest. On the 
clitellum the sete are invisible. 

Clitellum on Somites xiv to xvi or xvi, rarely only two 
or more than four; well defined, and altogether obliterating 
the intersegmental grooves. 

Spermiducal pores in Somite xvim, usually rather later- 
ally placed. 

Oviducal pores in Somite xtv very close together, or more 
usually single and median. 

Penial setz and various “ copulatory papille” are fre- 
quently present. 

Sperm-sacs, in Somites x1 and xu, two pairs, rarely more, 
and sometimes connected by median sacs enclosing testes. 

Prostates.—A pair in Somite xviii, lobed or greatly sub- 
divided, or even digitate; the duct after being joined by the 
sperm-duct is very muscular and probably protrusible; it 
may be called a “ penial duct.” 

[Worm cylindrical ; prostomium sometimes dovetailed into 
peristomium, sometimes not dovetailed. 

Dorsal pores present. 

Testes and ciliated rosettes in Somites x and x1, sometimes, 
at any rate, enclosed in the median portion of the sperm-sac. 

Ovaries in Somite x11. 

Spermathece, usually only two pairs, in Somites vir and 1x, 
opening anteriorly ; sometimes only one pair ; sometimes more 
than two pairs. Usually with an appendix which varies in shape. 

Gizzard occupies any position between Somites v and x.: 
usually occupying three Somites, vii1, 1x, and x. 

In most species a pair of tubular ceca in Somite xxvi are 
present. | 

Species 1. P. houlleti, E. P., 1872; Calcutta (and Nice) ; 

Bahamas (F. E. B.) ; Manila (F. E. B.). 

2. P. posthuma, Vaillant, 1869 = P. affinis, 
K. P., 1872 ; Cochin China; Java (Horst) ; 
Philippines (F. E. B.). 


AN ATTEMPT TO CLASSIFY EARTHWORMS. 235 


Species 3. P. robusta, E. P., 1872 = partly P. cingu- 


4,.:P 
bP 
6,; 'P 
/ iiaed ee 
$272 
oP 
10:2 
cn a 
12.) PB: 
13.0 P. 
14. P 
1550P: 
1G34P. 
AI pet 
133-2 
1? 
20.8 
Pal a 
22. 2 
23. P 
24. P. 
25. P. 
26.52. 
4 fe 
ao: Ps 
ooe 
30. P 
ole. 


lata, Sch. and Vaillant ; Mauritius, Manila, 
Nice, Bahamas. 


. aspergillum, H. P. 1872 (loc.?); Bermuda 


(F. E. B.). 


. quadragenaria, H. P. 1872 = partly P. 


cingulata, Sch. and Vaillant; Hast Indies. 


. elongata, EH. P., 1872; Peru (? indigenous). 
. indica, Horst, 1883; Sumatra; New Cale- 


donia (F. E. B.). 


. sumatrana, Horst, 1883 ; Sumatra. 
. hasseltii, Horst, 1883 ; Sumatra. 

. Sleboldii, Horst, 1883; Japan. 

. japonica, Horst, 1883; Japan, 


musica, Horst, 1883 ; Java. 
capensis, Horst, 1883; Cape of Good Hope. 


. annulata, Horst, 1883; Malay. 


cerulea, Templeton, 1844; Ceylon. 
ceylonica, F. E. Beddard, 1885; Ceylon. 


. armata, F. E. Beddard, 1883; Calcutta; 


Burmah (Rosa) ; Nias, near Sumatra (Rosa). 


. horsti, F. E. Beddard, 1886; Manila. 
. newcombei, F. E. Beddard, 1887; Aus- 


tralia. 


. upoluensis, F, E. Beddard, 1887; Upolu, 


Pacific Isles. 


. lawsoni, A. G. Bourne, 1886 ; India. 
. bivaginata, A. G. Bourne, 1886; India. 
. gracilis, A. G. Bourne, 1886 ; India. 


stuarti, A. G. Bourne, 1886; India. 
burliarensis, A. G. Bourne, 1886; India. 
hulikalensis, A. G. Bourne, 1886; India. 


. mirabilis, A. G. Bourne, 1886; India. 


salettensis, A. G. Bourne, 1886; India. 


. australis, Fletcher, 1886; Australia. 
. coxii, Fletcher, 1886; Australia. 


tenax, Fletcher, 1886; Australia, 


236 


Species 32. 
33. 
34. 


35. 


36. 
37. 
38. 
39. 
40. 
41. 
42. 
43. 
4A. 
45, 
46. 
47. 
48. 
“49. 
50. 
alt 
52. 
53. 
54. 
55. 
56. 


WH NWN 


W. B. BENHAM. 


. austrina, Fletcher, 1886; Australia. 
. barronensis, Fletcher, 1886; Australia. 


darnleiensis, Fletcher, 1886; Australia. 


. gracilis, Fletcher, 1886; Australia. 


peregrina, Fletcher, 1886; Australia. 


. queenslandica, Fletcher, 1886 ; Australia. 


bakeri, Fletcher, 1887 ; Australia. 
dorsalis, Fletcher, 1887; Australia. 


. canaliculata, Fletcher, 1887; Australia. 


exigua, Fletcher, 1887; Australia. 
fecunda, Fletcher, 1887; Australia. 
hamiltoni, Fletcher, 1887; Australia. 
monticolla, Fletcher, 1887 ; Australia. 
raymondi, Fletcher, 1887; Australia. 


. stirlingi, Fletcher, 1887; Australia. 


wilsoniana, Fletcher, 1887; Australia. 
birmanica, Rosa, 1888; Burmah. 


. fee, Rosa, 1888; Burmah. 


modigliani, Rosa, 1887; Nias (Sumatra). 
antarctica, Baird 1873; New Zealand. 
intermedia, Beddard, 1889; New Zealand. 


. attenuata, Fletcher, 1888; Australia. 
. enormis, Fletcher, 1888; Australia. 

. dissimilis, Fletcher, 1888; Australia. 
. macleayi, Fletcher, 1888; Australia. 


Doubtful Species.—Some of Perrier’s, viz. P. bicincta, 
P. luzonica, P. coerulea, P. biserialis, P. juliana. 
Schmarda’s P. leucocycla, P. viridis, P. brachycycla, 
P. cingulata. 
tima, Rhodopis, Lampito. 

See Perrier, ‘Nouvelles Arch. du Mus. d’Hist. Nat. de 
Paris,’ viii, 1872. Beddard, ‘ Ann. Mag. Nat. Hist.,’ 5th ser., 
vol. xvii, 1886; ‘Proc. Zool. Soc.,’ 1886; ‘ Proc. Roy. Soc. 


Edin.,’ xiv, 1887. 


Kinberg’s genera, Amyntas, Nitocris, Phere- 


Rosa, ‘Ann. d. Mus. Civico d. Storia 


Nat. di Genova,’ 2nd ser., vi, 1888, vol. vii, 1889; Fletcher, 
‘Proc. Linn. Soc. N.S.W.,’ 2nd ser., vols. i, ii, ii; A. G. 
Bourne, ‘ Proc. Zool. Soe.,’ 1886. 


AN ATTEMPT TO CLASSIFY HARTHWORMS. 237 


Remarks on the Perichetide. 


Although some fifty species of this genus have been formed 
within the last few years (besides those which have been cha- 
racterised only by their external anatomy, and which must be 
in many cases discarded), yet very frequently insufficient data 
have been given. On the whole it is a well-defined family, but 
the single genus may really be capable of subdivision. 

I have already mentioned my reason for removing Perionyx 
from the family, a proceeding which may at first appear 
arbitrary. 

The character of the prostomium and the presence or absence 
of the characteristic intestinal czeca, as well as the position of 
the gizzard, may prove to be of generic value. The observa- 
tions on the excretory system are in most cases very superficial 
and incomplete, and frequently no mention is made as to 
whether in a particular species large “ nephridia” or a small 
network of tubules is present. Where these observations have 
been carefully made the presence of a pair of large nephridia! 
appears to be associated with the absence of the intestinal 
ceca, a forward position of the gizzard in Somite v or v1, 
and with the existence of three pairs of spermathece. But 
there are too many apparent exceptions to generative on this 
point at present. 

Amongst the more peculiar species may be mentioned P. 
indica, Horst, where some of the more ventral sete are larger 
than the rest; P. hasseltii, Horst, in which the ventral sets 
are more closely placed; P. stuarti, Bourne, with two pairs 
of male pores and two pairs of prostates. P. bakeri and 
P. intermedia have prostates resembling those of Acan- 
thodrilus. 

The number of sete per somite, position of copulatory 
papille, extent of clitellum, number of spermathece and shape 
of appendix, and of the prostates, serve as the leading characters 
in which the species differ from one another. 

The worms figured by Schmarda are only described so far 

1 Probably accompanied by a network. 


238 WwW. B. BENHAM. 


as their external anatomy is concerned, and cannot be 
recognised with certainty. Kinberg’s genera must be relegated 
to oblivion. 


Branch II. Mrcanerurica. 

The excretory system is in the form of large, greatly coiled 
tubes unaccompanied by a network of small tubules. Each 
nephridium opening into the celom by a funnel: usually a 
pair in each somite, though the most anterior somites may be 
deprived of nephridia. 

a. A prostate is present. 

a. Male pores intersegmental, immediately behind Somite 
x or xr; clitellum developed around this and the 
adjacent somites. 


Family IV. Moniligastridz, Claus, Vejdovsky Rosa (= L. 
aclitelliens, E. P.). 
Genus 12. Monizicaster, E. P., 1872. 

Sete in four couples. 

Clitellum observed in only one species (M. sapphirina- 
oides, A. G. B.), where it occupies Somites x to x111; it isill- 
marked. 

Male pores between Somites x/x1 ; or x1/xil. 

Oviducal pore on Somite x11 (or xiv). 

S perm-sacs, one pair occupying Somite x1 (Horst), or 1x 
and x (Beddard). 

Ovisac in Somites xrv to xv1, or fewer somites. 

Nephridiopores in a line with the outer couple of sete. 

Prostates small, or large and tubular. 

Testes in Somite rx (Beddard). 

S permathece in Somite viir or rx. 

The nephridium has a long cecal prolongation of the duct 
beyond the point at which the short slightly coiled tubule 
enters. There is apparently no modification of the anterior 
nephridia. 

Gizzard moniliform, four-lobed, in all or some of the Somites 
x111 to xxii (sometimes there is an additional gizzard ante- 
riorly). 


AN ATTEMPT TO CLASSIFY EARTHWORMS, 239 


Species 1. M. deshayesii, E. P., 1872; Ceylon. 
M. barwelli, F. E. B., 1886; Manila. 
. houteni, Horst, 1887 ; Sumatra. 
. grandis, A. G. B., 1886; India, 
.uniquus, A. G. B., 1886 ; India. 
sapphirinaoides, A. G. B., 1886; India. 
robustus, A. G. B., 1886; India. 
. papillatus, A. G. B., 1886 ; India. 
9. M. rubens, A. G. B., 1886 ; India. 

10. M. minutus, A. G. B., 1886; India. 

See Perrier, ‘Nouv. Arch. du Mus. d’ Hist. Nat. de Paris,’ viii, 
1872; Beddard, ‘Ann. Mag. Nat. Hist.,’? 1886; ‘Zool. Anzieger,’ 
1887, No. 268, and 1889; ‘ Quart. Journ. Micr. Sci.,’ vol. xxix, 
1888 ; Horst, ‘ Notes from Leyden Mus., ix; Bourne, ‘ Proc. 
Zool. Soc.,’ 1886. 


So) GoD A Se) 
erie sie se 


Remarks on Moniligastride. 


The three authors who have studied the internal anatomy of 
the genus Moniligaster differ from one another in their 
statements as to the position of the male pores and other 
organs. 

Perrier, in his description of M. deshayesii, describes, as 
is well known, two pairs of male organs; the ducts of the first 
pair opening between Somites vii and vit, those of the second 
pair between Somites x and xr. In connection with each of 
the first pair of ducts is a double gland (his ‘ prostate’); and 
similarly there is a gland in connection with the other pair of 
ducts, which is fairly elongated (his “ seminal vesicle ’’). 

Beddard’s original description in 1886, as well as his more 
recent figures, shows considerable differences from this arrange- 
ment, apart from the question of numbering. He identifies 
Perrier’s first pair of “testes” as spermathece ; the “ prostates ” 
(which are not represented in M. barwelli) he suggests may 
be accessory sacs, which are so frequently found in connection 
with spermathecz, whereas Horst identifies these “prostates ”’ 
of Perrier as the true spermathece. 

Horst’s figures are much more like those of Perrier than 


240 WwW. B. BENHAM. 


are Beddard’s; and were it not that the spermathece and 
sperm-sacs in M. houteni occur one somite behind those 
of M. deshayesii we might believe that he was dealing 
with the same species. In fact, we have here another example 
of the difficulty of accurately counting the somites in earth- 
worms. Beddard has quite recently (October, 1889) altered his 
previous numbers for M. barwelli, owing to the discovery of 
a small setigerous somite following the peristomium, so that the 
male pores of M. barwelli are, as in M. deshayesil, 
between Somites x and x1. The spermathecal pores, too, 
which were previously given as between vi and vii, now agree 
with the pores of Perrier’s “ anterior sperm-ducts,” in being 
placed between Somites vir and viii. 

The diagram accompanying this paper is taken from Horst’s 
figure of M. houteni, and the position of the various organs 
differs somewhat from that in the other two species. As will 
be seen, the sperm-sacs are in Somite x1 (and probably also 
the testes and funnels of the sperm-ducts which open externally 
between Somites xr and x11). The ovipore is in Somite 
xrv, and probably the ovary is in Somite x111, these organs 
being therefore in the normal position. Here the prostate 
is a large structure, whilst in M. barwelli it is extremely 
small. 

The spermatheca in Somite 1x has a long duct opening 
anteriorly. 

The “ ovary”’ of Perrier’s species is not the true gonad, but 
the “ ovisac,” or receptaculum ovorum, and recalls the way 
in which the ova push their way back through several somites 
in Microdrili. The ovary is unknown. Beddard has figured 
(‘ Quart. Journ. Mier. Sci.,’ xxix, pl. xi) the oviduct with its 
funnel and external aperture; but the numbering here given 
is revised in the ‘ Zool. Anzeiger,’ No. 318, where the external 
aperture is placed on Somite x11, and the funnel in Somite 
x1, so that in all probability the gonad is in Somite x1, 

Prof. Bourne has given us a few facts about seven new 
species of the genus, chiefly as regards the position of the 
gizzard, but says nothing about the genital organs, The most 


AN ATTEMPT TO CLASSIFY EARTHWORMS. 241 


interesting point in this connection, however, is his descrip- 
tion of aclitellum in M. sapphirinaoides occupying Somites 
X—XIII, a structure previously denied to the genus. 

The recorded absence of a clitellum is probably due to the 
fact that, as in the water-worms, this structure is only deve- 
loped at the breeding season. 

The anterior gizzard, which Perrier described, has not been 
recognised in the later species. 

I believe Moniligaster to be more nearly related to the 
ancestors of earthworms than any other genus we know of, 
as I have pointed out in Part VI of this paper. 


b. Male pores on Somite xvit or xvitt. 
Clitellum occupies all or any of the Somites x111 to xviii. 


1. Eight setze per somite, in couples or separate. 


Family V. Eudrilidz, Claus (=Lumbriciens intraclitel- 
liens, E. P., in part=part of family Eudrilide, Vejdovsky, 
Rosa). 


The eight sete are in couples or separate; the clitellum, 
complete ventrally, extends over all or some of the Somites 
XIII tO XVII. 

The male pores are behind the clitellum, or just within 
its limits. 

The prostate is simply tubular, convoluted, or lobed. 

Spermathecze usually with diverticulum. 

Typhlosole absent. 

The duct of the nephridium is not produced into a cecum, 
nor is there any modification of the anterior nephridia. 


Genus 13. Evpritvs, E. P., 1872. 


Setz in four couples. 

Clitellum covers Somites (x111) xiv to xviII. 

Male pores large, on Somite xvir (from it the curved 
chitinous penis sometimes protrudes), in line with inner couple 
of sete. 


242 W. B. BENHAM. 


Female pores on Somite xiv, slit-like, large, dorsad of 
the inner couple of sete. 

Nephridiopores in line with outer sete (or inner sete 
in E. sylvicola, Beddard). 

Generative Apparatus.—Three pairs of sperm-sacs in 
Somites x, x1, x11. Testes and ciliated rosettes in Somites x, 
xr, enclosed in median sperm-sacs. The two sperm-ducts of 
each side run separately to the prostate, which is much elon- 
gated, and occupies Somite xvir and following somites. This 
communicates with a “ bursa copulatrix” in Somite xvi1, 
into which also open two small glands. The bursa contains a 
curved chitinous penis. 

There appear to be two pairs of ovaries (Beddard) in 
Somites x111 and xiv, enveloped in membranes which are con- 
tinuous with the wall of the spermatheca. Into the neck of 
the latter there also opens an albumen gland. The“ ovary” 
in Somite xrv is also an ovisac. 

[The gizzard occupies Somite v1. 

In Somites x, x1, there are ventral diverticula of the ali- 
mentary tube; in Somite x11, lateral calciferous diverticula. 

The nephridium consists of a slightly coiled tubule, the 
terminal portion of which is only slightly dilated to form a 
duct. | 


Species 1. E. decipiens, E. P., 1872; Antilles. 
2. KE. lacazii, E. P., 1872; Martinique. 
3. E. peregrinus, HE. P., 1872; Rio Janeiro, 
Surinam. 
4, E. boyeri, F. E. B., 1886; New Caledonia. 
5. E. sylvicola, F. E. B., 1887; British Guiana. 


Note.—Horst believes that the first four of these are in 
reality the same species, and proposes to retain the name of 
E. decipiens for them. 


See Perrier, ‘Nouv. Arch. du Mus. d’Hist. Nat. de Paris,’ 
viii, 1872; Beddard, ‘ Proc. Roy. Soc. Edin.,’ xiii, 1885-6; 
‘Proc. Zool. Soc. Lond.,’ 1886-7; ‘ Journ. Anat. and Phys.,’ xxii, 
1887; ‘Zool. Anzeiger,’ 1888, No. 293; ‘ Encycl. Brit.,’ 9th 


—~ 


AN ATTEMPT TO CLASSIFY EARTHWORMS. 243 


edition, “ Worms;” Horst, ‘ Notes from Leyden Museum,’ ix; 
Beddard, ‘ Quart. Journ. Micr. Sci.,’ xxx. 


Genus 14. TELEUDRILUS, Rosa, 1888. 


The eight set, in couples, are rather far apart. 

The clitellum includes Somites xtv—xvit. 

The male pore is median on Somite xrx; the pair of 
oviducal pores between Somites x1v and xv; a median 
spermathecal pore between Somites x11 and xtv. 

Nephridiopores in line with outer setz. 

The testes in Somites x and x1, enclosed in a sac-like con- 
tinuations of the sperm-sacs, which lie in Somites xz and x11. 
The ciliated rosettes are in the latter somites. 

The two prostates open into a median copulatory sac, 
communicating with the exterior and receiving another 
median sac. 

The ovary is continuous with the wall of the ovisac, into 
which the funnel of the oviduct opens. There is a commu- 
nication between the ovisac and the neck of the spermatheca 
on each side. 

[The gizzard occupies Somites vi and vit (? also v); there is 
a pair of lobed calciferous diverticula in Somite x111, and 
ventral diverticula in Ix, x, XI. 

Nephridia simple, as in Eudrilus.] 


Species 1. T. raggazii, Rosa, 1888; Africa. 


See Rosa, ‘Ann. d. Mus. Civico del Storia Nat. d. Genova,’ 
Series 2, vi, 1888. 


Genus 15. Pontopritvs, E. P., 1881. 


The eight set are separate. 

The clitellum, which is complete, occupies Somites x111 to 
XVII. 

The male pores in Somite xvitr. 

The prostate is tubular and convoluted. 

Sperm-sacs.—Two pairs, in xt and x11. Testes and funnels 
in 1x, Xx. Ovary and oviduct as usual. 


244 W. B. BENHAM. 


The nephridia do not commence till Somite xv; the pores 
are in line with the second seta. The “ duct ” of the nephri- 
dium is feebly marked. 

There is no gizzard, no typhlosole, no subneural vessel, no 
dorsal pores. 

[Found on the sea-shore. 

Two pairs of spermathece, which have small appendices in 
Somites vir and 1x, opening anteriorly. | 

Species 1. P. littoralis, Grube, 1855 ; Villa-Franca. 
2. P. marionis, E. P., 1874; Marseilles. 
See Perrier, ‘ Arch. d. Zool. Exp. et Gen.,’ ix, 1881. 


Genus 16. Puotopritvs, Giard, 1887 (=Lumbricus 
phosphoreus, Dugés). 

The eight sete are separate. No. 1 setais near the middle 
line. 

Clitellum on Somites x11 to xvit. 

The male pores on Somite xvm1. There are “ penial” 
sete in this somite, and anterior penial sete in Somites x11 
and xIII. 

Genital organs as in previous genus. 

The nephridia commence in Somite xiv; the pores are in 
a line with the second seta. 

There is no gizzard, no typhlosole, no subneural vessel. 

[One pair of spermathecz in Somite rx. 

The prostomium does not encroach on the buccal somite. 
“ Septal glands” in Somites v to 1x, probably open dorsally. 

Four esophageal swellings in Somites x to x11. 

Small, transparent, rose-coloured worm, clitellum orange ; 
phosphorescent. ] 

Species 1. P. phosphoreus, Dug., 1837; Europe. 

See Giard, ‘Comptes Rendus, 1887; Rosa, ‘Boll. Mus. 

Zool. ed Anat. Comp. Univ. Torino,’ iii, 1888. 


Genus 17. Microsco.ex, Rosa, 1887. 
The setze in four couples; those of outer couple further 
apart than those of the inner couple. 


AN ATTEMPT TO CLASSIFY EARTHWORMS. 245 


The clitellum, complete, covers Somites x111 to xvi (xvi). 

The male pores are in Somite xvi. 

Sperm-sacs, testes, ovaries, as in preceding genus. 

The prostates lobate; penial sete present. 

The nephridia commence in Somite 1v; nephridiopores 
in front of the third seta. 

There is no gizzard, no typhlosole, no subneural vessel, nor 
dorsal pores. 

[Small, transparent ; white clitellum. 

One pair of spermathecze in Somite 1x.] 

Species 1. M. modestus, Rosa; Italy. 

See Rosa, ‘ Boll. Mus. Zool. ed Anat. Comp. Univ. Torino,’ 

ii, 1887, and iii, 1888. 


Genus 18. Ruopopritvs, Beddard, 1889. 

The sete separate, in eight series. 

The clitellum occupies Somites x1v to xvit. 

The prostates are tubular; penial sete present; the male 
ducts open independently of the prostates—all in Somite xvi. 

The sperm-sacs in Somites x1, x11. 

Prostomium incompletely dovetailed into the peristomium. 

[Spermathecee.—Four pairs, in vi, vil, v1, 1x; each with 
appendix. 

A gizzard is present in Somite v. 

No calciferous glands. 

Nephridiopores in front of third seta. 

Dorsal pores are present. | 

Species 1. R. minutus, F. E. B.; New Zealand. 
See Beddard, ‘ Proc. Zool. Soc.,’ 1889. 


Genus 19. Piutetuuvs, E. P., 1878. 
Setz eight, equidistant. 
The clitellum covers Somites xiv to xvi1, complete ven- 
trally. 
The male pores on Somite xviit. 
Oviducal pores on Somite x (7). 
The nephridial pores in line alternately with sete two 


246 W. B. BENHAM. 


and four except anterior four pairs, which open in front of 
third seta; nephridia simple, slightly coiled tubule, lying 
entirely within one somite (?). 

The sperm-sacs in Somite x11. 

Prostate tubular, convoluted. 

[Spermathecze.—Five pairs, in Somites v to 1x; very small, 
with coiled diverticulum. 

Ovary in Somite x (?). 

Gizzard in Somite v1; cesophageal glands in Somites x, x1, 
XU, 

The dorsal pores begin behind Somite vi. Lateral hearts 
in Somites x, x1, x11.] 


Species 1. P. heteroporus, E. P.; Pennsylvania. 
See Perrier, ‘ Arch. d. Zool. Exp. et Gen.,’ 11, 1873. 


Remarks on Eudrilide. 

I have here united with the peculiar genera Eudrilus and 
Teleudrilus a number of other genera which are much more 
normal in the arrangement of their genital organs than are 
these two; for I think, with Rosa, that Eudrilus need not 
form a type of a separate family. 

It is only lately that we have had a thorough description of 
the female genital organs of Eudrilus; and though from 
Perrier’s descriptions, and the earlier ones of Beddard, it 
appeared as if we had to do with a very abnormal type, Bed- 
dard’s more recent papers on the subject, and Rosa’s descrip- 
tion of Teleudrilus, remove some of the apparent peculiari- 
ties. But they both remain very different from other worms, 
in that the ovary is not freely dependent in the cclom, but 
enclosed in a sac, the walls of which are continuous with those 
of the oviduct; a similar condition of things is present in 
Microcheta in regard to the testis. And no doubt both these 
cases are in reality similar to the enclosure of the testes and 
rosettes in a common sac in Lumbricus and other forms, 
Here, however, the portion of ccelom separated by the wall 
of the sperm-sac is very considerable, whereas in the case 
of the ovary of Eudrilus and the testis of Microcheta, 


AN ATTEMPT TO CLASSIFY EARTHWORMS. 247 


this separated coelomic space is smaller, and has appeared more 
peculiar than it really is. As above mentioned, Kudrilus 
possesses two pairs of ovaries according to Beddard, the 
posterior pair serving apparently as ovisacs. 

Rosa has already pointed out the close relation between 
Pontodrilus, Photodrilus, and Microscolex. ‘These 
three forms serve to show the invalidity of Claparéde’s charac- 
teristics of “Terricole.’’ The absence of a gizzard is, no 
doubt, connected with the character of the food. 

Plutellus is altogether a peculiar form; the only descrip- 
tion we have of it is that by Perrier. The position of the ovi- 
ducal pore and of the ovary is so abnormal that a renewed 
examination is desirable. 


2. Sete more than eight (30—40) per somite. 


Family VI. Perionycide. 
Genus 20. Perionyx, E. P., 1872. 

Setz thirty to fifty per somite. 

Prostomium dovetailed incompletely into peristomium. 

Clitellum on Somites x1v—xvi1 or less, complete ven- 
trally ; intersegmental grooves not completely obliterated. 

Male pores close together, in a depression on Somite 
XVIII. 

Oviducal pore median, in Somite xrv. 

Prostate flattened, rounded; its pore common with the 
spermiducal pore. - 

[Genital organs as in Pericheta, but without a median 
sperm-sac. 

Gizzard in Somites vi and vir; no ceca or other diverticula 
of the canal. 

Nephridia large, paired; the duct not provided with a 
cecum; apertures irregularly arranged in some species, as 
in P. saltans. | 

Species 1. P. excavatus, HK. P., 1872; Cochin China, the 

Philippines, and Burmah. 
2. P. MclIutoshii, F. E. B., 1883; Burmah. 
3. P. saltans, A. G. B., 1886; India. 


248 WwW. B. BENHAM. 


See Perrier, ‘ Nouv. Arch. du Mus. d’Hist. Nat. de Paris, 
vili, 1872; Beddard, ‘Ann. Mag. Nat. Hist.,’ 5th Series, 
vol. xii, 1883; ‘Proc. Zool. Soc.,’ 1886; Bourne, ‘ Proc. Zool. 
Soc.,’ 1886; Rosa, ‘Ann. d. Mus. Civico d. Storia Nat. d. 
Genova,’ vi, 1888. 


Remarks on Perionycide. 


In external characters Perionyx agrees exceedingly closely 
with Pericheta. In the former, however, the male pores are 
close together in a median pit, whereas in most species of 
Pericheta they are rather wide apart, and on papille. 
Again, the clitellum does not so completely obliterate the 
segments and the grooves in Perionyx (nor are its limits so 
distinctly defined) as in Perichzta. The absence of ceca, 
and of any other diverticula of the alimentary canal, and the 
presence of large nephridia, are characters said to be found in 
some species of Perichzta. The median position of the ovi- 
ducal pore has certainly a striking resemblance to that of 
Pericheta. It may be possible to transfer those worms with 
large nephridia, with forward position of the gizzard, without 
ceeca, and with closely approximated male pores, which are at 
present regarded as species of Pericheta, to the genus 
Periony x. 

At present only three species have been described: P. ex- 
cavatus, H. P:; P. MeIntoshi1, F>E. B.; and Py saliane 
A. G. B.—the last two very briefly. 


B. There is no hollow prostate in connection with or in 
the region of the male pore. 


1. The male apertures are behind Somite xvii1, within the 
area occupied by the clitellum. 


a. Eight sete, separate or even alternate in some part of 
the body. There is only one pair of sperm-sacs, which 
extend through several somites. 


AN ATTEMPT TO CLASSIFY EARTHWORMS. 249 


Family VII. Geoscolecidex, Rosa (=partly L. intraclitel- 
liens, E, P.=partly Eudrilide, Claus, Vejdovsky = partly 
Geoscolecidz, Rosa). 

The eight setz have a tendency to separate, or even to be 
arranged alternately in consecutive somites, either throughout 
the body or only posteriorly. 

The clitellum commences behind Somite x1v usually, and 
extends over nine or more somites, intersegmental grooves not 
being obliterated. 

The sperm-sacs are very long; there is but one pair of 
testes and rosettes; the genital pores are very small, and 
may be accompanied by glandular swellings. 

A few of the anterior nephridia are larger than the following 
ones, and may even be collected into a mass forming a pepto- 
nephridium. 

The typhlosole is a mere dependent fold. 


Genus 21. GzoscoiEx, Leuckart, 1841 (=Titanus, E. P., 1872). 


The separation of the setz occurs posteriorly, but no 
alternation seems to occur. 

Clitellum is incomplete ventrally, and extends over Somites 
Xv to XXIII. 

Spermiducal pores are intersegmental between Somites 
xviir and xix, surrounded by an internal thickening of epi- 
dermis. 

Oviducal pores are on Somite xtv. 

Sperm-sacs extend from Somites x11 to xx or xxv. 

[Testes and ciliated funnels are in Somite x11. 

No spermathece are known. 

Gizzard is in Somite vi; calciferous glands in Somite x111. 

Nephridia commence in Somite iv; the pores are in front of 
the inner couple of sete. The nephridium consists of a short, 
slightly and loosely coiled tubule, opening into a strongly 
developed duct, which is produced into a blind sac: this 
czecum varies in its proportions in different parts of the body. 
The first nephridium is rather different from the following, 

VOL, XXXI, PART 1I—NEW SER. R 


250 W. B. BENHAM. 


as the coil of the tubule is larger and more compact ; it serves 
probably as an extra-buccal pepto-nephridium. Both Leuckart 
and Perrier were unable to see the nephridiopores in front of 
the fourteenth somite, but nephridia are present although the 
pores are difficult to see. | 


Species 1. G. maximus, Leuckart, 1841 (=T. brasiliensis, 
E. P., 1872), Brazil. 
2. G. forguesii, E. P., 1881; La Plata. 

See Leuckart, ‘ Zool. Bruchstiicke, Stuttgart, part 11, 1841 ; 
Perrier (Titanus), ‘Nouv. Arch.,’ &c., viii, 1872; and ibid., 
ix, 1881, foot-note, p. 235; Rosa, ‘ Boll. d. Mus. Zool. ed 
Anat. Comp. Univ. Torino,’ i11, 1888, 


Genus 22. Urocuara, E. P., 1872. 

Sete eight; anteriorly in couples, then they gradually 
become separate; and finally, alternate in consecutive somites. 

Clitellum on Somites x1v to xx11, complete ventrally; inter- 
segmental grooves not obliterated. 

Spermiducal pores between Somites xx and xxi (on 
Somite xx, E. P.; between x1x and xx, Rosa). 

Nephridiopores in line with the 3rd seta. 

Sperm-sacs, one pair, occupying Somites x111 to xv, or 
evel more. 

[Prostomium appears to be absent. 

Penial setz on Somites xix, xx, xxi, and xxiI. 

Testes and ciliated rosettes in Somite x11. 

Three pairs of spermathecz, in Somites vi1, vi11, 1x (Rosa), 
VI, viI, vii (Beddard, Horst), or vi11, 1x, x (Perrier). 

The nephridia, except the anterior pair, are simple, slightly 
coiled tubes, without any or only with very feebly developed 
duct. 

The anterior nephridia are massed together to form “ pepto- 
nephridia,” the tubules of which open at one end into the 
celom by ciliated funnels, and at the other into a large duct 
which communicates with the exterior in front of Somite 111. 

The gizzard occupies Somite vir; and there are three pairs 
of flask-shaped calciferous diverticula, in Somites vil, Ix, X. 


AN ATTEMPT TO CLASSIFY EARTHWORMS. 251 


“ Pyriform ” sacs occur in posterior part of the body, on the 
ventral surface. | 

(N.B.—The enumeration of the somites is given differently 
by the three authors who have described Urocheta. Rosa 
has pointed out that there is some reason to believe that 
Perrier counted a portion of the extended buccal region as the 
first somite ; with the result that his first setigerous somite, 
instead of being the second, as in all other worms, is the third ; 
hence it becomes necessary to subtract one, in some cases, from 
Perrier’s numbers. Beddard has elucidated the position of 
the gonads by means of longitudinal sections—the only reliable 
means of deciding their position; and on this point I have 
followed him. ‘The position of the external organs, and some 
of the internal structures, I have been able to decide for myself 
by an examination of some specimens kindly given to me by 
Mr. W. Sclater, who obtained them in Demerara.) 


Species 1. U. corethrura, Fr. Muller, 1857; Brazil, Java, 
Martinique, Fernando Noronha, and Aus- 
tralia. 

2. U. dubia, Horst ; Sumatra. 


See Perrier, ‘ Arch. de Zool. Exp. et Gen., iii, 1874; Beddard, 
‘Proc. Roy. Soc. Edin.” 1887; ‘ Quart. Journ. Micr. Sci.,’ 
xxix; Rosa, ‘Ann. d. Mus. Civ.,’ Genova, viii, 1889. 


Genus 23. Dracnz#ta, Benham, 1886. 

Setz 8, separate, alternate from somite to somite through- 
out the body, except Seta 1, which always retains a linear 
arrangement. 

Clitellum complete ; covers Somites xx to xXxxIII, inter- 
segmental grooves distinct all round. 

Spermiducal pores on Somite xx11; no penial sete. 

Sperm-sacs extend through Somites xit to xxxvit.! 

Nephridiopores in front of the outer sete. 

[No prostomium. 

Testes (?) and ciliated rosettes in Somite x1. 


* In apaper by Beddard (a proof of which Prof. Lankester has kindly 
allowed me to see) on a new species of this genus, two pairs of sperm-sacs 
are described. 


202 W. B. BENHAM. 


Spermathece.—Three pairs in Somites vi, vu, and vil, 
opening at the posterior edge of the somites. 

Gizzard in Somite v1; no accessory glands or czca. 

Septa in anterior somites strong as in the other two genera. 

Nephridia large, the duct simple without a cecum. Those 
of the first pair, which open externally in Somite 111, are much 
larger than the rest; the coil of tubules compact, and having 
a glandular appearance. It no doubt serves as a “ pepto- 
nephridium.”” I have not observed any funnel to this first 
nephridium. | 

Species 1. D. thomasii, W. B. B., 1886; St. Thomas, W. 

Indies. 
See Benham, ‘ Quart. Journ. Micr. Sci.,’ xxvii. 


Remarks on Geoscolecide. 


I have divided Rosa’s family of this name into two families, 
retaining his name to include three genera which agree closely 
with one another, especially in having a single pair of testes 
and sperm-sacs. But the structure of the nephridia do not here 
serve as a family character, since the cecum of Geoscolex 
is not present in other two genera. 

The position of the male pores is noticeably different from 
that in most other families, and resembles that in the Rhino- 
drilide. 

The fact that Perrier’s worm Titanus is identical with 
a worm described by Leuckart some thirty years before was 
apparently discovered by Rosa, who pointed out the curious 
agreement even in the words used by these two zoologists in 
their description of the worm. 


b. The eight sete are in couples and exhibit no alternation 


in their arrangement. There are two or more pairs of 
sperm-sacs. 


Family VIII. Rhinodrilidz, mihi (= partly L. intraclitel- 
liens, E. P., partly Eudrilide, Claus, Vejdovsky, Rosa). 


The eight setz are in four couples, the individual set of 
each couple being close together. 


AN ATTEMPT TO CLASSIFY EARTHWORMS. P5933 


The clitellum, incomplete ventrally, commences in front 
of Somite xviii, and occupies ten or more somites. 

The spermiducal pores are behind Somite xv11r (with 
the exception of Hormogaster), and are usually nearly in 
the middle of the clitellum. 

There are two or more pairs of sperm-sacs, and two pairs 
of testes and rosettes. 

The spermathece are either small, or if large are quite 
simple, without appendices. 

The gizzard is in front of Somite x. 

Nephridia are provided with a large duct, usually produced 
into a cecum; nephridiopores are in a line with the outer 
couple of setz (except in Hormogaster). 


Genus 24. Rarnopritus, E. P., 1872 (= Thamnodrilus, 
Beddard, 1887). 

Prostomium is two or three times longer than the first 
somite [and can be withdrawn into the buccal cavity; at any 
rate, it is so in spirit specimens]. 

The setw are ornamented near their distal ends with 
several rows of crescentic ridges, which are slightly more 
marked in the clitellar set. 

The clitellum, which does not extend across the ventral 
surface, occupies seven or more somites, xv to xxv (XX—XxXVI, 
Horst). Along its ventral boundary, on each side, is a glan- 
dular band—tu bercula pubertatis—on Somites xx—xxy. 

The spermiducal pore is intersegmental between xx and 
XxI (according to my own observation) (x1x/xx, E. P.). 

The nephridiopores are in a line with the outer setz, 

Sperm-sacs are two pairs, in Somites x1, x11, with median 
sacs. 

[Two pairs of testes and ciliated rosettes in Somites x1, x11. 

Spermathece are long and club-shaped, in Somites vu, 
vii1, 1x (Horst), or globular, in Somites v to viii, in a species 
examined by myself. 

The nephridium has only a short and slightly coiled 
tubule ; the duct is produced into a cecum. The anterior six 


254. Ww. B. BENHAM. 


or seven pairs of nephridia are larger than the following 
ones; and the duct is simple. The first pair, or extra-buccal 
pepto-nephridia, opening externally on Somite 11, is parti- 
cularly large, and lies below the cesophagus. 

Gizzard in vit or VIII. 

Typhlosole is a small fold, with a spiral line of origin: 

(Hsophageal glands, six or eight pairs, in the next following 
somites. 

In addition to the dorsal vessel, there is a supra-intestinal 
trunk below it, from which two or three pairs of large “ intes- 
tinal hearts ” go to the sub-intestinal vessel. | 


Species 1. R. paradoxus, HE. P., 1872; Venezuela. 
2. R. tenkatei, Horst, 1887; Surinam. 
3. R. gulielmi, F. E. B., 1887; Brit. Guiana. 
4. R. ecuadoriensis, W. B. B., 1889 (MS.); 
Ecuador. 
See Perrier, ‘ Nouv. Arch. du Mus. d’Hist. Nat. de Paris,’ viii, 
1872; Beddard (Thamnodrilus), ‘Proc. Zool. Soc.,’ 1887 ; 
Horst, ‘Notes from Leyden Museum,’ ix. 


Genus 25. Microcu#ta, F. E. B., 1885. 

Setz extremely small, in couples. 

Clitellum occupies ten to twelve somites between x and 
XXV. 

Spermiducal pore on Somite xix or xx. 

Oviducal pore between Somites x11 and x1tt. 

Nephridiopores very large, in a line with outer setz. 

Sperm-sacs in Somites x and x1, each pair being connected 
by a median sac in Somites 1x and x. 

[Prostomium small. No dorsal pores. 

Testes in Somites 1x, x, in special sacs communicating with 
sperm-sacs. 

Ovary in Somite x11.1 

1 The peristomium in M. beddardi is provided with sete, and is therefore 
homologous with the peristomium + the second somite of Lum- 


bricus, hence the position of the ovary and other structures is typical, 
although dissection shows them one somite anterior to their real somites, In 


AN ATTEMPT TO CLASSIFY EARTHWORMS. 255 


Spermathece very minute and numerous, being from two to 
four pairs in all or some of the Somites x1z to xv. 

Nephridia very large, and of a peculiar form, each consisting 
of agreatly coiled tubule, arranged as a tuft, communicating 
with a large duct, which is produced into a great sac-like out- 
growth. This cecum is less developed in the anterior somites ; 
and is most feebly marked in the first pair of nephridia. 

Gizzard in Somite v1. 

A single pair of cesophageal glands in Somite rx, or partly 
in vir and partly in rx. 

The dorsal vessel is doubled anteriorly and specially enlarged 
in Somite viit. 

Lateral hearts, vi—x1.| 


Species 1. M. rappi, F. E. B., 1885; Cape of Good 
Hope. 
2. M. beddardi, W. B. B., 1886; Natal. 
See Beddard, ‘ Trans. Zool. Soc.,’ xii, 1886 ; Benham, ‘ Quart. 
Journ. Micr. Sci.,’ xxvi and xxvii. 


Genus 26. Urosenvs, W. B. B., 1886. 


Setz in couples. 

Clitellum occupies Somites x1v to xxv, incomplete ven- 
trally. 

Spermiducal pore on Somite xx. 

Nephridiopores in a line with the outer sete. 

Sperm -sacs two pairs, of which one pair is in Somites x11 
and x11, the second pair in Somite xtv. 

Peculiar “ pyriform sacs”? occur in pairs on the ventral 
surface of the body-wall opening externally ventrad of the 
inner setz, commencing in Somite x. 

Testes and ciliated rosettes in Somites x11 and x1mt. 

[Spermathece three pairs, in Somites vii, vi11, and rx. 

Gizzard in Somite viii. 

Three pairs of flask-shaped calciferous glands in Somites 
ix; x, 7k 
M. rappi I find no trace of sete in the peristomium—the fusion is com- 
plete. 


256 W. B. BENHAM. 


Intestinal pouches in Somites xvi to xxv. 

A pair of tubular intestinal ceca in Somite xxv1 resembling 
those of Pericheta. 

All the nephridia have a large ceecum; the duct being very 
long in the first seven nephridia. ] 


Species 1. U. brasiliensis, W. B. B., 1886; Brazil. 
See Benham, ‘ Quart. Journ. Micr. Sci.,’ xxvii. 


Genus 27. HormocastErR, Rosa, 1887. 


Set in couples; those of the inner couple rather far apart. 

Clitellum occupies Somites xv to xxv. 

Tubercula pubertatis along the edge of the clitellum on 
Somites xviii to xxiv. 

Spermiducal pores between Somites xv and xvi. 

Nephridiopores in line with Seta 2. 

Sperm-sacs in Somites x1 and x11. 

Gizzards three, in Somites v1, VII, VIII. 

[Globose intestinal czeca in Somite xxi, and smaller ones 
in following few somites. 

Spermathecz in Somites x, XI, XII. 

Testes and rosettes in Somites x, xI. 

Nephridial ducts provided with slight cxcal prolongations. | 


Species 1. B. redii, Rosa, 1887 ; Italy. 


See Rosa, ‘Sulla Struttura dello Hormogaster Redii,’ Torino, 
1888. 


Genus 28. Bracnypritvus, Benham, 1888. 


Setz very small, in four couples. 

Clitellum occupies Somites xvi to xxi (though probably 
more). 

Spermiducal pores in a deep fossa occupying Somite 
XVIII. 

Two pairs of sperm-sacs, in Somites x, x1, enclosing the 
ciliated rosettes. 

Large and muscular thickening of body-wall, through which 
the sperm-ducts pass to the exterior, occupies Somites xv 
to XX. 


AN ATTEMPT TO CLASSIFY EARTHWORMS, 257 


Spermathece small; two or three pairs on hinder margin of 
Somite x1. 

Two pairs of nephridia in each somite, each a simple 
tubule without a distinct duct. 

[Sixteen globular “albumen-glands” are present, as four 
sacs on each side of Somites x, x1. 

Ovaries in Somite x11.! 

The worm is very short in proportion to its width. ] 

See Benham, ‘ Zool. Anzeiger,’ 1888, No. 271. 


Remarks on Rhinodrilidz. 


I include in this family the remaining genera grouped by 
Rosa in his Geoscolecidz ; the two families together nearly 
correspond with Perrier’s “ intraclitellian worms.” 

The most aberrant form is Hormogaster, with its male 
pores far forwards, and nephridiopores in line with the inner 
couple of sete. In these two points, showing a decided 
affinity to Lumbricus, and perhaps it belongs to the family 
Lumbricide. 

Brachydrilus is of interest in possessing two pairs of 
large uephridia in each somite; evidently an intermediate 
condition between a network in which the tubules have 
become grouped, as in Cryptodrilus, into three masses on 
each side, and the ordinary condition of a pair of nephridia. 
It is quite conceivable that, as in Megascolides, one tubule 
becomes gradually larger, whilst at the same time the rest 
become fewer, in some other form two such tubules might 
increase in size, and so result in two pairs of nephridia per 
somite. 

The testes and spermiducal pores have abnormal positions 
very usually in this family; for instance, in Microcheta testes 
and ovaries are placed one somite further forwards than is 
normally the case;? in Urobenus and Rhinodrilus the 


1 The fusion between peristomium and first setigerous somite appears 
complete. 
* See foot-note on p. 254, 


258 W. B. BENHAM. 


funnels are further back, as I have ascertained by longitudinal 
sections, in addition to dissection. 


2. Male pores in front of Somite xvii, anterior to the 
clitellum. 


Family IX. Lumbricide, Claus and Rosa (=L. anteclitel- 
liens, E. P. = Lumbricide +Criodrilide, Vejdovsky). 


The eight sete are either in couples, the individual setz 
being very close together ; or they may gradually separate so 
as to give eight equidistant setz per somite. 

The clitellum, incomplete ventrally, usually commences 
behind Somite xx (in one case on Somite xv), and occupies 
from six to nine, sometimes more, somites. 

The spermiducal pores are on Somite xv, or on an 
anterior somite. 

There are three or four pairs of sperm-sacs in Somites 1x 
to XII. 

Testes and ciliated rosettes in Somites x, x1. 

The oviducal pores are on Somite xrv. 

Spermathecez may be absent, or when present are nearly 
spherical sacs without diverticula. 

Nephridiopores in a line with the inner couple of set. 

Each nephridium is a greatly coiled tube, terminating in a 
large muscular duct without a cecum. 

The gizzard when present lies behind Somite x. 

There are no pepto-nephridia. 


Genus 29. Lumpricus, Eisen (= partly Lumbricus, 
Limneus, &c.). 


Prostomium dovetailed completely into the peristomium. 

Setze always in couples, the individual sete of which are 
close together. 

Clitellum occupies six or seven somites, commencing some- 
where between Somites xxvi and xxxiI. 

Spermiducal pores on Somite xv. 

Sperm-sacs three pairs, in Somites rx, x1, x11, connected 


AN ATTEMPT TO CLASSIFY EARTHWORMS. 259 


across the middle line in Somites x, x1, by sacs enclosing the 
testes and ciliated rosettes. 

Tubercula pubertatis, four on each side, forming a band 
along the ventral limit of the clitellum. These, and the sper- 
mathece are absent in L. eiseni. 

[Colour reddish brown, iridescent. 

Form cylindrical, more or less flattened posteriorly. 

First dorsal pore may begin between Somites vir and vitt, 
or posteriorly to this. 

Anus terminal. 

Spermathecz.—Two pairs in Somites rx and x, opening 
posteriorly nearly in a line with the lateral sete. 

Spermatophores, in the breeding season, fixed to the body 
behind the genital pores. 

Gizzard occupies Somites xvi1 and xvitt. 

(Hsophageal calciferous glands x1 and x11.] 


Species 1. L. agricola, Hoffm., 1845; Europe, N. America 

(= L. terrestris, L., partly). 

2. L. rubellus, Hoffm., 1845; Europe, Newfound- 
land. 

3. L. castaneus, Sav., 1829; Europe, Newfound- 
land. 

4. L. melibceus, Rosa, 1884; Europe. 

5. L. eiseni, Levinsen, 1883; Europe. 

6. L. caucasicus, Kulagin, 1888; South Russia. 


See Hoffmeister, ‘ Die bis jetzt bekannten Arten aus d. Fam. 
der Regenwirmer,’ 1845; Hisen, ‘ Ofvers. af Kong. Vetensk. 
Akad. Forhandlungen,’ 1870, 1873, &c.; Rosa, ‘ Il Lombricidi 
del Piemonte,’ Torino, 1884, and later papers in ‘ Boll. Mus. 
Zool.,’? Torino. 


Genus 30. ALLoLosporHora, Hisen (= Lumbricus, L., partly). 


Prostomium only partially dovetailed into the peristomium. 
Setz either in four couples, or individual setz more or less 
widely separated. 


260 W. B. BENHAM. 


Clitellum occupies five to nine somites (rarely more), 
commencing somewhere between Somites xxvi and xxxII. 

Spermiducal pores on Somite xv. 

Sperm-sacs.—Four pairs, in Somites rx, x, x1, and x11, 
unconnected from side to side, so that the testes and ciliated 
rosettes lie freely in Somites x and x1. 

Tubercula pubertatis are two or three pairs, sometimes 
in consecutive somites, sometimes on alternate somites (ten 
pairs in one species): rarely absent as in A. subrubicunda. 

[Colour more varied thanin Lumbricus; from deep sienna- 
brown to light transparent grey, sometimes green. First 
dorsal pore may begin as far forwards as Somite rv, or more 
posteriorly. 

Spermathecz usually two pairs (sometimes more, or they 
may be absent asin A. subrubicunda), opening either ante- 
riorly or posteriorly, either near the lateral setz or near the 
dorsal line. 

Spermatophores fixed behind the genital pores. 

(Esophageal pouches in Somite x, and calciferous glands 
in XI. 

Gizzard as in Lumbricus.] 


Species 1. A. chlorotica, Sav., 1832; Europe, N. America 
(= L. riparius, Hoffm., 1845). 
2. A. foeetida, Sav., 1829; Europe, N. America, 
Australia (= L. olidus, Hoffm., 1845). 


3. A. submontana, Vejd., 1875; Bohemia. 

4, A. fraissei, Orley, 1881; Balearic Isles. 

5. A. mediterranea, Orley, 1881; Balearic Isles. 

6. A. nordenskjoldii, Eisen; Scandinavia, 
Siberia, Azores, Newfoundland. 

7. A. subrubicunda, Eisen, 1873; South Siberia, 


Europe, Magellan. 
8. A. tumida, Eisen, 1874; Denmark, N. America. 
9. A. parva, Eisen, 1874; Denmark, N. America. 
10. A. arborea, Eisen, 1874; Denmark. 
11. A. dubiosa, Orley, 1881 ; Europe. 


AN ATTEMPT TO GLASSIFY EARTHWORMS. 261 


Species 12. 
13. 


14. 
15. 


16. 
Wf 
18. 
Lo: 
20. 
21. 
22. 
20. 
24. 
25. 
26. 
27. 
28. 


29. 
30. 
él. 
32. 
30. 
34: 


A. norvegica, Hisen, 1873; Norway. 

A. mucosa, Eisen, 1873; Europe, Siberia, N. 
America. 

A. trapezoides, Dug., 1828; Europe. 

A. turgida, Eisen, 1873; Europe, N. America, 
Australia. 

A. longa, Uhde, 1885; Germany. 

A. hispanica, Uhde, 1885; Spain. 

A. profuga, Rosa, 1884; Italy. 

A. transpadana, Rosa, 1884; Italy. 

A. minima, Rosa, 1884; Italy. 

A. constricta, Rosa, 1884; Italy. 

A. alpina, Rosa, 1884; Italy. 

A. veneta, Rosa, 1886; Italy, Portugal. 

A. ninnil, Rosa, 1886; Italy. 

A. tellinii, Rosa, 1888 ; Italy. 

A. molleri, Rosa, 1889; Portugal. 

A. orleyi, Horst, 1887; Hungary. 

A. (Dendrobena) rubida, Sav. 1832; Europe, 
Siberia, N. America (= L. octohedra, Sav., 
= A. boeckii, Hisen, 1870; = L. puter, 
Hoffm., 1845). 

A. bagdonowi, Kulagin, 1888; Russia. 

A. nassonowi, Kulagin, 1888; Russia. 

A. celtica, Rosa, 1886; Brittany. 

A.camplanata, Dug., 1828; Europe. 

A. icterica, Sav., 1832 ; Europe. 

A. gigas, Dug., 1828; Europe. 


Genus 31. Criopritus, Hoffmeister, 1845. 


Prostomium not dovetailed into the peristomium. 

Sete in couples, which are so placed as to give the body a 
quadrangular outline in section. 

Clitellum, ill-marked, extends from Somite x1v to about 


Somite x.y. 


Spermiducal pores on Somite xv, on a large rounded 
papilla almost lateral in position. 


262 W. B. BENHAM. 


No tubercula pubertatis. 

[The worm is aquatic in habit; in colour, brownish green. 

In the breeding season one or more “ spermatophores ” are 
found fixed to the body in the neighbourhood, and in front, of 
the genital pores. 

Cocoons spindle-shaped, dark green. 

The anus is dorsal. 

Genital apparatus as in Allolobophora; the male duct 
passes through a glandular thickening of epidermis situated 
around the aperture. 

No spermathecee. 

No gizzard and no cesophageal glands are present. 

The typhlosole, frequently denied, is present. 

The nephridia commence in Somite x (according to Collin, 

Zeit. Wiss. Zool.,’ xlvi, 1888). ] 


Species 1. C. lacuum, Hoffm., 1845 ; Europe. 


See Benham, ‘Quart. Journ. Micr.Sci.,’ xxvii; Orley, ‘Quart. 
Journ. Mier. Sci.,’ xxvii; Rosa, ‘Sul Criodrilus lacuum,’ Torino, 
1887. 


Genus 382. AttuRus, Hisen (= L. tetraedrus, Dugés). 


Prostomium partially dovetailed into the peristomium. 

Setz in four couples, latero-ventral and latero-dorsal in 
position. 

Clitellum occupies Somites xx11 to xxvit. 

Spermiducal pores on Somite x11, lateral in position. 

Sperm-sacs as in Allolobophora; sperm-duct opens 
through a glandular thickening of epidermis as in Criodrilus. 

[Body posteriorly quadrangular. 

Spermathecee minute sacs (visible only in sections) in 
Somite vill; aperture not intersegmental, but close to the 
lateral sete. 

Gizzard in Somite xvii. 

Small esophageal glands in Somites x—xr1v, not very distinct. 

The nephridia commence in Somite rv. 

First dorsal pore between Somites rv and y.] 


AN ATTEMPT TO CLASSIFY EARTHWORMS. 263 


Species 1. A. tetraedrus, Sav., 1832; Europe. 


See Beddard, ‘ Quart. Journ. Micr. Sci.,’ xxviii. 


Remarks on Lumbricide. 

Hisen was the first to subdivide the genus Lumbricus into 
two sub-genera, according to the relative amount of dovetailing 
of the prostomium into the peristomium. This is accompanied 
by certain other characters, which have been held sufficient to 
characterise genera in other cases. So that I retain his sub- 
divisions Lumbricus and Allolobophora; but as his genus 
Dendrobena is only distinguished from the latter genus in 
having all the setze equidistant, and as all stages occurring in 
this separation are found in Allolobophora, I agree with 
Rosa that we ought not to recognise it. 

The anatomy of Criodrilus, recently worked out by Rosa 
and myself, and again by Collin, is not very greatly different 
from that of Allolobophora. The most important points of 
difference are in the extent of the clitellum—which, till my dis- 
covery of it, had been denied, and in which Collin confirms 
me—and in the fact that this glandular modification of the epi- 
dermis commences in Somite xv ; together with the absence of 
spermathece. This last character—which at first sight seems 
to mark it off from the rest of the family—serves in reality as 
a further link; for spermathecze are absent in Lumbricus 
eiseni, Levinsen,! and in Allolobophora constricta, 
Rosa.” This negative character is, as Rosa has recently * 
pointed out, accompanied by another negative character, viz. 
the absence of tubercula pubertatis—structures almost 
limited to the family Lumbricide, as they have only been 
mentioned or figured in the species of Rhinodrilus, and in 
Hormogaster. 

The spermatophores, so noticeable a feature in nearly 
every adult specimen of Criodrilus, are also known in many 

1 Levinsen, ‘ Syst. geogr. oversigt over de nordiske annulata,’ &c., Copen- 
hagen, 1883. 


2 Rosa, ‘Il Lumbricidi del Piemonte,’ 1884. 
3 Rosa, ‘ Boll. Mus. Zool. ed Anat. Comp.,’ Torino, vol. iv, November, 1889, 


264 W. B. BENHAM. 


species of Lumbricus and Allolobophora; and unknown 
elsewhere. 

Criodrilus, in fact, must be regarded as a degenerate 
Allolobophora, owing to its altered mode of life; its 
aquatic habit has no doubt a connection with the absence 
of a gizzard, and very likely with the absence of nephridia in 
the anterior somites, which may probably be used in ordinary 
earthworms as salivary glands—that is, for the purpose of 
moistening the food. At any rate, we find the same absence 
of anterior nephridia in another aquatic form, Pontodrilus; 
aud the fact that in so many worms the anterior nephridia are 
specially large, or modified in some way (as in Urocheta, 
Diacheta, &c.), and even open into the pharynx instead of 
externally, bears me out in this idea. 

In this connection it is interesting, though contradictory, 
to find that Allurus, which is also an aquatic form, but lives 
in the soii below the water, whilst Criodrilus lives actually 
in the water, has nephrida in the anterior somites. 

Allurus has no true spermathece. Beddard describes a 
minute sac embedded in the body wall, and opening exter- 
nally on the somite, but no spermatozon were observed in it; 
and it may perhaps be either degenerate, or of the nature of 
an albumen (“ capsulogenous”’) gland. 

The species both of Lumbricus and Allolobophora are 
in a state of great confusion; even modern authors make two 
species out of one, or split up one into two. The list I have 
given is taken from Vejdovsky’s ‘System und Morphologie,’ 
with additional species described since the date of his mono- 
graph. 

Incerte sedis. 
He opritus, Hoffmeister, 1845. 

Setz black, in couples. 

Clitellum absent. 

Spermiducal pores on Somite xv. 

Gizzard present. 

Pigment spots are present on peristomial somite, but are 
absent in young individuals. 


AN ATTEMPT TO CLASSIFY EARTHWORMS. 265 


Ecuinopritus, Vaillant, 1869 (= L. multispinus, Grube, 
1851). 
Set four bundles of 5 in each somite. 
Clitellum absent. 
Spermiducal pores in Somite x11. 


Anrtevs, E. P., 1872. 

Setze four couples. 

Clitellum ill-defined, on Somites xv to xxix. 

Spermiducal pores unknown. 

Sperm-sacs, two pairs, in Somites 1x, x. 

Gizzard, Somite v1. 

Nephridia large; pores in line with outer sete. 

Anterior septa very thick. 

Vaillant points out that in many respects Perrier’s descrip- 
tion agrees with that given by Beddard and myself for Micro- 
cheta rappi. Size: Anteus, 1:16 m. (i.e. 45 inches); 
Microcheta is 3 feet 6 inches to 6 feet. 

The arrangement of setz and indistinctness of the clitellum 
are also points of resemblance. To show the difficulty of 
deciding where the clitellum commences in Microcheta, it 
is noteworthy that whereas Beddard puts the extent of this 
structure as Somites x to xxx inclusive, I reckoned it as 
occupying Somites x11 to xxv. 

Both Beddard and I were unable to recognise the spermi- 
ducal pore externally. 

The anuulation of the somites rendered it difficult to count 
them ; thus Beddard figures the gizzard in Somite vi1, whilst 
I found it to be in Somite vr. He states that the spermi- 
ducal pore is on Somite xvi11; I found it to be on xx. 

In both Anteus and Microcheta the anterior septa are 
especially thick and infundibuliform. Perrier places the last 
of these thick septa behind Somite 1x; Beddard places it in 
Microcheta behind the eighth, and I found it behind the 
seventh. These discrepancies are no doubt due to the diffi- 
culty of counting the somites. 

A nephridium of Anteus is figured by Perrier. He repre- 

VOL, XXXI, PART II],—NEW SER. s 


266 Ww. B. BENHAM. 


sents it as a long, narrow tube, equal in diameter throughout, 
and thrown into a number of curves. It ends in what he 
regards as the celomic funnel—“ une sorte de houpe formée 
par une série de replis membraneux implantés sur sa portion 
terminale libre.” This I take to be in reality a tuft of loops 
of the coiled tube, such as exists in the nephridium of Micro- 
cheta (see my paper, ‘ Quart. Journ. Micr. Sci.,’ xxvi, pl. 
xvi, figs. 21, 25, 26). It is possible that the wide muscular 
duct there figured might in an ill-preserved specimen shrink, 
and have the appearance of such a duct as Perrier figures. 
Perrier states that behind the twentieth somite the nephridia 
are smaller and somewhat different from those anteriorly ; 
such is also the case in Microcheta. 

The fact that the spermathece in Microcheta are very 
small, and quite differently situated from what is the rule in 
other earthworms, might be suggested in explanation of their 
having been overlooked by Perrier. 

In Microcheta the dorsal vessel becomes doubled in 
each of the Somites tv, v, v1, vi1, and viii, and in the last is 
very much thickened. In Anteus Perrier figures and describes 
it as ampullate and bent aside in Somites x11—xvi1, and does 
not note any doubling. 

It would be exceedingly interesting to investigate more fully 
the anatomy of Anteus, for its locality, Cayenne, in Brazil, is 
so far removed from the home of Microcheta in South Africa 
that it seems scarcely credible that the two are identical. 


Kisenta, Vaillant, 1889 ( = Tetragonurus, Eisen, 1874). 
Prostomium does not dovetail into peristomium. 
Setz in couples. 
Male pores in Somite x11. 
No further details are given. 


Species 1. E. pupa, Hisen, 1874; Canada, N. America. 


See Hisen, ‘ Ofvers af. Kongl. Vetensk. Akad. Férhandl.,’ 
1874. 


AN ATTEMPT TO CLASSIFY EARTHWORMS. 267 


IV. Tasutar SumMMARY oF GENERIC CHARACTERS. 


I have here brought together the main characters of the 
various genera in a tabular form, the genera being arranged 
alphabetically. The information is, of course, condensed, and 
the terms employed are defined in the chapter dealing with 
nomenclature. 


V. InpEx To IDENTIFICATION OF GENERA. 


In addition to the following ‘‘ tabular summary” it has 
occurred to me that it would be useful to zoologists examining 
earthworms to have the genera arranged in such a manner 
that identification to some extent may be rendered less diffi- 
cult, as it is by no means an easy matter to distinguish many 
of the genera from one another, and I have found a table of 
this sort a great help to myself. 

In order to add to its usefulness I have appended to each 
genus the page in this memoir in which will be found further 
details and references to papers on the genus. 


OO eee 


oJLUIOS YqIANOF umrut RSOYy . 
Ur ulseq £ jos paryyy| opdurrs SOrrydoursoyy|-oystzod ogur poyieyoaoq LT aqodutoo (ZT) 9T-€T sodnod FAST ‘xXepoosorory_ “gT| 
GZ-OT W9EMYOq ALOT. AMONIOS ‘dd ‘a 
ajdnoo 104nQ} [wowo { ornydourSeyy Tews} 0% 10 6T ‘oyodumoour ‘sozimtos ZT} soydnod ur ‘qyems 1a, eggyT = “ey@@yoosor, “ZT 
Ajxo110ysod soqn4 
ASIVT YIM TOM {OU poxyreur Logo 
snoxourn yy AT LOLIegUB YyIOMION poreyoaop you “prorg) axypided uo “gt |A[qooy ‘ozojdutoo =[Z-gT| peaquaa ][w ‘satdnoo uy gygyT ‘sepryooseSoyy “gt 
uinruoysr1ed Ze-9z suo | | uuy 
a{dnoo asuuy) ajduns Orydoursopq|oqur £Joyo]Aut09 poyrezoaog, ST -uauIMMOD ‘saqtuLOs J 10 9 soydnoo urskemy yey, = ‘snolaquimy “eT 
garde rez ajdnoo Bsory 
Z BJOG) [wowo £ ornypdoursoayy poTreqoaop Jon 91/ST aqo[dutoour ez—eTcauur Jo asoyz ‘setdnog sggyT “layseSomsoy “PT 
p UL Surousurut09 AT[Ray Ayto1se4ysod | | gaeyonery 
a[dnoo aouuy)fyeowo formgdoursoyy prorg 61/81 wad ojojdmtoour ‘gz—Gqjayuaedas ynq ‘sajdnoo # TFT ‘xo[00s00x) “ET 
oid 
-noo xoyno [x0 xouuy |) opdutts “orydouvsoyy porreqoaop ATeyaeg Dt agatdmoo g{-FT [er] seqdnoo F.zyeT ‘dA ‘SQ[Upugq “ZT 
— stay dou0qoo] gq popreqoaop Ayeryqaeg) [*q "a ZT] ST jAtpeaquoea oyodutoo *4T—-pT sajdnoo Fizy QTd ‘Y “lO4seSiqT “TT 
pedoyaaep | ‘lA 
— ory douo0qoo, gq pereqoaop Ayeryaeg) appided uo “gt |Ayyvoas you ‘“gT-pt [eT] que4ystpinba Ajaven “g gggq “toysesouApig: “OT 
Aype.arquoa [ete ‘a ‘a a 
snorowM Ny ota dou0qo01q — LT podopoaop [eM Jou ‘OZ-ET|-e] pure [eaqua ‘satdnoo Fegey, “AaqyseSoyo1q “6G 
AT[eyuouSasaxoqur podojaa aqyru ee ep Top oyyy 
ajdnoo 104nG odunts ‘osarery aUON rad -ap jou ‘ajotduioo gge—oz|-499[8 “Gueystp ‘ayeredas ‘glgggT ‘jeyouiq “gs 
wHaATUL Ayyerq yueqsip a ‘aa 
_ day douoqoopq|-oystaod oguUI popieqoaoq ST -uaa ‘aqyayduios gf ‘eT ‘pT|-mba Apxeau ‘ayimos tad ZT\ggeyT ‘sn Ltpoutag “4 
‘qo q *soqng Savy tmnimoyst1od o7ut Blin) i | 
_ ‘sured ¢ 10 T 10 ‘sqgny|paprezoaop Aqenaed Tpeurg] aypided ou “gt |ATywaquaa ‘oqyo]dutoo » T-eT ayeandas ‘gigegT ‘snpapoyddug “9 
OT ut Surouout munturoyst.ad [e10}v]-os10p “ULOH 
aydnoo acu) -moo ‘elduns ‘asaerq|ozur paprezaaop jou ‘asuery cI GF-CT qhoge ‘poyxcvut-[]] puvyesoze[-orquoa SopdnopieFgy ‘sn IAMpoLig *e 
2] ‘da “GM 
“nod 1o9NO puw souUy) aydutts ‘saved omy, — ST gid doop v uy [erout ¢] [Z-9T sotdnoo F ut bicaciee| ‘snjupAyorig ‘p 
aydutts urmruroystaad [B194R]-Os.0p 
g[dnoo zouuy ‘F ur Surouomuopjoyur poprezoaop Ayeyaeg el LE-GF PUY [V19}zR]-OAZUOA ‘SoTdnog FET wast ‘snanITy “e 
uMr1m109st1410d | ‘a ‘a 
| sno.oun yy day dauoqoopgqjozur popreyaaop Aypeyavg ST A[xo1taqueoyo[Aut09 “GT-ET| opeaedas g 10 ‘sefdnoo F ZzgT ‘suplAupoyyUBoY “Zz 
| 
| uintutoystxzad 07UT ZE pure 9g wooayoq “Suurs aqvandas sso] 10 a0 uasiy 
gjdnoo aouuy} ayduus raydoursoyy| AToyoTdmoour payreyoAoq cL -oq ‘sopiutos G ‘g ‘Y ‘9 ‘Giio ‘saydnoo ut sammyautog FZgT ‘vtoydoqojo[[V “T 
ecctomatnan “erpayda Ny “WNTUOFSOLT ‘al0d 9B “mUNTPHTO “RPS ajeq | “AON YW WTA oR 


‘AUVWWAG UVINAVT— AT 


UIOM Juoredsuery [[VUIS 


fJassoa [RAnouqns ou “ajosopydky ou £G ur wooygeuracods ared oug OuONT eajos [etued £ poqory 9013 ‘TL ‘OT ZL ‘IL ‘OT 
‘snonordsuoour pue snorounu eooyyvutsedg sous 
‘snonordsuos Aroa vipraydou £G ur vpnotysoArp [eoSeydosw Jo eg 9 aUON |[eroads ur saqsoq “OT “6 IT ‘OL “LT 
‘aposorydsy on *erpraydou-oqdod [woonq-eaquy *gT-ZT seqtut0g eyos peruod PL 
jo Lue xo TV UL SUOTZR[Tp [VUTYSOJUT Jnq “eMoTZAL0ATp [va.cvTy dose ON 9 10 G ou £ poltod yonur ‘avpaqny, ee1J ‘TL ‘OT ‘EL ‘SL ‘IL ‘esomoovyy] “9T 
| “styeqaroqnd aes ovs URIpout 
Bnoteqny fTT ur yonod wv pur ZL ‘TT vMoyaoarp snosozwoye9 ST ‘LT OUON|UL pasojous “TT “OT/qIIA\ “ZT ‘TT G ‘saved ¢| ‘eT 
“PE-SL SozLULOg® 
uo siqyeqwoqnd vosoqny fTz ‘aurysozUL Jo eIMOYAATP asOqoTH 8 ‘2 *9 “oor, OuoN TL ‘OT ZI ‘TL ‘sared Z "PL 
“UMOUS 
eooyqeutiods ou foposopydhy fe, ‘Savy vowyaoarp snosoypoleg L oulo NT SL GZ 0} ZT ‘aed [ ‘eT 
‘xerpnood st snqevaedde peqyvues oyg £ eposopydAy ou oes UT aqyeaedos 
‘TT ‘OT Uf [waquoa pur “ZT ur snosozioyeo pur [e«oqLT VMOTAOAT 9 stuad snourqnyo fpoyesuopy|sjonp-wdods “TT ‘OT ZL pure ‘TL OL “OL 
— L“g ‘omy, eyos [vrued £ poyemqory dary SPOUUIN acelin On ctn 
[BUI UTGOSA0ZUT Jou sarngzode prooyyeutsodg 
“Pat {nop [assoa [estop £ poqstag ATpeads “gT-OT ur poyeyrp ourjsoquy L ‘9 ‘OMY, poqorztq Atpenbo “gT IL ‘OL SI 6 ‘OT 
‘wByos yetued ONY ‘aut, WeTpout avou ‘g Ur SUOTJOIOUUOD [VsIOp pue 
eoaygeuttads Jo ared [ Ayuo £ pT ‘OT “eT UL BMOII}OAIP SNOLOFOTVD| OT “G puv ‘8 ‘4 ‘OMT, |xefnqng “GT ‘ST ‘ZT ‘sured IL ‘OL ovs ueIpom “ZT ‘TL ‘OL! 6 
“RMOI4AOATP OU f vipraydou-oydad proonq-vaqxGy 9 eyes [eruod ou £ auON Ta SE 07 GT Sared T Te) 
“Spurs snoaozioTVo Ou fsoorpuedde yAra wooyyeuaodg ‘aqny Oru 
-0[9 [e1oads UT pasopoud puv ynoySno.yy o]qQnop Jassoa-poo]q [es10q 1 ‘9 poqynpoauoo “ET ‘ZT ‘sured z eon GL ob oy 
‘aided Axoqyetndoo 
[BtoAos ATTRusn £ eT-G SOPLMLOG UL SUOTZRALITP SHOLOFIO[LO [RUI4SOyUT [4 109|¢ ayos [erued ou £ ayeqory 901F ‘TL ‘OL ZL pur g ‘9 
ovs URIput 
‘oryenbe £ yesaop snue ‘oaenbs Kpoq “eoayqyeutsods oN OUON OuON IL ‘OT ‘0047 lou 21 ‘IL ‘OL 6 caited Pl °S 
‘OZ-ST Soqimtog ur [ea-Apoq Fo Suruayoryy £ TT “OT ‘spurys 
UOMUNGT VY “TT oqrurog ur sows [yems jo sared g 10 [ ‘wooyyeutsedg 9 OU N Tee CONE rye Woye a 
oes URTpout 
"PI-OL ‘suseydose zo uorzrg{e[Ip 4YSI[g}  oyLUIOS UO ‘ZT OuON IL ‘OL ‘00a, ou ZT ‘TL ‘OL 6 ‘sared F] *g 
‘sosupuodde 41M aooyyeutsodg eyos peruod 
‘quasard Apquonbay vrpraydou-oydod £ pavzzr3 puryeq vnora0arq 6 ‘8 £6L ‘ZT poqnjoauoo ‘saved g olf “TL ‘OL SI ‘IL 2 
*OTQBIABA oul] O[pprat 
Koozeuitods JO LquNNT “TT purx OT sopuog ut B[NO4AOAIp dOIF SO}JOSOL POYRIT [at] SSo1OR pojoouTOOUN 
[voseydosa £ SOPIULOS IVI[AYT[O ayy FO ots UO siyeqaoqnd vpnoaoqny, SI ‘ZI oUON|-10 pur soqsoz ‘TT “OTI*ZT ‘TL ‘OL ‘6 ‘sured FT 
*sycRUoy “panzzrg y “aqynysorg “s[ouUNy pue saqsoy, “sons-muiadg sen 


= SS 


2408 PANY, 


e[dnoo 104ng 


adnoo 1049ng 


BOS PAL, 


ajdnoo 104n¢Q 


Bas Puodag 
aos 


p pues zg ‘oyevurogyy 


BJIS PUOIIS 


APIs POE UO TENSaary 


SNOLOTUN NY 


aydnod 109nEQ 


eyduus ‘ormdoursa yy) 


yeowo £ orydoun soy 
otaydouoqoo[q 


oraydauoqoo[ 


afduus “ormydour.soyy| 


ory douvsoy, 
suo Aro 
saved J 10 9 TOTIOVUe 
fyeowo forrydoursoayy 
GT Ur ooueUt 


-u0o oyduus “ue.soyy) 


otaydoursayy *oy1u10s 
euo ulygia AyTaarquay 
FT oftutog ul odueuwt 

-utoo ‘opduis “ue.seyy 


oraydou0}02, gq 


oiduus ‘ormdourse yy 


dtaypdoau0qoe, 


yeowo £ orydoursoyy 


—YYIA oul, UT 
‘atodorpray da yy 


erp da NT 


auoNy 


qoUrysId 
polreqyoaop you “prorg 


tanrmoyst10d 
OFUL po[IVzoAop Jou ‘prorg 


twunru0ystz0d 
OUT pepteqoaop AT[eyABg 


poyesuolq 
want 
-oysttod = oJUT poprejaaog 


poTreqoaop JON 


UINTULOFS 
-uad 07Ur pop!ejaAop “OPT AA 


Aqeyaed poprezoaog 


umrutoystred 
OFUL POTIVJOAOp saTMIZEULOG 


Aypeaquoa 
ayo dunoo ‘AT[RyMOUISes.10} 


A{1011048s0d 
oyeuiej[e yurystpmba g 


Cad 02] 1z/oz|-ar podojoaop you “eg—pT aq} sepdnoa ur ApAo1toyUY Z/8T 


02 ojo[dun0our ¢Z-FT 


AT 
essoy doap Ur ‘gT 


LI-¥T 
oyotduoom! OF—FT 


6L “UeIpoy eqoidu0o £T-FT, 
LT 


(1z/0z) 02/6 


LI-¥1 
[42-02] =¢2-st ‘ood 


{[BJUOUISasiozUT|-WLOOUT SozTTMOS B1OUL 10 J 
ST eyotdu0o 2T-ST 
SI Ayperqzuoa ojopduroo ‘2 T—-FT 
ST aqoydutos J T-ST 
LT ‘9T 
‘aed xopury oy3 ur 4deo 
Celst —_|-xo ‘oyozdutoo “gt-#T [eT] 


10939504 aso7o “QT | ATpeaquoa ozopduoo ‘2 T-FT 
aqoid 


sopdnoo uy 


[erzU04 Te ‘sadnoo F 


| 
| 
| 


988T 
€88T 


[erzUaA [ev ‘sezdnoo F.9SST 


qande avy 
are qos YSnor}y ‘sojdnoo F 


ozerndas ‘g 


poyuemevato ‘sa_dnoo F 
ayeaedas *g 
queysiprnba “g 

euly O[Pprar 


wou [ ‘ON ‘oyeaudeas ‘g 


dOVFAINS [BIY 


S88T 


688T 


GL8T 


T881 
SL8T 
L881 


-U9A TO 10790504 980] [TV LEST 


snonutzuod “a1oul 10 Og 


SNONUIJUOOSIp 10 snONnuUTy 


[BAqUA-O1yRL‘QT|-wL0d ‘Loto 10] OT “ST “FT 
Z1/IT 40 TT/OT 


*WUNIMO}SOL 


“(BUI ULSasI0yUT peyrvut-[[t “ST-OT 


"a10d OVI “uTI}NTO 


‘(panuijuor) AUVWWAG AVTINAVT— AT 


-uod ‘saspit U0 QOT-0Z 


| 


éL81 


T98T 


ee 


“d “a 
“eqyeeqoory) 
‘TAM 
‘snueqo.g 
‘Ha 
‘sneayd sy, 
‘TAM 
‘JO|SBSIL], 


BSO 
‘sn[lipneyay, 

‘ad “a “a 
“sn[ixpopoyy 


‘d ‘a 
‘snyLIpouryy 

‘d ‘a 
‘sn[lIpojuog 


‘AH ‘SUypew 
paviy 
‘snpiqpozoy 


er 
£10]SVSOSSLIOg 


‘ag ‘xAuonog 
epareuypog 


eyeyoliag 
‘d ‘a 


sa[dnoo F ZJST| “toysestMOyy 


“BPS 


*97e7 


“6T 


“JOYJDY TILA oMeN 


erpraydou 


-ojded [eoonq-v1yxq = ‘aovzans [vajza uo Apoq Fo 4aed aor10480d TOyy.AMF 
ur sows waopAg “OT “6 ‘Q “eMo4AAtp snoagyto[vo Jo sared sory, ZL eyos perued £ auoNy ZI 10 “GT 09 ef ‘aed TI [ze 
"#408 UU Jo peaquoa ‘AT[VuIo4xo Suruedo sovs uto0zakd £ gz 
“goed [BUTJSOJUL $ TT “OT ‘G “vfnoraoAtp snoazozropvo Jo sared veayy, 8 OuUON, 901} ST ZT PL ‘ST ‘et ‘saved z | ‘Te 
*A[xo110juB Sutuodo *g ur woayyeutseds Jo oes-untods OJIULOS T UL} ot0Ur 
ted ougQ *suoryezeyIp [VUTysOJUT puB “ZT UI BlNOT4AOAIp snosozIO[eD L‘9 eyos [eruod £poynjoauop/uerpemt ur OT ‘aed [| Surkdnov0 10 ‘OT “aed T| ‘og 
[omg ejos [eruod ou £ poqnyt 
‘gjduns wooyyeuaodg -quosoad ore vrpraydou-oydag|to] 6 ‘g ‘2 ‘aarqg|-oatoo favnqny ‘EL ‘ZT — = "62 
“ST/PT OBS JOYJOUB SAATED 
‘saod ywonptao pored pue fpT/eT “aod yeooyyeutsods uerpom -01 OS[B PUB TOTIO}XA 04 ZT IL 
forstoqgovavyo pur ‘aernoed snqevavdde oayeiouox ‘sorts sossed yor ‘ows ueIpourt |spauung ‘ows peroods 
ALOM0ZUB UL Po[QNOp SI JASSOA [VSLOG, “ET Ul V[NOYAIAIP snoAOFIOTLA L‘9 eB doquo soqvqsord oOM/e UL posopous TT “OT ZL ‘IT ‘sated Z "8Z 
eos teraed £ aangaode 
‘soorpuodde oavy wooyyeuttodg epnorqazoatp Arvquoutrye ON G quopuedepur ‘yp axepnquy, — rare eye “1Z 
“GS-06 
‘sxyeqzoqnd epmoroqny, "“PT-G Ul vNoYAeATp Fo sated o10ul Io xIg 8 10 2 ouoN ZI ‘IL ZL ‘TT ‘sated g mo 
TL ‘OT 80}J0sS0n poqe 
*[OSSOA [BINoUQnS tou ‘oposopydhy ONY ‘arOYS-vos UO SOATT ouON IvTNGNy “poynfoauoy -1]10 pue “aay so4s0J, SI ‘IT "GZ 
*6-G Ul MoYQVUIIOds Fo sared 
aatg ‘[é] OT ar Arvao ogy, “ZT ‘TL ‘OT ‘emoyaoarp [voseydosny 9g Penpoauoo “remMqny, é a2Z1 "hZ 
‘JOSsaA TeMouqns ou £ oosopydsy ON *aoRVT TL ‘OT soqqosox 
-ans [estop wou usdo 07 «vodde soqruos cor107UB UT spuRLS [vqydog OUONT OS [VIUET|poyerpIo puw soqsay, ZI ‘IT 4 
jos 
[erued £ GT ojur spuoezxo 
"PI-G UI BMOTY|AOATp (2) snoaazoleo peuysoqyUy L °9 6g ‘ory, pure suo, aqoy, ouo “paqory eel} ‘TT ‘OT ai TiteOk G6 Ard 
*xipuedde qnoyqim wooyyeutateds 
feypided Axozetndoo ou ‘spur[s oarsoSrp Arossooov ou ‘vowo ON 19 poqoy APysrs ‘pous74eyy TROL ZI ‘IL om 
*xipuodde yt g pure 
wooyqvuttods £ Ajrorzoque spurs rossooov fpoyvodoar ATTeoItaUt |UOIMJoq syUOUTSes Z 
“BOUL SOMITJOMOS “OZ opLUOg UL vow [BUTysoqUT Jo aed v ATTensp|Aue.10‘OT “G6 ‘SAT[ENSQ| postout A[doop 10 ‘paqory II ‘OL rat SLUT ‘OZ 
6 10 g UT MooYZeUNALEds SOqTUOS [eurs OT Apgaed “6 
jo ured o[suis & fgT-pT ‘seqrutos saa} 40 omy sordnod0 ovstagQ|aomoy t0'Zg—eT ‘Suory|Ar0a to ‘Suoy pue avjnqny, 6 $0489, Ajgaed Surkdnoo0 ‘aed [) "ET 
*SyIBULOY *paezzry *aqeqsolg *S[OUUNT pue sojsaz, *sovs-uledg ‘ON 


‘Gan a snmeqoniy 9 haere AC ATO YT ye ene nee ee 
‘epg ‘d ‘snpiapnoeg, * * * *MAX—AaTx umpey fuerpeyy cial elas 
‘poz d eyeyoororq, * * * *° * AXX—X wn]jez]o £ porreg Me 
‘egg ‘d “(siynqzoqnd emosoqny £AXX—AX UINT[e}][9) SN[TApoury y * poyuoummuto egos: £ £ Moareu “‘poyusuorg * + urntu04s 
"epa ‘d “(11TXX—AxX UN][P}PIP) XOTOOSOOH * * + + + © grduis eyos fpeoig f-o1g  ‘xXIXx/IIIAX 
‘ogg ‘d ‘snitapAqowrg * ° ° ° “IXX—TAK unyjeyyp ° ° ° ° ° enON 
‘egg ‘d ‘sopijoosusoy ° °* * ° *(exoulao) IxxX—IIIX ) 
‘ozz ‘d ‘togsesossttag * * ° : SReIange ga eS oa eer - + + umt04s 
"922 “d ‘10 peu SouDiA aces fumruoystaod oyur AYSis popreyoaoq Rood MGA ATE ne ins as Seo Cee 
* UIMIUIOJsOIg f-Og UO UNTIED 
‘ITAX—AIX (I1IX) | ‘I1IAx ur “aed ou f* soqeysorg “1ITAX 
‘pred ‘aniizpoyogqd * *- * * © © * © © * PerreyeAop FON |. + 6+ mruroge * saqtutog *} 
‘peg d ‘snjripoydkrg * ‘oyeqog], * * * ° * * * O4Rzs0Ig fF); 7 uo sarod 
‘epg ‘d ‘SN[IIpozuog poynpoayon f ‘wmrmoysttod ojur paprezoaog d J] avy “out 
‘6c2 ‘d ‘sn[ltpoyqguvoy ° ° * Bssoy B Ur le avy “XIX c -JVU194[B 


‘epg “d ‘sutpoynt qd oyeandos aos f umruroysttod oqur poyreyoaod {: 


*ITAX—IIIX 


pues ITAX soqrut you ‘sMOL UT 
‘Lez di reg eus ray, * + + essoz doop uy J -og ut ‘sated omy, J 
"yee d fL04svSoyorq Xx ‘W1AxX ‘sared omy fsozvqsoad avynqny, * * * XX—IIIx 
‘ppg A ‘xopoosocorqy * °° * * ° aed ouo {saqeqsord poqory * * * ITAX—IITX 
‘TVS ‘d ‘s PL RE polteqoAop mantut0ysord fsajdnoo uraqyeg * * * IIIAK—AIX - 
omams se *: Paieieacp jou umnrmoysoad £sazdnoo uy DOSED. “es 


“Lez. ‘df sovs-u10ds 
‘rogsvsiq | yosared omy, 
"O2Z ‘df sows-utrods 


preps: Smee Nis tnputo4stiad * weyog ‘IIAX—AIX J tS 
‘snaydhy | youvd ougs “gpg ‘d SUIPOPONA{ Sou perreyaaop umrimmoqjsoad £ aqgvandag } 


‘gez ‘d ‘aogseSomioq °° ° * IAX/Ax 
‘Tog ‘d ‘sn[iaportg * * * * umrmoystsed ojur porezoAop JON 

"ecg ‘d ‘vaoydoqojo[[yV * * * * * * * Ajajeydtmoour perreyoAog p* UINTIUOJsOIg "AX 
‘geg ‘d ‘snotaqmny * umnrmojstrod oqur AJojo;dut09 pojreyoaoq, 


"8 *  saqrutog 


woe -d aNanTTy. |) Sf et 8 ce Se tere | 9qtu0s 
‘seg ‘d “ToqseattiuoyT ° °* * ° &Irx/rx 10 1x/x J rod 
‘oeg *d <(IIXX—AIX UNTEPTO fIXx/XxX os0d oyemt) Byw@yoorQ * * * * Ajxorsoysod AjuQ]* *seqrHI0s BOS 
dATINIsUO0D 


‘Tog ‘d “(iTxx oz0d oem) veyayourq * * Apog oy} qnoySnoaqy, ) wtoyeureypy J 
emer tcl (TAIN AcKiac Axe TIKOTTONLTO)) AN TARO Oi ee) 91 hee 
“ype °d “(vrprrydou oSavy) xXuotrag * soq}090} asojo A104] * sasod ayeyy 
“SULL B UL 
‘geez -d “(erpraydou jo synj [fems) vya@yottog * * * * gavdv avg J otou 109g0}9T) 


“VUANGH) FO NOILVOMILNAGT OL XPANT—' A 


AN ATTEMPT TO OLASSIFY EARTHWORMS. Par fe 


VI. Puytoceny. 


I will now endeavour to trace the phylogeny of the group of 
earthworms, but owing to the scanty information as to their 
ontogeny, it is impossible to found anything like a true phylo- 
genetic tree. 

First of all it will be desirable to say a few words as to 
what may be considered ‘ primitive characters,” as two widely 
different families have been regarded as the more primitive, 
viz. Pericheta by Beddard, and Acanthodrilus by Rosa. 
I hope to be able to bring forward sufficient reason for denying 
to either of them an archaic condition. 

The excretory system, the sete, clitellum, prostate, 
and sperm-ducts may be taken as the more important 
characters. 

The Excretory System.—The recent researches of Bed- 
dard and Spencer have resulted in the conclusion that the net- 
work of tubules is a more primitive state than the large 
nephridia ; that, in fact, the latter have been derived in some 
way from the former. 

In Megascolides the excretory system in the anterior 
region of the body consists in a network of delicate tubules, 
with numerous external apertures, but without ccelomic funnels. 
Further back, one of these tubules on each side increases in 
size, and the network diminishes in extent; whilst in the 
somites quite posteriorly there is on each side a large tubule, 
which possesses a coelomic funnel, and which still retains its 
connection with the network. Spencer regards the anterior 
plectonephric condition as more primitive, and differs from 
Beddard in considering the nephridial funnels as new struc- 
tures, and not as derivatives of the flame-cells of Platyhelmia. 
It is to be noticed that the modification begins in the posterior 
somites, whilst the anterior part of the body still retains a 
primitive condition. 

Other instances of the co-existence of large nephridia with 
the network of tubules have already been given, 


274 W. B. BENHAM. 


In Pericheta we have certainly a primitive condition, but 
more modified than in Megascolides, in that, at any rate 
in some species, the plectonephric tubules are provided with 
funnels, and in others co-exist with large nephridia.! 

2. The Setw.—Beddard considers the perichztous condition 
as antecedent to the octochetous. Now, I believe we have 
ample evidence that the reverse is the case. Firstly, it is a 
nearly universal character of the Chetopoda that the setz are 
in two bundles on each side of each somite; in the Polycheta 
there are many setz in each group, in the Oligocheta only a 
few, and in a very large number of cases only two. In the 
Archi-annelida sete may be absent or only in one bundle on 
each side in each segment, but it is not unlikely that this 
group contains degenerate and not primitive forms. 

In Pericheta itself it is very usual to find fewer sete on 
the anterior somites than posteriorly. Unfortunately, as far 
as I am aware, we are not in possession of actual details as to 
the mode of development of the setz in this genus. But if 
the modification of nephridia in Megascolides commences 
posteriorly and works forward, may we not assume that the 
same has happened in the case of the sete of Pericheta or 
Perionyx? If this were so, we should expect to find just 
what is actually existent, fewer setz anteriorly, i.e. less modi- 
fication than posteriorly where greater modification has taken 
place. 

In some of the species (P. attenuata and P. enormis) 
described by Fletcher (‘ Proc. Linn. Soc. N.S.W..,’ vol. v, 1888) 
there are only eight sete in four couples in the first few 
somites ; then twelve in some of the following somites; and 
posteriorly they become more numerous. In P. dorsalis, 
only 16 at first, more posteriorly 30. In P. monticolla, only 
16 per somite on first few rings, increasing to 27 about 
clitellar region, and behind to 50. 

Again,in Urocheta and inGeoscolex the setware arranged 


1 For a discussion of the subject see Baldwin Spencer’s monograph on 
Megascolides, and Beddard’s papers in ‘ Quart. Journ. Micr. Sci.,’ xxviii 
and xxix. 


AN ATTEMPT TO CLASSIFY HARTHWORMS. 275 


normally—i.e. in couples—anteriorly, but become separated 
posteriorly, or even, in Urocheta, alternate from somite 
to somite. That is, according to my view, modification has 
commenced posteriorly, but has not affected the whole of the 
body; whilst in Diacheta this change has extended all along 
the worm.! 

The perichzetous condition, according to my view, has arisen 
firstly by the separation of the individual sete, originally 
in couples, so as to produce eight equidistant sete (as in 
species of Acanthodrilus, in Plutellus, and Allolobo- 
phora boeckii) ; and then intermediate sete have appeared 
gradually fillmg up the spaces, leading on through Deino- 
drilus with twelve, to Perichzta with 20—100 per somite. 
I conceive this intercalation of setz to be effected by the 
gradual increase in length of the accessory sete (‘soies de 
remplacement” of Perrier), which are very usually found, one 
to each of the functional setze in many, perhaps in all earth- 
worms. Supposing all the accessory sete of a somite became 
thus fully developed contemporaneously with the existent 
sete, we should get a doubling of the sete, i.e. sixteen per 
somite. Hach of these would, later on, have an accessory 
seta, and these might develop into functional setz, and so 
on, till we get the perichetcus condition. 

Mr. Beddard would regard the penial setz in special sacs, 
found in many earthworms, as vestigial representatives of a 
perichetous condition. I would regard them, however, as 
secondary and as developed from ordinary accessory sete, 
which if carried to a greater extent would lead to a peri- 
chetous condition. If we look upon the perichztous condi- 
tion, then, in this light, the removal of Perionyx from its 
associations with Perichzeta merely indicates that the con- 
dition has been developed twice, and independently ; and if we 


1 In a new species, D. windlei, Beddard states that there are no sete 
on the first five somites. Here the modification has gone further, and the 
sete have disappeared altogether. Microcheta presents a somewhat 
similar case of disappearance of sete and fusion of somites. This condition, 
of course, may have resulted also from a perichztous condition. 


276 W. B. BENHAM. 


remember that the separation of the couples and that penial 
sete are present in various genera and families, I think it is 
allowable to so regard it. 


The Position of the Clitellum.—In the fresh-water 
worms (Microdrili) the clitellum is developed only during 
the breeding season, and around the somite carrying the male 
pore, or those immediately on each side of it. That is, the 
*intra-clitellian ” condition is the more primitive. 

Now, in Moniligaster sapphirinaoides the clitellum is 
on Somites x—x111, and the male pores between Somites x and 
x1. The reason that it has not been observed in other species 
of this genus is very likely due to the fact that it is present 
only for a short period, during the actual breeding season. 

When the male pores shifted backwards, as they have done 
in the rest of the earthworms, the clitellum probably accom- 
panied them, giving rise to what Perrier called “lombriciens 
intra-clitelliens :”’ in some cases the extent of the clitellum is 
small, at other times it is great. But apparently in some 
cases—Pericheta, Acanthodrilus, &c.—whilst retaining 
its limited extent, it has not kept up its relative position, 
coming to lie in front of the male apertures; whilst in the 
family Lumbricide it is still further removed from its primi- 
tive position, and lies far behind the spermiducal pores. 


The Sperm-ducts.—In the majority of the water-worms 
(except Lumbriculide) there is only one pair of sperm-ducts, 
and this I regard as the primitive condition—that is to say, 
when once the position of the genital glands had become fixed 
to definite somites, and the nephridia specialised for the pur- 
pose of conveying generative products to the exterior, there 
was only one pair serving as sperm-ducts, and one pair as 
oviducts ; previously to this state of things of course we should 
get a less limited specialization ; but from general considera- 
tions I believe one pair, and not two pairs (if so, why not 
three pairs or four pairs?), of sperm-ducts was the typical 
arrangement, 


AN ATTEMPT TO CLASSIFY EARTHWORMS. 277 


This condition is retainedin Moniligaster, where, too, the 
ducts are limited in length, passing through only one septum, 
and have their external aperture more nearly in the position 
common to the majority of water-worms than in any other 
earthworm. The single pair of sperm-ducts (and testes) is 
retained in the family Geoscolecide, in which, too, we find 
the sperm-sacs occupying, as in Tubifex, several somites. In 
Typheus, again, this primitive character is retained. The 
size of the ovisac in Moniligaster recalls the fact that the 
ova in water-worms after separation from the ovary push the 
septa back, and come to occupy several somites. 

When a second pair of sperm-ducts appeared, each would 
have its separate external aperture ; but (except in Pericheta 
stuarti, A. G. B.) the two pairs of apertures have dis- 
appeared; the two sperm-ducts become more or less fused 
together; and as in the case of sete and nephridia this 
fusion commences posteriorly and gradually ex- 
tends forwards. Thus in Acanthodrilus, and in Eu- 
drilus and Megascolides, the two ducts remain separate 
till they join the prostate; in Microcheta they remain 
separate through several somites; finally, in Lumbricus and 
others, the two unite immediately behind the second rosette. 


The Prostate.—In the majority of water-worms there is an 
enlargement of the sperm-duct near its pore, and this enlarge- 
ment may have glandular walls; this condition is retained 
in Moniligaster barwelli. In the rest of the earthworms, 
when present, we have either (a) a diverticulum of the sperm- 
duct, (4) a single pair of sacs opening independently of the 
sperm-ducts, or (¢c) a couple of pairs of separate prostates. In 
all the prostatiferous earthworms except in Acanthodrilidex 
we find either (a) or (0). Dichogaster has prostates of both 
varieties. No doubt the tubular prostates, as seen in these 
latter and in other genera, are more primitive than the branched 
prostates of Perichta, the flattened condition seen in 
Cryptodrilus and Perionyx leading towards this. 

Moniligaster barwelli is, in this matter, more primi- 


278 W. B. BENHAM. 


tive than the remainder of the earthworms, and closely 
resembles Stylaria in the condition of its prostate. 

As I said above, we have practically no embryological data 
on which to found our theories as to ‘“‘ primitive” and 
“secondary ” characters in the earthworms. But there is 
one organ on which we have definite information, and that is 
that the dorsal blood-vessel is in Criodrilus formed by the 
fusion of a double vessel. Now, in several earthworms we 
find this double condition of the vessel. 

In Acanthodrilus multiporus and in Deinodrilus 
benhami there is a pair of dorsal vessels; in A. dis- 
similis this vessel is doubled in every somite, fusing at the 
septa: this condition is also present in the anterior somites 
of Microcheta rappi, and according to Beddard in Peri- 
cheta cerulea (Pleurocheta moseleyi), and this seems 
to have been the chief reason, in addition to its plectonephric 
condition, for regarding Acanthodrilus as the more primi- 
tive genus. 

In which worm are any of these organs retained in their 
most primitive condition? I think that Moniligaster sup- 
plies the answer in most points. The setz, clitellum, sperm- 
ducts, and prostate are all in agreement with the above-formu- 
lated conditions. The gizzard, too, is very different from what 
we find in other worms; its walls appear to be much less 
muscular than is usually the case; it is less marked, extends 
through several somites, and recalls the enlarged intestine of 
water-worms, with its wall only slightly thicker than the pre- 
ceding cesophagus.! 

Thus, on the whole, I am inclined to regard Moniligaster 
as the most primitive living earthworm, or rather as 
approaching most nearly to their original ancestor. At the 


1 T should add that the anterior gizzard mentioned by Perrier has not been 
found in any of the species recently described—seven by Bourne, one by 
Horst, one by Beddard; and it is probable that he mistook for gizzard a 
mere dilatation of esophagus, as was the case in his description of Perionyx. 
Here he stated that the gizzard was in Somite x11; Rosa found here a 
swelling only, the true gizzard being in vil. 


AN ATTEMPT TO OLASSIFY EARTHWORMS. 279 


same time, in the condition of its excretory system and in 
the matter of the dorsal blood-vessel, Moniligaster is in 
a less primitive condition than many other worms, which, 
whilst advancing in respect of certain of their other organs, 
retain the primitive network of tubules more or less com- 
pletely. 

My idea as to the relation of the various families is as fol- 
lows:—From some of the earlier “ Limicolous” forms—Lum- 
bricomorpha minora—the earthworms have been derived 
along two lines.! Along one branch (a) the more primitive 
plectonephric condition has been retained from some Platy- 
helminth ancestor of the whole Chetopoda. Along the other 
(B) this has been replaced by the meganephric condition 
more usually found in the group. 

The Typhzide, having a single pair of prostates, stand 
at the end of the main branch of the first line (a); but from 
this line a branch has given rise to the Acanthodrilida, 
to which Dichogaster has some affinity. 

The Perichetidz appear to have arisen from the Typhzid 
stem—from some form with flattened prostates, by multipli- 
cation of sete. Deinodrilus, having a dozen set, would 
not necessarily be related to the Perichetidz, but might 
point to the possibility of the development of a perichetous 
condition in the family Acanthodrilide. 

The branch (8) leads through Moniligaster to the Geo- 
scolecidz, which retain the single pair of testes, &c., and 
exhibit amongst the genera stages in the separation of the 
sete, but which have lost the prostate, a primitive character 
of the group. Springing from this branch is another, leading, 
after the appearance of the second pair of testes, &c., through 
the Eudrilide to the Perionycide. 

The loss of prostates and the extension of the clitellum gives 
us a new line leading to Rhinodrilide, which, through 
Hormogaster, presents some affinity to the Lumbricide. 


1 It is quite possible, of course, that earthworms have not been derived 
from water-worms; the latter may have been developed from earthworms, 
but I think the evidence is in favour of the statement in the text. 


280 W. B. BENHAM. 


Indeed, Hormogaster might perhaps be included in the 
latter family but for the existence of only two pairs of sperm- 
sacs. 


Per ae Rhinodrilide 
Perichetidz 
; ~—~Lumbricidze 
Typhede alt 
Acanthodrilide——_—-- x gure 
“sli 


Ferg ee minora 


I would here record my thanks to Professor Lankester for 
his help and advice on many points during the progress of 
this paper. 


Postscript, April 30th. 


While this paper was in the press, I received from Dr. 
Michaelsen his recently published memoir! describing two 
new species and six new genera from the neighbourhood of 
Zanzibar. 

The two new species are Trigaster stuhlmanni and T. 
affinis, which are evidently very closely similar. He suggests, 
with good reason I think, the removal of Horst’s species 
Acanthodrilus schlegelii, A. biittikoferi, and A. bed- 
dardi, as well as Rosa’s A. scioanus, from the genus under 
which they have been placed in the present paper to my genus 
Trigaster (= Benhamia, Mich.), since the male pores are 


1 «Beschreibung d. v. H. Dr. F. Stuhlmann im Miindungsgebiet des 
Sambesi gesammelten Terricolen,” ‘Jahrb. d. hamburg. wiss. Anstalten,’ 
vii, 1890. 


AN ATTEMPT: TO CLASSIFY EARTHWORMS. 281 


placed in a deep median fossa, and with the exception of 
T. schlegelii they each have two gizzards. It is to be 
regretted that the “law of priority” is so frequently dis- 
regarded. The name Trigaster, though losing its struc- 
tural significance, has every right to be retained as a generic 
title. The new genera appear to belong to my family 
Eudrilidz with one exception; but it is a pity that more 
figures, in illustration of the very curious arrangement of the 
genital system in some of these, have not been given. I have 
diagrammatised four of these genera. 


1. Pygmeodrilus. 


Setze in four couples. 

Clitellum complete, round Somites xiv, xv, XVI. 

Male pores paired on Somite xvit. 

Nephridiopores in front of the outer couples of sete. 

The male apparatus resembles that of Eudrilus, but the 
female system is not aberrant [for further characters see 
fig. 33]. 

Species.—P. quilimanensis; from Quilimane. 


2. Eudriloides. 


Sete in four couples. 

Clitellum complete, on Somites xiv to xvitt. 

Male pore single, median, on Somite xvit. 

Spermathecal pore single, median, between Somites 
X111/XIv. 

There is no direct communication between oviduct and 
spermatheca. 

Penial setz and dorsal pores are present. 

The details given are insufficient to diagrammatise. 

Species.—E. parvus and E. gypsatus. 


3. Nemertodrilus. 
Setz in tour couples. 
Clitellum on Somites xu to xvitt. 
Male pores paired between Somites xvi1/xviil. 
“Spermathecal” (?) pores paired on Somite x11. 
VOL. XXXI, PART II,—NEW SER. T 


282 W. B. BENHAM. 


Ovipores on Somite xiv. 

Nephridiopores in line with inner couples of sete. 

No penial sete. 

The septa, x111/x1v and x11/x111, are fused together except 
below the intestine, so that Somite x111 is almost obliterated. 

There is a connection between the oviduct and ‘‘ sperma- 
theca” of each side; in fact, the so-called spermatheca appear 
to be a greatly elongated ovisac, which has, as Michaelsen 
suggests, taken on the function of spermatheca (see Poly- 
toreutes). Apart from this the genus somewhat resembles 
Rhododrilus. 

Species.—N. griseus. 


4, Callidrilus. 


Setz in four couples. 

Clitellum only developed ventrally on Somites xvi1 to xx. 

Male pores paired on Somite xvii. 

Spermathecal pores numerous, between Somites x111/xrv. 

Ovipores on Somite xrv. 

Nepkridiopores in line with inner couple of setz. 

Numerous paired copulatory pits are present on Somites 
XI tO XXIV. 

Spermathece in the form of small sacs; a dozen in 
auterior of Somite xiv [cf. Brachydrilus]. 

Species.—C. scrobifer. 


5. Polytoreutes. 


Setze separate, eight. 

Clitellum on Somites x111 to xvii. 

Male pore median on Somite xvii. 

“ Spermathecal” pore median, single, on Somite xrx. 

Ovipores paired on Somite xiv. 

No penial sete. 

Each prostate is very long, and provided with two rows of 
small contiguous diverticula along its whole length. The two 
prostates unite on Somite xvir. The “‘spermatheca” isa 
median sac passing forwards from its pore in Somite x1x to 


AN ATTEMPT TO CLASSIFY EARTHWORMS. 283 


Somite xiv, where it divides into two short processes, one to 
each oviduct. InSomites xvi and xviii there are long, paired, 
blind processes from the spermatheca. These remind one of 
the elongated ovisacs of the water-worms. ‘This is the only 
form known with spermathecal aperture behind the male pore, 
and it appears to me to be doubtful whether this sac is homo- 
logous with the spermatheca of the ordinary type. Michaelsen 
makes no statement as to its contents. It may be, though he 
gives no grounds for this supposition, that the oviducts are in 
a state of degeneration, and that the “‘ spermatheca ” serves as 
an enlarged ovisac, in which perhaps the ova are fertilised and 
retained during development. In fact, the worm may be ovi- 
parous. 
Species.—P. ceruleus. 


6. Stuhlmannia. 


Sete four couples. 

Clitellum on Somites xiv to xvit. 

Male pore median in Somite xvit. 

Spermathecal pore median, single, in Somite x11. 

Ovipores in Somite xiv. 

Two long prostates extending from Somites xvir to xxiv, 
and uniting in xvit. 

Median spermatheca; from its proximal end a pair of out- 
growths surround the intestine and meet dorsally. 

Oviducts communicate with the spermatheca. 

These structures are so complicated, and so brief a descrip- 
tion of them is given, that I have not attempted to construct a 
diagram. 

Species.—S. variabilis. 


Of these six genera, all but one—viz. Callidrilus—are 
probably referable to the family Eudrilide, mihi. The 
exception appears to belong to the family Rhinodrilide, 
mihi, although it presents one or two points in which it does 
not agree with my diagnosis of the family, e.g. position of 
male pore and of nephridiopores. Michaelsen gives no details 
as to the structure of what he called “ prostate,” and it may 


284 W. B. BENHAM. 


very likely be merely a thickening of the body-wall. If it be 
a prostate, then Callidrilus, like the rest, will belong to 
Eudrilide; but even here it will form an exception to the 
general characters of the family in the position of the clitellum, 
and in the character of the nephridia and spermathece. 

The chief point in which Pygmezodrilus approaches 
Eudrilus is in the possession of elongated prostates, and of a 
penis within a “bursa.” The muscular part of the sperm- 
duct very probably corresponds to a portion of the prostate of 
Eudrilus, in which the duct enters the prostate some distance 
along its length. 

Polytoreutes and Stuhlmannia present so many ab- 
normal characters that it is desirable that we should have more 
detail before deciding on their affinities. Apparently they are 
most nearly related to Teleudrilus. 


VII. Expianation oF DraGRrams. 


These diagrams, representing the genital, alimentary, and 
excretory systems of the genera recognised in the accompany- 
ing paper, have been constructed in most cases from drawings 
published by the various authors mentioned below, or, where no 
figures have been given, from the descriptions of the different 
worms. I have usually selected those species which have been 
most fully and most recently described, as types of the genera. It 
must be borne in mind that the accompanying figures are merely 
diagrams, and are not accurate copies from previous figures. 
I have arranged them in the same order as that in which the 
genera have been described in the body of the paper, and the 
numbering of the figures agrees with the numbering of the 
genera. 

In every case the clitellum, if it occurs within the first 
twenty somites, is indicated by the thickened boundary of the 
diagram. In all cases the upper figure (a) represents genital 
system ; the middle (4), alimentary and excretory systems ; the 
lowermost figure (c), a few somites seen externally. 

In the diagrams of the genital system the testes and ovaries 
are in black, the genital ducts in outline, the sperm-sacs and 


AN ATTEMPT TO CLASSIFY EARTHWORMS. 285 


ovisacs are dotted; the spermathece are outlined; and the 
prostate and sac with penial sete are also represented in 
outline. 

In the alimentary canal, the extent of the buccal cavity as re- 
presented is not intended to indicate the actual extent, as we 
have no sufficient data for determining this point; hence the 
position of the commencement of pharynx is hypothetical ; but 
its posterior limit is true in most cases, although in many genera 
this point is left vague in the drawings and descriptions of 
authors: the limit here assigned is, in these cases, deduced 
from analogy with better known genera. The gizzard is 
indicated by its thicker outline, and by the transverse line in 
front and behind it. — 

In the excretory system the black dots represent the funnels, 
the thicker line the ‘duct,’ and the narrow wavy line the 
coiled tube: there is no attempt to indicate the arrangement 
of this coil. In the “ plectonephric” genera the short lines 
are intended to indicate the passage of the duct through the 
body-wall to the exterior. The external openings are shown 
in the lowermost diagram as small circles,' the sete as short 
lines, and any genital apertures as black dots. In cases where 
the nephridiopores are not indicated their arrangement is not 
known. 

The representation of the excretory system is purely con- 
ventional, especially in forms with “tufts” or network, and 
merely serves to indicate whether the genus is “ plectonephric ” 
or “ meganephric,”’ and in the latter case whether the duct is 
“simple” or “ cecal.” 

The modification of the anterior nephridia or excretory net- 
work to serve to moisten the food, and act as ‘‘ pepto-nephridia,” 
is indicated. 

The diagrams represent the worms slit open along the middle 
line of the dorsal surface, and cut edges of the body-wall pinned 
aside. 


1 Unfortunately the ‘“‘ process” has in many cases not reproduced the 
circle, so that the nephridiopores appear as small black dots. The index 
lines, too, are often not reproduced. 


286 W. B. BENHAM. 


Family I.—Typazipz. 


Fic. 1.—Typhezus. The drawings, both the alimentary and 
the genital systems, are taken from the descriptions and figures 
of T. gammii by Beddard, ‘ Quart. Journ. Micr. Sci.,’ xxix, 
p- 111, and pl. xii. 


In fig. 1 a, the small sac (Ps.) in Somite xvi1 represents the special sac 
with penial sete, the external aperture of which is shown in fig. 1 ¢, in front 
of the male pore on each side, which is indicated by the large black dot. Pro. 
is the prostate, which joins the sperm-duct before the latter opens externally. 
The outlined structures in Somite vir are the spermathece. 

In fig. 1 4, the gizzard is shown in Somite vir; the calciferous diverticula 
in Somite xir (Ca.). The sacculated intestine commences in Somite xvz. 
EPN. is the extra-buccal pepto-nephridial network, which is shown in thicker 
lines than the remaining network. 

Fig. 1 ¢, represents Somites xvI, xvi, and xv1iI seen externally, when 
flattened out; showing the arrangement of the sete, the numerous nephridio- 
pores (as small dots), the penial sete in Somite xv replacing the ordinary 
sete; and behind these the large black dots indicate the male pores. 


Distribution of the genus: India. 


Fie. 2.—Megascolides. Modified from the drawings of 
M. australis, McCoy, given by Baldwin Spencer in ‘ Trans. 
Roy. Soe. Victoria,’ vol. i, pl. i. 

In fig. 2 @, the four pairs of lobed, dotted structures marked s., in Somites 
XI, XII, XIII, and XIv, represent the sperm-sacs. PRo. is the coiled prostate. 

In fig. 2 4, the plectonephric condition is seen, the network in Somites 1 
to IV communicates with the pharynx by small ducts, and forms an “ intra-buccal 
pepto-nephridial network ” (1PN.). G. is the gizzardin Somite v. In Somites 
x11 to xvi the pouched intestine is shown (Pp.), the sacculated, non-typh- 
losolar intestine commencing in the next somite. 

Fig. 2 c, represents Somites XVII, XVIII, and xrx externally, showing sete, 
nephridiopores, and the male pores which are in XVII, 


Distribution of the genus: Australia. 


Fic. 3.—Cryptodrilus. a. Genital system, modified from 
the figure of C. fletcheri, given by Beddard in ‘ Proce. Zool. 
Soc.,’ 1887, p, 547. pro.,the prostate. 6. Alimentary system, 
composed from Beddard’s description of C. fletcheri in 
‘Proc. Zool. Soc.,’ 1887, and from Fletcher’s description of 
various species in ‘ Proc. Linn. Soc. N.S.W.’ 


The gizzard occupies two somites. oa. The first of the four calciferous 
diverticula on each side. The sacculated intestine commences in Somite XVI. 

The plectonephridia are not continuous from somite to somite. 

Fig. 2 c, represents any three somites behind the male pores, showing the 
sete: the nephridiopores are probably numerous and irregular, 


Distribution : Australia. 


= 


BF PPEEE Ss *® 


del | | LL SS 


fell 


FLaaB0ss 


ra ee egre 


CRYPTODRILUS 


Dat 


it 


(esa 


BY GIS, 
Z 


= 


CRYPTODRILUS 


ry vs SriCiiep ie - 
TYPHEUS 


MEGASOOLIDES 


288 Ww. B. BENHAM. 


Fic. 4.—Didymogaster. a. Genital system. The clitel- 
lum, xtv to xvi. sF.!, sp.°, are the first and third sperma- 
thece. s. The digitate sperm-sac of Somite rx. ps. The sac 
of penial sete ; and pro., the prostate. 6. Alimentary system. 
cl, «2 The two gizzards lying in Somites vi and vir. Both 
diagrams modified from Fletcher’s figures of D. sylvaticus 
in ‘Proc. Linn. Soc. N.S.W.,’ i (2nd ser.), 1886. c. repre- 
sents three somites from any part of the body, and shows the 
arrangement of the sete; the nephridiopores are probably, as 
in other plectonephric genera, numerous. 

Distribution: Australia. 


Fic. 5.—Perissogaster. a. Genital system. ps. Sac 
with penial sete. pro. The unequally lobed prostate. 0. 
Alimentary system. a.!, c.?, Gc. The three gizzards. p. The 
first dilatation of the intestine in Somite rx, followed by five 
others, all of which are doubtfully calciferous. s. The com- 
mencement of the sacculated intestine in Somite xv. ¢. shows 
the arrangement of the sete. Constructed from the descrip- 
tion given by Fletcher for P. excavatus in ‘ Proc. Linn. 
Soc. N.S.W..,’ vol. ii, ser. 2, 1887. 

Distribution : Australia. 


Fie. 6.—Dichogaster. a. Genital system. vPro. The 
first prostate, communicating with the sperm-duct. pro.” and 
pro.’ The second and third prostates, which are independent 
of the sperm-duct. 6. Alimentary system. c.', c.? The two 
gizzards, each occupying two somites. Ca. The first of the 
three calciferous diverticula on each side. s. The commence- 
ment of the sacculated intestine. ren. The intra-buccal pepto- 
nephridium with its duct. c. represents the Somites xvi, 
xvir, and xvi externally. The large black dots in xvir are 
the male pores, those in xv1i1 the prostate pores of this somite. 
In these two and in Somite xrx the inner couples of setz are 
absent. Modified from the figures given by Beddard for D. 
damonis in ‘ Quart. Journ. Micr. Sci.,’ xxix, pls. xxiii, xxiv, 

Distribution: Fiji. 


=| 
i 


\ 
6, 


MULT TY PT 


in 


PERISSOGASTER 


DIDYMOGASTER 


T DICHOGASTER 
DIDYMOGASTER PERISSOGASTER 


290 WwW. B. BENHAM. 


Fie. 7.—Digaster. a. Genital system. ps.) The anterior 
sac with penial sete. The posterior sac, which is described as 
lying in this somite, is omitted to prevent crowding. Pro. 
Prostate. 6. Alimentary system. «a.!, gc.” The two gizzards 
lying in Somites v and vir. The intra-buccal pepto-nephridia 
are indicated at 1pn. c. The exterior of Somites xvi, xvitt, 
and x1x to show the sete; male pores represented by black 
dots, and the openings of the two pairs of sacs with penial 
sete, ps.!, ps.”, lying one in front of the other. 

Modified from the figures of D. lumbricoides given by Perrier in 
‘Nouv. Arch. du Mus. d’Hist. Nat. de Paris,’ viii, 1872, pl. ii; the modifica- 
tions being in accordance with Fletcher’s description of D. armifera in 
‘Proc. Linn. Soc. N.S.W..,’ vol. i, p. 943. 

Distribution: Australia. 


Family II1.—AcantHopRILip#. 


Fie. 8.—Acanthodrilus. a. Genital system. Modified 
from Beddard’s figure of A. dissimilis in ‘ Proc. Zool. Soc.,’ 
1887, p. 388, in accordance with his more recent observations 
in regard to the independent opening of sperm-duct and pros- 
tates. ps.t, ps.2 The sacs of penial sete in Somites xvir and 
xvill. PRo.!, pro.” The prostates in Somites xvii and xviii. 
6. Alimentary system. 

Modified from Beddard’s figure of A. multiporus in ‘ Proc. Zool. Soc.,’ 
1885, pl. liii. Ca. The first pair of calciferous diverticula. s. Commence- 
ment of sacculated intestine in Somite xvi. pn. Intra-buccal pepto- 
nephridia. The grouping of the nephridial network so as to form eight groups, 
six of which only are represented, is indicated in Somites xv1, et seq. See 
Beddard, ‘ Proc. Zool. Soc.,’? 1885. The clitellum in A. dissimilis and A. 
multiporus commences in Somite x11. I have represented it as beginning 
at Somite xIII, which appears to be more usuaily the case. c. External view 
of two Somites, a, B, of A. multiporus, and others in which the eight 
setee are separate. c, somite of A. nove-zealandiz, and other species in 
which the sete are in couples. 

Distribution: New Zealand, Australia, West Africa, New 
Caledonia, Kerguelen Island, the shores of Magellan Strait. 


Fic. 9.—Trigaster. a. Genital system. sp. Sperma- 
thece. o. Ovary. pro.!, pro.? The prostates. Sperm-ducts, 
sperm-sacs, and testes unknown.' 9. Alimentary system. 
Slightly altered from my own figures of T. lankesteri in 
‘Quart. Journ. Micr. Sci.,’ xxvii, pl. ix, after a renewed 
examination as to the position of some of the structures. 

c.1, G.2, 6.3 The three gizzards. s. Commencement of sacculated intestine. 
The anterior nephridial network is more evident than that more posteriorly, but 
it is uncertain whether there is any communication with the pharynx. c. Three 
somites seen externally to show the arrangement of the sete. 


Distribution: St. Thomas, West Indies. 


Michaelsen, however, has in his last paper described the testes and 
ciliated rosettes in Somites x and x1, and sperm-sacs in Somites xI and x1. 


eaten 


TS 


TRIGASTER 


ACAN THODRILUS - ‘TRIGASTER 


292 W. B. BENHAM. 


Fic. 10.—Deinodrilus. a. Genital system. sp. Anterior 
spermatheca. pro.!, pRo.? The two prostates of one side. 46.! 
Alimentary system. oc. Gizzard occupying two somites. 
EPN. The extra-buccal pepto-nephridia. A connection exists 
here and there between the nephridial network of neighbouring 
somites. Beddard says but little of the alimentary canal or 
the nephridia, but special groups (ePn.) of the latter occur in 
Somites 11, 111, and iv. c. Exterior of these somites, showing 
the characteristic twelve sete and numerous nephridiopores. 
Constructed from Beddard’s description of D. benhami in 
‘Quart. Journ. Micr. Sci.,’ xxix, p. 105. 

Distribution: New Zealand. 


Family I1I.—Prricnarips. 


Fic. 11.—Pericheta. a. Genital system. Modified from 
Rosa’s figure of P. fez in ‘ Ann. Mus. Civico d. Stor. Nat. di 
Genova,’ vi, 1888, pl. i. 5. Alimentary system. oc. Gizzard, 
occupying more or less of three somites. vp. The first, and p.1 
the last, of the thirteen pouches. Nine somites are cut away, 
as they are merely a repetition of these pouches. c. The 
characteristic cylindrical czcum of one side. [These are 
absent in some species.] s. The commencement of the sac- 
culated, typhlosolar intestine. Composed partly from Perrier’s 
and Rosa’s description. The excretory system is taken from 
Beddard’s figure of P. aspergillum in ‘ Quart. Journ. Micr. 
Sci.,’? xxix, pl. xxiv. The black dots represent the funnels 
of the nephridial network. c¢. represents three somites from 
different regions of the body of P. monticolla, Fletcher, 
‘Proc. Linn. Soc. N.S.W., vol. ii. In Somite a, from anterior 
part of body, only sixteen sete are present ; in Somite B, from 
about Somite xv, there are thirty setz, and posteriorly in Somite 
c fifty or more. This same variation in numbers occurs in 
other species. The nephridiopores (represented as small 
dots) are drawn from Beddard’s description of P. asper- 
gillum. 

Distribution of the genus: India (with Ceylon), Malaya, 
Australia, and islands between the two continents. 


DELNODRILUS 


294 W. B. BENHAM. 


Family [V.—Moni.icastTripa. 


Fie. 12.—Moniligaster. a. The left upper figure, genital 
system. Modified from Horst’s figure of M. houteni in 
‘Notes from the Leyden Museum,’ ix, pl. i, fig. 1, in accord- 
ance with Beddard’s more recent figures and descriptions 
(‘Quart. Journ. Micr. Sci.,’ xxix, pl. xii, fig. 12; and ‘ Zool. 
Anz.,’ No. 818, 1889). sp. Spermatheca, with its long duct, 
in Somite 1x. These are the “anterior testes” of Perrier. 
s. Sperm-sacs—Perrier’s “ posterior testes.” sp. Sperm-duct 
and ciliated rosette. pro. Prostate. o. Ovary. op. Oviduct. 
os. Ovisac. 6. The right figure, alimentary system. Modified 
from Perrier’s figure of M. deshayesii in ‘ Nouv. Arch. d. 
Mus. d’Hist. Nat.,’ viii, 1872, in accordance with the more 
recent descriptions of Beddard, Horst, and Bourne, who have 
not observed Perrier’s anterior gizzard. «. The characteristic 
elongated gizzard region. The nephridia are after Horst. ca. 
The cecum of the nephridial duct. c. The left lower figure, 
the exterior of Somites x to x111I—the region of the clitellum, 
showing the sete, nephridiopores, and male pores, which lie 
between Somites xi and x11. 


Distribution: India, Ceylon, Sumatra. 


AN ATTEMPT TO CLASSIFY EARTHWORMS 295 


MONILIGAS TER, 


296 W. B. BENHAM. 


Family V.—EvpriLip#. 

Fic. 13.—Eudrilus. a. Genital system. Modified from 
Beddard’s figure of E. sylvicola in ‘ Proc. Zool. Soc.,’ 1887, 
p. 381, in accordance with his subsequent descriptions and 
figures of the female apparatus, which are peculiarly arranged. 


Tt will be seen that the ovary (o'.) in Somite x11 is enclosed in a sac, the 
neck of which communicates with the true oviduct (oD.); into the latter also 
open an albumen-gland (GL.), a large spermatheca (sp.), and the ovisac (0.),’ 
which functions also as a second ovary, and lies in Somite xiv. PRo. is the 
prostate, into which the sperm-ducts open about halfway along its length. 
x. is one of two glands communicating with B.c., the bursa copulatrix, which 
contains a chitinous penis: the bursa opens externally at Pp. in xvu, a dotted 
ring? indicating the male pore. 4. Alimentary system. Modified from 
Perrier’s figure of E. decipiens in ‘ Nouv. Arch. Mus. d’Hist. Nat.,’ viii, 
1872, pl. ii, fig. 26, in accordance with Horst’s more recent description in 
‘Notes from the Leyden Museum,’ ix. oc. The gizzard. ca. Calciferous 
diverticula. Commencement of the sacculated intestine in Somite xv. The 
nephridia are original; they present no cecum, but the duct is continuous 
with the tubule asin Lumbricus. c. External view of a normal somite of 
E. decipiens, showing (A) sete and nephridiopores ; (8) Somite xvir of the 
same species, showing the male pores—as black oval dots—occupying the 
position of the inner couples of sete, which are absent; (Cc) is a normal 
somite of E. sylvicola, in which the nephridiopores are in line with the 
inner couple of sete, instead of with the outer couple, as in other species. 

Distribution : South America and New Caledonia. 


Fic. 14.—Teleudrilus. a. Genital system. Notice the 
peculiar recurved ciliated rosettes ; the median position of the 
male pore in Somite x1x (represented as a dotted semicircle P.) 
and spermathecal pore in Somite xvii is peculiar. 

The female organs present some difference from the typical arrangement, 
analogous to that in Kudrilus in that there is a connection between the 
oviduct and spermatheca, but a less direct communication in that genus. 
o. Ovary. op. Oviduct. os. Ovisac.? sp. Spermatheca. 3B.c. Bursa copu- 
latrix, opening externally at P in xIx. It is in communication with the sperm- 
ducts, prostates (PRo.), and asac (s.), probably glandular. 4. Alimentary 
system. After Rosa’s figures of T. raggazil in ‘Ann. Mus. Civ. d. Stor. 
Nat. di Genova,’ vi, 1888, pl. ix. G. Gizzard. ca. Calciferous diverticula. 
s. The commencement of sacculated intestine. c. represents the exterior of 
Somites XVIII, XIx, and xx, showing sete, nephridiopores, and median male 
pore in Somite xix. 


Distribution: Scioa, Africa. 


Fie. 15.—Pontodrilus. a. Genital system. pro. Pros- 
tate. sp. Spermatheca. 6. Alimentary system. Note absence 
of gizzard. s. Sacculated but non-typhlosolar intestine. Note: 
the nephridia do not commence till Somite xv. c. shows 
the exterior of Somites xvi, xvir, and xvii, with sete, 
nephridiopores, and male pores. After Perrier’s figures for 
P. marionis in ‘ Arch. de Zool. Expér. et Gén.,’ ix, 1881. 

Distribution: Europe (France). 

1 The index line does not go quite far enough; it should extend to the 
small round sac at the end of the tortuous tube. 


2 This is very feebly indicated in the diagrams. 
3 The index line should extend to the round dotted area. 


xa 
ce ee 
Sie eee 


EUDRILUS TELEUDRILUS 


VOL. XXXI, PART II.— NEW SER. 


PONTODRILUS 


298 W. B. BENHAM. 


Fic. 16.—Photodrilus. a. Genital system. sp. Sperma- 
theca. pro. Prostate. 6. Alimentary system. Both diagrams 
are constructed from the description given by Giard for P. 
phosphoreus in ‘ Comptes rendus,’ 1887. Here the nephridia 
do not commence till Somite xrv. Their pores are between 
the outer and inner couple of sete as shown in fig. ¢. 

Distribution : Europe (France). 


Fie. 17.—Microscolex. a. Genital system. Pro. Prostate. 
ps. Sac with penial sete. 6. Alimentary system. Constructed 
from Rosa’s descriptions of M. modestus in ‘ Boll. Mus. Zool. 
ed Anat. Comp. Torino,’ ii and ii. Although closely allied to 
the two preceding genera, the nephridia, as usual, commence 
far forwards. c., taken from Rosa’s woodcut, represents Somites 
XVI, XVII, xvi1I. He does not state that the sete are absent 
on this somite, but he does not figure them. The male pore 
is in line with the seta 1. 

Distribution : Italy. 


Fie. 18.—Rhododrilus. a. Genital system. sp. The first 
of the four spermathece on one side. pro. Prostate opening 
independently of the sperm-duct. 6. Alimentary system. 6G. 
Gizzard. s. Commencement of sacculated intestine. c¢. Ex- 
terior of three somites. Diagrams constructed from Beddard’s 
brief diagnosis of R. minutus in ‘Proc. Zool. Soc.,’ 1889, 
p. 381. 

Distribution: New Zealand. 


18. 


ft 


( BEBE 
NULL PREP 


PRP Pea Ae eee ee ao 


RHODODRILUS 


MICROSCOLEX 


PHOTODRILUs 


RHODODRILUS 


MICROSCOLEX 


LUS 


RHODODRI 


MICROSCOLEX 


PHOTODRILUS 


300 W. B. BENHAM. 


Fie. 19.—Plutellus. a. Genital system. The position 
of the genital organs is very abnormal, and requires confirma- 
tion. s. The sperm-sacs in Somite x11. sp. The first of the 
five spermathece. pro. Prostate. The sperm-funnels and 
ducts are unknown. o. Ovary in Somite x; and op., the ovi- 
duct opening externally in Somite x1. 6. Alimentary system. 
e. Gizzard. ca. Calciferous diverticula. s. Commencement of 
sacculated intestine. The nephridia are shown alternating in 
position, the first four pores being in line with the third seta, 
the rest alternating with second and fourth sete. Moreover, 
the funnels are said to be in the same somite as the coiled 
tubule and external aperture. n.! The first of the series of 
nephridia which open in line with the third seta. wN.2 The first 
of the series which open in line with the fourth seta. n.3 
The first of the series in line with second seta. c. An external 
view of Somite v1, to show spermathecal pore (black) in line 
with second seta, and nephridiopore (n‘) in line with third 
seta; of Somite x11, to show the normal arrangement in the 
even numbered somites; and of Somite x111, to show normal 
arrangement of the odd numbered somites. Composed from 
Perrier’s description of P. heteroporus in ‘Arch. de Zool. 
Expér. et Gén.,’ 11, 1873, p. 381. 

Distribution : Pennsylvania. 


Family VI.—Perionycip#. 


Fic. 20.—Perionyx. a. Genital system. o. Ovary. op. 
Oviduct. pro. Prostate. 6. Alimentary system. oc. Gizzard. 
s. Commencement of sacculated intestine. Modified from 
Perrier’s figures of P. excavatus in ‘ Nouv. Arch. d. Mus. 
d’Hist. Nat.,’? 1872, in accordance with Rosa’s description of 
the same species in ‘ Ann. Mus. Civ. d. St. Nat. di Genova,’ 
vi, 1888. c. External view of three somites, showing the 
numerous set and the paired nephridiopores. 

Distribution: India, Burmah, Philippines. 


PERJONYX 


~ PLUTELLUS PERIONYX _ 


302 W. B. BENHAM. 


Family VII.—GeroscoLecip™. 

Fie. 21.—Geoscolex (Titanus). a. On the left genital 
system. o. Ovary. op. Oviduct. No spermathece are known. 
Modified (in accordance with my own observations) from 
Perrier’s figure of G. maximus, Leuckart (T. brasiliensis, 
E. P.), in Nouv. Arch. Mus. d’Hist. Nat.,’ viii, 1872, pl. i, fig. 
15. 6. On the right alimentary system. «G. Gizzard. Ca. 
Calciferous diverticulum. s. Commencement of sacculated 
intestine. The first nephridium is slightly different from the 
rest, and forms an extra-buccal pepto-nephridium, EPN., the 
coiled tubule being more compact, and the cecal part of the 
duct shorter. Nn. The anterior nephridia, in which the 
tubule leaves the duct about halfway along its length. w.! 
The posterior nephridia, in which the tubule joins the cecum 
near its external aperture. Composed from my own observa- 
tions. c. Exterior of four somites: a.a. from the anterior part 
of the body, where the sete are in couples; B.B. from the 
posterior region, where the setz are separate. 

Distribution: Brazil. 


Fie. 22.—Urocheta. a. Genital system. Composed from 
Beddard’s description in ‘ Quart. Journ. Micr. Sci.,’ xxix, 
p- 246. sp. The first spermatheca. 4. Alimentary system. G. 
Gizzard. ca. Calciferous gland. s. Commencement of the saccu- 
lated intestine. Composed from my own observations. The 
nephridia are modified from Perrier’s figures of U. corethrura 
in ‘Arch, Zool. Exp., iii, 1874. The first nephridium (rpwn.) 
is much larger than the following ones, both the tubular 
portion and its duct being greatly developed; there are at least 
three funnels to this extra-buccal pepto-nephridium. c. View 
of four somites, namely, v1, xx, and two consecutive more 
posterior somites (pp.), in order to show the couples of sete 
anteriorly, and the scattered and alternate arrangement of these 
posteriorly. In Somite xx, in which the spermiducal pore is 
situated but not shown, the ventralmost sete (No. 1) are 
replaced by groups of larger penial sete. After Beddard, 
‘Proc. Roy. Soc. Edinb.,’ xiv, 1887, p. 162. 

Distribution: South America and neighbouring islands 
also Australia, Sumatra, 


21, 


GEOSCOLEX 


| UROCBE&TA . 


304 W. B. BENHAM. 


Fic. 23.—Diacheta. a,b. Genital and alimentary sys- 
tems. sp. Spermatheca. «a. Gizzard. s. Commencement of 
acculated intestine. Modified from my own figures in ‘ Quart. 
Journ. Micr. Sci.,’ xxvii. n. The normal nephridia commence, 
as in Urocheta, immediately behind the extra-buccal pepto- 
nephridium (PN.). c. Somites XXI, XXII, XXIII, seen exter- 
nally, showing the alternation of the setz with exception of 
No. 1 on each side, the position of the nephridiopores (rings), 
and of the spermiducal pores, which are represented as black 
dots. 

Distribution: St. Thomas, West Indies. 


Family VIII.—Ruinoprivipaz. 


Fie. 24.—Rhinodrilus. a, 6. Genital and alimentary 
systems and nephridia. Composed from my own observations 
on, and from Beddard’s description of, R. (Thamnodrilus) 
gulielmus, ‘ Proc. Zool. Soc.,’ 1887, p. 154. Following the 
greatly modified first or extra-buccal pepto-nephridium (nPw.) 
are seven pairs of nephridia (N.) which differ from the more 
posterior ones (N.!) in Somite x in having no cecal prolon- 
gation of the duct. ca. The cecum of the posterior nephridia. 
sp. Spermathece. oe. Gizzard. ca. The first calciferous diver- 
ticulum. s. Commencement of the sacculated intestine. c. 
External view of Somites xrx, xx, and xxi, to show sete, 
nephridiopores, and intersegmental spermiducal pores. 


Distribution: North of South America. 


Epa a 


DOO uk im 
NA GCC | 


PrP RERTM ARERR ERA E 


RHINODRILUS 


> DIACILETA 


RHINODRILUS 


DIACHETA 


RIINOGRILUS 


306 W. B. BENHAM. 


Fic. 25.—Microcheta. a, 6. Genital and alimentary 
systems and nephridia. Copied from my own figures of 
Microcheta rappi, ‘Quart. Journ. Micr. Sci.,’ xxvi, 
pl. xv. The cecum of the nephridium is less marked in the 
anterior nephridia (N.). In the posterior nephridia (N.!), in 
Somite x, et seq., the czecum (c@.) is very large. In @, o. is 
ovary ; OD. oviduct ; os. ovisac; sp. the numerous small sper- 
mathece. As mentioned in the body of this paper, the 
peristomium in M. beddardi is provided with setz, so that 
the ovaries and testes are morphologically in their normal 
position. In M. rappi, however, there is no trace of setz in 
the peristomium, so that the gonads appear in one somite in 
advance of the normal position. In 4, «. is the gizzard; ca. 
the calciferous diverticulum; s. the commencement of the 
sacculated intestine. Inc, three normal somites are shown. 

Distribution : Cape of Good Hope and Natal. 


Fic. 26.—Urobenus. a,b. Genital and alimentary systems 
and nephridia. Copied from my own figures in ‘ Quart. Journ. 
Mier. Sci.,’ xxvil, pl. viii. The anterior nephridia (N.N.) are 
larger than the following ones (N.'), and the cecal portion 
(cm.) of the duct is less developed. In a, sp. marks the first 
spermatheca. In 4, G. is the gizzard; ca. the first calciferous 
diverticulum ; P. the first of the series of pouches of the intes- 
tines; p.! the last of the series,—the intermediate eight somites 
are removed ; c. the peculiar cecum in Somite xxvi, and s. 
the commencement of the sacculated intestine. c¢ represents 
Somites xrx, xx, and xx1, to show the setz, nephridiopores, 
and spermiducal pores. 

Distribution : Brazil. 

Fic. 27,.—Hormogaster. a, b. Genital and alimentary 
systems and nephridia. Modified from Rosa’s figures of H. 
redii in ‘ Sulla Struttura dello H. redii,’ Torino, 1888. We 
have no information as to any variation of nephridia. sp. 
Spermatheca. os. Ovisac. G.!, c.?, «.3 The three gizzards. c. 
Globose cecum in Somite xx1. Six somites are removed. ¢ 
represents three ordinary somites. 

Distribution ; Italy. 


R 


" HORMOGASTE 


int 


BOrGrcr (EE We Aehomae eocos A 


gs 
w 


AY 
UROBENUS 


26 
Ss 
= 


cal ED ty 


eanLIaa0 


royetrrpsbenrehanedcered 
ia 


25. 


HF 
if . 3 Cl 


MICROCLETA 


SAN RN G 


Se See goa aa EE ee EE s 


~ MICROCHATA 


‘ uw 
HORMOGASTER 


UROBENUS 


MICROCHA.TA 


308 W. B. BENHAM. 


Fic. 28.—Brachydrilus. a, 6. Genital and alimentary 
systems. Original. See the description, ‘ Zool. Anzeig.,’ 271, 
1888. There are two pairs of nephridia which are very simple 
—iN. the inner nephridia; on. the outer nephridia in every 
somite. In the Somites x and x1 are four sacs on each side, 
which are not represented, as they le underneath the sperm- 
sacs; these probably correspond to the “ albumen-glands” of 
Lumbricus. I believe the same fusion of somites has 
gone on here as in Microcheta. Ina, sp.is one of the small 
spermathece, which in position is abnormal. o. is the ovary. 
In 6, c. is the gizzard. ca. Calciferous gland. s. Commence- 
ment of sacculated intestine. In c, the sete and nephridio- 
pores are shown. 

Distribution : Unknown. 


Family [X.—Lumsricip2. 

Fic. 29.—Lumbricus. a, 6. Genital and alimentary sys- 
tems. Original. See Hering, ‘ Zeit. f. wiss. Zool.,’ viii, 1856, 
pl. xviii; Lankester (alimentary canal), ‘ Quart. Journ. Micr. 
Sci.,”” 1865-6. In a, sp. the first spermatheca of one side. 
os. Ovisac. In 6,G.is gizzard. pr. Proventriculus. s.! Com- 
mencement of sacculated intestine. ca. Calciferous glands, 
bilobed, one lobe in Somite x11, the other in Somite x1; these 
communicate with a pouch (cr.) which opens into the ceso- 
phagus in Somite x1. c¢ shows the exterior of Somites xiv, 
xv, and xvi, with setz, nephridiopores, oviducal pore in xtv, 
and spermiducal pore in xv (as black dots). 

Distribution : Europe. 

Fic. 80.—Allolobophora. a. Genitalsystem. sp. Sper- 
matheca: three pairs are shown, as is the case in A. chloro- 
tica; in other species more, in others less than three pairs are 
present. os. Ovisac. Modified from Bergh’s figure of A. 
turgida, ‘ Zeit. f. wiss. Zool.’ xliv, pl. xxi. 6, Alimentary 
canal. a. Gizzard. pr. Proventriculus. ca. Calciferous 
gland in Somite xr, opening into the pouch (cr.). Original. 
c. External view of a somite from three different species: a. 
of A. chlorotica; Bs. of A. subrubicunda; and c. of A. 
boeckii.! 

Distribution : Europe. 

1 The nephridipores are too feebly indicated, 


LUMERICUS 


me oe se 
| Baio 


XVI 


LUMBRICUS ALLOLOBOPHORS 


310 W. B. BENHAM. 


Fic. 51.—Criodrilus. a, d. Genital and alimentary sys- 
tems. os. Ovisac. w. Thick-walled region of cesophagus. 
From my own figures, ‘Quart. Journ. Micr. Sci.,? xxvii, 
pl. xxxvill. The nephridia, similar to those of Lumbricus, 
commence in Somite x. 

Distribution : Europe. 


Fie. 82.—Allurus. a. Genital system. Modified from Bed- 
dard’s figure, ‘ Quart. Journ. Mier. Sci.,’ xxviii, pl. xxv, fig. 2. 
The small white circles (sp.) in Somite vi11 represent certain 
microscopic spermathece according to Beddard, but no sper- 
matozoa were found, and perhaps, from their abnormal position, 
they may be “albumen-glands.” os. Ovisac. In 3, G. Giz- 
zard. s. Commencement of sacculated intestine. ca.,ca. The 
first and last of the four calciferous glands. x. A pouch 
(? corresponds to the pouch [cr.]in Lumbricus). 4. Alimen- 
tary system. Composed from Beddard’s description, ‘ Quart. 
Journ. Micr. Sci.,’ xxviil, p. 368. 


Distribution : Europe. 


ALLURUS 


CRIODRILUS 


CRLODRILUS 


ALLURUS 


CRIODRILUS 


312 W. B. BENHAM 


[The following figures are placed out of their proper place, as they have 
been constructed during the passage of the text through the press. ] 

Fic. 833.—Pygmeodrilus. a. Genital organs, partly from 
Michaelsen’s figure of P. quilimanensis [‘ Jahrb. d. ham- 
burgh wiss. Anstalten,’ vii, 1890]. sp. Spermatheca, with its 
numerous diverticula. s. Sperm-sac. Mm. Muscular thickening 
of the sperm-duct. sc. Bursa copulatrix, which contains a penis. 
pro. The lay prostate ; the narrow portion is muscular, the distal 
region is glandular. 6. Alimentary canal, from Michaelsen’s 
description. ca. Calciferous diverticulum. There is some douht 
experienced as to the existence of a gizzard. ‘The nephridia are 
not described beyond the statement that there is a pair in each 
somite. c. Exterior of three somites, showing sete in couples ; 
the nephridiopores ; and male pores in xv1ith Somite. 

Distribution : Quilimane, near Zanzibar. 


Fic. 34.—Nemertodrilus. a. Genital system, from 
Michaelsen’s description of N. griseus in ‘Jahrb. d. ham- 
burg. wiss. Anstalten,’ vii, 1890. s!. Anterior sperm-sac. 
s*. Greatly elongated posterior sperm-sac. os. Ovisac, which 
is prolonged backwards (os'.) and is regarded by Michaelsen as 
“ spermatheca.” pro. Prostate. 6. Alimentary tract and 
nephridia, from Michaelsen’s description. «. Gizzard.  s. 
Commencement of sacculated intestine. No details as to 
nephridia are given, except that they are a pair to each somite, 
and have, apparently, a dilated ‘‘ duct.” c. Exterior of Somites 
XVI, xviI, and xviiI, to show the couple of sete, nephridio- 
pores, and male pores. 

Distribution: Quilimane, Zanzibar. 


Fic. 85.—Callidrilus. From Michaelsen’s description of 
C. scrobifer, in ‘Jahrb. d. hamburg. wiss. Anstalten,’ vii, 
1890. a. Genital system. s.s.* The first and fourth sperm- 
sacs. sp. The numerous, small spermathece. x. is a struc- 
ture which Michaelsen identifies as a prostate. He states 
that it is small, and from the general anatomy of the worm I 
fancy that it may be merely a thickening of the body, such as 
is present in Brachydrilus. 6. Alimentary and excretory 
system. mM. A thickening, which is, according to Michaelsen, 
not muscular. It probably represents a gizzard. s. Com- 
mencement of sacculated intestine. The nephridia are merely 
said to be paired, and to be provided with a bladder, which I 
take to mean a “ cecal” outgrowth of the duct. c. Exterior 
of Somites xvi, xvi1, and xvi11, to show couples of setz, 
nephridiopores, and male pores. [N.B.—I believe this worm 
belongs to my family Rhinodrilide]. 

Distribution: Quilimane, Zanzibar. 


PYGMEODRILUS NEMERTODRILUS 


CALLIDRILUS 
VOL, XXXI, PART II.—NEW SER. D. ¢ 


314. W. B. BENHAM. 


Fic. 36.—Polytoreutes. A very brief description, with a 
figure, is given by Michaelsen of the hinder portion of the 
genital system of P. ceruleus, in ‘Jahrb. d. hamburg. wiss. 
Anstalten,’ vii, 1890. No details are given of the alimentary 
system. a. Genital system. sp. Sperm-duct. 0. Ovary. 
m. A large sac, into which open (op.) the oviduct; the ovisac 
(os.) and the so-called “spermathece” (w.). This has a 
unique position and shape; it is produced into lateral pouches 
(w.!, w.*), and opens externally by a median pore (wo.) in 
Somite xrx. The oviducal pore is shown at op. The prostate 
(PRO.) is long, but how long Michaelsen does not say ; and is 
beset with numerous secondary sacs. Its aperture is in front 
of that of the spermatheca on Somite xvu1, and is represented 
by a dotted circle, and labelled pv. c. Exterior of Somites 
XVII, XVIII, and x1x, to show the sete and median genital 
pores; that of the spermatheca in x1x, and that of the ducts 
in XVIIe 


Distribution : Zanzibar. 


AN ATTEMPT TO CLASSIFY EARTHWORMS. 3815 


POLYTOREUTES 


Aly 


‘Se, 


a 


rae aeik $e] 


ON THE ORIGIN OF VERTEBRATES FROM ARACHNIDS. 317 


On the Origin of Vertebrates from Arachnids.! 
By 


William Patten, Ph.D., 
Professor of Biology in the University of North Dakota, Grand Forks. 


With Plates XXIII and XXIV. 


“In the growth of each science, not only is correct observation needful for 
the formation of true theory, but true theory is needful as a preliminary to 
correct observation.”—H. SPENCER. 


*The “ Annelid theory,” after fifteen years of dexterous 
modelling, is now as far as ever either from fitting the facts of 
Vertebrate structure, or from shedding any direct light on the 
great problem of the origin of Vertebrates. It certainly is not 
without significance that, of all those who with willing eyes 
and minds have grappled with the Annelid theory, not one has 
discovered a distinctively Annelid feature in Vertebrates: 
mesoblastic somites, nephridia, segmental appendages, and seg- 
mental sense-organs are found in nearly all segmented 
animals. 

1 Tinclude in the Arachnida the Spiders, Scorpions, Limulus, Trilobites, 
and Merostomata. 

2 Most of my observations on Acilius, Scorpio, and Limulus 
were made in the Lake Laboratory, Milwaukee, Wis. I am greatly indebted 
to the founder of that institution, Mr. E. P. Allis, for generously placing at 
my disposal the excellent facilities for research which his laboratory affords. 

As a full description of my observations could not be published without 
considerable delay, it seemed advisable to present my theoretical conclusions 
first, at the same time giving a short account of those facts bearing directly on 
the subject-matter. 

A few simple diagrams have been introduced to make the text more 
intelligible. These are throughout referred to as Figs. 1, 2, 3, &c. The 
reference to figures in the two plates is always indicated by the addition of 
the letters Pl. XXIII or Pl. XXIV. 

VOL. XXXI, PART III.—NEW SER, Y 


318 WILLIAM PATTEN. 


In failing to add materially to what the anatomy and 
embryology of Vertebrates themselves can demonstrate, the 
Annelid theory not only is sterile, but is likely to remain so; 
because unspecialised segments being characteristic of Annelids, 
it cannot hope to elucidate that profound specialisation of the 
Vertebrate head which it is the goal of Vertebrate morphology 
to expound. Moreover, since Vertebrate morphology itself 
reflects as an ancestral image only the dim outlines of a 
segmented animal—but still not less a Vertebrate than any 
now living,—it is clear that the problem must be solved, if at 
all, by the discovery of some form in which the specialisation 
of the Vertebrate head is already foreshadowed. 

Since of all Invertebrates, concentration and specialisation 
of head segments is greatest in the Arachnids, it is in these, on 
a priori grounds, that we should expect to find traces of the 
characteristic features of the Vertebrate head. Finding from 
time to time confirmation of this preconceived idea as the 
unexpected complexity of the Arachnid cephalothorax revealed 
itself, I now feel justified in formulating a theory that Verte- 
brates are derived from Arachnids. 

I have presented the facts as they appear to me, and have 
hazarded an interpretation of them; not, however, without 
a lively sense of the difficulties of the task, certainly not 
without the conviction that I may have fallen into errors which 
greater experience and a better knowledge of the intricacies 
of Vertebrate anatomy might have avoided. 

In the following preliminary sketch of the structure of 
Limulus, and especially of the Scorpion, I shall attempt to 
prove—(1) That in the Scorpion the cephalothoracie neuro- 
meres, nerves, sense-organs, and mesoblastic somites present, in 
a general way, not only the same specialisation and the same 
numerical arrangement in groups, but also the same difference 
as a whole from the body-segments, as do the corresponding 
parts in the Vertebrate head; (2) that the Arachnid cartila- 
ginous sternum represents the primordial cranium of Verte- 
brates ; (3) that in the Trilobites and Merostomata the internal 
structure of the cephalothorax resembles in some respects that 


ON THE ORIGIN OF VERTEBRATES FROM ARACHNIDS. 319 
of Scorpio and Limulus; (4) that the remarkable fish-like 
Pterichthys and related forms, judging from their external 
structure, are closely related to the Merostomata, and serve to 
connect Arthropods with Vertebrates ; and (5) that the em- 
bryology of Vertebrates in its main features can be reduced 
to the Arthropod type. 

First let me state certain conclusions that have been reached 
concerning segmentation in Arthropods. 

In Scolopendra each neuromere has four pairs of spinal 
nerves ; the first two pairs in each neuromere are larger and 
darker, and probably contain more sensory fibres than the two 
following pairs. 

Certain facts indicate that this condition is the ground plan 
of the nervous system in all Arthropods, and that the various 
modifications of it found in other Arthropods are produced by 
fusion of the nerves. The two sensory nerves tend to fuse 
with each other first ; afterwards the two motor nerves; and 
finally the double motor and the double sensory nerves unite, 
thus producing, in different groups of Arthropods, neuromeres 
with four, three, two, and one pair of nerves. 

In Scolopendra the neuromeres appear to be double ; and, if 
what we have indicated above is true, it follows that in all 
Arthropods the neuromeres, and consequently the segments 
themselves, are double. In support of this view we mention 
the following facts :—(1) In all Arthropods carefully studied 
two cross commissures have been found in each neuromere. 
(2) In Acilius the median furrow between these cross com- 
missures is similar to that between the successive neuromeres. 
(8) In Acilius, according to my observations, there are two 
pairs of tracheal invaginations in each segment: one pair, that 
which is always readily seen, is situated near the anterior 
edge of the segment ; the other, which is very rudimentary and 
difficult to distinguish, is situated in the same line as the first, 
but near the posterior edge of the segment. (4) In all the 
insect embryos I have examined, and in almost all figures where 
the tracheal openings were represented, the stigmata were 
situated near the anterior edge of the segment. (5) The 


320 WILLIAM PATTEN. 


frequent presence in Arthropods, especially Crustacea, of 
bifurcated appendages ; this condition is due, we may suppose, 
to the partial fusion of two originally distinct appendages. 
(6) The frequent occurrence of insect monsters having double 
pairs of legs. (7) According to Heathcote’s important obser- 
vations, the segments in Julus are certainly double, as shown 
by the duplication in each segment of the somites, cardiac ostia, 
arteries, neuromeres, trachez, and legs. (8) In Scorpio the 
neuromeres are distinctly double, each one being composed of 
a large anterior portion and a small posterior one. Large pit- 
like invaginations of the median furrow are found between 
the halves of the anterior portions, and faint indications of a 
second series of pits between the halves of the posterior por- 
tions (Pl. XXIV, fig. 3). But in Scorpio the most singular 
feature of all is that the parts of each abdominal neuro- 
mere finally separate, the posterior portions uniting with 
the anterior portion of the neuromere just behind it (PI. 
XXIV, figs. 3 and 4, and Fig. 11, p. 348). This process 
may be followed with ease and perfect certainty in surface 
views. 

All these facts point to the conclusion that the segments in 
all Arthropods are double, and are derived from those of 
diplopod-like ancestors. 

If Vertebrates are derived from Arthropods, they are also, 
in all probability, composed of double segments. It may be 
worth mentioning in this connection that in many fishes the 
spinal nerves, and especially the cervical ones, split up into 
two, three, and sometimes four pairs of nerves for each neuro- 
mere. 


I. Tue Grovupine oF THE CRANIAL NEUROMERES OF Scor- 
pio is a result of the varying union of the first thirteen 
neuromeres. 

From the cephalic lobes three neuromeres arise which fuse 
completely to form the fore-brain of the adult (figs. 1—4, 
Pls. XXIII and XXIV, and Fig. 1, F. B., p. 821). 

The first neuromere of the six thoracic segments pushes its 


321 


ON THE ORIGIN OF VERTEBRATES FROM ARACHNIDS. 


‘snseydosa 04 soposnm r0jowrjoxT “mw “4 “4UOULSaS-UIeAG 4SAy IO ‘oqo] avUNTIWAg “7 ‘s “aansstuU0D ureiq 


IOMO4SOg "09 'H *XRAOY} OY} JO SOAIOU [RUIVY IOWO\SOg ‘y-\~ "YM “SOAIOU snsea [eMoU 91} JO SJOOY 4-2 "'Y ‘*syuauSes 
OOVIOY} Of} JO SAATIU [RAMAN “g~(%" “SPAIU [LINAU O1[} JO Oseq 94} 4B VISURd JeNaNT *y-\ 4° ‘oasou Jeunoad 40 
(UO ‘2 “UW “OAIOU [BAYSO 10 [RRpOWIOS WeIpayYy “wv “7s mw ‘aatou oho URIPO| “w'a"m “YyNOT ‘mw ‘oAroU of [eIOyery “wa “7 
‘SAIWOINIU SNSVA PAY} puke PUODS ot} JO saarou yeuULBY posny ATfeEsed MO] OY, “42 *y pue *,a y ‘oroWMOANoW snSea qsiy ony 
JO Soarou OMejsod pus IoMojue yuspuadepur oy], *,2°y pue*a"y “quouISes [eIDOI[OYO VY} JO oAIOU [RMA O44 YITA pasny 
Ayeyaed ‘orowomou urerq-a10y party} oy} JO oAtou [ema oy} Surywasoador A[qvqoad “yoos ayeorfap AI9A YW "uy = “UN yeLys *3 
1o ‘umsopou worsuey “pow 4 ‘oursoyisny uorpsuen fF “ajayo Jo UTyS UI suBS10-asuas po19yyvos a4 04 soyoueiq Aueut yo 
SUIAIS ple ‘sueS.10-9sUas [exOd pUe [e|UaWSas of} WO, poattop uoT[Sued osiey *,6 °(¢) oyaqyeduids [elueig “4s ‘a *satouoinou 
NOBOY} OT} JO SOATOM [BUTBI] LOMUY “yy “Y “SOAMSSTULUMOD TOMY ‘09 -M —"UlRAQ-a10T “GT eS“ UTeAQ-PIW “YT “TL 


“UIeIg-PULET “ZH “Wreaq Aiossaody “gq Y—'suoroassip puv suoiyoos jo suvaut fq pajonaysuoo ‘uordtoog y[upe Jo uEIG—T “dI\T 


ws, 


S22 WILLIAM PATTEN. 


way in front of the mouth, forming a sharply defined region, 
that I shall call the mid-brain (Fig. 1, m. B.). 

The remaining five thoracic neuromeres are imperfectly 
fused; they constitute the hind-brain (#. B.). Finally, four 
very intimately fused abdominal neuromeres are added to the 
preceding ones, forming an accessory brain (A. B.). 

A very similar grouping is found in Vertebrates. (1) As 
shown by the segmental character of the optic, pineal, and 
olfactory nerves, the fore-brain probably contains at least three 
completely fused neuromeres. (2) The mid-brain, as is now 
generally recognised, contains but a single neuromere, which, 
judging from the character of its nerves and somite, probably 
belonged originally to what Gegenbaur calls the six primitive 
head-segments, and which, just as in Arachnids, has subse- 
quently become separated from them, forming an independent 
region. (3) The hind-brain is composed of five or six neuro- 
meres, which Gegenbaur, omitting the fore and accessory brain, 
regards as the primitive brain; the large size of these neuro- 
meres in Vertebrates, their incomplete fusion, and the distinct 
swellings at an early stage in this brain region, are facts to 
be expected on the Arachnid theory, for these features are also 
characteristic of the six thoracic neuromeres of Scorpio and 
Limulus. (4) According to Balfour and Van Wyhe, there is 
an accessory brain in Vertebrates composed of four body 
neuromeres, secondarily added to the head. 

There is difference of opinion as to the exact number of 
neuromeres in each brain region of Vertebrates; but as the 
matter now stands it would not violate these views more than 
they do one another to assume that the grouping of cranial 
neuromeres in Vertebrates is exactly the same as in Scor- 
pions. 

Il. Sprvau Nerves.—In embryo Scorpions each neuromere, 
except those of the fore-brain, has three pairs of nerves; one 
pair is mainly motor, another mainly sensory, and the third is 
probably sympathetic. 

In the abdominal region the nerves to each neuromere fuse 
to form the spinal nerves of the adult; but the distal and 


ON THE ORIGIN OF VERTEBRATES FROM ARACHNIDS. 323 


proximal ends of the primitive motor and sensory nerves 
remain unfused, forming for each spinal nerve two distinct 
branches and two roots. The abdominal sympathetic nerves 
are very small, and their relation to nerves of the adult has 
not been in all cases determined. 

The sensory root of the adult spinal nerve arises near 
the neural surface of the neuromere. Besides the ordinary 
fibres, it contains an axial bundle of coarse and deeply stain- 
able nerve-tubes, surrounding which is an elongated mass of 
small ganglion-cells (Fig. 2, sp. g.). 

The motor root arises near the hemal surface of the neu- 
romere, and is distinguished by its light colour and by the 
absence of the dark nerve-tubes and ganglion-cells. A short 
distance from the neuromere the motor and sensory roots 
unite to form a single nerve, which, on reaching the sides of 
the bedy, divides into two branches, one extending backwards, 
the other laterally (Fig. 4). 

The above features are not so clearly defined in the caudal 
segments. 

Thus the abdominal spinal nerves of Scorpio resemble the 
spinal nerves of Vertebrates—(1) In their origin from two or 
more originally separate nerves ; (2) in the failure of the distal 
and proximal ends of the nerves to unite; (3) in the motor 
and sensory roots arising respectively from the hemal and 
neural surfaces of the nerve-cord; (4) in the presence of two 
kinds of nerve-tubes in the sensory root; (5) in the presence 
of a collection of ganglion-cells in the sensory root, between 
the nerve-cord and the point where the two roots unite ; (6) in 
the origin, as will be shown later, of this ganglion from a spe- 
cialised part of a dark lateral border of the ventral cords, 
comparable with the neural crest of Vertebrates. 


III. Tue Tworacrc on Craniat Nerves of Scorpio remain 
separate throughout life; hence they differ from the abdominal 
nerves in the same way that it is supposed some of the Verte- 
brate cranial nerves differ from the spinal ones. 

Examined more closely, we find that of the three pairs of 


324. WILLIAM PATTEN. 


nerves to the first hind-brain neuromere, that supplying the 
chele is much the largest. We shall call it the neural or 
pedal nerve; it undoubtedly corresponds to the sensory 
roots of the abdominal segments, and agrees with them in 
being a mixed motor and sensory nerve, in containing two 
kinds of nerve-tubes, and in having at its base a ganglionic 
swelling that we shall call the neural ganglion (Figs. 1 and 
3, n.g.). The latter is serially homologous with the spinal 
ganglia, as shown by its development from the neural crest ; 
but it differs from them in being very much larger, in having 
the ganglion-cells arranged upon the surface of the nerve- 
root, and in being more intimately fused with the nerve-cord. 
In Limulus there are at the base of the pedal nerves similar 
swellings; they are here more clearly ganglia of pedal nerves, 
because they are more independent of the nerve-cord than in 
Scorpio (Fig. 10, g. n*.). 

In Scorpions about ready to hatch, a short distance beyond 
the neural ganglion is a purely sensory and richly ganglio- 
nated coxal nerve; it is distributed to a number of sense- 
organs on the median basal side of the legs; one of these organs 
is very much larger than the rest, and from it is split off a very 
large coxal ganglion (Fig. 8, cv.g. and cx.n.). Each of 
the sense-buds (s. 6.) also gives rise to one or more ganglion-cells, 
which pass into the nerve that supplies the bud. There is a 
similar set of coxal sense-organs in the spiny mandible-like 
swellings in the coxal joints of Limulus. 

The main nerve is continued beyond the coxal nerve into 
the chela. Near the base of the chele it expands into a 
ganglionic swelling, formed by an inward proliferation from a 
true segmental sense-organ ( S$. 8. 0.). 

The exact fate of the coxal and segmental ganglia I have 
not been able to determine. The large coxal sense-organ 
seems to disappear, but the ganglion produced by it wanders 
inward, forming a swelling on the coxal nerve. The segmental 
sense-organ also disappears, and its ganglion probably unites 
with the coxal ganglion, At any rate in the adult, I find a large 
lateral ganglion united by several branches not only with 


ON THE ORIGIN OF VERTEBRATES FROM ARACHNIDS. 3209 


the pedal and hzmal nerves of the chele, but also with the skin 
(Fig. 1, g.7.). I believe this ganglion is formed by the fusion 
of the segmental and coxal ganglia. 

All the pedal nerves of the thorax are built on the above 


Fic. 


2.—A. Section through the anterior portion of a neuromere of Scorpio 
in Stage z, fig. 3. B. Section, at same stage, through the posterior 
portion of a neuromere. C. Section of nerve-cord in anterior portion 
of abdomen of a Scorpion embryo, with pigment just appearing in 
the body-wall. D. Section through the third abdominal neuromere of 
embryo about ready to hatch. Z#. Section through first abdominal neuro- 
mere of adult. #. Section through the occipital ring of the endocranium 
or sternum of Scorpio. G. A segmental sense-organ of Stages F, G. 
bt. c. Botryoidal cord. c. cr, Cartilaginous endocranium. c. cent. = g. m.c. 
Imperfect canalis centralis formed by the ganglionic portion of the 
median furrow. 4g. m.c. Ganglionic portion of median furrow. 4g. s. 0. 
One of the large marginal sense-organs that give rise to the spinal ganglia. 
h, m. Heemal nerve. 7. m.c. Interganglionic portion of median furrow, 
or “anlage” of spinal artery. 7. 2/. Inner neurilemma. &. st. Wedge- 
shaped cord, a remnant of the median furrow, out of which a branch to 
the spinal artery is formed. x. cr. Neural crest. md. Medulla. 
m. m. Neural nerve. sp. a. Spinal artery. sp. gy. Spinal ganglion. 
sp. 2. Spinal nerve. sz. o. Sexual organs. 


3826 WILLIAM PATTEN. 


plan, the only difference being that in the other segments the 
coxal sense-organs and consequently the coxal nerve are 
smaller, and there seems to be no lateral ganglion in the 
adult. 

Two pairs of hemal nerves arise from each of the six 
thoracic neuromeres ; they are small and light coloured, and are 
probably entirely motor, supplying the innermost muscles, and 
probably some of the anterior viscera (Figs. 1 and 3, a. h. n. 
and p. h. n.). | 

Segmental Sense-organs and Ganglia.—One of the 
most important evidences of the Annelid origin of Vertebrates 
has been the similarity between the segmental sense-organs of 
fishes and Annelids. The value of this evidence has recently 
been destroyed, because it is now known, from the researches 
of Beard and Allis, that the lateral line-organs of fishes are 
formed by a backward growth of cranial sense-organs, and 
that their segmental arrangement is only secondarily acquired. 
Moreover, Beard’s researches show such an unsuspected com- 
plication of cranial ganglia, sense-organs, and nerves, that it is 
difficult, if not impossible, to compare them with similar 
parts in the body. To say the least, his observations do not 
strengthen the Annelid theory, because the latter cannot 
explain this extraordinary difference between the cranial and 
spinal nerves, its aim and only hope being the reduction of the 
ancestral Vertebrate to a collection of like, not unlike meta- 
meres. We certainly do not have this difficulty with the 
Arachnid theory, because the distribution and history of the 
thoracic sense-organs, ganglia, and nerves of Scorpio and 
Limulus, resemble in a striking way those of the corresponding 
parts of Vertebrates. For example, in Scorpio (1) the pedal 
nerve, its neural and lateral ganglia, and its purely sensory 
branch, or coxal nerve, the coxal and the segmental sense-organs, 
and the anterior and posterior hemal nerves,—all these features 
produce in each thoracic neuromere a complex condition 
similar to that found in a typical cranial neuromere of Verte- 
brates. (2) Omitting the fore and accessory brain, and using 
the cranial ganglia as guides, there are in the head of 


ON THE ORIGIN OF VERTEBRATES FROM ARACHNIDS. 827 


Vertebrates six sets of sense-organs, or exactly the same num- 
ber as in Scorpio and Limulus. In Vertebrates the ganglia, 
presumably derived from segmental sense-organs, are the ciliary, 
Gasserian, facial, auditory, and glossopharyngeal, and the first 
free vagus; or omitting the latter, and counting with some 
authors the facial ganglion as double, we would still arrive at 
the same conclusion. (3) In both Scorpio and Vertebrates these 
sense-organs give rise to “ lateral ganglia.” (4) In both cases 
neural ganglia are developed in the head, which are serially 
homologous with spinal ganglia. (5) In both cases the neural 
and the spinal ganglia develop from aspecial modification of 
the edge of the nerve-cords, the “ neural crest.” In Scorpio the 
crest consists of a row of large dark sense-organs extending the 
whole length of the nerve-cord (Pls. XXIII and XXIV, figs. 
1—3, nc. = sp. y.; and text, Fig. 2,a and p, ne. and sp. g.; also 
text, Fig. 11, sp.g.). (6) The manner in which the coxal nerve 
unites with a coxal sense-organ and receives ganglion-cells 
from it, and the way it becomes connected by small branches 


Fic. 3.—Semidiagrammatic section through the base of a leg and a 
thoracic neuromere of an embryo Scorpion.—a. 4. z. Anterior hemal 
nerve. cw. s. 0. Coxal sense-organ. 4. s. s. Ganglion-cells arising from 
segmental sense-organ. z.g. Large neural ganglion, serially homologous 
with spinal ganglia of abdomen. jp. %. z. Posterior hemal nerve. p. x. 
Pedal nerve. s. 4. Sensory buds (comp. fig. 4, Pl. XXIV, s. 0.). 8. 8. 0. 
Segmental sense-organs. 


328 WILLIAM PATTEN, 


with innumerable sense-buds scattered over the skin of the 
legs and ventral surface of the body (Fig. 3, s. ., p. 327, and 
Pl. XXIV, fig. 4) is comparable with the growth of the “ supra- 
branchial”? nerve of Vertebrates. (7) Moreover, although 
we have not determined with certainty the history of the 
ganglia arising from the segmental sense-organs of Scorpio, 
there is reason to suppose they represent the ganglia which, 
according to Van Wyhe, are connected with the ventral 
branch of cranial nerves. We should thus, in another way, 
arrive at and confirm the conclusion of Froriep, that the 
ventral root-ganglion is the most primitive; for in Scorpio 
and Limulus the segmental sense-organs and ganglia are 
undoubtedly more primitive than the coxal ones. Beard 
denies that there is any ganglion to the ventral root, so it is 
difficult to determine whether the coxal sense-organs or the seg- 
mental ones, or both, correspond to the supra-branchial sense- 
organs described by Beard; but for several reasons I am 
inclined to think they are the coxal sense-organs. 

If we accept Beard’s scheme of the cranial nerves, the enor- 
mous transitory sense-organs of Limulus would come in exactly 
the same place as the ear of Vertebrates—that is, reckoning 
three segments to the fore-brain, on the seventh cranial 
segment. It is also worth mentioning that the general 
appearance of the two organs at an early stage is very much 
alike. 

The lateral cord of ganglion-cells and nerve-fibres of Limulus 
may be compared with the “ ganglien-zellen-strang” described 
by Vejdovsky in the Oligochzeta, and may be regarded as having 
the same morphological value as the lateral cord of the central 
nervous system. It is not improbable that the longitudinal 
nerves of the Vertebrate head, such as that, for instance, 
uniting the seventh and fifth nerves, are remnants of a lateral 
nerve-cord like that in Limulus. 


IV. Tue Vacus Nerves of Scorpio, as I shall call them, 
or those arising from the accessory brain, are intermediate in 


ON THE ORIGIN OF VERTEBRATES FROM ARACHNIDS. 329 


character between the cranial and spinal nerves; at the same 
time they present remarkable features not found elsewhere. 

The neural nerves to the four vagus neuromeres fuse com- 
pletely to form the large pectinal nerve ; but the neural ganglia 
at the base of the nerves retain to a certain extent their inte- 
grity, forming what I have called the ganglion nodosum or 
ganglion laminatum (owing to the remarkable concentric 
laminz composing its medullary core), the ganglion fusi- 
forme, and the ganglion minus (Fig. 1). 

The hemal nerves to the first vagus neuromere form two 
distinct pairs, as in the typical cranial segments (A. v!. and 
h. v?.). In each of the succeeding neuromeres the hzmal 
nerves have united with each other, forming three nerves with 
double roots; the latter decrease in length from the first pair 
to the third, passing gradually into a condition like that in 
the abdominal hemal nerves. 

A short distance from the brain all five hemal nerves form 
a compact bundle extending backwards, some of the nerves 
passing through the hemal wall of the cartilaginous cranium 
or sternum, others passing out of the neural canal. Thesecond 
and third double nerves (v*. and v*., Fig. 4), some distance 
from the brain, fuse to form a single nerve supplying the first 
and second lung-books and the ventral surface of the body ; 
on its way to these organs it passes over the ventral surface of 
the liver, to which it possibly gives branches. The anterior 
heemal nerve of the first vagus neuromere (v!.) runs close to the 
coxal gland, and, dividing into numerous branches, is lost on 
the surface of a thick peritoneum-like membrane. The poste- 
rior nerve (v?.) extends along the arthrodeal membrane supply- 
ing numerous sense-organs in the skin of the sides and back 
of the abdomen. The fourth vagus (v‘.) supplies the skin 
and longitudinal muscles on the ventral surface of the 
abdomen. 

A small nerve arises from the ventral surface of the accessory 
brain, and supplies the distal portion of the sexual ducts 
(Figs. 1 and 4, 2). I could find no way of ascertaining 
to what neuromere this nerve belongs. 


330 WILLIAM PATTEN. 


Hence the term vagus is applicable to these nerves, for, 
owing, as we shall see, to the almost complete disappearance 
of their proper field of distribution, they have not only wan- 
dered into other segments, but to organs which they do not 
normally supply. 

Little is certainly known about the vagus nerves of Verte- 
brates, but at present I see no serious objection to supposing 
they are derived from the vagus of Scorpions. The most im- 
portant resemblance between these remarkable groups of 
nerves are the following :—(1) The vagus nerves in both Scor- 
pions and Vertebrates extend backward (although the neuro- 
meres to which they belong have been pushed forward), and 
supply muscles and internal organs to which the corre- 
sponding nerves of the other segments are not normally 
distributed. (2) This wandering of the nerves in both Verte- 
brates and Scorpio is probably due to the same cause, i. e. 
the great concentration of their neuromeres and the absence of 
their mesomeres; the result is that the nerves must also 
disappear or wander to other tissues. This point is an im- 
portant one, because these conditions are not found in any other 
animals besides Vertebrates and Arthropods. (8) In the Scor- 
pion the main vagus nerve is formed by the early and remark- 
ably complete fusion of four neural nerves of an accessory brain. 
In Vertebrates the vagus is formed in the same way; but 
there is nothing to show whether these fused nerves, either in 
Scorpio or Vertebrates, represent neural nerves or only their 
sensory branches, or both. (4) The hemal vagus nerves of 
Scorpio form a compact and isolated group of nerves evidently 
undergoing profound secondary changes; already they are 
partly fused with one another, and their roots have moved 
backward at the same time that the neural roots have moved 
forward. Since in Vertebrates there has probably been a similar 
movement in the vagus region (Gegenbaur), it is possible that 
in Petromyzon the four posterior vagus roots of the eight de- 
scribed by Ahlborn represent hemal vagus roots which have 
moved backward along the medulla oblongata only a little more 
markedly than the hemal vagus roots in Scorpio. <A_ part 


ON THE ORIGIN OF VERTEBRATES FROM ARACHNIDS. 3881 


or all of these hemal nerves in both Scorpio and Petromyzon 
supply segmental respiratory organs. (5) One large nerve in 
the vagus group of both Scorpions and Vertebrates is con- 
spicuous on account of its sensory nature and lateral position. 
(6) In Scorpio the neural vagus nerves supply a specially 
modified appendage called the comb. There is reason to sup- 
pose that originally the vagus of Vertebrates also supplied a 
specially modified appendage—the pectoral fin. The resem- 
blance between these appendages will be considered later. (7) 
In Scorpion the neural roots of the vagus and their ganglia 
decrease in size from before backwards ; this is remarkable, 
since we should naturally expect the third or comb root to be 
the largest. Van Wyhe has discovered a similar condition in 
the ‘“‘anlage”’ of the vagus of Selachians. (8) The vagus nerves 
of Vertebrates and Scorpions are derived from four neuro- 
meres not belonging originally to the brain, which are more 
intimately fused with one another than are those in front 
of or behind them, while the nerves themselves are the most 
complex and most modified nerves in the whole body. This 
condition is all the more extraordinary since, from their posi- 
tion, we should naturally expect these nerves and neuromeres 
to be intermediate in character between those of the head and 
trunk. Finally, (9) if we count three segments to the fore- 
brain, the vagus neuromeres in Scorpio and Vertebrates fall 
in exactly the same place in the series—that is, in the tenth to 
thirteenth segments inclusive. Nowhere else in the animal 
kingdom do we find four segments in the middle of the body 
with these extraordinary characters. 

There are, of course, important differences between the 
vagus of Scorpio and that of the lowest Vertebrates; but we 
are content to show here that there are very decided resem- 
blances between them, and that the differences are not greater 
than those found among the Vertebrates themselves. It is 
evident that a still further modification of the vagus of Scor- 
pio, in the direction along which it has already advanced so 
far, would lead up naturally to the most primitive condition of 
the vagus in Vertebrates. 


332 WILLIAM PATTEN. 


(4, i 


Fic. 4.—Brain and portion of nerve-cord of adult Scorpion, seen from the 
neural surface, constructed from sections and surface views.—/. B. Ac- 
cessory brain. H. B. Fore-brain. V’. and V". Heemal nerves of first 
vagus neuromere. V*, and V%. Hemal nerves of second and third vagus 
neuromeres. V‘, Hamal nerves to fourth vagus neuromere. 


ON THE ORIGIN OF VERTEBRATES FROM ARACHNIDS. 383 


V. Tue Forr-srain.—The structure of the fore-brain and 
its nerves of Scorpio and of Limulus cannot be fully under- 
stood without a knowledge of the development of the brain and 
optic ganglia of some forms like Acilius. Here the cephalic 
lobes consist, at a very early embryonic period, of three seg- 
ments, each segment bearing two pairs of eyes, a pair of optic 
ganglia, and a segment of the brain. Each optic ganglion 
arises from a separate invagination on the median edge of the 
optic plate, just opposite the pair of eyes to which it belongs! 
(Bigs9:, 4). 

After carrying the ganglion inwards the invagination closes, 
leaving the ocelli in their original position on the outer surface 
of the optic plate (Fig. 5, a). 

The cephalic lobes of Scorpio represent merely a modi- 
fication of the Acilius type. In both cases the relative posi- 
tions of the optic plate, optic ganglia, and segments of the brain 
are, at first, the same (Pls. XXIII, XXIV, figs. 1—4). Accord- 
ing to my interpretation, the ganglionic invaginations of the 
first segment unite with each other across the median ventral 
line to form a deep transverse furrow (Fig. 5, a), the thickened 
walls of which are probably derived from the rudiments of 
that part of the optic plate, optic ganglia, and brain belong- 
ing to the first segment. It is possible that this furrow 
contains an unsegmented portion, comparable with the preoral 
lobe of Annelids ; but this is a question which cannot be dis- 
cussed here. No eyes are developed in this segment. 

The second ganglionic invagination is at first like the 
corresponding one in Acilius, except that the optic plate is 
rudimentary and the eyes are not at first discernible. The in- 
vagination becomes so large that it involves the optic plate, 
which then forms the outer wall of a ganglionic sac (Fig. 
5, 6) ; the crescent-shaped openings to the invaginations, mean- 
time, move backward and inward until they unite with each 
other over the median line, forming a single sac with a poste- 

1 The optic ganglion to the convex eyes of Vespa, in which no larval 


ocelli are developed, arises in a similar manner on the median edge of the 
optic thickening. (See “Eyes of Vespa,” Patten.) 


VOL. XXXI, PART III.—NEW SER. Z 


334 WILLIAM PATTEN. 


rior median opening. Owing to its mode of formation the sac 
is at first very broad and bilobed, but it is rapidly reduced to 
the size of the future median eye. The lines a and 4 in 
Fig. 5, 8, on the right, show the successive positions assumed 
by the lateral limb of the ganglionic invagination. 

The outline of the cavity itself is only shown in this figure 
in the last two stages; in the earlier stage the cavity is 
shaded, in the later surrounded by a dotted line. At this 
period the outer wall of the sac is formed by the closely 
united and well-developed eyes (Fig. 6, c). It is evident 
that this sac is in no sense an optic vesicle, nor, strictly 
speaking, the cavity of a ganglionic invagination, although 
derived from one. I shall call it the optico-ganglionic 
vesicle. . 

Almost to the time of hatching it extends backwards a short 
distance as a rather thin-walled tube, opening outward by a 
narrow pore (Fig. 5, B, n. p., and Fig. 6, Fr). By this time 
the optico-ganglionic vesicle is secondarily shut off from the 
brain ; otherwise the pore would lead directly into the cerebral 
cavity. 

The carrying of the originally lateral eyes toward the 
median line also affects the optic ganglia to this segment, 
folding them over the brain as shown in Fig. 5, B and d. 

The ganglionic invagination of the third segment is in all 
respects like a typical ganglionic invagination in Acilius ; it is 
deep and well defined, but does not involve the optic plate ; 
hence the difference between the lateral and the median eyes 
of Scorpio. It is important to notice that while all the 
ganglionic invaginations of Scorpions are deeper and larger 
than those in Acilius, in both cases they decrease in size and 
depth from the first to the third. 

Although the entire anterior portion of the fore-brain of 
Scorpio is practically invaginated, so that the brain and its 
epithelium form the floor of a great complicated sac, the 
brain itself takes no active part in the invagination. It is 
enclosed solely by the extension of the lateral ganglionic in- 
vaginations, and the consequent inward and backward growth 


ON THE ORIGIN OF VERTEBRATES FROM ARACHNIDS, 335 


of the optic plate and ganglia over the brain. This fact is of 
great theoretic importance, especially when compared with 


Fic, 5.—A. Plan of cephalic lobes of Scorpion at an early period; the 
arrows indicate the directions along which the lateral lip of the ganglionic 
pits advance. B. Semi-diagrammatic view of fore-brain of Scorpion in 
Stage F. The left side shows the position of the deeper portions ; the right 
side, the superficial ones. The lines a, 4, c, and d represent successive 
positions of the rim to the ganglionic pits. The shaded portion in front 
of ¢ represents the area of the optico-ganglionic pockets about to unite ; 
the dotted line back of d, the sac after the union has taken place; the 
lateral eye-plate, /. e., represents the only part of the original cephalic lobes 
now on the surface. 4. Section through cephalic lobes of Acilius, or 
through third brain-segment of Scorpio; compare 4, a. B. Section 
through second segment of Scorpio, in direction 4, 4. C. Same at later 
period. D. Ditto, still later. «@, 4, and d. Direction of sections a, 4, and 
d. 6-3. Lobes of brain. Br. Brain. c.p. Capsule of eye. c¢. Cuticula. 
e. Eye. ec. Ectoderm, g. v'—*. Ganglionic pits. g. v°., a, d. Succes- 
sive positions of third pit. 7. e. Lateral eyes. m. Mouth. me. Median 
eyes. 2. /. Nerve-fibres. z. 7. e. Nerve to lateral eyes... Neuropore. 
op. g'—8. Optic ganglia. op. x. Optic nerve. o. pl'—%. Optic plates. 
r. Primitive retina. 7. Secondary permanent retina. 


what takes place in the ventral cord, where only a narrow 
median furrow, the only part lined by primitive ectoderm, is 
invaginated (see Fig. 9 and Fig. 2, c BE). 


330 WILLIAM PATTEN. 


The development of the cephalic lobes in Limulus 
represents a still greater modification of the Acilius type. 
There is a well-marked transverse furrow at the anterior end 
of the cephalic lobes, probably representing, as in Scorpio, 
the fused ganglionic pits of the first segment. Behind the 
furrow, on either edge of the cephalic lobe, is a small pore 


Fic. 6.—F. Median eye of Scorpion about ready to hatch, seen from 
above as a transparent object (compare Fig. 5, B). d'-Z£’ indicate 
planes of sections 4-2. cy. Corneagen. cp. Capsule. c?. Cuticula, 
on ends of primitive retinal cells. Dr. Diverticulum (?). ec. Ectoderm. 
np. Neuropore. op.x. Optic nerve. pg.c.and7’. Pigment-cells, derived 
from primitive retina. yp. 7. Primitive retina. 7. Ditto. 7. Secondary 
retina. 


leading into a long narrow tube. The latter may be regarded 
as extremely deep optico-ganglionic invaginations ; they corre- 
spond, in part at least, to the invaginations of the second 
segment of Scorpio. The position of the pits is the same in 
both cases ; besides, just as in Scorpions, the mouths of the pits 
move toward the median ventral line, forming an unpaired 
pore leading into the cavity of the optic tubes. The latter 
have, meantime, united to form an unpaired tube with a swollen 
blind end from which the median eyes are finally developed 
(Fig. 7, 8B). The walls of the tube give rise to the optic nerve 
to the median eye, and the basal portion of it to the optic 
ganglia (Fig. 10, g'. and g’.). 

There is a third ganglionic invagination, much lke the 
corresponding one in Scorpio, which gives rise to the optic 
ganglion of the lateral eyes (Fig. 7, g. v*.). This gang- 


ON THE ORIGIN OF VERTEBRATES FROM ARACHNIDS. 337 


lion, strictly speaking, does not belong to the lateral eyes, 
for they develop on the third thoracic segment, but to a 
small sense-organ lying on the outer edge of the invagination 
(Fig. 7, a, e°.); consequently this sense-organ and not the 
lateral eye, is homologous with a lateral eye of Scorpio. In 
the larvee it is connected by a nerve with the ganglion of the 
lateral eyes (Fig. 10, e3.); it is there deeply pigmented, and 
connected with a branching plexus of pigment-cells. In this 
stage it has been seen and described as a mere pigment-spot 
(Brooks). In the adult this simple eye stands close to its 
fellow in the median line, in front of the mouth ; the overlying 
cuticula is there clear and transparent, forming two rudi- 
mentary lenses. These facts establish beyond doubt the 
visual character of the organ and its serial homology with the 
other eyes. 

Passing backward into the thorax, we find that each line of 
the ganglionic pits just described is continued into a seg- 
mentally deepened furrow extending the whole length of the 
thorax (Fig. 9, 5, /. f., and Fig. 16). On the lateral side of 
this furrow there is a thickened band of ectoderm, which in 
each thoracic segment contains a broad shallow depression, 
undoubtedly of a sensory nature. The sense-organ of the 
second or third thoracic segment (I could not determine with 
certainty which) gives rise to the lateral eye, J. e.; that in the 
fourth is very large, and has erroneously been supposed to give 
rise to the lateral eyes (Kingsley), or to represent a dorsal 
organ (Watase). It is a true sense-organ, and is connected by 
a nerve-bundle with the cord of ganglion-cells arising from 
the lateral furrow ; it persists until the end of the first larval 
moult, and in this period is covered by a great disc-like thick- 
ening of the cuticula, which, judging from its shape and 
transparency, undoubtedly represents a rudimentary lens, 

The nerve to the lateral eyes arises, like the lateral cord of 
ganglion-cells, from or near the lateral furrow, and may be 
regarded as a specialisation of that part of the cord extending 
from the third segment of the brain to the second or third 
segment of the thorax. It is accordingly unlike any other 


338 WILLIAM PATTEN. 


nerve of the body, resembling rather one of the lateral cords 
of the central nervous system; or it may be compared with 
the ‘‘Ganglien Zellenstrang” of Vejdovsky. 


Ca op st 


Fic. 7.—A. Diagram of fore-brain of Limulus embryo, constructed 
merely to show probable condition from which the parts in the older 
embryos are derived, and the relation of these parts to similar ones in 
fore-brain of Scorpion. 2. Diagrammatic view of fore-brain at later 
stage, showing the union of ganglionic pits and eye-tubes. 1-4. Dia- 
grammatic sections, showing origin of eye and unfolding of brain and optic 
ganglion. 5. Section through the third eye-plate and ganglionic inva- 
gination. .B, 5 indicates direction of section. 1%. Lobes of fore-brain. 
br. Brain. e!—*. Eyes of the three brain-segments. g. v'—*. Ganglionic 
pits. m. b. Mid-brain, or fourth neuromere. m. e. Median eye of first 
segment. m.¢'. Medianeye of second segment. 2. . e. Nerve to lateral 
eye. zp. Neuropore. @.g. Ganglion to stomodeal nerves. op. g'—*, 
Optic ganglia. op. g.+ Optic ganglion to lateral eye and to e*. 


The segmental thoracic sense-organs and the lateral furrow 
of Limulus, as shown by their position and mutual relations, 
are, respectively, serially homologous with the cephalic sense- 


ON THE ORIGIN OF VERTEBRATES FROM ARACHNIDS. 389 


organs or eyes and their ganglionic invagination; and the 
lateral cord of ganglion-cells may be regarded as a cord of 
segmental ganglia serially homologous with the optic ganglia. 
Each segmental sense-organ of Limulus is represented in 
Scorpio by a pair of sense-organs at the base of each thoracic 
appendage (Pls. XXIII and XXIV). 
If now we compare Acilius, Scorpio, Limulus, and Verte- 


larva, seen from the side (compare Woodcut 10). 4-F. Section of the 
eye and its stalk at points 4-F. G. Surface view of median eyes. ant. 
Anterior side of eye. e!, Eye of first brain-segment. ¢?. Eye of second 
brain-segment. ec. Ectoderm. '. Nerve to eyes of first segment. 
n*, Nerve to eyes of second segment. zp. Neuropore. o. g'—*. Optic 
ganglia. 


brates, we shall find they present four distinct stages of one 
process of enclosing the fore-brain—that is, four stages in the 
extension of the ganglionic invaginations (see Fig. 9). 

In Acilius we have three distinct pairs of invaginations on 
the outer edge of the cephalic lobes. In Scorpio and 
Limulus the invaginations of the first two segments unite, 
forming a continuous, amnion-like fold, the free edge of which 
grows medianly and backward, infolding not only the optic 
ganglion, but the eyes and the anterior part of the brain, 


340 WILLIAM PATTEN. 


The third ganglionic invagination of Scorpio and Limulus is 
not yet extensive enough to involve the adjacent eyes, but it is 
much deeper than in Acilius; and in Limulus it shows a tendency 
to repeat the condition found in the preceding segment, for 
the sense-organ on its lateral edge is finally carried toward 
the median ventral line, and just falls short of being invagi- 
nated with the ganglion. Hence it is evident that the next 
step in the extension of the ganglionic invagination would 
result in the union of the three pairs of invaginations, and the 
formation of a continuous fold which would finally grow over 
and enclose the entire fore-brain, which, with the median and 
lateral eyes and the optic ganglia, would form the walls of a 
single sac. The result would be a condition practically lke 
that in Vertebrates. If, as in Scorpions, the appearance of 
the eyes was deferred until the infolding of the brain was 
completed, and if no secondary optic tubes were formed, as 
would naturally be the case if the eyes were degenerate, then 
the eyes would appear as thickenings of the outer wall of the 
brain-sac, as in Ascidians and Amphioxus. If there was a 
tendency to develop the eyes in long tubes as in Limulus, and 
the formation of the tubes postponed until after the brain was 
enclosed, the eye would appear at the end of a long tube pro- 
duced apparently by an evagination of the outer wall of the 
brain-sac (pineal eye and lateral eye of most Vertebrates). 
On our hypothesis it is not surprising such variations should 
occur, for similar ones are found in Arthropods. For ex- 
ample, in Scorpions and Limulus the involved eyes do not 
appear, as such, until long after they have been carried 
inward, although they ought to appear much earlier if onto- 
geny gave a complete picture of phylogeny. 

A careful examination of the diagrammatic figures will, I 
think, make clear what has preceded, and will, no doubt, sug- 
gest a number of other interesting comparisons which we have 
not space to consider here. 

The Median Eyes of Arachnids and the Pineal Eye 
of Vertebrates.—If we have progressed so far on solid 
ground, it is evident that the pineal eye of Vertebrates must 


ON THE ORIGIN OF VERTEBRATES FROM ARACHNIDS, 341 


be derived from the median eye of Scorpio and Limulus. As 
we shall now show, the structure of the median eye of Limulus 
supports this conclusion. In Limulus the optic tubes of the 
median eye unite in such a way as to form a J-tube, with the 
median pore at the junction of the arms and upright (compare 
Figs. 7 and 10). This condition is easily understood if we 
imagine that as each tube was bent toward the median line its 
mouth was gradually carried toward the distal end of the tube. 
The median pore represents not only the last trace of the coa- 
lesced ganglionic invaginations, but also the last point 
where the enclosed brain is attached to the surface 
ectoderm or communicates with the exterior. Nos. 
1—4, Fig. 7, represent, in a diagrammatic way, the forma- 
tion of the optic tube of the median eyes; at first they are 
apparently not connected with the brain at all, but later the 
invagination involves both the brain and optic ganglion. It is 
important to notice what a great variety of conditions might 
arise by slight variations of this process. The base of the 
L-shaped optic tube soon splits, in a way not thoroughly 
understood, into three nerves: (1) a delicate anterior pair 
(Fig. 7, B, and Fig. 10, n’), which arise from the very anterior 
lobe of the brain and unite with the eye-stalk some distance 
in front of the neuropore ; after joining the stalk they may be 
followed along its sides to the eyes (Fig. 8, C., n’.); (2) a 
posterior impaired nerve (Fig. 10, n%.), which for a long 
time is tubular at its base. Here it divides into two smaller 
arms (g.), which join the sides of the second brain-segment. 
Towards the eyes the tube is gradually converted into a small 
bundle of pigmented nerve-fibres (Fig. 8,*.). The structure 
of the median eye and its stalk is well shown by the series 
of cross sections in Fig. 8. 

The eye-stalk extends through the anterior part of the body 
to the dorsal surface, and then expands into a bulb-like thick- 
ening. The outer wall of the bulb develops two groups of 
cells filled with black pigment; they give rise to the median 
eyes proper, and are supplied by the median pigmented nerve, 
n*. The inner wall of the bulb is filled with white pigment 


342 


WILLIAM PATTEN. 


granules, and forms an unpaired mass of cells below and a 
little in front of the eyes proper; it is supplied by the paired 


Fic. 


¥ 
9.—Diagrammatic views of Arthropod and Vertebrate embryos, to 
illustrate the stages in the infolding of the brain, optic ganglia, and eyes ; 
the heavy black lines indicate the ganglionic pits, and the shaded por- 
tions the parts that are infolded.—A. Acilius. , G, and D. Three 
stages in the development of Scorpio. 4. Limulus. /. Hypothetical 
transitional form. G. Vertebrate. aw. Ear. ¢%. Hye on third segment 
of fore-brain of Limulus. yg. »'—*. Ganglionic invaginations. /,e. Lateral 
eye. J.f. Lateral furrow. m.e. Medianeye. xp. Neuropore. 0. pi'—, 
Optic plate. s,s. o!—°. Segmental sense-organs of thorax. x ands, s. of. 
Very large ear-like sense-organ. 


colourless nerves, 2!. Toward the late larval stages the white 
mass becomes constricted off as a solid diverticulum of the 
primitive bulb, and finally lies some distance from the surface 
as a cylindrical mass of cells filled with white pigment; in all 
stages the walls of the cells composing this body develop 


ON THE ORIGIN OF VERTEBRATES FROM ARACHNIDS. 343 


refractive rod-like thickenings, which resemble those in the 
superficial median eyes. 
The extraordinary condition described above can be explained 


ios . 


Nit 


TH 
Ny 


y nte 


Fic. 10.—Fore- and mid-brain of young larva of Limulus (just hatched), 
seen from neural surface; constructed from sections and dissections.— 
c. #. Posterior commissure of brain. e. Hye to third brain-seg- 
ment. g'—*, Optic ganglia. yg. a‘. Spinal ganglion on fourth neural 
nerve. g.«@.z. Ganglion to stomodeal nerves. 4. x*—-4. Heemal nerves 
to second, third, and fourth neuromeres. /'-*, Three ganglionic lobes 
to ganglion of lateral eyes. V’—*. First four neuromeres. x. e. Nerve 
to rudimentary eye of third segment. wz. /. e. Nerve to lateral eyes. 
n. nm, Neural nerve to fourth neuromere. zp. and g. v?. Neuropore. 
m. @. Stomodeal nerves. @. Gsophagus. z. Small nerve, extending 
backwards from posterior brain commissure. 


by supposing that two distinct pairs of segmental eyes are fused 
in the primitive eye-bulb; one pair belonging to the first, the 
other to the second brain-segment. Compare the diagrammatic 
figs. A and B, Fig. 7. In confirmation of this view it may be 
added—(1) at an early stage there are traces of two pairs of 


344 WILLIAM PATTEN, 


ganglionic pits, one of which was small and disappeared 
quickly, so that its history could not be determined. (2) 
Although there are only three distinct nerves in the eye-stalk, 
they represent four completely fused nerves, for the colourless 
nerves are still clearly enough paired, while the median one 
was undoubtedly paired originally, as shown by the diverging 
arms terminating in paired ganglia. (3) The anterior pair of 
nerves arise from a distinct crescent-shaped brain-lobe (with 
small dark nuclei), which is undoubtedly homologous with a 
similar lobe in Scorpio, and which there represents the first 
brain-segment ; the posterior pair arise from another lobe, 
which probably represents the second brain-segment. Thus 
the diverticulum of the median eye-bulb represents, in all 
probability, a pair of eyes belonging to the first brain-segment. 

We will merely note in passing that in some Trilobites on 
the anterior portion of the glabellum are three eye-like spots, 
which may possibly represent the three fused ocelli of Limulus. 

In comparing the median eye of Limulus with the pineal 
eye of Vertebrates the following points are important—(1) 
In both cases the eye is situated at the end of a long median 
tube. (2) In both cases the tube is apparently an evagination 
of the roof of the brain. (3) In both cases the position of the 
tube relative to the rest of the fore-brain is the same. (4) In 
both cases the manner in which the distal end of the tube 
grows forward away from its point of attachment is the same ; 
the exception to this manner of growth in some Vertebrates is 
probably due to the subsequent enlargement of the cerebrum. 
(5) In Limulus the bulb-like swelling at the distal end of 
the tube, which gives rise to a pair of median eyes lying 
close beneath the ectoderm, may be compared with the ter- 
minal sac of the pineal eye-stalk. (6) The remarkable diver- 
ticulum of the under wall of this bulb, giving rise to a peculiar 
whitish body not connected with the surface ectoderm, but 
which contains undoubted retinal cells, may be compared 
with a very similar diverticulum in Vertebrates. (7) The 
proximal end of the eye-stalk in both Limulus and Hateria 
(Spencer) contains three distinct nerves, two anterior paired 


ON THE ORIGIN OF VERTEBRATES FROM ARACHNIDS. 345 


ones and a totally different unpaired one. (8) In both 
cases the unpaired nerve may at certain stages be a mere 
hollow tube, the wall of which is composed of ordinary 
columnar cells, showing no trace of nerve-structure. (9) 
In both cases the proximal end of the eye-tube may re- 
present either the last point where the brain is connected 
with the ectoderm (Leydig), or the lips of a round, median 
opening or neuropore, leading into the cavity of the pineal 
eye or into the brain, or into both, according to the stage 
and method of development. (10) The manner in which, in 
Vertebrates, the fore-brain is enclosed and the median eye- 
tube formed represents a modification of a fundamentally 
similar process found in Limulus, and the Arachnids generally ; 
and this fact is of the utmost importance on account of the 
peculiar and complicated nature of the process, and its total 
absence in other groups of animals. 

The lateral eyes of Scorpion and Limulus are not in- 
volved by ganglionic infoldings so as to lie at the ends of brain- 
tubes, but, as shown above, the next step in the changes 
already accomplished would probably lead to that condition. 
Suppose a kidney-shaped eye, such as is usually present in 
and best adapted to forms like Limulus and other Merosto- 
mata, came to lie at the end of an optic tube; then a kidney- 
shaped retina would be produced with its concave edge 
directed hzmally, instead of in the opposite direction, as in 
Limulus, &c. As a kidney-shaped retina would no longer be 
required, it would naturally become circular; and, owing to 
the peculiar distribution of nerve-fibres, this would be most 
easily and economically accomplished by bringing the halves 
of the concave edge together, thus producing a choroid fissure, 
the direction and position of which would be like that in 
Vertebrates. 

Now the ommateum of Limulus consists of circles of from 
fifteen to twenty retinal cells surrounding a single central 
one of a little different character (Watasi). If such an omma- 
teum were converted into a true retina the arrangement 
of the cells would probably be retained, and we would have a 


346 WILLIAM PATTEN. 


retina like that in many Vertebrates. Again, the lateral eye of 
Limulus arises from the third or fourth thoracic segment, 
although its nerve and optic ganglia are united with the 
third brain-segment; if such an eye were involved by ganglionic ~ 
invagination, it would lie at the end of a long backwardly 
directed tube, like that of the lateral eye of Vertebrates. 

In other words, we can explain the most remarkable and 
characteristic features of the lateral eyes of Vertebrates, 
such as (1) the shape of the retina, (2) its histological 
structure, (3) the formation of the choroid fissure, (4) the 
backwardly directed eye-stalk (in marked contrast with the 
forwardly directed pineal eye), by supposing them to be derived 
from kidney-shaped thoracic eyes like those in Limulus, 
Trilobites, and Merostomata. 

The proximal ends, at least, of the optic nerves of Scorpio 
and Limulus, are probably serially homologous with the pedal or 
neural nerves of the thorax; this is shown by their histological 
structure and by their position. For example, in Limulus 
both the second and third segments of the fore-brain (and 
possibly the third in Scorpio) are provided with segmentally 
arranged motor nerves exactly like the motor nerves of the 
post-oral segment; the optic nerves taking the place of, and 
resembling in structure, the pedal or neural ones of the 
thorax (Fig. 10). 

In Limulus, one of the motor-like nerves of the fore-brain 
is connected with the skin about the upper lip; this fact, 
together with its position, suggests that it may represent an 
incipient olfactory nerve (Fig. 10, hn’.). 


VI. Tae Craniau Frexure of Vertebrates has been ex- 
plained by supposing that the fore-brain represents an inverte- 
brate supra-cesophageal ganglion. But, judging from Kleinen- 
berg’s view and Beard’s speculations, this theory must be 
abandoned, since the Annelid brain arises independently of the 
remaining central nervous system, while the Vertebrate fore- 
brain does not. It is therefore tacitly admitted that the 
Annelid theory cannot explain the cranial flexure, that 


ON THE ORIGIN OF VERTEBRATES FROM ARACHNIDS. 347 


extraordinary feature of the Vertebrate brain which has 
always been justly regarded as the strongest evidence in favour 
of the origin of Vertebrates from segmented animals. But if 
we still attach the same significance to the cranial flexure that 
many of the ablest zoologists have done in the past, then we 
exclude the Annelids from any direct genetic connection 
with the Vertebrates; at the same time we strengthen the 
Arachnid theory, because the Arachnid brain, not only in its 
flexure, but also in its primitive continuity with the nerve- 
cord, fulfils completely the condition demanded by a rational 
theory of the origin of Vertebrates. 

There is another neural flexure of equal importance with 
that mentioned above; I shall call it the somatic flexure. 
At certain periods it is present in nearly all Arthropods as a 
ventral flexure of the tail, producing in such forms as Scor- 
pio and others an S-shaped embryo. Now, in many mam- 
malian embryos there is a strong flexure just behind the 
brain, the ‘‘ cervical flexure’? of His, and in many fish and 
amphibian embryos there is at one period a strong upward 
curve of the tail. In fact, the Vertebrate embryo is also some- 
what S-shaped, a condition to be expected on the Arthropod 
theory, but otherwise inexplicable. 

Beard attempts to explain how the distal portion of the 
hypophysis cerebri may represent an Annelid cesophagus, but 
he only succeeds in showing how the difficulties are thickening 
around the Annelid theory. He does not recognise that if 
there is no Annelid brain in Vertebrates there 
should be no cranial flexure, and no reason what- 
ever for regarding the present Vertebrate mouth 
as secondary, or the hypophysis or any part of it 
as the remnant of a former esophagus. Moreover, 
his view necessitates the assumption that the fore-brain repre- 
sents a single segment, whereas its complex nature can hardly 
be doubted. 

On the other hand, (1) the strong cranial flexure in Arach- 
nids, (2) the primitive continuity of the fore-brain and ventral 
cord, (3) the absence of the mesoblastic somites in the fore- 


348 WILLIAM PATTEN. 


brain region, (4) its remarkable complex structure, as well as 
the development of its nerves and sense-organs,—all invite a 
detailed comparison of the csophagus and fore-brain of 
Arachnids with the fore-brain and hypophysis of Vertebrates. 


rm ody 
ae ODER ND NO 


+ aes 
OTT lela | 


3B 

Fic. 11.—The fourth vagus and first abdominal neuromeres of Scorpion, 
seen from the neural surface.—A. Stage Fr. 2B. Stages Gc, H. By 
comparing with figs. 1-4 in the Plates we see how the parts of the 
primitive double neuromeres have recombined to form new ones. a. n'—*. 
First and second abdominal neuromeres. 4g. m. c. Ganglionic portion of 
median furrow. sp.g. Spinal ganglia. sp.z. Spinal nerve. v.24, Fourth 
vagus neuromere. JZ. Tail. 


The arguments in favour of supposing that the infundibulum 
cerebri represents an Arachnid cesophagus with its nerves are 
much stronger, I believe, than those advanced by Beard in 
favour of his view. (1) Since in most Arachnids the ceso- 
phagus is very small, owing to their blood-sucking habits, and 
the canal for its passage through the brain is extremely small, 
the Arachnid theory can give some explanation of how and 
why the old mouth has disappeared. (2) In Limulus and 
Scorpio two large parallel nerves extend along the whole 
length of the sides of the cesophagus. Hach nerve arises from 
a thickening, which persists for a long time, of the distal 
end of the esophagus. As the latter increases in length, 


ON THE ORIGIN OF VERTEBRATES FROM ARACHNIDS. 349 


the thickening is carried inwards; when the process is once 
started we get between the inner and outer ends of the ceso- 
phagus all stages in the separation of the nerve from the ecto- 
derm. The proximal ends of the nerves are united in the 
middle of the fourth neuromere, and then terminate in a pair of 
large ganglia. The latter arise not from the esophageal walls, 
but from an invagination of the median surface ectoderm, 
and so close to the neuromere that there is no distinct boun- 
dary between them (Figs. 7 and 10, g. w., and Fig. 2, sp. n.). 
I will also add that the unpaired frontal ganglion of insects 
arises as an evagination of the cesophageal wall. Thus we 
would naturally expect that if the infundibulum represents an 
Arachnidan cesophagus, it would contain a large portion of 
nervous tissue, and be directly continuous, through the rudi- 
ments of the stomodeal ganglia, with the brain. (3) The 
dorsal organ of some Arthropods is derived from the em- 
bryonic membranes, and is finally invaginated into the yolk 
and absorbed. In some Crustacea a dorsal organ is retained 
throughout life, and is often used as a sucking or adhesive disc 
of attachment. If an animal attached itself by such an organ 
to the soft tissues of another, aud if the inner portion of the 
disc were absorbed by the yolk, a way would be opened for 
the alimentary canal to communicate with the exterior by a 
new sucking mouth, which would lie in about the same place 
as the embryonic mouth of Vertebrates. In Dytiscus larve, 
where the points of the sickle-like mandibles are perforated 
and serve as mouths, the old one becomes temporarily closed 
and functionless. 

In Scorpio I have found nothing like a dorsal organ, but 
in Limulus embryos there is a great mass of loosely connected 
cells near the anterior dorsal surface of the yolk, or about 
where the dorsal organ in other forms is present. In most of 
my sections this portion was cut out, so that I know little 
about it. It is not impossible that it represents a rudimentary 
dorsal organ, all the stages in the formation of embryonic 
membranes, &c., being omitted. 

VOL. XXXI, PART 11].—NEW SER. AA 


350 WILLIAM PATTEN. 


VII. Tue Mepian Fvurrow (or Mittelstrang of Hatschek) 
AND THE NorocHorp.—The ventral nervous system of Arthro- 
pods consists of three longitudinal cords (fivein Limulus), 
two lateral and one median. The latter, in its indifferentiated 
state, is represented by the so-called “ sympathetic” or “ median 
nerve.” 

In Acilius, between the ganglionic swellings of the lateral 
cords the median furrow is almost tubular and somewhat 
swollen, forming what may be regarded as the ganglia of 
the median nerve. In the abdomen the interganglionic 
portions of the median furrow give rise to the median nerve 
proper, which is therefore merely a longitudinal connective, 
comparable with those of the lateral cords. 

In the thoracic region the interganglionic parts of the 
furrow produce great ectodermic thickenings (the furce), to 
which muscles are attached. 

In Scorpio, the median furrow, in the centre of each neuro- 
mere forms a deep pit, the thick walls of which, epithelium 
and all, are converted into ganglion-cells (Pls. XXIII and 
XXIV, and Fig. 2, 4, c, and ©). By the crowding together of 
the thoracic neuromeres the pits in some places unite, forming 
a temporary central canal. 

Between successive abdominal neuromeres the median 
furrow proliferates inwards, producing solid spindle-shaped 
clusters of cells, which grow forward and backward until they 
meet, forming beneath the nervous system a continuous 
longitudinal cord (Fig. 2, G, 7. m.c.). Owing to its method of 
formation, the latter is for some time segmentally swollen. In 
embryos about to hatch, the cord is hollowed out, and forms 
the “spinal artery.” The whole cord is sometimes filled 
with large vesicular cells with small nuclei (Fig. 2, c), and in 
half-grown specimens the wall of the artery may be extremely 
thick and hyaline (p). Thus the whole organ presents a strik- 
ing resemblance to a Vertebrate notochord. 

At the anterior end of the embryo, beneath the vagus neuro- 
meres, the interganglionic part of the median furrow forms a 
great solid ball of tissue, composed of a confused mixture of 


ON THE ORIGIN OF VERTEBRATES FROM ARACHNIDS. 351 


coiled muscle-fibres and gland-like cells. As this remarkable 
body represents an isolated and specialised segment of the 
spinal artery, I have called it the merochord. In the adult 
it is arather compact ball of tissue lying between the brain 
and the posterior portion of the cartilaginous sternum, or 
endocranium (Fig. 12, mc.). 

In Lepidoptera (Cecropia) the “lemmatochord” is 
derived in part from the neurolemma of the persistent median 
nerve, and in part from the neurolemma of the lateral cords; 
but in some parts of the thorax of Cecropia the lemmatochord 
is entirely derived from the neurolemma of the median nerve, 
the nerve itself having disappeared. The median nerve in 
some cases runs in the centre of a spinal artery (S. Selvatico, 
‘Zool. Anz.,’ Aug., 1887, p. 562). 

These facts show that the median furrow is a much more 
important organ, morphologically, than has been supposed. 
It is certain that the interganglionic portions of the furrow, or 
at any rate something that cannot be distinguished morpho- 
logically from them, may give rise to an extraordinary variety 
of structures—to the furce (in the thorax of Acilius), the 
median nerve (in the abdomen), the lemmatochord, or at 
least a portion of it, in Lepidoptera, and to the spinal artery 
of Scorpio. Practically it makes little difference whether we 
regard the spinal artery of Scorpio as derived from the 
median nerve itself, its sheath, or a lemmatochord-lke organ 
with traces of the median nervein the centre. The important 
fact remains that, in a great many Arthropods there is a 
median cord, which in position and general character bears 
such an extraordinary resemblance to the notochord of Verte- 
brates, that the burden of proof lies with those who deny that 
the two cords are of the same nature. If it is urged that in 
Arthropods the median cord arises from the ectoderm while 
the notochord arises from the endoderm, we may safely answer 
that there is nothing in the embryology of Vertebrates to show 
to what germ layer the notochord belongs. It is never con- 
tinuous with functional endoderm; there is no evidence that 
it ever exercised, itself, any alimentary functions; it is never 


302 WILLIAM PATTEN. 


connected in any way with an alimentary canal. Only a 
strong faith in enteric diverticula, and in the red, white, and 
blue gastrules of embryological treatises, can lead one to 
believe in the endodermic origin of the notochord. On the 
other hand, its growth at both ends from superficial cells, and 
the manner in which it is frequently wedged in between the 
nerve-cords, indicate its ectodermic origin. 

Owing to the origin of the spinal artery of Scorpio from 
the interganglionic portions of the median furrow, temporary 
communications are formed between the central canal and the 
artery. The communications between the notochord and the 
neural canal of some Vertebrates (lizard, duck, &c.), may be 
of a similar nature. 

The indefinite anterior termination of the notochord beneath 
the hind-brain, just behind the pituitary body, andits segmental 
swellings in this region, may be compared to the gradual dis- 
appearance of the spinal artery in the same region just behind 
the merochord, and to the segmental swellings of the artery 
between the successive neuromeres. 

The “ godets” of Moreau found in Amphioxus may be 
remnants of arterial branches; and the origin of the chorda 
cells from the sheath may be compared to the origin of blood- 
corpuscles from the wall of the spinal artery. 

The fact that the notochord does not at first lie for its 
whole length in the ectoderm may be regarded as a secondary 
condition, all its very early phylogenetic stages being passed 
over hurriedly in the primitive streak, or growing-point. 


VIII. Tur Borryorst Corp.—Beneath the spinal artery of 
Scorpio lies a remarkable rod-like body which I shall call the 
botryoidalcord. It develops as a forward growth from a 
great primitive-streak-like group of cells, in much the same 
way the notochord of Vertebrates does. The primitive streak 
itself is at first situated at the posterior end of the body, but 
when the tail fold arises it lies just at the junction of the tail 
with the posterior end of the abdomen; the primitive streak 
then seems to produce tissue in both directions. At first it is 


ON THE ORIGIN OF VERTEBRATES FROM ARACHNIDS. 358 


short and massive, but it finally extends forwards (partly, per- 
haps, by an actual forward growth, but mainly owing to its 
being left behind by the backward growth of the posterior end 
of the body) the whole length of the abdomen as a rather large 
cylindrical cord ; the latter, about the time of hatching, splits 
into two parts, an outer one composed of a thin layer of small 
dark nuclei, representing the “anlage” of the botryoidal 
cord, and an inner one composed of large polygonal cells, repre- 
senting the anlage of the sexual organs (Fig. 2, c). 

The anlage of the botryoidal cord soon develops, at irregu- 
lar intervals, spindle-shaped enlargements, each intermediate 
portion heing reduced to a delicate hyaline fibre. The spindles, 
eight to ten in number, become attached to the wall of the 
spinal artery, and in most cases an imperfect communication 
is established between the cavity of the artery and the interior 
of the spindle (Fig. 2, £, also d. dt. c.). 

In the adult the spindles are composed of botryoidal masses 
of fibroid tissue densely packed with small, deeply stained 
nuclei. The organ is, perhaps, a gland for the production of 
blood-corpuscles. 

If the spinal artery of Scorpio represents a notochord, then 
the botryoidal cord probably represents the “ subchordal rod.” 


IX. Tue Cartitacinous STERNUM AND THE PRIMORDIAL 
Cranium.—A characteristic organ of the Arachnida is the 
cartilaginous sternum or “endocranium.” In Scorpio 
(and in Limulus it is about the same) it is a broad, lyre- 
shaped bit of fibroid cartilage, with forwardly directed arms, 
lying beneath the hind-brain (Fig. 12). Its posterior portion 
is more massive, and completely surrounds the posterior 
portion of the brain, forming a kind of occipital cartilaginous 
ring (Fig. 2, Fr). It develops from the mesoderm as a mem- 
branous diaphragm underlying the thoracic neuromeres. In 
size, shape, position, structure, and manner of development— 
in fact, in every particular except its chemical composition, the 
endocranium of Scorpio corresponds with the primordial 
cranium of Vertebrates. 


304 


WILLIAM PATTEN. 


There are in primitive fishes no head-muscles sufficiently 
powerful to account for the development of the primordial 


Fic. 


mbar . At 
‘ 
\ 
i 


12.—The endocranium or cartilaginous sternum of Scorpion, seen 
from the hemal surface. The figure has been constructed entirely by 
carefully plotting a complete series of transverse sections, only the 
purely cartilaginous parts being drawn.—d. p/. Basal plate, forming a 
thick floor to the central canal. m.c. Merochord, an hypophysis-like 
segment of the spinal artery. 7. c/. Neural canal or passage through 
the occipital ring for the nerve-cords (compare section of this region 
in Fig. 2, F). dé. Trabecula. z. Thickened portion of trabecule 
x. Thick bar of cartilage extending forwards from the neural surface 
of the occipital ring, and forming an imperfect roof to the endocranium. 
y. Membrane for support of merochord. 


cranium. As the cranial segments were very early immove- 
ably united, the cranium could not have served for the attach- 
ment of muscles moving the segments on each other for 
purposes of locomotion as the vertebral column does in the 
body. In fact, we can best account for the primordial cranium 
of Vertebrates by supposing that it has been evolved independ- 
ently of the spinal column, serving originally for the attach- 


ON THE ORIGIN OF VERTEBRATES FROM ARACHNIDS. 395 


ment of powerful muscles of locomotor visceral arches or 
Arachnoid appendages. 

The nature of the trabecule, the occipital ring, the diaphragm- 
like membrane, the absence of segmentation, and the relation 
of the primordial cranium to the vertebra are obscure points 
which the anatomy and embryology of the Vertebrate head 
failed to elucidate, but which the Arachnid theory resolves 
into the comparatively simple question as to the origin of the 
cartilaginous sternum. Vertebrate embryology, it seems, 
told all there was to tell: the fault was not in the answer, but 
in the interpretation; or rather, in the conviction that if 
ontogeny did not show what was expected, it was due to the 
imperfections of the ontogenetic record, not of the expec- 
tations. 


X. Grii-stits AND GrLi-ArcHEs.—In Scorpio there are 
coils of mesoderm (?) cells in the coxal portion of each pair of 
thoracic appendages (Pls. XXIII, XXIV, figs. 1—3). All these 
cell-coils disappear except that in the fifth coxa, which 
develops into the adult nephridium-like coxal gland; hence 
each of these coils probably represents a rudimentary nephri- 
dium. 

An ectodermic invagination, which appears on the outer side 
of the base of the fifth pair of legs, gives rise to the outlet of 
the gland. There are similar invaginations at the bases of the 
other appendages ; but they give rise to chitin-lined tubes, 
which serve for the support of muscles. The chitinised tubes 
are comparable with the three or four pairs of tracheal invagi- 
nations which in insects give rise to the tentorium. Since in 
Acilius some of the abdominal trachee at first communicate 
with the cavities of the mesoblastic somites, it is probable that 
all the trachez represent the ectodermic portions of nephridia. 
I regard the lung-books of Scorpio and the chitin-lined tubes 
described above as belonging to the same category, for after 
careful study I have found nothing to indicate that they arise 
as modifications of rudimentary abdominal appendages. 

If we suppose that all the thoracic and vagus appendages are 


356 WILLIAM PATTEN. 


reduced, by parasitism or a variety of other causes, to trans- 
verse ridges (like the pectens, for example, Pl. XXIV, figs. 3 
and 4; or like the thoracic and abdominal appendages of 
Limulus), and the segmental tubes enlarged to great trans- 
verse respiratory slits, like those in the abdomen of Scorpio, 
then we should have a condition much like that of the gill- 
slits and gill-arches of Vertebrates. Such an Arthropod 
appendage would resemble a gill-arch (1) in being supplied 
with a neural nerve; (2) in containing an artery following the 
nerve first mentioned ; (3) in possessing a great sense-organ, 
from which a ganglion to the neural nerve arises; (4) in the 
origin of its muscles from a diverticulum of a mesoblastic 
somite; (5) they would agree approximately in number with 
true gill-arches ; (6) they would agree with gill-arches in their 
serial physiological differentiation, for in both cases the ante- 
rior pairs are of great size, forked, and serve as mouth parts, 
the posterior ones being associated with respiratory organs, and 
showing a tendency to degenerate. 

Segmental respiratory sacs or tubes are eminently charac- 
teristic of Arthropods; and, as they are probably derived from 
the outlets of nephridia, they represent just the kind of respira- 
tory organs required, according to Dohrn’s theory, in ancestral 
Vertebrates. 

In Scorpio the wandering backward of vagus nerves to 
abdominal lung-books is important, and shows that we may 
not, without other evidence, assume that in Vertebrates the 
true gill-arches agree in number with or belong to the same 
segment as the nerves that supply them. Accepting Van 
Wyhe’s views as to the structure of the Vertebrate head, we 
offer the following tentative conclusions :—(1) The chele and 
first, or perhaps first and second, walking legs of Scorpio cor- 
respond to the mandibular and hyoid arches. (2) The remain- 
ing two or three pairs of thoracic appendages and somites are 
not present in Vertebrates. (3) The rudimentary vagus ap- 
pendages of Scorpions and the corresponding somites, except 
the muscles extending to the pectoral or pectinal arches, have 
in Vertebrates disappeared. (4) The true gill-arches represent 


ON THE ORIGIN OF VERTEBRATES FROM ARACHNIDS. 357 


the tissue between segmental respiratory organs or lung-books, 
and not modified appendages; consequently they differ from 
the true appendages, like the mandibular and hyoid arches, in 
not containing somatic diverticula(Van Wyhe). (5) The che- 
liceree are represented in Vertebrates by adhesive pre-oral 
papillz ; this is not so improbable as at first might appear, for 
in many Crustacea the first antenne, which are probably 
homologous with the chelicere, aided by the secretion of a 
sticky substance, serve as sucker-like organs of attachment. 


XI. Musctzes.—Since longitudinal muscles serve to move 
the segments on one another, the complete fusion of the first 
thirteen segments to form the cephalothorax is, without doubt, 
the cause of the disappearance of the dorsal and ventral longi- 
tudinal muscles of Scorpio. 

In the Vertebrate head the dorsal and ventral longitudinal 
muscles are also absent; therefore we conclude that in the 
ancestral Vertebrates the segments were immoveably united to 
form a hard outer skeleton like the thoracic shield of Arthro- 
pods ; the partial union of soft, flexible parts, as in an Annelid 
thorax, would not explain the absence of these muscles. As 
longitudinal muscles, when present, are very large, we can 
thus account, in a measure, for the great difference between 
the development of the head and trunk somites of Vertebrates 
and Arachnids. 

A more detailed comparison shows further that, (1) in both 
Scorpio and Vertebrates, there is very little mesoblastic 
tissue, and no distinct somite at all in the fore-brain. It is 
even possible that all the original fore-brain mesoblast has 
disappeared, that which is now present being derived from a 
forward growth from the post-oral segments. (2) The mid-and 
hind-brain region in both Vertebrates and Arachnids contains 
six mesoblastic somites, from which diverticula are formed 
leading into the appendages (gill-arches), and giving rise to 
muscles passing from these appendages to the cartilaginous 
cranium. (8) In the accessory brain region of Scorpions and 
in the vagus region of Vertebrates the mesoblastic somites dis- 


358 WILLIAM PATTEN. 


appear completely with the exception of a few longitudinal 
muscles passing from the occipital region of the cranium to the 
pectoral (= pectinal) arch. 

It is important to observe that in Arthropods, according as 
the anterior portion of the body becomes more specialised, the 
appearance of segmentation in those regions is retarded. For 
example, in insects, segmentation appears in the cephalic lobes 
and in the region of the mandibles and first maxille later than 
in the rest of the body; being the exact reverse of what we 
should expect, since the anterior part of the body is the oldest. 
It is the same with the cephalic lobes of Scorpio. In 
Limulus the fourth, fifth, and sixth (?) thoracic segments 
appear first; the third, second, and first later, and those of the 
fore-brain last of all. 

In view of these facts it is probable that in Vertebrates what 
has been taken for an intercalation of new head-segments is 
really a retarded segmentation. 

Eye Muscles.—In Scorpio, and probably in most Arach- 
nids, there is a small number of muscles which belong neither 
to the system of leg muscles nor to the longitudinal ones. 
They are dorso-ventral, and in Scorpio the largest one is 
attached to the hemal wall of the head on either side of, and 
close to, the median eyes. Although I have not been able to 
follow the development of these muscles, it is almost certain 
they develop from the first two or three thoracic somites. No 
such muscles are found in the posterior part of the thorax. 
They thus agree to some extent with the Vertebrate eye 
muscles, for the latter arise, as Van Wyhe has shown, inde- 
pendently of the eye from the first three somites, and belong 
neither to the primitive gill-arches nor to the longitudinal 
muscles. 


XII. Pecrorat Frns.—There are four completely fused 
segments in the vagus region of Scorpions, the comb belonging 
to the third. As there is no external evidence of this con- 
dition in the adult, we must, in attempting to determine the 
homologies of the appendages in such forms as Merostomata, 


ON THE ORIGIN OF VERTEBRATES FROM ARACHNIDS. 3859 


be guided in the main by comparison with those related forms 
whose internal anatomy is better known. 

If the metastomum of Pterygotus is homologous with the 
similarly-named organs of scorpions, it must be derived from 
the first two abdominal segments. If, as its position indicates, 
it arises in front of the swimming legs, the latter would belong 
to the third abdominal segment, and would, therefore, be homo- 
logous with the pectines of Scorpio. If this be so, then the 
four pairs of appendages in front of the oars would correspond 
in number and uniformity of structure with the four pairs of 
walking legs of Scorpio; consequently the great chelate 
appendages would be homologous, as indicated by their struc- 
ture and function with the chele of Scorpions. The small 
size of the four appendages of Pterygotus and the large size of 
the anterior ones is not difficult to understand, because if a 
Scorpion-like animal should gradually adopt the habit of 
swimming on its back by means of its comb or any oar-like 
appendage, the walking appendages would naturally decrease 
in size, while the grasping ones would not be so readily affected 
by such achange. There should be in Pterygotus, according 
to the above view, a pair of chelicere in front of the chele; 
but, as in all Arachnids these appendages are very small, it is 
not strange there is no trace of them in the fossil forms under 
consideration. 


XIII. Aracunip Features or Prericutnoys.—We have 
been impressed with the way certain fossil fishes, such as 
Pterichthys and allied forms, resemble the Merostomata,! 


1 That this external resemblance is real may be shown by the fact that 
those most familiar with the subject, and the ones best able to judge, were 
also greatly impressed by the same fact. 

The “genialer” Hugh Miller, the discoverer of Pterichthys, says (‘Old 
Red Sandstone,’ p. 50), in comparing a Tribolite with Cephalaspis, “The 
fish and the Crustaceans are wonderfully alike. ... . They exhibit the 
points at which the plated fish is linked to the shelled Crustacean.’’ 

Also Sir Roderick Murchison, when first shown specimens of Pterichthys, 
wrote regarding them that, “if not fishes, they more clearly approach to Crusta- 
ceans than to any other class,’ Again, “ They [Cephalaspis and Pterichthys] 


360 WILLIAM PATTEN. 


especially as regards size, shape of the body, the hard outer 


4 = ae 

Fie. 13.—(1) Ocular or hemal surface of the cephalic shield of a Tri- 
lobite. (2) Same of Pterichthys. (8) Same of Bothriolepis. These 
figures are intended to show the similarity in the arrangement of the 
cranial plates, and of the cervical and facial sutures. (4) Diagram of 
the neural surface of a merostomatous Arachnid, showing paired arrange- 
ment of coxal plates, and the eyes passing from the bemal to the neural 
surface. (5) Hemal surface of a Vertebrate head.—m. Metastomum. 
s. 8. Cervical suture. w.y.z. Facial suture. ca. s. Coxal sense-organs. 
s. 0. Sense-organs. 


skeleton and its peculiar sculpturing, its concentration at the 


form the connecting links between Crustaceans and fishes.’’ Agassiz himself 
was at first in doubt as to whether Pterichthys was a fish or Crustacean. 

To show still further the doubtful position of some of these primitive fishes, 
it may be stated that recently a distinguished paleontologist advanced the 
opinion that forms like Bothriolepis were Ascidians ! 


ON THE ORIGIN OF VERTEBRATES FROM ARACHNIDS. 361 


head end to form a compact cephalothoracic shield or cranial 
buckler—the post-cranial segments being in some cases pro- 
bably soft and disconnected,—the shape of the swimming 
appendages, and the probable method of locomotion. And, 
moreover, since the concrescence of vagus segments and their 
union with the true thorax is shown in the adult Scorpion and 
also in Trilobites and Merostomata by the cervical suture, and 
since a similar suture is seen in Pterichthys and Bothriolepis, 
we may infer that the last two forms also possess vagus seg- 
ments, and that their swimming appendages are homologous 
with the combs of Scorpio and with the swimming legs of 
Pterygotus. But as Pterichthys is also fish-like, its swimming 
appendages are probably homologous with pectoral fins, and 
consequently the combs of Scorpio are also homologous with 
pectoral fins. 

In support of this view we may add (1) that the fundamental 
structure of the pectens of Scorpio and the pectoral fins of 
embryo Selacians is the same, for the framework of both con- 
sists of a longitudinal bar, attached by its median end to the 
body, which gives off at right angles a series of rays. (2) 
In Scorpio the pectens, which develop differently from any 
other Arthropod appendages, first appear as an enormously 
long transverse ridge, divided into a series of small lobes (PI. 
XXIV, fig. 4). The important point is that the lateral end 
is gradually separated more and more from the body, until 
finally the whole appendage is only supported by a slender 
stalk at its neural extremity. This method of development 
agrees with the well-known development of the pectoral fins of 
Selacians, and the result is a long basal bar attached at one 
end only, and which may thus be swung forwards or backwards 
as on a pivot, or even rotated on its long axis. (3) They 
probably belong to the same segments, i.e. to the vagus. That 
the pectoral fins arose very far forward is certain; their 
probable origin from the vagus region is shown by the fact 
that in Protopterus (Wiedersheim) they receive nerves from 
the vagus, and also by the fact that the remarkable longitudinal 
muscles uniting the pectoral arch with the cranium are derived 


362 WILLIAM PATTEN. 


from the vagus mesomeres. (4) They show the same back- 
ward migration. (5) They are associated with transverse 
cartilaginous bars or pectoral arches, to which are attached 
longitudinal muscles arising from the cartilaginous cranium. 
The transverse bar in Scorpions is quite small, owing to 
the tactile function of the pectens; but in the gigantic 
Merostomata the muscles to their oar-like appendages 
must have been enormous, and the cartilaginous bar for their 
support not a bit inferior in size to the pectoral arch of Verte- 
brates. 

The hemal surface of the cephalothoracic shield of Trilobites 
is divided into a number of distinct plates, which resemble in 
shape and arrangement those on the cephalic bucklers of 
Pterichthys and Bothriolepis (Fig. 13, 1—3). 

The most important resemblances are shown—(1l) In the 
size and general shape of the shields. (2) In the posterior 
line or row of small plates, which form a cervical suture, 
and which indicate the presence of vagus segments. (3) In 
the great semicircular sutures extending parallel with the edge 
of the shield around the front and sides. (4) In the roughly 
triangular ocular plates, with the eyes on their median edges. 
(5) In the facial suture, z.z. (6) In the median row of plates 
or lobes in which the median eyes are situated. If the reader 
is not impressed with the resemblances above indicated, I am 
sure a careful comparison will convince him that the resem- 
blance is at least much greater than that between the ocular 
surface of Bothriolepis and Pterichthys and that of a true 
Vertebrate. 

Now let us compare the neural surface of Pterichthys, or 
the neural surface of a true fish, with the neural surface of 
Scorpions and Merostomata, and we shall see that in all these 
cases the median cranial plates are arranged in pairs, termi- 
nating in a posterior unpaired plate (Fig. 18, Nos. 4 and 5), 
If these facts mean anything, they show that whenever distinct 
plates are developed in the cephalothoracic shield of Arachnids 
and lower Vertebrates the hemal surface will contain an un- 
paired median row, and the neural surface a paired median 


ON THE ORIGIN OF VERTEBRATES FROM ARACHNIDS. 363 


row, terminating posteriorly with an unpaired plate or meta- 
stomum. 

As in all true Vertebrates, the eyes are found on the same 
side as the paired cranial plate, we may conclude that the 
real break between Arthropods and Vertebrates is made by 
the transference of the eyes to the neural surface. Since their 


Fie. 14.—Diagram of the circulation (¥) in the thorax of an Arachnid, 
and (V) in the head of a Vertebrate. 


eyes are situated on the heemal surface, we may conclude that 
Pterichthys and Bothriolepis, &c., are nearer related to the 
Arachnids than to the Vertebrates. The position of the eyes 
undoubtedly depends largely on the position in swimming. In 
Pterygotus, for example, where locomotion was probably largely 
effected by swimming on the hemal surface, the eyes have 
already become lateral—a position very unusual in Arachnids, 
This change is readily explained, since the original position of 
the eyes in the embryo of all Arthropods is neural ; moreover, 
this history of Arthropod eyes shows conclusively that they can 
assume any position the method of locomotion may demand. 

The above view of the Vertebrate cranium explains why 
the cephalic bucklers of many primitive fossil fishes are fre- 
quently divided into distinct dorsal and ventral shields, and 
also why the abocular or neural surface is so imperfectly 
preserved. 


364 WILLIAM PATTEN. 


It follows from what has preceded, that in Vertebrates and 
Arachnids, those surfaces of the head bearing a median row of 
paired plates and those bearing a median row of unpaired 
plates are homologous. This conclusion is supported by the posi- 
tion of the nervous system, and also by the relation these plates 
bear to ganglionic sense-organs. For example, each of 
the median cranial plates of Vertebrates contains a primary 
group of sense-organs supplied by a ganglionated “ramus 
dorsalis ;’’ they therefore agree perfectly in this respect with 
the coxal plates, each of which also contains a group of coxal 
sense-organs, supplied by a nerve which, on other grounds, we 
concluded was homologous with a ‘‘ramus dorsalis.” Thus, 
since the coxal plates of Arachnids and the paired cranial plates 
of Vertebrates are homologous and segmentally arranged, the old 
Goethe-Oken theory of the bony cranium appears in a new light. 


I trust in all that has preceded I have succeeded in showing 
that there are important resemblances between the cephalo- 
thorax of Arachnids and the head of Vertebrates. There are 
other resemblances of a more general character, a few of 
which we shall merely mention here:—(1) The division into 
cephalothorax, body, and tail. (2) The tadpole-like larva of 
Vertebrates, with its enormous head, and small body and tail, 
is explicable on the Arachnid theory. (3) The Trilobites 
probably swam, if at all, on their backs; and it is still more 
probable that the Merostomata, from their shape and the posi- 
tion of their oar-like appendages, swam in the same way. 
The larve of Limulus, according to my own observations, always 
swim on their backs. Thus the way is prepared for the manner 
of locomotion in fishes. (4) The development of the Arthro- 
pod heart from the concrescence of paired mesodermic folds 
is like that in some Vertebrates; and the aortic, arch-like 
blood-vessels in the appendages may be compared to the 
branchial arteries (Fig. 14, 5.) (5) Moreover, the enor- 
mous liver of Arachnids, (6) the blood, (7) the histological 
structure of the nervous system, and the manner of nerve 
terminations, (8) the minute structure of the muscles, and (9) 


ON THE ORIGIN OF VERTEBRATES FROM ARACHNIDS. 865 


the development of the ova and spermatozoa, all supply us 
with abundant evidence of the great morphological and _histo- 
logical specialisation of the Arachnids, and their structural 
similarity to Vertebrates. 

There is another point which deserves more than passing 
notice. In insects the sexual organs are developed (Acilius) 
from the wall of one, or at most two, mesoblastic somites, and 
are carried by the growth of these organs to the hzmal surface; 
the adult organs are never reticulated. In Scorpio (and 
Limulus ?) it is just the reverse ; the sexual organs arise as a 
median longitudinal band of cells underlying the nervous 
system, and extending the whole length of the primitive 
abdomen—that is, over at least seven segments ; moreover, in 
the adult the sexual organs are remarkable for their reticula- 
tion. In the adult Scorpion the sexual gland is neural in 
position, and is composed of three longitudinal tubes united 
by transverse ones. Now in Protopterus the testis has exactly 
the same arrangement of longitudinal and transverse bars as 
in Scorpio, the only difference being the position of the outlets 
and the small size of the median longitudinal tube. In other 
words, it is a most significant fact, when viewed in connection 
with all that has preceded, that the sexual organs of Scorpio 
and Limulus, in their exceptional position, structure, and 
development, should resemble in these very features the sexual 
organs of primitive Vertebrates. 


XIV. Emsryotocy.—Kleinenberg’s admirable observations 
on the development of Lopadorhynchus afford the first secure 
foundation for the interpretation of the embryology of the 
higher segmented animals. They teach us not only exactly 
what the gastrula of a segmented animal is, but also what it is 
not. As long as such forms show no trace of concrescence, of 
celomic diverticula, or of any connection between an un- 
doubted gastrula and an undoubted “ primitive streak,” we 
must, in order to explain the facts of Arthropod and Verte- 
brate embryology, follow other paths than those laid down by 
the concrescence theory, the ccelom theory, or any other theory 

VOL. XXXI, PART Il1I].—NEW SER. BB 


366 WILLIAM PATTEN. 


that regards the “ primitive streak” and “‘blastopore” as 
remnants of a gastrula—unless, indeed, we expect to prove that 
the Vertebrates out-Celenterate the Celenterates. The ab- 
sence of an Annelid pre-oral lobe, and the formation of the 
head by the pushing forwards of three post-oral body segments, 
show that the Arthropod head and body are comparable only 
with the post-oral portion of the Annelid. The Arthropod 
body represents an outgrowth from the trochosphere, but the 
trochosphere itself, the Ccelenterate stage, has disappeared. 
Hence there is no such thing as a gastrula in Arthropods, and, 
strictly speaking, no germ layers. The germ-layer theory 
requires, as one of its ablest expounders, Balfour, explicitly 
states, that the entire ectoderm, mesoderm, and endoderm be 
derived directly from the primitive layers. Now in Lopado- 
rhynchus it is certain that the greater part of the mesoderm 
arises from the ectoderm at the growing tip of the tail, and has 
nothing to do with primitive mesoderm. In Arthropods the 
mesoderm and also part of the endoderm may arise in an 
exactly similar manner. Besides, mesoderm may arise (ACci- 
lius) at a late embryonic period from a great variety of places, 
just as ganglionic cells are formed from the general surface of 
the body wherever a new sense-organ is formed. Hence it is 
highly probable that all the endoderm, except possibly a small 
portion at the inner end of the cesophagus, and the mesoderm 
have arisen independently of, and have finally supplanted, the 
primitive layers ; just as in Arthropods the pre-oral lobe and 
brain of Annelids have been replaced by other organs. 

If we suppose that the body of segmented animals repre- 
sents, not the elongated body of a Celenterate, but only one 
of its tentacles, we can explain why a segmented animal 
always grows at one end only—something the concrescence and 
other theories cannot do; we can also explain why the Annelid 
and molluscous body is a lateral outgrowth from the subum- 
brella ; and this supposition would be in perfect harmony with 
the position and history of the “ Prototroch” nerve, as de- 
scribed by Kleinenberg. Moreover, the imperfect division of 
Celenterate tentacles into joints provided with lateral pro- 


ON THE ORIGIN OF VERTEBRATES FROM ARACGHNIDS. 367 


cesses may be regarded as a forerunner of metameric segmen- 
tation (Fig. 15). 


Fig. 15,—Diagram of an Anmnelid (4), Mollusc (B), primitive Annelid (C), 
and Celenterate (D) larva. 


To show, further, that a reasonable explanation of some of 
the salient points in the embryology of segmented animals 
may be given on the above view, without resorting to impos- 
sible gastrulas, I suggest the following. Since in Ceelente- 
rates mesoderm cells are constantly forming from the ectoderm, 
it is probable that their growing tentacles contain at least 
some mesoderm and endoderm cells uot derived from the 
primitive layers. If such a tentacle were transformed into the 
body of a segmented animal, its elongation and the formation 
of the layers in the embryos would naturally take place by the 
rapid division of large terminal! cells like those found in all 
rapidly elongating organs of both plants and animals, and, 
indeed, just as it probably takes place in the Celenterate ten- 


368 WILLIAM PATTEN. 


tacle itself; moreover, there would be no necessity for regard- 
ing these terminal cells or teloblasts as any part of a gastrula. 
Bearing in mind the above conclusions, it is evident that two 
surface teloblasts or mesoblasts multiplying by tangential and 
radial division would produce two long mesoblastic bands 
lying beneath the ectoderm, as in most Annelids. Many 
surface teloblasts dividing by radial division alone would pro- 
duce a band of mesoderm lying its whole length on the sur- 
face. Then, if tangential division began, a longitudinal cord 
of cells would be either proliferated or invaginated inwards ; if 
this median cord divided into two lateral ones, which then 
became segmented, we should have the essential features in the 
development of the mesoblastic bands of many insects. The 
important point is that the segmented mesodermic bands of 
insects have been produced in exactly the same way as those 
of Annelids, except that the tangential division of the teloblast 
and of its products is slightly postponed. That the median 
furrow of insects is merely an ontogenetic adaptation is suffi- 
ciently evident from the fact that it may be present or absent 
in closely related forms. When it is absent the resemblance 
of the mesoblastic bands and their growth, to those of Anne- 
lids is more evident. 

In Doryphora, Acilius, Musca, and others, the great mass 
of terminal proliferating cells also gives rise to two cords of 
endoderm extending forward, one on either side of the 
median line. In some insects the median furrow produced by 
the invagination of mesoderm tends to close from before back- 
wards, leaving a terminal pore, the anterior wall of which con- 
tinues to proliferate endoderm and mesoderm : in this case the 
mesodermic bands would appear to grow forwards from the 
anterior median wall of the pore, while the endodermic bands 
would appear to arise from its anterior lateral walls. 

Ordinarily the mesodermic bands become segmented while 
the endoderm is yet only two narrow bands (Fig. 16, c), 
but later the endodermic bands begin to spread out ventrally 
and dorsally to enclose the yolk. Suppose the segmentation 
of the mesoderm is retarded and the extension of the endo- 


ON THE ORIGIN OF VERTEBRATES FROM ARACHNIDS. 369 


derm accelerated, we might then have a continuous sheet of 
mesoderm and endoderm, apparently a forward continuation of 
the anterior wall of the terminal pore or telopore (Fig. 16, 
bp and F). Suppose the hemal edges of the mesodermic bands 


Fie, 16.—Diagrams to illustrate the formation of germ layers.—2. Cross 
section of the posterior end of telopore of an insect just before it closes. 
B. Section a little in front of the last, showing the cords of mesoderm 
and endoderm. C. One still farther forward. JD. Section through an 
embryo in which it is supposed mesoderm and endoderm form a continuous 
layer. #. Section of an embryo in which it is supposed the lateral 
endodermic bands have grown completely round the yolk before the 
mesoderm became segmented and separated fromthe endoderm. F, G, H. 
Three longitudinal sections, showing successive stages in the formation 
of atelopore by the invagination of teloblasts. #. Insect. G. Crus. 
tacean, Astacus. H. Amphioxus. 


unite before the mesoderm is separated from the endoderm, 
we should have a condition like that in Amphioxus, where 
the endoderm and mesoderm form a sac opening outward 
posteriorly by the telopore (Fig. 16, = and #). All the 
so-called gastrulas and blastopores of Arthropods and Verte- 
brates can be explained on the same principle. Accepting 


370 WILLIAM PATTEN. 


the above view, it is plain that no invagination at the 
posterior end of a segmented embryo can be justly 
regarded asa gastrula; and conversely, in all segmented 
animals the gastrula or any remnant of it must lie 
at the very anterior end of the body. The conclusion 
is obvious that no trace of a gastrula, any more than of 
Annelid pre-oral lobes, is to be found in Vertebrates and the 
higher Arthropods. 

Just as in the blastosphere the invagination of 
endoderm toformatrue gastrula is the ontogenetic 
way of repeating what was originally a mere local 
proliferation, so in the segmented animals the in- 
vagination of endoderm and mesoderm to form a 
telopore is an ontogenetic modification of acluster 
of proliferating cells or teloblasts. The telopore and 
the gastrula are thus to a certain extent analogous, but in no 
wise homologous. The formation of the telopore is com- 
plicated by the increase in length, and by the presence of 
both mesoderm and endoderm; but we have, I believe, in 
Arthropods and Vertebrates a complete history of its various 
phases. 

It is a fact of great importance that in Vertebrates and 
Arthropods the functional endoderm exists at one time as two 
longitudinal lateral bands—a condition, as far as I know, not 
found in Annelids. 

The importance of yolk in modifying development has been 
greatly exaggerated. If its presence obscures the primary 
characters, it is extremely probable that in animals normally 
having large eggs its absence would result in a still farther 
modification, not in a reversion to the simpler condition. 
Such has been the case, I believe, in Amphioxus, where the 
simplicity in development is apparent, not real. In fact, the 
celom theory proves too much, for it cannot explain why Am- 
phioxus in the development of its body-cavity falls little short 
of being a typical Ceelenterate, while the Annelids themselves 
do not in this respect show the slightest trace of Coelenterate 
characters. 


ON THE ORIGIN OF VERTEBRATES FROM ARACHNIDS. 3871 


In Cymothoa the multiplication of the teloblasts goes 
on with great regularity (Fig. 17). This condition is un- 
doubtedly a mere modification of that found in most insects. 
But it is incredible that these teloblasts can have anything to 
do with a gastrula. There is no segmentation of a continuous 


2 BOOE ,0500000030° 
(Sy 19H OOD) HOOSSQOO oe 
) ©H © PSOQODIPBOOO 5 
boneges ©000984 On 


EGOS 099999809 
; JOODOS 0099290 90 D979 
) ~ 8895 ate ik 
ae 226) © ®©O9S2 e) & 

ee 209 
FELD ODO © IO 


Fic. 17.—Formation of layers in Cymothoa.—. Median longitudinal 
section of posterior end of germ band. J. Posterior end of embryo, 
seen from below; the rows of inner layer cells are not represented on 
the right. The last row of ectodermic teloblasts often divides simul- 
taneously along its whole length. 


sheet of mesoderm, but each somite arises from the repeated 
division of a single row of eight cells. Each somite consists of 
two lateral parts arising from the six lateral mesoblasts, and a 
median one arising from the two median mesoblasts. 

I long ago pointed out that in Phryganids the neural furrow 
is developed in the anterior portion of the germ band before 
the terminal pore has disappeared, and that its posterior end 
extends into the terminal pore. I then stated, and still main- 
tain, that an incipient neurentric canal is thus formed, 
similar to that in Vertebrates. 


372 WILLIAM PATTEN. 


But there are certain structures in Arthropods which may 
represent remnants of gastrulas. For example, if the mouth 
and cesophagus of Arthropods is primitive—and there is no 
reason to suppose it is secondarily acquired,—then we must 
look for primitive endoderm at its inner end. I have figured 
in “Eyes of Acilius,” at the very anterior end of the embryo, 
a great sac of endoderm cells which probably arise by in- 
vagination, although the process was not directly observed. 
The sac, which soon opens outward by the cesophagus, after- 
wards becomes solid, and finally is converted into two longi- 
tudinal bands, one on either side, extending backwards 
towards the middle of the body, where they become con- 
tinuous with similar bands extending forwards from the 
posterior end of the embryo. 

There are several other cases where great vesicular cells 
appear at the inner end of the esophagus, and they probably 
are of a similar nature to those just described. 

In Limulus there is a great lump of endoderm-like cells at 
the inner end of the esophagus; they grow fainter, and are 
quickly absorbed without forming either yolk-cells or any part 
of the definite endoderm. 

These endoderm-cells are the only structure in Arthropods 
which I can see any reason for regarding as remnants of a 
gastrula. This evidence consists solely in their position at the 
inner end of the cesophagus, and their speedy absorption in 
Limulus, and possibly other Arthropods. But even this 
evidence is of no great weight, since in Doryphora—according 
to Wheeler, whose preparations I have had the privilege of ex- 
amining—a great mass of cells arising from the posterior end 
of the body is absorbed in very much the same manner as the 
cesophageal ones in Limulus. 

The history of the layers in Scorpio and Limulus is com- 
plicated, and I have not yet obtained satisfactory results ; 
much of the endoderm and mesoderm is produced from the 
cluster of cells at the posterior end of the body, but there are 
no such distinct endodermic bands as in insects, It is 
probable that some of the endoderm, and perhaps meso- 


ON THE ORIGIN OF VERTEBRATES FROM ARACHNIDS. 3878 


derm, is split off from the blastodisc before the germ-band is 
formed. 

In surface views of the germ-band of Scorpio, we see in 
the median line at the posterior end of the body a great 
thickening, from which cells grow forward and laterally. From 
the median portions of the band of tissue thus formed arise 
the sexual organs and the botryoidal cord, and from the more 
lateral portion the mesoblast and endoderm, 

In Limulus there is, at the posterior end of the embryo, a 
short, but distinct slit-like invagination, which, in surface views 
and in sections, is exactly like the primitive streak of many 
birds and reptiles (Fig. 18, B and pb). 

The lateral edges of the wedge-shaped mass of cells pro- 
duced by this invagination spread out on either side between 
the yolk and the ectoderm to form poorly defined mesodermic 
bands. Atirregular intervals, but principally near the primitive 
streak, great masses of cells wander inwards from the inner 
layer cells, and are scattered throughout the yolk. 

Moreover, in a number of different places great masses of 
mesoderm and endoderm cells may be proliferated inwards 
from the ectoderm of the ventral plate, in a manner very 
similar to what occurs in the primitive streak ; the proliferating 
points differ from the primitive streak in being nearly trans- 
verse instead of longitudinal, and in not being accompanied 
by any overlying depression of the ectoderm! (Fig, 18, 
A x, x). These facts prove conclusively, I think, that the 
formation of mesoderm and endoderm in the ventral plate of 
Limulus does not follow any method which can possibly be 
regarded, in the concrescence theory or any other, as a modifi- 
cation of a gastrula. I have examined a great many embryos, 
and have never seen any traces of a median furrow connecting 
the mouth and anus such as is described by Kingsley, and 


1 It is possible that in some places at least these proliferating points may 
be cross sections of the somites, which are much curved at the posterior 
end of the body. I have seen indications in young scorpion embryos that 
some somites are formed by a distinct transverse proliferation 
of the ectoderm. 


374 WILLIAM PATTEN. 


which he regards as the remnants of a medianly coalesced 
gastrula. According to my observations, the mouth, the 
primitive streak, anus, and the neural canal, are totally inde- 
pendent organs, and appear at widely different stages. 

Out of a large number of embryos, a few are provided at 


Fic. 18.—4. Diagram to show distribution of mesoderm and endoderm 
in young embryo of Limulus. m. 6. Peripheral zone of mesoderm (?)- 
m. s. Mesoderm underlying the ventral plate. o.e. Mass of transient 
endoderm at inner end of cesophagus. JB. Slightly diagrammatic view 
of Limulus embryo. ©. Cross section of same through primitive streak. 
D. Ditto, a little older stage than C. /.e. Lateral eyes. m. 6. Meso- 
dermic (?) bands. m.s. Mesoderm. — s. 0!—°, Thoracic sense-organs. 
p.s. Primitive streak. y. Masses of cells passing inward from the 
ectoderm. 


the posterior end with an enormous depression, either circular, 
oval, or pyramidal in outline, at the bottom of which, or on 
its anterior wall, lies the primitive streak. It is not accom- 
panied by any special modification in the development of the 


ON THE ORIGIN OF VERTEBRATES FROM ARACHNIDS. 875 


germ layers, and it certainly is not a gastrula, although in 
surface views it is suggestive of such an organ." 

In Limulus there is a remarkable ring of mesoderm cells 
extending completely around the embryo. It may be seen 
in surface views as two dark bands diverging from the region 
behind the primitive streak, and uniting with each other 
in the neighbourhood of the esophagus (Fig. 18, a). In 
sections the bands appear as two clusters of oval cells, 
formed by an inner proliferation of the ectoderm (Fig. 18, 
c and vp, m.b.). They are soon freed from the overlying 
ectoderm, and finally, after imcreasing enormously in size, 
meet each other along the median dorsal surface. By the 
time the bands are well formed the cells are oval, and contain 
an enormously long, brilliantly refractive filament, which is 
either wound or bent back and forth a great many times. The 
small remaining space in the cell is filled with a watery fluid, 
and a small, laterally placed nucleus. Some of the cells 
become elongated, and the coiled fibres give rise to the striz 
of the longitudinal dorsal muscles. Some of the cells appear 
to degenerate and disappear. The resemblance of this ring, 
at an early stage, to the “ Keimwall” of Vertebrate embryos 
need not be enlarged upon here. I have not with certainty 
found anything similar in Scorpions. 

The following expresses what I conceive to be the most 
natural arrangement of the segmented animals :— 


1 This condition is probably produced by some peculiarity in the division 
of the cells of the posterior end of the body. All foldings of cellular mem- 
brane are due, I believe, to local variations in cell division. Wherever a cell 
layer tends to increase in thickness by tangential division a simple thickening 
will be produced. But if the cells thus produced at the same time multiply 
by radial division, the inner surface of the membrane will be larger than the 
outer, and the whole membrane curved or warped. The direction of 
curvature will depend entirely on the relative rapidity of division at certain 
points. All foldings may then be regarded as the expression of certain 
methods of cell-growth, and may not have themselves any morphological 
significance. 


376 WILLIAM PATTEN. 


Vertebrata 
Urochorda Cephalochorda 
as ——— = 
SIPHONEURA 


Central nervous system tubular. 
Eyes on neural surface. Retinas to both median and lateral eyes, inverted. 


. Cephalaspis, Pterichthys, 
Eeliphigie dats Pteraspis, Bothriolepis, &c. 


Arachnida Merostomata me 
urypterus 
Trilobites 
I TH 
PELTACEPHALA. 
Head shield-shaped. Eyes on heemal surface. 
ae 
SYNCEPHALA. 


Divided into Head, Body, and Tail. 


Head-segments 9--13; in four groups; first three head-segments 
devoid of appendages. Terminal segments more or less fused to form 
aurostyle. Eyes median and lateral. Median eyes 2—4, in a common 
sac or tube; retinas inverted. 


Insecta 


Crustacea 
Myriapoda 
Rea ES a oe ee Se 
PSEUDOCEPHALA. 
Brain segmented ; originally post-oral. 
Annelida 
—-—_—~_ 
ENCEPHALA. 
Brain unsegmented ; pre-oral. 
Ree oe oe RiaeerNeeerterere mere 
SEGMENTATA 
he \ 
CHLENTERATA ECHINODERMATA ? 


as 
PROTOZOA 


ON THE ORIGIN OF VERTEBRATES FROM ARACHNIDS. 377 


EXPLANATION OF PLATES XXIII and XXIV. 


Illustrating Mr. William Patten’s paper “On the Origin of 
Vertebrates from Arachnids.” 


a. . Abdominal neuromeres. dr. Brain. c.g. Coxal glands or nephridia. 
ch. Chele. chi. Chelicere. c.s. 0°. Indicated position of coxal sense-organs. 
e. s. Outline of optico-ganglionic sac. f. 6.= s./. Frontal or semilunar 
lobe. y.m.c. Ganglionic portion of median furrow. g.v'—*. Ganglionic 
invagination. 7. m.e. Interganglionic portion of median furrow. J. '-*. 
Lung-books. /.¢. Lateral eyes. m. Mouth. m.e. Median eye. mai. 
Maxillaria. x. ph. Nephridia. 2.c. = sp.g. Neural crest or spinal ganglion 
* Anlage.” op. g'-. Optic ganglia. op. p'-*. Optic plates. p!—*. Walking 
legs. s./. Semilunar lobe.  s. o'. Sense-organs near joints of legs. ss. 0”. 
Scattered sense-organs. sp. z. Spinal nerve. s,s. o. Segmental sense-organs. 
st.n. Anlage of stomodeal nerves. 7. 7—®. Thoracic neuromeres. v. 2!—*, 
Vagus neuromeres. v. p'—*. Vagus appendages. 


PLATE XXIII. 


Fic. 1.—Stage B: surface view of embryo Scorpion removed from the 
yolk. Observe the forward position of the mouth, and the absence of nervous 
or other tissue immediately in front of it; the absence of abdominal append- 
ages ; the post-oral position of the chelicers ; the divergence of the abdominal 
nerve-cords; the segmental sense-organs, s.s.0.; the absence of the eyes, 
aud of the rudimentary optic plate; the innumerable sense-organs from 
which the nervous system arises ; the large sense-organs, 2. c. = sp. g., which 
later form an incipient “neural crest,” and from which the spinal ganglia 
develop ; the “ anlage ” of the optic ganglia, op. g., and of the brain, dc., which 
is not yet segmented. (Obj., oc. 2.) 

Fic. 2.—Stage c: surface view of detached embryo of Scorpion, show- 
ing the three brain-segments ; the beginning of the second and third gan- 
glionic invagination, g. v. and g. v°.; the paired median sense-organ-like in- 
vagination in the brain region, those opposite the fourth neuromere giving rise 
to the ganglia of the stomodeal nerves ; the continuity of the paired pits with 
the median furrow ; the great frontal lobe, / d., produced by the invagination 
of the first brain metamere, and possibly made up in part of an Annelid pre-oral 
lobe (?); the large optic ganglion of the second and third brain metameres, 
op. g*. and op. g®.; the nephridia-like cell-coils in the basal portion of each 


378 WILLIAM PATTEN. 


pair of thoracic legs; the double-lobed neuromeres; the four pairs of rudi- 
mentary vagus appendages, v. p'—*.; and the bundle of nerve-fibres lying in 
the skin from which the spinal nerves are developed, sp. x. 


PLATE XXIV. 


Fic. 3.—Stage ©: surface view of detached embryo of the Scorpion. The 
left side of the head is represented as an opaque object, and shows the growth 
of the optic ganglia over the brain-segment, and the advance of the lateral 
lips of the ganglionic invagination over the optic ganglia. The right side is 
represented as transparent; the dotted line, ¢. s., shows the form of the sac, 
the outer wall of which is formed by the eye; the frontal, now nearly semi- 
lunar lobe, s. Z., lies beneath the brain and optic ganglia, op. g®. and op. g°. 
We also see the position of the coxal sense-organs indicated at c. s. 0°., al- 
though 1 have not been able to detect them in surface views; the appendage- 
like maxillaria, ma/.,in the third and fourth thoracic segments; the double 
neuromeres, the anterior portion of each containing in the median line a very 
distinct pit-like invagination of the median furrow, g. m.c.; the posterior 
portion, a very small and indistinct one, the interganglionic portion of the 
median furrow, 7. m.c.; the four lung-books, /. 4'-*., the first one belonging 
to the last vagus segment, the rudimentary appendage of which has dis- 
appeared. I have found no evidence that it is infolded to form the lung-book, 
the development of this lung-book being the same as the following segments 
where no appendages are ever present. 

Fie. 4.—Stage r: surface view of a detached embryo of Scorpion, showing 
the almost complete union of the two optico-ganglionic sacs; the lateral eye- 
plates, /. e.; the numerous sensory buds, s. o!., scattered over the surface of 
the segmental sense-organs, s. s. 0. There are also usually one or more quite 
large sense-organs on the basal joints of the appendages, s. o'. The first three 
vagus neuromeres have completely fused with one another, but the fourth is 
still quite independent. 


N.B.—The Scorpion referred to in the present memoir is probably Buthus 
carolinianus. 


VERTEBRATES FROM A CRUSTACEAN-LIKE ANCESTOR. 379 


On the Origin of Vertebrates from a Crustacean- 
like Ancestor. 


By 
Ww, H. Gaskell, M.D., F.R.S. 


With Plates XXV, XXVI, XXVII, XXVIII. 


Cuap. I.—Tue Evipencre GIVEN BY THE CrEnTRAL NERVOUS 
System anp Pineau Eyes or tHE AMMOC@TES OF PETRO- 
MYZON PLANERI. 


In a former paper (1) I have described how my investiga- 
tions into the Vertebrate nervous system have led me to the 
conclusion that that system is composed of nervous material 
grouped around a central tube which was originally the alimen- 
tary canal of the Invertebrate from which the Vertebrate arose ; 
and in a second paper (2) I have suggested that the physiology 
and anatomy of the Vertebrate nervous system both fit in best 
with the assumption that the Invertebrate ancestor was of the 
Crustacean’ type. In both papers I have promised to point 
out the confirmation of this theory which is afforded by the 
study of the lowest Vertebrate nervous system, viz. the 
Ammocetes form of Petromyzon. This promise I propose 
now to redeem, and in order to bring out as prominently as 
possible the theory which I hold to be true I will discuss the 
nervous system of the Ammoccetes in terms of the Crustacean; 

1 By Crustacean I include such a form as Limulus, without intending to 
pronounce any opinion on the question whether Limulus is an Arachnid or a 


Crustacean ; it would, perhaps, be better to use the term proto-Crustacean, 
meaning thereby a form from which both Crustaceans and Arachnids may 


have been derived, 


380 W. H. GASKELL. 


in other words, I will take separately the prominent features 
of the alimentary canal and central nervous system of a Crus- 
tacean-like animal, and point out how each one exists in the 
nervous system of the Ammoceetes. 

Before commencing the study of the special nervous system 
treated of in this paper I will briefly recapitulate the general 
arguments which I have put forward in my previous papers. 

The central nervous system of all Vertebrates consists of two 
parts—(1) a nervous portion, and (2) a non-nervous epithe- 
lial portion, which is in part free from admixture with nervous 
elements, and partly helps to form the supporting tissue of the 
nervous elements. This non-nervous part of the central ner- 
vous system forms a canal around which the nervous material 
is grouped, in the same manner as the nervous system of the 
Crustacean is grouped around the alimentary canal. This 
similarity of grouping is not merely anatomical, but is also 
physiological; the functions of the supra-cesophageal ganglia 
of the infra-cesophageal and of the ventral chain correspond to 
the functions of those parts of the Vertebrate central nervous 
system which are situated in the same anatomical position, 
with respect to the non-nervous tube, as the corresponding 
ganglia of the Crustacean with respect to the alimentary 
canal. 

The natural conclusion to draw from this striking coincidence 
is that this non-nervous tube of the Vertebrate central nervous 
system is the altered alimentary canal of the Crustacean ancestor 
of the Vertebrate. 

If this conclusion is right, then not only should the evidence 
upon which it is based come out more clearly when the lowest 
Vertebrate nervous system is examined, but also fresh evidence 
of Crustacean characteristics may reasonably be expected to be 
found in such nervous system. We should expect, in fact, to 
find a nearer approach to the Crustacean type than in the 
higher Vertebrates. 

We should expect, therefore, to find that the vestiges of the 
mouth, cesophagus, and cephalic stomach were more conspicuous 
than in the higher forms; that such an ancestral characteristic 


VERTEBRATES FROM A GCRUSTACEAN-LIKE ANCESTOR. 381 


as the pineal eye was easily recognisable and of a Crustacean 
type, and that the proportion of nervous material to non- 
nervous approached nearer to the proportion found in the 
Crustacean than in the higher Vertebrates. 

In addition, we should expect to find the vestiges of other 
Crustacean organs which are connected with the alimentary 
canal, especially that large and important organ the so-called 
liver ; and, indeed, seeing that the so-called liver represents the 
whole of the digestive glandular apparatus of the alimentary 
canal of the Crustacean, the cephalic stomach and straight in- 
testine being simple epithelial structures for the purpose 
respectively of holding food, perhaps also of absorbing it, and 
of passing out the indigestible residue, it follows that any 
theory which compares the simple epithelial non-nervous tube 
of the central nervous system of the Vertebrate with the non- 
glandular epithelial stomach and intestine of a Crustacean is 
imperfect, unless some explanation is given of the fate of the 
large glandular liver which must have existed in the cephalic 
region of the ancestor of the Vertebrate if that ancestor was 
formed after the Crustacean type. 

Conversely, the theory which I have put forward is im- 
mensely strengthened if the examination of the lowest type of 
Vertebrate brain brings to light the vestiges of a large glan- 
dular organ, which bears the same anatomical relation to the 
ventricular cavity of the brain as the so-called liver bears to 
the cephalic stomach of the Crustacean. 

I have chosen the Ammoceetes stage of Petromyzon as the 
subject of investigation because it undoubtedly represents a 
very primitive form among living Vertebrates, and also because 
it is easy to obtain in large quantities at different ages, and 
will live well in the laboratory: all my specimens were ob- 
tained by me at Hind Head, in Surrey—the first batch, to the 
number of about forty, in the beginning of September of 1888; 
and the second batch, amounting to 100, at the end of October, 
1889. They were all Ammoceetes of Petromyzon Planeri, 
and varied in size from 25 mm. in length up to full-grown 
specimens of 130 mm. or more. 

VOL. XXXI, PART 1II.—NEW SER. Cre 


382 W. H. GASKELL. 


Those which I brought with me to Cambridge in 1888 were 
kept alive in a small basin containing sand and weed from 
their native stream, the water in which was renewed occa- 
sionally. They lived in this way perfectly well, and some of 
them are still living; as far as I can see they have not grown 
at all, or made the slightest attempt at transformation. The 
others, which I have obtained lately, are kept in a large tank 
with sand and weed in it. 

The method of observation which I have mainly relied on is 
that of serial sections, the whole head of the animal having 
been embedded in paraffin, or else the brain dissected out and 
then embedded. For hardening and staining I have used 
various methods; the whole head has been placed either in 
osmic acid | per cent. solution, or else in Perenyi’s fluid and 
then alcohol, with subsequent staining in boro-picro-carmine 
or picro-carmine and eosin: I have also stained on the slide 
with aniline colours and with hematoxylin. 

As arule the sections are cut through the whole head and 
mounted in order. By this means it is certain that all the 
structures within the cranial cavity will be fixed on the slide 
in their right position; but, on the other hand, there are in 
many specimens hard parts in the region of the basilar plate 
which turn the edge of the razor and tear through the neigh- 
bouring soft tissues; also so large a section is apt to be 
crumpled when cut into ribbons by the Rocking microtome. 
I have succeeded in getting over the folding of the sections by 
a method which I have now used for some years past, but have 
never as yet published ; it consists simply in floating the series 
of sections on the surface of warm water and then transferring 
them to the slide, which has been previously coated over with 
albumen and glycerine. By this means the sections are 
flattened ; they are then dried by pressure between blotting- 
paper, the wax is melted, removed by xylol, and the series of 
sections mounted in Canada balsam. 

The figures in Pls. XXV, XXVI, XXVII, XXVIII, are, for 
the most part, drawn by myself, and by Mr. Wilson, of the 
Scientific Instrument Company. I am indebted to Mr. 


VERTEBRATES FROM A CRUSTACEAN-LIKE ANCESTOR. 383 


Durham and Mr. Griinbaum for making me photographs of 
various specimens, which have been of great assistance in 
making the drawings; to Mr. Shipley for making the series 
of drawings in Pl. XXVI, figs. 13a—13e; to Mr. Hardy and 
Mr. Shore for assistance in the preparation and staining of 
material ; and to Dr. Hunter for valuable advice in connection 
with the formation of pigment. To all these gentlemen I beg 
here to offer my sincere thanks. 

The description of the brain of the adult Petromyzon as 
given by Ahlborn (8) is so thorough and so excellent, that it 
is hardly necessary for me to re-describe the arrangement of 
its different parts, especially as the brain of the Ammoceetes 
resembles very closely that of the adult form; the main dif- 
ference being, as Ahlborn has pointed out, that the internal 
cavities are more roomy, and that the optic lobes are not so 
fully developed. I shall, therefore, refer my reader to his 
descriptions and illustrations, and only give such illustrations 
from my own specimens as seem to me necessary to make clear 
the arguments brought forward in this paper. 

In Pl. xiii of Ahlborn’s paper (3) drawings are given of 
the brain from different points of view: as, however, he ‘has 
removed the choroid plexuses in all his drawings I have 
thought it best to give a drawing of the dorsal aspect of the 
brain as it appears in the fresh condition when the brain-case 
is removed. The specimen from which fig. 1, Pl. XXV, was 
taken measured 120 mm. in length, and the brain was care- 
fully dissected out under a dissecting lens, and then the draw- 
ing was made. The greater portion of the dorsal surface is 
taken up with the transparent folds of the choroid plexuses 
li and ili; the folds of these two plexuses meet, as 
described by Ahlborn, over a deep fissure, which indicates the 
position of the band of fibres constituting the cerebellum. 
In front of the folds of the choroid plexus ii are seen, in the 
middle line, the large r. gang]. habenulz, and the white opaque 
pineal eye with the masses of the cerebral hemispheres on 
each side. On the ventral side (see Ahlborn) the brain is 
divided by a marked constriction into a pre-chordal and an 


384 W. H. GASKELL. 


epi-chordal portion. ‘The epichordal portion terminates in a 
projecting prominence, termed by Fritsch the conus post- 
commissuralis, and recognised by Ahlborn as the region 
of the ganglion interpedunculare. The ventral appearance of 
the brain is shown in Pl. xiii, fig. 6, of Ahlborn’s paper (8) 
after removal of the hypophysis and saccus vasculosus. If the 
brain is carefully dissected out so as to leave the saccus vascu- 
losus intact, and then examined in the fresh state, it is seen 
that the projection formed by the infundibular region, and 
also that of the conus post-commissuralis, stand out most 
prominently from the rest of the nervous material owing to 
their peculiar appearance: instead of appearing white and 
opaque like the rest of the nervous matter, they look like 
transparent jelly-like masses imposed upon the opaque white 
mass of the rest of the brain; so transparent is the material 
here that the brain-cavity shows through, as can be imagined 
from the inspection of Ahlborn’s fig. 6. 

Further, the brain lies in a mass of glandular-looking 
material composed of cells, and characterised by the amount 
of pigment which ramifies between the celis. This tissue 
clings largely to the brain when it is dissected out, so that the 
surface of the brain presents an appearance as in Pl. XXV, 
fig. 1. This thick glandular material is arranged in a 
definite manner on the surface of the brain; it forms a 
bilateral pigmented cellular organ, which may be looked upon 
as forming two lobes symmetrically placed on each side of the 
ventral constriction which marks the separation between the 
epi- and pre-chordal portions of the brain. From here, as 
starting-point, the tissue spreads in a fan-shaped fashion over 
the lateral portions of the epichordal brain, never reaching 
the mid-line dorsally, but stopping short at the choroid plexus 
iil. In the prechordal portion it passes towards the dorsal 
surface, round the infundibular projection, up to the choroid 
plexus 11, and surrounds the pineal eye and cerebral hemi- 
spheres, so as to fill up the brain-cavity anteriorly to the 
nervous system. In most of the sections I have shown the 
arrangement and structure of this tissue. 


VERTEBRATES FROM A CRUSTACEAN-LIKE ANCESTOR. 385 


Sxecr. 1.—The Cephalic Stomach. 


The brain of the Ammoceetes, then, differs markedly in ex- 
ternal appearance from that of higher Vertebrates; and one of 
its most striking peculiarities consists in the fact that nearly 
the whole of its dorsal surface is composed of membranous 
folds, and not of nervous material. The large amount of sur- 
face which these folds represent is manifest upon inspection of 
the external appearance of the brain, and is clearly visible in 
transverse sections such as are represented by figs. 2, 3, 9, Pl. 
XXV. These membranous folds form the choroid plexuses 11 
and iii of Ahlborn, and may be looked upon as a folded crumpled 
globular bag which stretches from the commencement of the 
fourth ventricle to the posterior commissure and ganglion 
habenule ; this folded bag has been constricted into two 
parts, which form the roofs of two cavities, viz. the fourth 
ventricle and the cavity of the mesencephalon, by the presence 
of a narrow constricting band formed by the fourth nerve and 
the incipient cerebellum (see also Ahlborn). This bag, if un- 
folded, would, in accordance with the view put forward in 
my previous papers, form the main part of the large globular 
cephalic stomach of the Crustacean-like ancestor; it is only 
possible to imagine its expansion on the dorsal side, for ven- 
trally the epithelial walls of this bag form the lining mem- 
brane of the ventral nervous masses of the brain, and in the 
epichordal portion of the brain at all events the internal 
surfaces of the most ventral portion are brought close together 
by the pressure of the two lateral nervous masses, thus forming 
the raphe (see p. 394). In other words, the infra-cesophageal 
and upper ganglia of the ventral chain, which, on my theory, 
form the epichordal portion of the brain, have by their growth 
nipped and compressed the ventral surface of the cephalic 
stomach which was lying dorsal to them. 

Lying on the anterior part of the cephalic stomach we see, 
in the position of the supra-cesophageal ganglia (see fig. 1, 
Pl]. XXV), the nervous masses known as the cerebral and olfac- 
tory lobes, optic thalami, ganglia habenulz, together with the 


386 W. H. GASKELL. 


pineal eye. These nervous masses have enclosed the walls of 
the stomach in this region to such an extent that only a small 
portion of its surface remains free, and is not utilised for the 
purpose of lining the cavities of the brain in this region. Its 
anterior rounded extremity forms the lamina terminalis, 1. ¢., 
the surface of which continued backwards up to the ganglia 
habenule is slightly folded, and spans the space between the 
two optic thalami forming the choroid plexus 1 of Ahlborn. 
Passing round to the ventral surface, the stomach wall forms a 
bulging known as the recessus chiasmaticus. Laterally it 
forms the lining of the optic thalami, and is compressed by 
the growth of the nervous matter of the cerebral and olfactory 
lobes so as to form the diverticula called by Ahlborn: the 
lateral ventricles (fig. 6, Pl. XXV; also Ahlborn’s fig. 37, Pl. 
xv). The appearance presented by a longitudinal vertical sec- 
tion in the middle line through the brain of an Ammocetes is 
given by Ahlborn in fig. 41, Pl. xvi; and if the choroid plexuses 
were dilated we should have an appearance somewhat as in the 
diagram on Pl]. XXVIII, fig. 29, where the red line represents the 
contour of the cephalic stomach of such an animal as Limulus, 
and the yellow line represents the modifications of contour 
which have occurred by the growth and compression of nervous 
material in the case of the Ammoccetes. In fact, it is perfectly 
clear that in this primitive form of Vertebrate the non-nervous 
epithelial bag which was originally the cephalic stomach of the 
Crustacean-like ancestor of the Vertebrate is more conspicuous 
and stands out more prominently with respect to the nervous 
system than in the higher Vertebrates; in other words, the 
relation between the size of the nervous system and of the 
cephalic stomach approaches to that which occurs in the Crus- 
tacean as we descend from the highest to the lowest Verte- 
brates. 


Sect. 2.—The Mouth and Gsophagus. 


Again, we see in this animal more clearly than in any other 
Vertebrate the remains of the old Crustacean cesophagus. 
From the central cavity a broad passage (the infundibulum) 


VERTEBRATES FROM A GRUSTACEAN-LIKE ANCESTOR. 387 


exists on the ventral side leading into the saccus vasculosus or 
saccus infundibuli, the thin walls of which come to the surface 
of the brain on the ventral side. The bag of the saccus vas- 
culosus is limited in front by the optic chiasma, which separates 
it from the recessus chiasmaticus; its cavity is continuous 
backwards into that of a well-defined tube which lies in the 
middle line and extends nearly up to the large ventral fissure 
which separates the prechordal and epichordal portions of the 
brain. This tube is apparently called by W. Miller (5) the 
processus infundibuli; it is called by Ahlborn (8) the thick- 
walled lobus infundibuli; and is compared by him with the 
unpaired lobus infundibuli of the Amphibia. It is figured 
£.%. in his drawings, and is shown in transverse section in 
figs. 27 and 28, Pl. xiv. The walls of this tube are composed 
of ependyma continuous with that of the central cavities of the 
brain, and as it approaches its termination its lumen becomes 
occluded by the coming together of its walls. Near its ter- 
mination it is in very many cases found to be lying outside the 
brain, being separated from the nervous matter by lines of 
pigment continuous with the pigmented tissue around the 
brain. The appearance which it then presents is given in fig. 3, 
Pl. XXV, which is a section through a carmine preparation of 
the brain of a specimen which had just undergone transforma- 
tion. In fig. 2 I give another section from the same series a 
few sections nearer the infundibulum, to show the commencing 
separation of this tube from the brain-substance. In fig. 2 the 
section has cut the last portion of the right ganglion habenule, 
and the notochord is not yet involved in the section; it corre- 
sponds very nearly with fig. 27, Pl. xiv, of Ahlborn. In fig. 3 
the notochord has just made its appearance, and the section 
passes through the posterior commissure ; it corresponds so 
closely with Ahlborn’s fig. 26 that the very next sections 
contain the two large nerve-cells pictured by Ahlborn in that 
figure. The comparison of his figure and mine shows that 
he had really before him the closed end of this tube, but 
he has misinterpreted its meaning and labelled it J/. B., or 
Meynert’s bundle. 


388 W. H. GASKELL. 


In his description (p. 285) of this figure he says, speaking 
of Meynert’s bundles, “ Dicht iiber der Haubeneinschniirung 
[i.e. the large ventral fissure which separates the pre-chordal 
and epichordal portions of the brain] lésen sich die Biindel 
auf; ein grosser Theil der Fasern fahrt pinselformig gegen die 
Mediane aus einander und bildet hier einen asymmetrischen, 
eigenthiimlich hellen und ausserst feink6rnigen Korper, welcher 
der Haubeneinschniirung direct aufgesetzt ist.” 

This description expresses very well the appearance of the 
termination of this occluded tube as it appears in my speci- 
mens, with the exception that I have not noticed any appearance 
of asymmetry in the peculiar transparent termination of this 
tube. The meaning of this tube is to my mind perfectly 
plain ; we see here more clearly than in any other Vertebrate 
that a median tube formed of ependyma, continuous with that 
of the central cavity of the brain, is lying freely on the 
surface of the brain in the exact position where the cesophagus 
ought to be found if the central cavity of the brain is the 
cavity of the cephalic stomach of a Crustacean-like ancestor. 
This tube, to my mind, cannot be called lobus infundibuli; 
it is not homologous with the paired lobi infundibuli of other 
fishes : it is the same in position and in meaning as the median 
tube which I have described (1) in the tuber cinereum of the 
sheep; it corresponds with the median prolongation from the 
saccus vasculosus which I have described in the brain of the 
dog-fish, where this same tube exists in addition to the paired 
lobi infundibuli. The difference in its apparent position in the 
dog-fish and in the lamprey is due simply to the more extensive 
cranial flexure which has taken place in the former animal. 

In fact, the paired diverticula from the canal of the infundi- 
bulum which form the cavities of the lobi infundibuli open 
into the infundibular canal between the saccus vasculosus 
and the third ventricle ; while the so-called lobus infundibuli 
of the lamprey, and I presume also of the Amphibian, is a 
posterior median prolongation from the saccus vasculosus 
itself, exactly as is found also in the dog-fish, in the sheep, &c. 

We see, then, in the lowest Vertebrate, the presence not only 


VERTEBRATES FROM A CRUSTACEAN-LIKE ANCESTOR. 389 


of the cephalic stomach, but also of the mouth and cesophagus, 
is much more clearly visible than in the higher forms. 
Another point worthy of note is the shape of the cavity of 
this median tube; it is always either round or compressed 
into a horizontal slit, never into a vertical slit; showing that 
the pressure to which it has been subjected by the growth of 
the nervous system is not lateral, but is in the dorso-ventral 
direction, as would naturally follow if the cesophagus of such 
a form as Limulus were compressed by the growth of the infra- 
oesophageal ganglia under which it lies. 

Embryologically, this tube and its termination form the 
anterior neuropore, which, as Ahlborn (4, p. 333) has pointed 
out, must be looked for in the neighbourhood of the hypo- 
physis rather than of the epiphysis. 


Sect. 3.—The Relation of the Infra-csophageal and 
Thoracic Ganglia to the Walls of the Cephalic 
Stomach. 


If the choroid plexuses are the free dorsal walls of the cephalic 
stomach thrown into folds, it follows that the continuation of 
the tissue of the choroid plexuses which lines the cavities of 
the brain must be the ventral portion of the stomach walls. 
Also the nervous masses which lie on the outer side of the 
lining epithelium, and form the grey and white matter of the 
brain, must, on the same theory, be composed of elements 
arranged in the same way, and of the same intimate structure 
as the nerve-masses which form the supra-cesophageal, infra- 
cesophageal, and thoracic ganglia of the Crustacean-like ancestor. 
Such appears to me to be undoubtedly the case ; and although 
I must wait for the proof of the complete homology of the 
different parts in the two nervous systems until the completion 
of this series of papers, yet it is well worth while to point out 
how strikingly the microscopic appearance of the brain of the 
Ammoceetes bears out the theory. I will first deal with the 
epichordal portion of the brain, i.e. with the relations of the 
infra cesophageal and thoracic ganglia to the stomach walls ; 


390 W. H. GASKELL. 


and then in a subsequent section treat of the supra-cesophageal 
ganglia. 

The intimate structure of the brain has been described so 
fully and well by Ahlborn (8) that it is unnecessary for me to 
treat of the different parts, section by section, as he has done. 
I will therefore confine myself to the description of general 
features as much as possible. 

In the whole of the epichordal portion the white and grey sub- 
stances form two well-defined layers, the former being compact, 
dense, and fibrous, while the latter is loose, transparent, and 
composed mainly of cellular elements. Further, the grey 
matter is itself divided into two very distinct layers in all 
Ammoccetes which are at all fully grown. The innermost of 
these two layers is formed by the columnar cells which con- 
stitute the lining epithelium of the central cavities; they 
are separated in the most marked way from the cells of the 
outermost layer of the grey matter, not only by difference of 
structure, but also by a distinct band of tissue, which presents 
an appearance as of a limiting membrane. In sections 
stained with hematoxylin the line of separation between the 
columnar cells with their dark-staining oval nuclei, and the 
cells of grey matter with round nuclei which stain less darkly, 
is very well shown. In fig. 4 I give a section through the 
commencement of the fourth ventricle, and in fig. 5 a 
magnified representation of a part of the same section. In 
fig. 7 I give a magnified section through a portion of the 
cerebral hemispheres. In both figures we see how clearly the 
lining layer of columnar cells forms a distinct membrane, 
which is marked off from the rest of the grey matter by a 
clear space. This clear space separates the oval nuclei of the 
lining epithelial cells from the neighbouring mass of cells 
with round nuclei which form the greater portion of the grey 
matter ; it looks as though the walls of the cells of the internal 
membrane were distended with some transparent substance. 
The nature of the contents of these cells is shown by treat- 
ment with osmic acid. I have cut acontinuous series through 
the brain of Ammoceetes at different ages after the whole head 


VERTEBRATES FROM A ORUSTACEAN-LIKE ANCESTOR. 391 


of the animal has been treated with osmic acid (1 per cent.). 
The result of such treatment is most striking, as is seen in figs. 
6, 11, Pl. XXV, and figs. 12, 13, Pl. XXVI. 

We see in all cases that the white matter of the brain stains 
of a greyish colour, and that the cells and cell nuclei of the 
grey matter stain also of a greyish colour, their contour being 
much less well defined than with carmine or hematoxylin pre- 
parations. The tissue of the cellular layer is less dense 
than that of the white substance, and, being different also in 
character, it stands out most conspicuously from the darker 
coloured white-fibred layer. Further—and this is what is so 
striking—the whole of this cellular layer is sprinkled over with 
intensely black fatty granules, which in the younger stages are 
seen to be more especially congregated in the neighbourhood 
of the lumen, as represented in figs. 13a—18e, Pl. XXVI. 

These fat-granules are confined to the grey matter ; and if 
a very few black dots are to be seen in the white matter, they . 
are invariably found to indicate the position of one of the 
nuclei similar to those found in the grey matter. At a 
later stage, when the Ammoceete is near the time of its 
transformation, the lumen of the brain-cavity is found to be 
marked out in the most extraordinary way by this layer of 
black fat-granules, which now form a dense black lining to the 
central canal, marking out not only that part of the cavity 
which is patent, but also the line of closure, wherever the 
original cavity has become closed by its walls coming into 
contact. The general appearance is shown in fig. 11, Pl. XXV, 
which represents a section, through the fourth ventricle, of an 
Ammoccete which was near the time of transformation. Exami- 
nation with a higher power (fig. 12, Pl. XX VI) shows that only 
the cell Jayers near the lumen are crowded with this dense mass 
of fat-globules, and that a very distinct limiting line of tissue 
separates the layer of fat-globules from the deeper lying cells 
of the grey matter. In among the radially arranged cells of 
this deeper layer fat-globules are seen scattered about here 
and there up to the edge of the white-fibred tissue of the 
medulla. 


392 W. H. GASKELL. 


A comparison of the osmic specimens with those stained by 
hematoxylin shows that the main portion of the fat-globules 
are situated in the position of the clear space which is so con- 
spicuous in the latter specimens ; in other words, the epithelial 
layer of the central nervous system, which is, upon my theory, 
formed by the columnar epithelium of the original stomach, 
is clearly differentiated from the rest of the grey matter at a 
certain stage in the development of the Ammoceetes by its cells 
becoming distended with fat, so that a marked limiting line 
between these cells and those of the grey matter is brought 
prominently into view. 

This accumulation of fat is due, in my opinion, to fatty de- 
generation of the old stomach walls, rather than to the absorp- 
tion of fatty particles, such as yolk-granules, which, I under- 
stand, have been found in connection with the central canal of 
the nervous system in the embryos of various Vertebrates. I 
have come to this conclusion from the consideration of the 
following facts. In the first place, these globules which stain 
so black with osmic acid are really of a fatty nature, because 
they dissolve away when the slide is placed in turpentine over a 
water-bath for some time. In the second place, they are due 
to degenerative changes, and not to absorption of yolk-particles, 
because they increase in number as the animal grows; in the 
youngest specimens which I possess—about 23 mm. long—the 
globules of fat are sparse and scattered about in the neighbour- 
hood of the lumen, as shown in figs. 13a—e, Pl. XX VI; while 
in the nearly full-grown Ammoceetes osmic specimens show the 
dense black line of fat at the base of the columnar epithelial 
cells which is so conspicuous in figs. 6, 8, 11, Pl. XXV. Also 
I have made an osmic preparation of one of the Ammocecetes 
(70 mm. long) which had been kept alive in the laboratory 
over a year without having increased in size. In this specimen 
the layer of fat-globules is not only present, but also the fatty 
change is much more marked than in fresh-caught specimens 
of the same size. 

Further, the changes which take place in the cells of the 
lining epithelium as the animal grows into its adult form are of 


VERTEBRATES FROM A OCRUSTACEAN-LIKE ANCESTOR. 393 


such a kind as to confirm the view that we are dealing here 
with true degenerative changes. We find, in the first place, 
evidence that the formation of the substantia centralis 
gelatinosa is partly due to the transformation of the cylin- 
drical cells which line the central cavities. In the second 
place, we find that, in those parts where the lining surfaces come 
into contact, a process of stitching together of the two sur- 
faces takes place, so as to forma seam, or raphe, as it is called 
by anatomists. This fusion of the two layers of cylindrical 
cells so as to form a seam is largely due to the fatty degenera- 
tion and breaking down of the cells. 

(1) The Formation of the Substantia Centralis 
Gelatinosa.—lIn the adult Petromyzon, Ahlborn describes 
(3, p. 247) the changes which take place in the grey substance 
as we pass from the spinal cord to the medulla oblongata ‘as 
consisting in part of an extraordinary increase in the so- 
called connective tissue or non-nervous elements, so that 
these, embedded in a thick tangle of the finest fibres, now 
form a broad zone around the lining epithelium and spread 
laterally in between the ganglionic elements of the grey sub- 
stance. They form the greater and compact central part of 
the grey substance, on the outside of which the ganglion-cells 
are imposed as a peripheral cortex.” 

In the adult condition, then, we find that the cellular 
portion of the grey substance is separated from the lining 
epithelial layer by a space which is composed of a complicated 
mass of the finest fibrils, some of which form supporting 
structures, others are probably nervous. The appearance of 
a basement membrane to the columnar epithelial cells has 
disappeared, and these latter cells themselves have altered 
very much in appearance; they can no longer be called 
columnar epithelium, for, as Rohon (6) has described and 
figured, and as is seen in fig. 10, Pl. XXV, the basal part of 
the cell is drawn into a long-tailed process, which is lost in 
the fibrillar tissue of the substantia centralis gelatinosa; 
the protoplasm of the cell is very scanty, so that the nucleus 
fills up the greater part of the body of the cell, and the nuclei 


394 W. H. GASKELL. 


themselves are no longer so closely packed together as in the 
younger stages, so that the appearance of carmine preparations 
of the brain in the animal after transformation is as is repre- 
sented in figs. 9, 10, Pl. XXV. This transformation of a 
columnar cell into a tailed cell which forms part of the support- 
ing tissue in, at all events, the substantia centralis gela- 
tinosa, the shrinking of the cell body, and the disappearance 
of a marked limiting layer at the base of the columnar epithe- 
lial cells, are due mainly to the fatty degeneration already 
described ; so, also, I imagine that the lining cells themselves 
are no longer so crowded together, because some of them 
have shrunk to such an extent that they no longer form properly 
developed cells, but present rather the appearance of strands 
and shreds of tissue. 

In so far, then, as the columnar cells of the old stomach wall 
assist in forming supporting tissue for the nervous elements 
of the brain, I am not prepared to say more than that they 
assist in forming the supporting tissue of the substantia 
centralis gelatinosa by their elongation into long-tailed 
processes in consequence of a process of fatty degeneration 
which they have undergone. 

(2) The Formation of the Raphe.—Further, the result 
of this breaking down of protoplasm into fat brings about not 
only the formation of a lining epithelium with long processes 
directed into the nervous matter, but also causes a raggedness 
in that part of the cells which is turned towards the lumen, 
so that wherever the cavity is closed by the coming together 
of its walls, as is the case in the medulla oblongata, the original 
lumen is bridged over by a ragged reticulum, the interlacement 
of the fibres of which tend to permanently close the lumen ; 
in other words, owing to the degenerative changes going on, 
the walls of the original cavity become, so to say, stitched 
together in the middle line, and in this way the raphe, 
which is so characteristic of the epichordal portion of the 
brain, is formed. 

It is especially instructive to notice that throughout the 
medulla oblongata the closed tube which forms the raphe, 


VERTEBRATES FROM A CRUSTACEAN-LIKE ANCESTOR. 395 


and the limits of which are clearly marked out by the layers 
of blackened fat-cells, extends close to the ventral surface, so 
that the nerve-matter with its groups of ganglion-cells giving 
origin to the segmental cranial nerves is situated on each 
side of, and ventrally to, the original large bag; in other 
words, a small ventral portion of the cephalic stomach has been 
nipped in this region by the growth of the two laterally situated 
masses formed by the upper ganglia of the ventral chain. By 
this lateral compression the original cavity has become first 
converted into a vertical slit, and then by the fatty degenera- 
tion of its cell-walls into an interlacement of fibres, known 
in anatomy by the name of the raphe. 

This formation of a raphe or seam by the lateral compres- 
sion of two nerve-masses can be followed up to the infundibu- 
luny; in other words, not only the upper ganglia of the ventral 
chain, but also the infra-cesophageal ganglia, have compressed 
a small ventral portion of the cephalic stomach, thus forming 
the raphe of the aqueduct. In this region, however (fig. 
3, Pl. XXV, r. aq.), the raphe no longer extends to the ventral 
surface of the brain, but, on the contrary, the ventral surface is 
occupied by a tube, the sides of which have been compressed as 
already mentioned, not so as to form a vertical slit, but, on the 
contrary, so as to form a horizontal slit. In other words, the 
growth dorsalwards of the bilateral infra-cesophageal mass has 
nipped and compressed a ventral fold of the cephalic stomach, 
with the result of forming a vertical slit, which ultimately be- 
comes the raphe of the aqueduct; and at the same time the 
increase in size of the whole infra-cesophageal nervous mass has 
compressed (as already explained) the cesophagus which lies 
between it and the wall of the cranial cavity, with the natural 
result of causing the lumen of the esophageal tube to take 
the form of a horizontal and not a vertical slit. Also we see 
that the presence and position of the horizontally compressed 
cesophagus in the infundibular projection is the reason why 
the raphe comes to the ventral surface in the middle line only 
in the epichordal portion of the brain. 

The appearance, then, of an interlacement of fibres in the 


396 W. H. GASKELL. 


ventral middle line of the epichordal portion of the brain of all 
Vertebrates, which is known by the name of the raphe, is due 
to the breaking down of the two apposed layers of cylindrical 
epithelial cells which originally lined the central cavity in this 
position. Doubtless the closeness of the apposition of these 
cells by which the lumen was obliterated was due very largely 
at first to the interlacement of the cilia with which these cells 
are provided, and not simply to commencing degenerative 
changes. 

With respect to the question whether the lining cells of the 
cavities of the central nervous system are ciliated or not, the 
literature of the subject shows that a considerable discrepancy 
of opinion exists. In a recent paper from Johns Hopkins 
University, Wightman (8) gives a list of observers for and 
against the occurrence of cilia in this region, and concludes 
from his own experiments that cilia are undoubtedly present 
in the ventricles of the frog. As to Petromyzon, Ahlborn (8, 
p- 249), speaking of the transition of the central canal of the 
cord into the cavity of the fourth ventricle, says that “the 
epithelium of the central canal, where it passes into the epen- 
dyma of the brain-cavities, undergoes a considerable elonga- 
tion of its cells ; the cilia are conspicuous to the view, and one 
can easily recognise a long oval nucleus in the body of the cell, 
as well as a tailed prolongation which passes deep into the brain 
substance.” Rohon (6) figures isolated cells of P. marinus 
with cilia attached ; and Owsjannikow (7) describes the epithe- 
lium in the neighbourhood of the optic thalami as composed of 
cells, each of which bears a single cilium. 

My preparations, especially those stained with osmic acid, 
present appearances as of ragged, irregular projections protrud- 
ing into the lumen of the cavity ; in some places these appeared 
to resemble a close-set fringe of cilia, in others they presented 
single projections or a loose sort of network. Such appearances 
might be caused by a breaking down of the edge of the cells 
turned towards the lumen, or they might be due to a ciliated 
edge, the cilia of which were more or less matted together and 
cut at different angles, as well as mixed up with the shreds of 


VERTEBRATES FROM A ORUSTACEAN-LIKE ANCESTOR. 397 


the coagulum of the albuminous fluid which the central cavities 
contain. 

Mr. Hardy kindly undertook to make specimens for the ex- 
press purpose of solving this question, and found conclusively 
that cilia are present in the cells lining the brain-cavities, and 
also on those of the choroid plexuses. The brain-cavities were 
opened, and a small portion of the lining epithelium, as well as 
of the choroid plexuses, examined in normal saline solution in 
the living condition ; in various places the action of the cilia 
upon small fatty particles and even red blood-corpuscles was 
very clearly visible. Other portions were taken, washed out 
with normal saline solution, and exposed to the action of osmic 
vapour ; sections then showed clearly a fringe of short cilia 
standing out from the lining epithelial cells. 

So far, in considering the relations between the cephalic 
stomach and the infra-cesophageal and thoracic ganglia, I have 
dealt mainly with the changes of the stomach wall in conse- 
quence of its altered function, and have merely incidentally 
mentioned that the compression of a ventral fold of the 
stomach by which the raphe is formed was due to the growth 
of the nervous masses of the fused infra-cesophageal and 
thoracic ganglia; I will now shortly consider the structure of 
that portion of the brain of the Ammocetes which I look 
upon as formed by these ganglia: at the same time I feel 
that I cannot, as yet, compare the structure of the two nervous 
systems in such minute detail as I feel convinced further 
investigation will enable me todo. I desire at present merely 
to point out that there is nothing in the structure and arrange- 
ment of the nervous elements in the epichordal portion of the 
brain which is inconsistent with the supposition that we have 
here the fused infra-cesophageal and thoracic ganglia of the 
Crustacean-like ancestor. 

The structure of the Arthropod nervous system may be 
described as consisting of nervous elements, held together by 
a framework of supporting tissue, the nervous elements con- 
sisting of—(1) nerve-fibres which connect the different gan- 
glia together and form the outgoing nerves; (2) groups of 

VOL. XXXI, PART I1l,—NEW SER. D D 


398 W. H. GASKELL. 


large pear-shaped ganglion-cells, some of which have relation 
to the outgoing metameric nerves; (3) masses of closely packed 
small cells which are presumably nervous in nature; and (4) 
a peculiar mass of fine nerve-fibrils, known by the name of 
the reticulated substance, or ‘‘ Punctsubstanz.”? These two 
latter elements increase in amount as we approach the supra- 
cesophageal ganglia. Corresponding to these separate elements 
we find in the epichordal portion of the brain of the Ammo- 
coetes externally—(1) the white substance composed of nerve- 
fibres; internally the grey substance composed of (2) groups 
of large ganglion-cells, which are partly the nuclei of the seg- 
mental cranial nerves, and partly are connected with the large 
Miillerian fibres ; (3) a thick layer of closely packed small cells 
in which the groups of larger cells are embedded (see figs. 
12, 14, Pl. XXVI). These small cells are arranged like 
clusters of berries, and are called therefore by Ahlborn “ beeren- 
formige Zellen ;”? their numbers increase enormously, as he 
points out, when the spinal cord passes into the medulla oblon- 
gata. (4) Internally, of all in the adult condition, a thick 
layer of very fine fibrils, interlacing in all directions, which 
forms the substantia centralis gelatinosa, and which, as already 
mentioned, increases markedly as we pass from the spinal cord 
into the medulla oblongata. 

If, as I believe, these four structures are the same as in the 
nervous system of the Crustacean, then it follows that the 
supporting tissue of the Ammoccetes’ brain is derived from the 
supporting tissue of the Crustacean nervous system mainly, 
and only to the slight extent already mentioned, from the 
modification of the walls of the old stomach. In the sub- 
stantia centralis gelatinosa we have, on the view put forward, 
the meeting-place of so many different structures; here we 
must find the remains of the connective-tissue elements sur- 
rounding the stomach, and those which surround the nervous 
system ; here are the modified extremities of the columnar 
stomach-cells as already mentioned ; here also is that whorl of 
fine fibrils partly nervous (corresponding to Gerlach’s plexus), 
partly supporting, known by the name of the reticulated sub- 


VERTEBRATES FROM A CRUSTACEAN-LIKE ANCESTOR. 399 


stance, or ‘‘ Punctsubstanz ;” and it is very suggestive to read 
Ahlborn’s description of the relation between the masses of 
small berry-like cells, and the reticulated tissue of the sub- 
epithelial spongiosa. Speaking of the grouping of the small 
cells in the central grey matter, he says (3, p. 253) they 
can be observed especially well about the middle of the aque- 
duct where they are not mixed up with cells of other kinds, 
and where they are arranged in regular rows, hanging together 
like the berries on a grape-vine. He goes on to say, ‘ Die 
Spitzen der Zellen sind nach aussen gewandt, und die feinen 
daraus hervorziehenden Fadchen treten in die zwischen den 
Reihen befindlichen engen Zwischenraume, wo sie sich mit 
Fortsatzen aus der benachbarten Zellreihe zu vereinigen 
scheinen. Querschnitte (Osmium) Zeigen daher an dieser Stelle 
Bilder, als seien die kleinen Zellen wie die Beeren einer 
Traube mit einander verbunden. Die Traubenspindel ist 
hier jedoch meist kein einfacher Faden oder ein glattes, 
gleichsam durch die Komposition der Beerenstich entstand- 
enes feines Biindelchen, sondern sie zeigt die Higenschaften 
der peripherischen Neuroglia (oder der subepithelialen Spon- 
giosa), mit welcher sie unmittelbar zusammenhangt; nur an 
besonders giinstigen Stellen kann man beobachten, dass sich 
aus den Zellfortsaitzen zunichst ein feiner Faden bildet, der 
dann in ventraler Richtung gegen die Spongiosa zieht und 
sich in derselben aufzulésen scheint.” 

Seeing, then, that the whorl of fine fibrils which form the 
substantia gelatinosa centralis (or sub-epithelial spongiosa) is 
known to consist, in the higher Vertebrates, largely of Ger- 
lach’s plexus of fine nerve-fibrils, Ahlborn’s description is 
wonderfully like a description of the relation between the 
small cells supposed to be nervous, and the reticulated tissue 
(Punctsubstanz) of the nervous system of an Arthropod. 


Sect. 4.—The Relation of the Ventral Ganglia to 
the Walls of the Intestine. 


In the region of the spinal cord where, according to my 
theory, the nervous matter of the ventral chain of ganglia has 


400 W. H. GASKELL, 


surrounded the straight intestine of the Crustacean-like animal 
from which the Vertebrates arose, we find a decided difference 
of structure from what occurs in the higher parts of the central 
nervous system. Here the central grey matter is very much 
subordinated to the surrounding white matter. The white 
matter is remarkably free from any scattered nuclei in it, and 
contains a large number of giant nerve-fibres which run 
longitudinally. The grey matter contains but little of the 
spongy reticulated substance, and very few of the small 
nerve-cells with circular nuclei as compared with the higher 
parts of the central nervous system. 

It is significant, in connection with the meaning which I 
ascribe to these small nerve-cells, to find that Ahlborn, after 
criticising the view ascribed to Freud, that fibres of the 
posterior roots are in connection with certain large cells, 
speaks as follows (8, p. 242): “ Dagegen glaube ich nicht mehr 
bezweifeln zu dirfen, dass die dorsalen Nervenwurzeln wenig- 
stens zum Theil thatsiichlich ihren Ursprung in den ‘ kleineren 
Zellen’ Reissner’s nehmen.”’ In addition to these small nerve- 
cells, Reissner (9) describes two groups of large ceils—(1) an 
inner group situated bilaterally close to the central canal on 
the dorsal side, and (2) an outer group situated near the ex- 
tremity of the elongated band of grey matter. This latter 
group is recognised by Ahlborn and others as corresponding 
to anterior horn-cells. The inner group were the “ Hinter- 
zellen,” supposed to be connected with the fibres of the 
posterior roots. Ahlborn, however, can find no trace of such 
connection, and he cannot say what their connections are; 
all he finds on sagittal sections are (8, p. 242) “ kurze, 
starke, nach vorn (nasalwarts) gerichtete Zellfortsitze und 
eben solche feinere, die sich sehr schnell in rein dorso-ventraler 
Richtung (nach oben) auflésten.” 

In addition to these three kinds of cells, which are all 
nervous, a fourth kind is found which are recognised by all 
observers as non-nervous. These cells are described by Ahlborn 
as lying thickly crowded near the central canal, and diffused 
over the whole of the grey matter. He then goes on to say 


VERTEBRATES FROM A CRUSTACEAN-LIKE ANCESTOR. 401 


(8, p. 244), “Sehr auffallend ist ihre grosse Ahnlichkeit mit 
den Epithelzellen des centralkanals die sich von ihnen nur 
durch die oberfliichliche Lage unterscheiden.”? Although he 
looks upon it as highly probable that they are derived directly 
from the epithelium surrounding the central canal, he con- 
siders that the proof must be left to embryological investiga- 
tion. 

Shipley (10) does not deal with this point. Scott (11, 
p. 274), however, describes how the shape of the spinal cord 
is at first like that of other Vertebrates with a large central 
canal, and how a portion of the canal closes up when the cord 
by its elongation forms its well-known flattened ribbon-like 
shape. He does not, however, attempt to trace the fate of 
the cells of this closed-off portion. In connection with this 
paper of Scott I may mention that he still holds to the belief 
that the epithelium of the central canal is formed by the 
epidermic layer of epiblast passing down into the nervous 
layer, notwithstanding the criticism of Shipley. Naturally, as 
far as I am concerned, Scott’s explanation of the appearances 
presented by his sections is entirely in accord with my theory, 
that this epithelium lining represents the original intestinal 
wall. Although these observations of Scott require extension, 
they make it possible, if not probable, that the diminution of the 
central canal gives rise to the formation of groups of outlying 
elements, derived from its epithelial walls in a similar manner 
to that noticed by Corning (12), who describes the substantia 
gelatinosa Rolando as formed by a lateral and nipped-off 
fold of the original embryonic tube. A still further piece of 
evidence in favour of considering these cells to be derived from 
the epithelium of the central canal, i.e. from the walls of the 
old intestine, is afforded by osmic specimens, which show the 
presence of fat-globules in the epithelial cells of the central 
canal, and the extension of those fat-globules in two lateral 
lines, corresponding exactly to the position of these non- 
nervous cells. 

Here, then, if anywhere, we find a tissue corresponding to 
the myelospongium of His (18), and it is possible that in the 


402 W. H. GASKELL. 


spinal cord a system of supporting elements derived from the 
walls of the old intestine, as I have suggested in my former 
papers, may be formed after the fashion described by His. 
At the same time I agree with Ahlborn that we cannot look 
upon the epithelial layer of the central canal as many-layered 
until the origin of these small cells has been demonstrated 
embryologically. Certainly, as far as can be judged from 
His’s (14, fig. 30) figure of a section of the cord of a larva of 
Petromyzon fluviatilis 6 mm. long, the epithelial layer at 
that age is only one cell thick, and also from his figure and 
description there is no sign of any supporting tissue elements 
(myelospongium) outside the layer of epithelium, but only 
of nervous elements, so that any formation of supporting 
structures from this epithelium layer must take place at a 
later stage than in an Ammoceetes of 6 mm. long. In Scott’s 
description of the closure of the canal his figures are drawn 
from larve of 7 mm. in length and upwards. 


Sect. 5.—The Cephalic Liver and its Duct. 


We have already noticed how clearly the fatty degeneration 
of the columnar epithelium cells which line the cavities of the 
brain enables us to trace out the original limits of those cavities 
even when the lumen has been closed by the approximation of 
their walls, as in the formation of the raphe. In the same way 
any diverticula from the main cavity will be disclosed by the 
black staining of the fat-globules in its walls, even if its lumen 
has entirely disappeared. 

In the medulla oblongata the vertical slit which forms the 
raphe comes close to the surface on the ventral side (figs. 
13a, 14, Pl. XX VI); on the other hand, in the region of the 
aqueduct, as seen in fig. 3, Pl. XXV, the raphe is short, and 
only extends a short distance on the ventral side. Now the 
series of sections show that the most ventral portion of the 
fold which forms the raphe ceases abruptly at the projection 
of the conus post-commissuralis, its termination coincid- 
ing with the position of the so-called ganglion interpe- 
dunculare. 


VERTEBRATES FROM A ORUSTACEAN-LIKE ANCESTOR. 403 


Sections through this region show indications of a separa- 
tion of the raphe into two parts—a ventral one which terminates 
in the conus post-commissuralis, and a dorsal one which 
continues without interruption as the raphe of the aqueduct. 
The appearance (fig. 11, Pl. XXV) is as though the ventral 
part was formed from a separate small tube lying in the middle 
line ventrally, while the dorsal part was formed by the com- 
pression of the ventral portion of the large cephalic stomach 
as already mentioned. In some cases the separation of this 
ventral tube from the rest of the raphe is much more con- 
spicuous than in others, and in one instance it was remarkably 
evident. 

In figs. 13a—18e, Pl. XXVI, I give a selection out of the 
series of sections, to show the formation of this tube. The 
Ammoceetes from which these sections were cut was by no means 
fully grown ; its length was not measured, but judging from the 
size of the sections I should estimate it somewhere about 40 
mm.in length. The sections are all through the epichordal 
portion of the brain, and, as is seen by the diminishing size of 
the notochord, are numbered in the headwards direction. 

In the first section of the series, fig. 13a, the raphe of 
the epichordal portion of the brain is seen to extend close to 
the ventral surface, and we see from the shape of the section 
the commencement of a constriction on each side of the 
external surface of the white matter; this constriction marks 
the beginning of the separation from the rest of the brain of 
that portion of nervous matter which forms the ventral projec- 
tion known as the conus post-commissuralis. 

In the succeeding sections we see how this indentation be- 
comes deeper and deeper, until at last, as in figs. 13d, 13¢e, this 
projecting mass is entirely separated from the rest of the brain 
by the strongly pigmented tissue which surrounds the brain, 
and marks out most clearly the extent of the lateral indenta- 
tions. Further, the sections show clearly that this portion, 
which is thus cut off, contains the continuation of the ventral 
portion of the raphe in the shape of a tube whose walls are 
marked out with the usual fat-globules, and whose lumen has 


4.04. W. H. GASKELL. 


disappeared in consequence of the closure of its walls by the 
lateral compression of the nervous material. 

On the one hand, then, this particular specimen shows what 
is only indicated in others, that the remarkable ventral projec- 
tion known by the name of the conus post-commissuralis 
is due to the growth of nervous material around a ventral 
median tube, which arises from the cephalic stomach near its 
pyloric orifice ; on the other hand, it shows that the closure of 
this tube has taken place in such a manner as to form a 
vertical slit by the compression of the nervous material on 
each side, just as in the case of the formation of the raphe, 
and that the tube comes to the surface of the brain in a very 
similar manner to the tube which forms the old cesophagus 
already described in a former section. In both cases the 
structure of the actual end of the tube is a reticulum filled 
with transparent material, so that in the fresh condition the 
end of the conus post-commissuralis and the infundi- 
bular projection are so transparent as to allow the central 
cavity to be dimly seen through their walls. 

The nervous material immediately around the raphe is 
similar to that surrounding the ependyma, and is composed 
therefore mainly of the small berry-like cells already spoken 
of; some of these cells continue onwards around the ependyma 
of this ventral tube or duct, as it may be called, and so in com- 
bination with the broken-down epithelium of the duct walls 
gives rise to the structure known as the ganglion inter- 
pedunculare. The connection of the ganglion interpedun- 
culare with the cells of the central grey matter is noticed by 
Ahlborn, who says (8, p. 254) “es ist nicht zu bezweifeln, 
dass die Zellen des Ganglion interpedunculare thatsachlich aus 
dem centralen Bodengrau hervorgegangen sind.” 

We see, then, that the extraordinary shape of the brain on 
the ventral side, with its two projections in front of and behind 
the large ventral constriction, is due to the presence in the 
middle line of two tubes, both of which pass from the central 
cavity to the surface: the one tube, compressed to form a 
horizontal slit, leads out of the third ventricle, i.e. out of the 


VERTEBRATES FROM A CRUSTACEAN-LIKE ANCESTOR. 405 


anterior end of the cephalic stomach, and is directed tailwards, 
terminating at the commencement of the notochord; this 
forms the remnant of the original csophagus. The other 
tube, compressed to form a vertical slit, leads out of the fourth 
ventricle, i. e. out of the pyloric end of the cephalic stomach, 
and is directed headwards, terminating also near the commence- 
ment of the notochord. This latter tube is in my opinion the 
remnant of a single median duct of the cephalic liver, which in 
the Crustacean ancestor opened into the pyloric end of the 
stomach on the ventral side. In the diagram already referred 
to (Pl. XXVIII, fig. 29) I have indicated how the liver-duct 
and cesophagus from the Crustacean cephalic stomach would 
form the projections known as the conus post-commissuralis 
and the infundibular projections. If this interpretation is 
true, we ought to find evidence of the vestiges of the large 
cephalic liver in connection with the termination of this duct. 

In the description of the external surface of the brain I 
have already described how the brain is covered over with a 
glandular-looking organ containing a large amount of pigment 
arranged in irregular clumps and lines, and I have described 
also the general arrangement of this tissue on the brain surface. 
It forms what is called by Ahlborn the “arachnoidale Fiill- 
gewebe,” and is looked upon by him and other observers, 
e.g. Sagemehl (15), as a peculiar kind of connective tissue 
somewhat allied to fat tissue. 

Peculiar the tissue undoubtedly is, but that it bears the 
slightest resemblance to any form of connective tissue, whether 
fat-bearing or otherwise, I entirely deny. It presents the 
exact appearance of a compact mass of large gland-cells belong- 
ing to some such organ as a liver, each cell containing a 
central nucleus surrounded by the remains of the protoplasmic 
cell-contents. These cells are pressed closely to each other in 
all directions, so as to form a bilaterally symmetrical organ as 
described on p. 384, the glandular nature of which is in my 
opinion fairly evident. The arrangement of the cells is seen 
in fig. 16, Pl. XXVI, which represents a magnified portion 
of an osmic preparation from the series represented in figs. 


406 W. H. GASKELL. 


6, 8, Pl. XXV, and 20a—20d, Pl. XXVII. The body of 
each cell contains a homogeneous semi-fluid substance, which 
stains but slightly with osmic acid. In the centre of this 
is seen in some cells an irregular brownish mass, in others 
‘a more clearly defined large round nucleus, surrounded with 
some of the same irregular brown granular-looking débris. 
In other cells nothing is seen but the clear slightly-staining 
fluid. These appearances evidently mean that the cells com- 
posing this substance have so much thickness that some sections 
pass through the nucleus; others through the scanty granular 
strands of material surrounding the nucleus, which represent 
the remains of the protoplasm of the cell ; and others still nearer 
the edge of the cell, where nothing but clear fluid is to be 
found. The same appearances are seen in whatever direction 
the section is made, showing that these cells are solid poly- 
gonal bodies pressed together in all directions, without any 
sign of being in connection with specially arranged connective- 
tissue elements either in one direction or the other. Again, 
that they have no connection with adipose tissue is evident, 
for in the first place they do not stain like fat. A small fat- 
globule may be seen here and there, but that is all. All the 
deep blackness seen in the osmic preparations is due to the 
presence of pigment in between the cells. In the second place, 
the nucleus lies in its natural position in the centre of the cell, 
and the semi-fluid contents lie simply in between the meshes 
of the degenerated-looking protoplasmic material which sur- 
rounds the nucleus and spreads thence towards the periphery. 
Again, staining with hematoxylin or carmine gives appear- 
ances as drawn in fig. 15, Pl. XX VI. Here, again, we see that 
the nucleus stains but poorly, and the staining material in the cell 
consists of a broken-down network which starts from the nucleus 
and forms scanty strands which pass towards the periphery. 
In fact, the structure and massing of this “ arachnoidale Fill- 
gewebe ” is precisely what one would expect if it represented 
the degenerated remains of the cephalic liver of the Crustacean- 
like ancestor of the Ammoceetes. 

I have already described its general arrangement on the 


VERTEBRATES FROM A ORUSTACKAN-LIKE ANCESTOR. 407 


surface of the brain, and I will now briefly point out how the 
examination of sections still more clearly defines the limits of 
this tissue. 

A series of sections, both transverse and horizontal, through 
the whole brain show that it forms two lateral masses which 
do not spread dorsally beyond the choroid plexuses, so that 
the latter, with the blood space surrounding them, are pressed 
against the roof of the cranial cavity; anteriorly at the 
termination of the choroid plexus ii it reaches to the dorsal 
roof, and fills the spaces round the pineal eye and ganglion 
habenule, as seen in figs. 20a—20d, Pl. XXVII, spreading 
over and in front of the cerebral lobes, even to the space between 
the two olfactory nerves, as in fig. 6, Pl. XXV. On the ventral 
side it does not exist in the infundibular region, for here the 
pituitary body and the infundibular projection come into close 
contact with the floor of the cranial cavity. In the region of 
the medulla oblongata it is at first very scanty on the ventral 
side, but thickens as the termination of the conus post-commis- 
suralis is reached, until at last in the region of the large ventral 
commissure it attains a considerable thickness, so that here, 
and here only on the ventral side is there anything in the 
nature of a starting-point, from which it spreads so as to 
form two symmetrically placed bilateral organs; at the junc- 
tion of the medulla and the cord the cells pass round to the 
dorsal side, but here (fig. 4, Pl. XXV) they are no longer com- 
pact and thick, but isolated and scattered. We are here clearly 
on the fringe of the compact organ, isolated cells from which 
are found far down the spinal canal. 

At the termination of the conus post-commissuralis, in the 
deep ventral fissure where alone anything approaching to a hilus 
is to be found, we see, as in fig. 13e, Pl. XX VI, how the lines of 
pigment in between these large cells are massed together ; and 
we see how the termination of the tube which I have spoken of 
as the liver-duct is lost in among this pigment mass at the very 
place where it ought to come to the surface if the cells of this so- 
called arachnoidal tissue are in reality the cells of the so-called 
liver of the Crustacean-like ancestor, and this duct is the 


408 W. H. GASKELL. 


liver-duct opening into the pyloric end of the cephalic stomach. 
Further, as I have mentioned in my paper in ‘ Brain’ (2), 
it is absolutely necessary to explain why so large a portion of 
the cavity of the brain-case is occupied with this peculiar tissue 
in the lowest Vertebrates only. The one explanation which 
is offered, as far as I know, is no explanation at all, viz. that 
the fat-tissue fills up the space between brain and cranial walls 
in order to support and steady a brain which is too small for 
its case. The natural rejoinder to this is, according to modern 
views of evolution, why is the brain too small for its 
case ? and why also is not the spinal cord also supported by 
the same kind of tissue’—for here, as well as in the brain, 
we find the cavity in which the cord lies is very much larger 
than the cord itself; and yet, apart from a few large cells, the 
only solid structures which are found in the space around the 
cord are irregular strands of connective tissue, as seen in 
fig. 19, Pl. XX VII, formed, according to Ahlborn, of elastic 
fibres and “ Schleimzellen.” 

In accordance with evolutionary ideas, the natural answer 
to this question is that this peculiar tissue, which is different 
from any other, the cells of which appear degenerated, which 
contains lines of pigment between its cells, which is found 
only in the lowest Vertebrates, and is gradually pushed out of 
existence in the higher classes as the brain increases in size, 
fills up the space around the brain because it represents some 
pre-existing organ which was of importance to the animal 
from which the Vertebrate sprang. Such an organ is clearly 
the cephalic liver of the Crustacean-like ancestor—a conclusion 
which is in perfect harmony with not only the rest of the 
arguments in this paper, but also with those which I hope to 
bring forward in the series of papers following this one. As 
to the tissue round the spinal cord, it is natural that the cells 
of the liver should not be found in any great quantity there, 
and it is possible that the loose connective tissue which is 
there found is the remnant of the connective and muscular 
tissues which originally were situated in the neighbourhood 
of the intestine, along the length of the body of the Crustacean, 


VERTEBRATES FROM A ORUSTACEAN-LIKE ANCESTOR. 409 


One of the most striking characteristics of this degenerated 
liver-tissue is the presence of so much pigment in between its 
cells and lining the surface of the brain. What, then, is the 
meaning of this pigment ? 


Szect. 6.—The Formation of Pigment. 


The almost universal source of pigment in the Vertebrate 
kingdom is the colouring matter of the blood; we must, there- 
fore, look especially to the destruction of red blood-corpuscles 
in order to account for the presence of pigment in the majority 
of instances. 

Recently a paper has appeared by Hunter (16), who has 
been occupied for a long time in the investigation of the nature 
of the destruction of the red blood-corpuscles. He concludes 
that pigment is formed from the hemoglobin of the blood- 
corpuscles in two distinct ways, called by him respectively 
Active and Passive destruction, and defined as follows : 

Active destruction is that form of disintegration of the 
red blood-corpuscles where the hemoglobin is liberated and 
escapes into the plasma ; it is the change in the blood-corpus- 
cles which is caused by the action of water, various salts, &c. 
The chief evidences of this active destruction are—(1) the 
formation of bile-pigments by the liver, to which organ the 
hemoglobin thus set free is mainly carried ; (2) to a secondary 
and altogether subsidiary extent in health, the formation of 
blood-pigment. 

The chief characters of this blood-pigment are the small 
uniform size and spherical shape of its individual particles. 
The size of the red corpuscles have no influence on the size of 
the pigment particles resulting from active destruction. It is 
the same in mammals as in birds or amphibians. The pig- 
ment is formed from free hemoglobin. 

Passive destruction of the blood-corpuscles, on the other 
hand, is a slow and gradual decay of the red blood-corpuscles. 
The hzmoglobin remains in the corpuscle to the last, and 
becomes gradually converted into a globule of inert blood- 
pigment. 


410 W. H. GASKELL. 


It is the mode of death of the red corpuscles which would 
occur were the blood subjected to no other changes than those 
involved in carrying on its respiratory functions. This change 
can be best studied in extravasated blood. 

The chief evidence of passive destruction is the presence of 
pigment, and in health the chief organs in which this evidence 
is to be found are the spleen and hone marrow. The chief 
characters of this pigment are the large and varying size of 
the pigment particles and their irregular shape. The larger 
the red corpuscles of the animal the larger are the pigment 
particles which result from their passive destruction. He 
attaches special importance to this difference in the character 
of the pigment in the two varieties of blood destruction. From 
it alone he is able to say in any particular case whether the 
pigment has been formed in situ from red corpuscles, or from 
free hemoglobin carried thither in solution. When once 
formed this blood-pigment is remarkably resistent to the 
action of reagents. It can be recognised long after the death 
of the original blood-corpuscles. Hence the amount of pigment 
found in the organs specified affords a most reliable indication 
as to the amount of passive destruction that has occurred. 

Active destruction depends on certain changes in the blood, 
most marked during digestion, and occurring chiefly within 
the blood of the portal system. The spleen, however, is par 
excellence the seat of active blood destruction. Active 
destruction in this organ is specially favoured by certain struc- 
tural features, viz. (1) slowness of circulation, (2) closeness of 
relation of the cells of the pulp to the blood flowing through 
it, (3) capacity for accommodating large and varying quantities 
of blood. 

The same features favour also the seizure of red corpuscles 
in process of becoming effete by slow and gradual decay. The 
spleen is thus also the great seat of passive destruction, and is 
hence the chief seat of pigment accumulation resulting from 
this process. Next to the spleen the red bone marrow is the 
most important seat of passive destruction, the structural 
features here being closely similar to those in the spleen. 


VERTEBRATES FROM A CRUSTACEAN-LIKE ANCESTOR. 411 


The observations of Hunter, then, point to the conclusion 
that in health wherever the blood tends to stagnate the con- 
ditions are most favorable for passive destruction, and therefore 
for the accumulation of pigment. As he shows, all active cells 
in immediate relation to the stagnating blood play a part in 
converting the hemoglobin of the original blood-corpuscles 
into blood-pigment. Around extravasated blood it is the cells 
of leucocytes and of connective-tissue nature which carry on 
this process ; in the spleen and bone marrow the still more 
active splenic and marrow cells. 

The same conditions as are present chiefly in the spleen and 
red bone marrow in the case of the higher Vertebrates occur 
freely in different parts of the body in the Ammocetes owing 
to the peculiarities of its circulation, for the characteristic of 
the vascular system of the Ammoceetes, especially before trans- 
formation, is the abundance of large blood spaces in different 
parts of the body. These blood spaces have been somewhat 
strangely described as lymph spaces containing red _ blood- 
corpuscles. In them, as in the spleen pulp, the blood circu- 
lates slowly and with difficulty. In addition, important and 
extensive changes in the arrangement of various parts of the 
body take place during metamorphosis. In fact, all the con- 
ditions necessary for passive blood destruction are present in 
the most pronounced manner. 

The examination of the Ammoceetes by dissection and sections 
shows that masses of pigment are found in very definite arrange- 
ments in many different regions of the body. Apart from such 
places as the eye and the skin, we find pigment most markedly 
in connection with the branchiz and with the pronephros, In 
both these cases the pigment is found in close connection with 
the vascular arrangements. Thus the tubules of the pronephros 
of the Ammoceetes are enclosed in a large lacunar blood space, 
the walls and septa of which are strongly pigmented. In the 
branchial regions masses of pigment are found forming a 
coarse network, the meshes of which are full of blood-cor- 
puscles, so that the pigment here, just as in the pronephros, is 
situated in the walls of the vascular system. The heart itself 


412 W. H. GASKELL. 


and the large aorta springing from it are absolutely free from 
pigment. It is the walls of the large venous spaces, where the 
blood flow is slow, that are so markedly pigmented. 

Further, in the Ammoceetes, especially near the time of 
transformation, we find evidence of changes going on in the red 
blood-corpuscles, remarkably like those which Hunter dis- 
tinguishes as passive destruction. Here and there, in the 
neighbourhood of the layers of pigment which have been 
already formed, especially in the region of the upper branchie, 
we see red corpuscles, the nucleus of which does not present 
the usual finely granular appearance, but, on the contrary, is 
more or less hidden by large strongly refracting globules of a 
yellowish-brown tint: in cases where these globules are numerous 
the blood-corpuscle appears to carry upon its surface an irre- 
gular-shaped mass of dark brown, almost black pigment, which 
presents the same appearance as the thinner, more translucent 
portions of the neighbouring pigment masses. Such appear- 
ances point strongly to the conclusion that the pigment in the 
branchize and other places is due to the accumulation in these 
places of blood-corpuscles which have undergone passive de- 
struction similar to the formation of pigment in the spleen. 

This conclusion is rendered certain by the examination of 
individuals in which metamorphosis has been delayed. In 
September, 1888, I brought with me from Hind Head, in 
Surrey, a large number of Ammoceetes of different sizes, vary- 
ing in length from about 25 to 130 mm. These were placed ina 
basin containing sand and weed, brought from the stream in 
which they were living, and the water in the basin was renewed 
occasionally. A considerable number of these were killed soon 
after their arrival at Cambridge ; about twenty, however, were 
left alive, and have continued to live in the sand up to the 
present time, October, 1889, i.e. more than a year after they 
were caught. They show no sign of transformation, and have 
apparently not grown in length, some being still as small as 
28 mm., others from 40 to 70 mm. I have made series of 
sections of these Ammoceetes through different regions of the 
body, the animal having been stained and prepared for section- 


VERTEBRATES FROM A ORUSTACEAN-LIKE ANCESTOR. 413 


cutting as a whole, so that all parts of it had undergone 
the same preliminary treatment. These sections show that the 
heart and large blood-vessels in connection with it are full of 
blood-corpuscles of normal character, amidst which it requires 
careful search to find a single pigment-bearing corpuscle; 
while, on the other hand, the blood spaces of the branchie, 
especially the upper ones, are so loaded with corpuscles bear- 
ing masses of pigment of all shapes and sizes as to give 
the appearance of an injection of a pigmented mass into the 
branchial vessels. 

In fig. 18 I have drawn a group of corpuscles as they 
appeared in a section of one of the upper branchial blood 
spaces. a and 6 are normal blood-corpuscles ; in the others the 
nucleus is more or less concealed by the dark-yellow brown irre- 
gular clumps of pigment: the colour is not shown in the figure. 

These pigmented blood-corpuscles are found lying singly or 
massed together, especially in the upper branchial vascular 
spaces, often in close contiguity to the pigmented walls of 
these spaces ; and it is frequently difficult to decide whether a 
mass of pigment belongs to a blood-corpuscle or to the pig- 
mented walls. 

Further, the comparison of these sections with those of a 
fresh-caught Ammoceetes of the same size shows that the former 
is much more pigmented than the latter, and that the increase 
of pigment is due to the accumulation of pigment in those 
places where pigment is always found, viz. in the blood spaces 
of the branchiz, and other places where stagnation is liable 
to occur. ‘The result, then, of preventing metamorphosis by 
placing the animal ina condition of impaired nutrition—a con- 
dition which, according to Hunter, is the one most favorable 
to passive blood destruction in health—is to increase largely 
the passive destruction of the red blood-corpuscles in those 
places where the blood is liable to stagnate, and concurrently 
to increase the deposition of pigment in those places. This 
thickening of the pigmented walls and septa of the blood 
spaces naturally must lead to a diminution of their size, and 
so to a quickening of the current through them, and con- 

VOL, XXXI, PART II]L—NEW SER. EE 


414, W. H. GASKELL. 


sequent diminution of the tendency to stagnation. In addi- 
tion to this possible cause of narrowing of the bed of the 
stream another factor in all probability comes into play, viz. 
the formation of fibrous tissue around and amidst the pig- 
ment masses. The presence of pigment in the walls and 
trabeculee of the blood spaces is due to the circumstance 
already noted, that all active cells seize on inert pigment with 
which they are brought into relation. If, as may sometimes 
happen, there is entire stagnation of a mass of these effete 
blood-corpuscles, then the changes in this mass will correspond 
in all respects to those met with in the organisation of a 
thrombus, viz. the penetration and final replacement of the 
blood mass by cells of connective-tissue nature ; the subsequent 
contraction of this newly formed tissue, just as in the forma- 
tion of a cicatrix would convert the original space which 
contained the blood-corpuscles into a strand of pigmented con- 
nective tissue. 

The process which is going on seems to me to be somewhat 
as follows:—The effete blood-corpuscles heavy with pigment 
do not circulate with the general circulating normal blood- 
corpuscles, for the heart and ventral aorta are free from them. 
They remain stranded in those parts where the current is 
slow, and where, owing to the absence of fine capillary channels, 
eddies and back currents must take place. In such places 
they are apt to congregate and remain stationary against the 
walls and septa of these spaces, thus forming a thrombus of 
dead pigment-bearing corpuscles. By the growth and spreading 
inwards of the connective tissue of the walls and septa this 
thrombus becomes part and parcel of the walls against which 
it is formed, in the same way as a thrombus becomes organised 
in any other Vertebrate. With this organisation constriction 
occurs, so that the original blood space is narrowed or oblite- 
rated, and at the same time its walls become more and more 
full of pigment-granules. These pigment-granules are partly 
taken up by the cells of the connective tissue, but many are 
doubtless lying free in the interspaces of the fibrous 
bundles between the time of the death of the blood-corpuscles 


VERTEBRATES FROM A CRUSTACEAN-LIKE ANCESTOR. 415 


which contain them, and their absorption into the cells of 
the connective tissue. Ultimately the formation of regularly 
arranged pigmented connective tissue is complete, and in the 
adult animal the circulation has become so modified that such 
marked stagnation no longer occurs. 

In the whole life of the animal, as far as I have had an oppor 
tunity of judging at present, the time when most pigment is 
accumulated in the walls of the blood spaces of the branchial 
and other regions is just before the time of transformation. 
In the embryo, as shown by Shipley’s specimens, very little 
pigment is to be found. In the adult, again, the pigment in 
the internal regions is much less apparent than in the full- 
grown Ammoceetes; and indeed it seems to me probable that 
not only are the pigment-granules arranged in an orderly manner 
in the connective-tissue cells after the metamorphosis of the 
animal, but also that a considerable proportion is got rid of, 
perhaps after the fashion described by Hisig (17), by excretion 
into the skin. 

This process of the formation of pigment, as I have 
imagined it to be taking place in the Ammocetes, is in strict 
harmony with what is known to occur in the liver of the 
frog, and in pathological cases of cirrhosis of liver accompanied 
by pigmentation. Thus Hunter describes the occurrence of 
cirrhosis in the liver of a rabbit as a result of the presence of 
pigment. In one experiment he found a form of perilobular 
cirrhosis, some of the lobules being entirely replaced by small- 
celled connective tissue. In each case the seat of formation 
of this tissue corresponded with the presence of larger or 
smaller heaps of pigment. In a case of cirrhosis in man he 
found similar appearances, and also in the liver of pigeons. 
The large size of the red corpuscles in birds renders their 
arrest within the capillaries of the liver very easy after they 
have undergone passive destruction, and become converted 
into rigid pigment masses. Wherever they become arrested 
a proliferation of the cells of the capillary walls takes place, 
and the small clumps soon become surrounded by asmall-celled 
growth of connective-tissue cells. He concludes that in all 


416 W. H. GASKELL. 


such cases the pigment is in the position of a foreign body, 
and acts as a mild irritant te the surrounding connective- 
tissue cells, with the result of increased proliferation and in- 
creased fibrous-tissue formation. 

I conclude, then, that in the Ammoceetes the formation of 
pigment which is so characteristic of this animal is due to the 
large amount of spleen-like tissue in this animal in the spaces 
of which passive destruction of blood-corpuscles occurs, with 
the result of the deposition of pigment in the walls of these 
spaces, and the diminution in the size of the spaces themselves. 
Conversely I conclude that lines and irregular masses of pig- 
ment in any locality indicate the existence of former spleen- 
like tissue—in other words, are evidence of the position of 
obliterated lacunar blood spaces. 

If we now revert to our original question—what is the 
meaning of the pigment found so markedly in among the 
degenerate cells surrounding the brain ? the answer would be: 
These lines and collections of pigment are the remains of the 
blood channels which supplied the cells of the old cephalic 
liver with blood. With the loss of function of this organ these 
blood channels have become filled more or less with pigment- 
bearing corpuscles, thus obliterating the greater part of the 
original vascular space, certain blood-vessels only being left in 
the tfssue itself, as shown in fig. 2, Pl. XXV (d.v.). The 
remnant of the blood space surrounding this tissue is still to be 
seen in the shape of the so-called venous sinus found both on 
the dorsal and ventral sides of the brain. The dorsal sinus is 
figured in figs. 9, 14, and is a large blood space which surrounds 
the choroidal plexus, and separates the two lateral masses of 
arachnoidal tissue from each other. The pigment in between 
the folds of the choroidal plexus (figs. 9, 14), which Ahlborn 
considers to represent included arachnoidal tissue, is, to my 
mind, much more probably due to the passive destruction of the 
blood-corpuscles of this sinus ; and the vessel described by him 
in the extremity of the pigmented folds is the remnant of the 
original blood space. This sinus itself is perhaps homologous 
with the longitudinal sinus of the higher Vertebrates. 


VERTEBRATES FROM A ORUSTACEAN-LIKE ANCESTOR. 417 


On the ventral side, in the region of the medulla oblongata, 
the obliteration of the original blood space by the deposition 
of pigment has also left a blood sinus which may be divided 
into two, one on each side of the middle line, as shown in 
fig. 14, These sinuses are perhaps homologous with the 
cavernous sinuses of the higher Vertebrates. 

Such I imagine to be the history of the pigment in the 
tissue around the brain; and this view is confirmed by its 
appearance in those Ammoccetes which have been kept over a 
year in the laboratory. As already mentioned, the accumula- 
tion of large irregularly shaped pigment masses in the branchial 
regions, in close connection with the occurrence of a great 
number of pigment-bearing corpuscles, is a most striking result 
of such confinement. Precisely the same kind of increase of 
pigment is apparent in the pigmented tissue around the brain, 
as is seen in fig. 17, Pl. XX VI. 

The length of the Ammocetes from which fig. 17 is taken 
was 28 mm.; it had remained in the laboratory from the 
end of September, 1888, until October, 1889, and I doubt 
whether it had increased in size at all. The increase of pig- 
ment in connection with the branchia was very marked, and 
the pigmented tissue around the brain had undergone most 
marked modifications. In fig. 17 I have drawn a part of one 
of the sections through the epichordal region of the brain, 
and the difference between it and the appearance it would 
have presented in a fresh-caught Ammoceetes of the same 
size is most striking; instead of closely packed liver-cells 
with lines of pigment in between, we see that the cells have 
disappeared, leaving only irregular-shaped clumps and shreds 
here and there, while the pigment lines have become large 
irregular clumps on a coarse network of connective tissue. 
Close against some of these masses of pigment are seen blood- 
corpuscles (a) lying in the ventral sinus, which bear masses 
of pigment round their nuclei; and the pigment in some of 
them is indistinguishable from projections of the pigment 
masses themselves. 

It is, then, clear that an increase of the pigment between 


418 W. H. GASKELL. 


the cells of the arachnoidal tissue just as the increase of the 
pigment in the branchial region, can be produced by a pro- 
cess of malnutrition which causes at the same time extensive 
passive degeneration of the red blood-corpuscles; and the 
evidence here is just as striking as in the branchial region, 
that this increase of the pigment is due directly to the 
presence of pigment-bearing corpuscles. The natural conclu- 
sion is that the pigment which is normally found in this situa- 
tion is also due to the passive destruction of blood-corpuscles, 
and that therefore these lines of pigment represent closed-up 
vascular spaces in between the cells. 

If, then, these cells represent the cells of the old cephalic 
liver of the Crustacean-like ancestor, a straightforward explana- 
tion of the pigment between them, and of the formation 
of the dorsal and ventral sinuses, is afforded by the supposi- 
tion that the whole represents the blood spaces and blood 
channels by which the liver-cells were originally supplied with 
blood. 

The interpretation of the pigment found in other places, 
such as that in connection with the branchiz, upon the 
hypothesis that such pigment denotes the locality of previously 
existing lacunar blood spaces, will be dealt with in the next 
chapter, where I shall consider the formation of the present 
alimentary canal. 


Sect. 7.—The Relation of the Supra-csophageal 
Ganglia to the Walls of the Cephalic Stomach. 


Upon the supposition that the Ammocetes is derived 
directly from a Crustacean-like ancestor, we ought to find that 
the supra-cesophageal ganglia are situated in front of the 
cesophagus, close against the anterior rounded termination of 
the cephalic stomach on each side of the middle line; and 
these supra-cesophageal ganglia ought to form—(1) olfactory 
lobes giving origin to olfactory nerves, (2) cerebral hemispheres 
giving origin to no outgoing nerves, and (3) an optic portion 
which gives origin to the optic ganglia and nerves of eyes of 
an Arthropodan type. Further, seeing that the nervous 


VERTEBRATES FROM A ORUSTACEAN-LIKE ANCESTOR. 419 


matter in the infra-cesophageal region has in the course of its 
growth become closely united with the walls of the cephalic 
stomach, and has compressed and folded those walls in various 
ways in different regions, it is reasonable to suppose that 
here also the elements of the supra-cesophageal ganglia will 
be found in close connection with the walls of the anterior 
termination of the cephalic stomach, and will also surround 
diverticula produced by the compressing growth. 

The mere inspection of the dorsal view of the brain of the 
Ammoceetes (fig. 1) almost compels us to recognise the supra- 
cesophageal ganglia in the separate masses of nervous matter 
which are situated in front of the folds of the choroid plexuses, 
and form the cerebral and olfactory lobes connected by means 
of the optic thalami with the ganglia habenule, which in their 
turn are connected with the pineal eyes. 

Here, also, we find, as in the lower regions of the brain, that 
the walls of the cavity are partly free from the invasion of 
nervous matter, and partly form the lining epithelium of the 
nervous masses which are lying outside them. Thus, as already 
mentioned, we can trace the free anterior wall of the cephalic 
stomach in the middle line in front of the choroid plexus ii 
as forming the recessus infrapinealis (Ahlborn), then the 
choroid plexus i which bridges over the space between the two 
optic thalami (Hirnschlitz), continuing onwards as the lamina 
terminalis to reach the ventral side of the brain, where it 
forms a bulging known by the name of the recessus chias- 
maticus. 

On each side of the lamina terminalis le the simple 
olfactory lobes and cerebral hemispheres of the Ammoceetes. 
They have grown round the anterior wall of the cephalic 
stomach, so as to include two diverticula called by Ahlborn the 
lateral ventricles of the brain. In fig.6, Pl. XXV,I give a section 
through an osmic preparation of the brain of an Ammoccetes 
100 mm. long. The section is a horizontal one, and is one of 
a consecutive series through the whole brain. It shows the 
olfactory nerves passing from the olfactory lobes, and the 
lateral ventricles of Ahlborn connected by the ventriculus com- 


420 W. H. GASKELL. 


munis, which is bounded in front by the lamina terminalis, 
and is continuous behind with the third ventricle. Again, we 
see that here also in those parts of the walls of the cephalic 
stomach which form the lining epithelium of these anterior 
ventricular cavities the same marked fatty degeneration of the 
cells has taken place as already mentioned in the region of 
the infra-cesophageal ganglia. 

This figure is much the same as fig. 1, Pl. II, in Edinger’s 
paper (18) on the ‘‘ Comparative Anatomy of the Brain,” in 
which he shows that the simple cerebrum of Ammoceetes cor- 
responds to the olfactory lobes and corpus striatum of the higher 
Vertebrates. In this most instructive paper he points out that 
the cortical grey matter of the brain does not appear until we 
reach the reptiles. In fishes, as in Ammoceetes, the cerebral 
hemispheres consist simply of an olfactory part, and the basal 
ganglion or corpus striatum, with the grey matter inside and 
the white matter outside. The peripheral cortex of grey 
matter is only formed much later in phylogenetic development. 
The connecting commissure between the two basal ganglia is 
described by Ahlborn as the anterior commissure, and it is 
recognised as such by Edinger through the whole ascending 
series of Vertebrata. If, then, the olfactory lobes and basal 
ganglia are the cerebral portion of the supra-cesophageal 
ganglia, then the anterior commissure is the commissure 
which originally connected together the corresponding parts of 
the supra-cesophageal ganglia ; and as evidence of its antiquity 
we see that it is stated to be the earliest formed of all the 
commissures of the brain. 

Further, we see that these two nervous masses which form 
the brain proper and the olfactory lobes are not only in the 
position of the supra-cesophageal ganglia with respect to the 
walls of the cephalic stomach, but also are in connection with 
a special optic portion which is also supra-cesophageal in situa- 
tion, and gives origin to eyes of a strictly Arthropodan type. 

Lying between and dorsal to the two cerebral lobes we see 
in fig. 1 the large rounded right ganglion habenule ; much 
less conspicuous is the smaller left ganglion habenule. These 


VERTEBRATES FROM A ORUSTACEAN-LIKE ANCESTOR. 421 


two masses of nervous matter are connected, as Ahlborn has 
shown, with each other by a band of fibres forming the com- 
missura tenuissima, and with the cerebral lobes by strands 
of peculiarly coloured fibres which he calls the tenia thalami, 
represented in fig. 8, Pl. XXV, by ¢. th. That part of the 
nervous system connecting together the cerebral lobes and the 
ganglia habenulz is recognised by Ahlborn as the optic thalami 
(the. in fig. 8)./ gc, Z 


Sect. 8.—The Median Eyes and their Optic Ganglia. 


In close connection with the ganglia habenule are the 
structures known as the dorsal and ventral pineal eyes. Of 
these the dorsal eye is the large and conspicuous intensely white 
object which is seen in front of the right ganglion habenule ; 
and the descriptions which have been given of the pineal eye 
refer in the majority of instances to this eye, and not to the 
much more insignificant ventral eye. Before entering upon a 
criticism of the statements which have been already published 
respecting the pineal eye, I will describe the appearances pre- 
sented by my sections which have forced me to the conclusion 
that this eye is, as I have already stated in my paper in ‘ Brain ’ 
(2), an Invertebrate eye of the Crustacean or Arachnidian 
type. 

I have cut sections through a large uumber of heads of 
Ammocetes of sizes varying from 25 to over 100 mm. 
These sections have been cut either transversely to the long 
axis of the animal or horizontally in the direction of that 
axis, and I have found that the transverse sections do not cut 
the eye in the direction of its optic axis, but that the hori- 
zontal sections give a much nearer approximation to sections 
parallel to the true median plane of the eye, so that we must 
imagine the eye has been bent forwards so that its optic axis is 
directed somewhat forwards as well as upwards. Also by com- 
pression the eye has been distorted out of shape sufficiently to 
prevent any plane of section accurately passing through the 
median plane of both eye and nerve. In figs. 20a—d, Pl. 
XXVII, I give selections only of a series of horizontal sections 


422 W. H. GASKELL. 


through the pineal eye of a full-grown Ammoceetes, the whole 
head of which was placed in osmic acid soon after it was caught. 

These sections show very clearly the arrangement of the 
pigment in the eye, the shape and position of the eye, the 
course and relations of its nerve, and the manner in which 
its supposed central cavity is filled up with masses of definite 
protoplasmic material. 

1. The Pigment Layer and Nerve-end Cells.—With 
respect to the pigment, we see that Wiedersheim (19) and 
Ahlborn (38) describe the eye as containing, in all cases, intensely 
white pigment, while Beard (20) asserts that in all his speci- 
mens of the eyes of Ammoceetes pigment was not present 
except in three sections which were given to him and came 
from another locality; in these specimens the pigment was 
black, not white. He concludes, therefore, that pigment is 
only rarely present in the eyes of Ammoccetes Planeri, 
that the white pigment of Ahlborn and Wiedersheim if it is 
present is dissolved away in the process of preparation; and he 
claims to have discovered the presence of black pigment. 

By the careful study of the pineal eye I am able to abso- 
lutely clear up these apparent discrepancies of fact, and to 
show that Ahlborn is entirely right in his interpretation, while 
Beard has misinterpreted what he saw. The eye of every 
Ammocetes, without exception, presents an intensely white 
appearance, due to the presence of white pigment as described 
by Ahlborn ; never in one single instance have I seen a speci- 
men free from this glistening white substance. At first, 
specimen after specimen which I cut presented on section lines 
and markings of apparently black pigment without exception, 
so that I was utterly unable to understand Beard’s assertion 
that the eye in Ammocecetes was only rarely pigmented; at 
the same time it was difficult to understand how such a white, 
glistening little mass could be so full of what was apparently 
intensely black pigment. Afterwards I began to find that 
some of my preparations were entirely free from pigment, and 
looking through the whole series which I possessed, it was 
immediately evident that all those which were killed with 


VERTEBRATES FROM A CRUSTACHEAN-LIKE ANCESTOR. 423 


osmic, and had then been mounted after passing through the 
series of alcohols in the ordinary way, were full of pigment, 
as in fig. 20, Pl. XX VII; while all those which contained no 
pigment had been killed in Perenyi’s fluid or in picric acid, 
and then stained with various staining reagents. Clearly the 
pigment had been removed during the preparation of the speci- 
mens in these latter cases; and doubtless one agent in its 
removal is the nitric acid, which is so important a constituent 
of the Perenyi’s fluid. I have dissected out the eye and placed 
it in Perenyi’s fluid, and watched how the white pigment was 
dissolved away. Ahlborn says that these white granules are 
composed of calcium phosphate, and are the same as the brain- 
sand found in the pineal body of man and the higher Verte- 
brates. Ahlborn also lays stress on the absolute untrans- 
parency of these particles. 

Clearly, then, all the eyes are pigmented, and it is equally 
clear that there is no black pigment ; it is always white. Beard’s 
observations of the presence of black pigment are due to a 
want of care on his part; what he figures as black is in 
reality white, as he will see at once if he will look at his 
specimens with reflected light instead of transmitted light. 
The white shining particles of pigment are so opaque that they 
appear, when the section is viewed by transmitted light, quite 
black, as in figs. 20a—d. If now the sub-stage be darkened, 
and a ray of light from a condenser be thrown on the section 
so as to view it as an opaque object, the only part of the 
section which is visible are these pigment particles, and they 
shine and glitter as white particles on the uniform blackness 
of the rest of the section. Very curious is it to follow, in 
this way, the changes of the distribution of the pigment in a 
series of sections through the eye, and to see how, as we pass 
from section to section, the compact group of brilliant white 
crystalline-looking particles flashes suddenly into view from 
out of the uniform blackness. 

Wesee, then, that the pigment which Beard considered to be 
black is really white, and that this white pigment is always pre- 
sent, but is easily dissolved out by reagents such as nitric acid. 


424, W. H. GASKELL. 


As to the arrangement of the pigment, the nature and 
symmetry of its arrangement are seen much better in hori- 
zontal than in transverse sections. In fig. 22, Pl. XXVIII, 
which is a magnified portion of fig. 200, Pl. XX VII, we see 
how the pigment is arranged in definite, somewhat conical 
masses, which terminate at a very distinct limiting edge, and 
are curved round towards the middle line in certain parts of 
the eye in a very striking and characteristic manner. 

In the specimens where the pigment has been removed we 
see that the pigment region is made up of a number of lines 
arranged radially in the same manner as the lines of pigment ; 
at the base of these lines numbers of nuclei are seen, which 
follow in their arrangement the shape of the pigmented region. 
These nuclei are of two kinds, large and small; the large ones 
are situated at the base of the lines, while the smaller ones are 
scattered about in between the large ones, and extend to the 
edge of the eye (figs. 26, 27, Pl. XXVIII). In the interior 
of the eye the limit of the pigment layer is still seen, and no 
nuclei occur between this limit and the row of basal nuclei. 
We may look upon this radial appearance as due to the 
presence of elongated cells, the nuclei of which are situated 
at the base, and form the more or less regular row of large 
nuclei seen in all the specimens. The appearance is exactly 
what one would expect in an eye which is no longer functional 
if these large nuclei belonged to the nerve-end cells of an 
Arthropodan eye, as described by Lankester and Bourne (21), 
Grenacher (22), and other observers. The pigment also corre- 
sponds in position to that always described in connection with 
the layer of nerve-end cells. 

As to the smaller nuclei, it is possible that they belong to 
intrusive connective-tissue cells, as described by Lankester 
and Bourne (21) in the median eye of Limulus. 

2. The Termination of the Nerve-end Cells with 
their Attached Rhabdites.—The pigmented layer which 
forms the posterior wall of the eye is separated from the anterior 
non-pigmented layer by a central cavity, which according to 
Beard (20), following the suggestion of Spencer (23), contains 


VERTEBRATES FROM A ORUSTACEAN-LIKE ANCESTOR. 425 


nothing except the coagulated remnants of an albuminous 
fluid. In none of my specimens, whether cut horizontally or 
transversely, is there any sign of such a central cavity as figured 
by Beard; in all cases this space is filled with the remains of 
what is clearly a tissue possessing a well-defined structure. In 
the older Ammoceetes the substance of this tissue is very apt 
to be fissured irregularly (as in fig. 20, Pl. XX VII), or else to 
be vacuolated, the spaces so formed being of different sizes 
and somewhat circular in outline. 

In all cases, however, square-shaped blocks and clumps of 
tissue, in which there is no sign of nuclei, are seen lying 
between the pigmented posterior part of the eye and the 
anterior non-pigmented part; and it is significant, as showing 
how this vacuolated appearance has misled Beard, that the 
younger the animal the more regular and compact is the 
arrangement of these square-shaped masses of tissue. Accord- 
ing to Shipley (10), the eye at its commencement possesses no 
sign of an internal cavity, but presents an appearance of a regu- 
lar solid mass of cells ; this statement is denied by Owsjannikow 
(7), who says that the anterior and posterior walls come so 
close together at an early stage of development as to almost 
obliterate the appearance of a central cavity. Everything 
seems to me to point to the conclusion that the appearance 
of a large central cavity is brought about by the partial de- 
generation of elements which originally filled it, and that their 
remains have given rise to the impression held by Spencer 
and Beard, that a large cavity exists which is filled with a 
coagulated albuminous fluid. 

In addition, it may be remarked that it is perfectly easy to 
see the appearance presented by the coagulated albuminous 
fluids within an eye by simply looking at sections of the lateral 
eyes of the same animal; the appearance presented resembles 
that of a blood-clot, and is totally dissimilar to the square- 
shaped masses of protoplasmic-looking material seen in the 
cavity of the pineal eye. 

The examination of these square-shaped masses by a high 
power in osmic preparations, especially in the neighbourhood 


4.26 W. H. GASKELL. 


of the pigment layer, reveals an extraordinary sight; scattered 
about over the whole field of view of the microscope are seen 
rod-like bodies, some crescent-shaped, some nearly straight, 
some shaped like a hook, others like an elongated S, all 
apparently attached to the edge of one or other of these square 
or oblong-shaped masses, and differentiated from them by 
their greater refracting power. They lie in connection with 
the bodies to which they are attached at different angles to 
each other, but occasionally a number of them appear to be 
regularly arranged in respect to the pigment layer. In this case 
they present very closely the appearance given in fig. 25, Pl. 
XXVIII, which is taken from Grenacher (22), and represents 
the pigment free ends of the nerve-end cells of an eye of an 
Acilius larva with their attached rhabdites; in other cases the 
appearance is very like that represented in fig. 24, Pl. XXVIII, 
which is a reproduction of Lankester and Bourne’s (21) picture 
of a section through the nerve-end cells and rhabdomes of a 
Euscorpius eye. 

In fig. 22, Pl. XXVIII, I give an accurate copy of a 
magnified portion of the section in fig. 20c, Pl. XX VII. 

These bodies, then, which appear on section somewhat 
square or conical, according to the direction in which they 
happen to be cut, with their rod-like, highly refractile 
attached pieces, are by no means mere coagulated albumen, but, 
on the contrary, are the terminal parts of the nerve-end cells, 
with the rhabdites attached to them, which project freely beyond 
the pigment layer, and form the layer of rods as pictured and 
described in de Graaf’s original paper (24) on the pineal eye 
of Anguis. The arrangement of the pigment of the nerve-end 
cells, and of the rods, calls to mind very forcibly the figures 
of the larval eye of Dytiscus and Hydrophilus as figured in 
Grenacher (22) and Patten (25). In specimens stained by 
hematoxylin no such distinct appearance of rod-like bodies 
can be seen; but here we find, in between the square or poly- 
gonal cell-like masses, lines and strands of substance which 
stain very much darker than the rest of the tissues of the eye, 
and in this respect resemble cuticular structures. 


VERTEBRATES FROM A ORUSTACEAN-LIKE ANCESTOR. 427 


The most anterior part of the substance which fills up the 
central cavity is composed of masses which are less definite 
in shape than those near the pigment layer, and are free or 
nearly free from any appearance of possessixg cuticular rods ; 
they often closely resemble fig. 23, Pl. XXVIII, which is 
taken from the paper of Lankester and Bourne (21), and repre- 
sents a transverse section of a clump of nerve-end cells from 
the eye of Euscorpius beyond the region of rods. When 
stained with picro-boro-carmine it is seen (fig. 24, Pl. XX VIII) 
that the middle portion of these polygonal masses stains red- 
dish, while the edges stain yellow. 

3. The Cells of the Hypodermal or Vitreous Layer. 
—The anterior part of the eye is free from pigment, and is 
composed, as is seen in hematoxylin or carmine specimens 
(figs. 26, 27), of an inner layer of nuclei which are frequently 
arranged in a wavy line. These nuclei are continous laterally 
with the nuclei at the base of the layer of nerve-end cells. 
From this nucleated layer strands of tissue, free from nuclei, 
pass to the anterior edge of the eye. 

In the horizontal longitudinal sections it is seen (figs. 26, 
27) that these strands are confined to the middle of the eye; on 
each side of them the nuclear layer reaches the periphery; so that 
if we consider these strands to represent long cylindrical cells, 
as described by Beard (20, p. 62), then the anterior wall 
may be described as consisting of long cylindrical cells, which 
are flanked on each side by shorter cells of a similar kind. 
The nuclei at the base of these cylindrical cells are not all 
alike. We see, in the first place, large nuclei resembling the 
large nuclei belonging to the nerve-end cells; these are the 
nuclei of the long cylindrical cells. We see also smaller nuclei 
in among these larger ones, which look hke nuclei of intrusive 
connective tissue, or may, perhaps, form a distinct layer of 
cells, situated between the cells of the anterior wall and the 
terminations of the nerve-end cells already referred to. The 
appearance presented is drawn as accurately as possible by 
Mr. Wilson, of the Scientific Instrument Company, in fig. 27, 
Pl. XXVIII, from aspecimen stained with picro-boro-carmine, 


428 W. H. GASKELL. 


It shows between the layer of square-shaped, non-nucleated 
masses (7), which I consider to be the terminations of the 
nerve-end cells, and the layer of large nuclei (v) of the anterior 
wall, a more or less definite layer of small cells with distinct 
nuclei (0.w.). 

In the diagram (fig. 28, Pl. XXVIII) I have depicted the 
different elements which I see in my sections as they would 
appear if the eye were restored; and without making any 
positive statement upon the meaning of the smaller nuclei in 
the anterior wall, whether they are connective-tissue elemeuts, 
or form a distinct layer as figured by Patten (26), it seems to 
me clear that the large nuclei belong to the hypodermal layer 
of cells known by the name of the vitreous layer in the Arthro- 
pod type of eye. 

According to the views of Patten (26, p. 165), the ‘‘ ground 
plan of all the variations in the eyes of both Molluscs and 
Arthropods is a three-layered eye consisting of an invaginated 
optic vesicle, the inner wall of which becomes the retina, and 
an overlying layer of hypodermis, the corneagen.” 

According to this view the pineal eye would consist as in 
the diagram of the following layers: 

1. (R) The inner or posterior wall of the optic vesicle or 
retina, composed of the large nerve-end cells with their asso- 
ciated pigment. 

2. (oc.) The cavity of the optic vesicle, containing the termi- 
nations of the nerve-end cells with their attached rhabdites. 

3. (ow.) The outer or anterior wall of the optic cavity, com- 
posed of a thin layer of small cells. 

4, (v) The overlying layer of hypodermis or corneagen. 

If, on the other hand, the small cell-elements be looked 
upon as intrusive connective tissue, as described by Lankester 
and Bourne (21) in the central eye of Limulus, then the pineal 
eye would be described as consisting of a retina with nerve-end 
cells associated with pigment and bearing rhabdites near their 
terminations, and in front of the retina the vitreous layer of 
large hypodermal cells, so that the eye would be included in 
the group of diplostichous eyes as defined by these authors. 


VERTEBRATES FROM A CRUSTACEAN-LIKE ANCESTOR. 429 


In either case, whether Patten’s or Lankester and Bourne’s 
description be accepted as the more correct, the type of 
eye is clearly that of an Arthropod, and indeed of an ancient 
form, for the arrangement of the nerve-end cells, the shape of the 
internal cavity, the position and simplicity of the rhabdites, all 
point to larval characteristics, and therefore to an ancient type. 

So far no mention has been made of any lens. One of the 
characteristics of the Arthropod eye is the cuticular lens. In 
the pineal eye of the Ammoceetes no observer has been able up 
to this time to clearly point to any lens-like structure in that 
eye. I entirely agree with Beard that the anterior wall of the 
eye does not represent alens. In my opinion, as just mentioned, 
it does represent the corneagen. Where, then, is the lens ? 

4. The Cuticular Lens.—Upon the supposition that 
we are dealing here with an eye of an Arthropodan type, it 
follows that the lens must have been cuticular in structure, 
and simply a local modification of the general cuticular covering 
of the front part of the body. 

In all the horizontal sections through the eye it is very 
plainly visible that the anterior wall of the eye is closely 
pressed against the tissue which forms the wall of the cranial 
cavity at this spot ; and in many cases it 1s very striking to see 
how (as is represented in fig. 26, Pl. XX VIII) the anterior wall 
of the brain-case in this one place dips inwards, and is thickened 
so as to form a projecting knob which is closely attached to the 
pineal eye. The closeness of this attachment is seen when the 
brain and eye are dissected out under a dissecting lens. I have 
prepared many specimens in this way, and have found that with 
care it is easy to cut away and remove the brain-case with the 
overlying skin, and yet leave the eye in position and in con- 
nection with the brain. In every case the eye then appears as 
a white opaque round mass, in the centre of which anteriorly 
the surface is not flat, but is hollowed out to form a distinct 
manifest cup. The hollow of this cup was filled with the 
projecting knob of the cranial wall, which was therefore removed 
with the rest of that wall, and indeed can in some cases be 
plainly seen under a dissecting lens as a slight projection on 

VOL, XXXI, PART III].—NEW SER, FF 


430 W. H. GASKELL. 


the inner surface of the portion of the wall which has been 
removed. In Ahlborn’s fig. 5, Pl. xiii (3), this hollow appear- 
ance is perhaps represented. 

In horizontal sections the manner in which the eye is 
attached to the brain-case is clearly visible, and in some cases, 
as in fig. 26, Pl. XXVIII, the projecting knob of the tissue of 
the brain-case into the pineal eye so as to form its most anterior 
wall is very conspicuous. We see how the tissue of the cranial 
wall is thickened at this one spot, and dips down to meet the 
pineal eye which is lying beneath it. We see how easy it is 
to understand that the removal of the brain-case must leave a 
shallow depression in the anterior surface of theeye. Further, 
the structure of this tissue is very peculiar. When dissecting it 
out it feels more like cutting through cartilage than through 
fibrous tissue. Upon section it presents a curiously homo- 
geneous appearance, with an evident tendency to split into 
fibrous-looking laminz. It is remarkably free from any sign 
of nuclei in it, and is very apt to contain fine globules of 
yellowish refractile substance, somewhat like those which appear 
also in the sheath of the notochord and other places. It stains 
very deeply with hematoxylin, similarly to the bodies in the 
eye already mentioned, just as the cuticular lens and rhab- 
dites of the eye in Arthropoda are known to stain deeply with 
heematoxylin. 

In its reaction to staining fluids and in its general appear- 
ance it resembles that curious laminated layer of tissue which 
lies just below the epithelium layer of the skin in the Ammo- 
coetes. 

If, then, as I believe, this thickened portion of the tissue of 
the cranial wall represents the position of the cuticular lens of 
the pineal eye, it follows that the brain-case itself is partly a 
modified portion of the integument of the Crustacean-like 
ancestor—a conclusion which is entirely in harmony with my 
theory, and which will be considered by me when I come to 
treat of the manner in which the skeletal tissues of the Verte- 
brates have arisen. 

5. The Optic Nerve and Optic Ganglion.—From the 


VERTEBRATES FROM A CRUSTACEAN-LIKE ANCESTOR. 481 


eye itself arises the nerve, which is most beautifully shown in 
the series from which figs. 20a—d are taken. It is seen 
to be a delicate thin nerve, which is not mixed up with any 
other structures, and passes freely from the eye to the 
large ganglion habenule. It is clearly not hollow, and at the 
same time is filled with something which does not look like 
nerve-fibres, but more like somewhat elongated cells closely 
packed end to end. ‘The only place where there is any appear- 
ance of a cavity is where the central cavity is prolonged in 
the direction of the place of entrance of the nerve. This con- 
tains the remains of nerve-end cells and rhabdites, just as the 
rest of the central cavity. 

The shape of the central cavity is given in Ahlborn’s fig. 44 
(3), which represents a sagittal section through the eye. Its 
shape is due to the arrangement of the structural elements of 
the eye, and resembles the picture of the eye of the larva of 
Hydrophilus given in Patten’s paper (25). Horizontal sections 
show that the axis of the cavity has been twisted upwards in 
the dorsal direction at the end where the nerve enters, so that 
the shape of the cavity may be likened to a cornucopia, the 
bent part of the stalk of which lies in the dorso-ventral plane. 
As the result of this distortion all the most dorsal of a series 
of horizontal sections must cut through the central cavity 
twice, first through the expanded mouth of the cornucopia, 
then through the stalk end, and in these sections the entrance 
of the nerve must be shown (see figs. 20a, 206, Pl. XX VII); 
then in the middle sections one cavity will be shown consisting 
of asection through the expanded mouth and commencing stalk 
of the cornucopia (fig. 20c, Pl. X XVII), and finally on the 
ventral side of the middle section one cavity only will be shown, 
viz. the cavity of the expanded mouth of the cornucopia (fig. 
20d, Pl. X XVII). 

In this way the appearances seen on section receive a 
simple explanation, and in all probability the small circular 
cavity depicted by Owsjannikow in figs. 7, 8, and 9 of his 
paper (7) is due to the direction of his sections in a precisely 
similar manner, 


432 W. H. GASKELL. 


Tracing the nerve in the series of sections, it is easy to see 
how it at first lies free in among the liver-cells until it reaches 
the left-hand side of the rounded mass of the right ganglion 
habenule. It then passes along the whole length of this 
ganglion, lying close against the side of it, successive sections 
showing that it is becoming less and less superficial, until near 
the posterior end of the right ganglion habenulz it is lying 
between the left and right ganglia. It then, still passing deeper 
and deeper, curves round the posterior border of the right 
ganglion habenule just in front of the commissura pos- 
terior. From this point it is not easy to trace it further. 
All the latter part of its course it has been shifting its direction 
so as to become more and more vertical, and at this point it 
appears to me to pass straight downwards into the substance 
of the right ganglion habenule. 

Ahlborn (8, p. 232) describes the thread-like portion of the 
epiphysis as a white hair-like thread, which passes from the 
upper snow-white vesicle (pineal eye) to terminate a little in 
front of the posterior commissure. He describes it as lying 
closely over the left ganglion habenulz, and speaks of it as 
having undergone an unequal obliteration, so that its proximal 
part has disappeared to such an extent that “die umhullenden 
Piablatter meist vollstandig kollabirt sind.” This description 
he illustrates with fig. 44, Pl. xvi, which is a sagittal section 
through the dorsal eye and the left ganglion habenule. Now 
in this very picture it is clear that the nerve has been cut by 
the section at the spot where he supposes this obliteration of 
its proximal part to commence, so that in reality the proximal 
part of the nerve is not in his section at all; the reason being, 
as already described, that it clings close to the side of the 
right ganglion habenule, and would therefore be found in the 
sections to the right of the one he has figured. 

As is figured in my sections, the nerve lies free in among the 
cells of the arachnoid tissue, and passes along the face of the 
right ganglion habenule to its posterior border without any 
sign of connection with either the optic thalamus or the left 
ganglion habenule. 


VERTEBRATES FROM A CRUSTACEAN-LIKE ANCESTOR. 4933 


I have come, then, to the conclusion that the nerve is in 
connection with the right ganglion habenule, and originally 
passed straight to the surface from the posterior part of that 
ganglion. By the shifting forward of the eye and brain wall 
the nerve also was bent forward until it became almost hori- 
zontal in position. That such a shifting of position actually 
took place is nearly certain from the observations of Scott 
(11), who describes how the eye which was originally behind 
the ganglion habenule has shifted forwards during the deve- 
lopment of the young animal. Ahlborn also mentions the 
same shifting of position. 

Further, in the eyes of the Arthropoda the optic nerve 
passes into the optic ganglion, which is a well-defined struc- 
ture separate from the rest of the supra-cesophageal ganglia. 
The optic ganglion is composed of a cortical layer of small 
nerve-cells closely packed together, and an internal medullary 
portion composed of nerve-fibres which form, according to 
Patten (26), the medullary stalk of the optic ganglion by 
which it is connected with the brain, and form also the medulla 
of the ganglion itself, which bears special relations to the eye 
belonging to that optic ganglion. 

In my opinion the right ganglion habenule is the optic 
ganglion of the dorsal pineal eye. It is composed of an internal 
medullary portion of nerve-fibres, and a cortical portion of 
small nerve-cells closely packed together of the same kind as 
the berry-like cells of the grey matter of the brain, resembling 
therefore the cells of the optic ganglion as given by Patten 
(25). If I am right in my belief that the nerve of the pineal 
eye can be followed into the right ganglion habenule, it appears 
to me to lose itself in the central medullary mass of fibres. In 
the innermost portion of the ganglion the medullary part is 
the main portion From it are formed the large Meynert’s 
bundle which is traceable to the ventral side of the brain in 
the neighbourhood of the large ventral fissure, also fibres which 
connect together the two ganglia habenule forming the com- 
missura tenuissima of Ahlborn (which perhaps becomes 
the commissura mollis of human anatomy). A third set 


434 W. H. GASKELL. 


of fibres which arise from it are the conspicuous strands which 
pass by way of the optic thalami to the cerebral lobes. These 
latter form the medullary stalk of the optic ganglion, the optic 
thalamus being that part of the supra-cesophageal ganglion 
which connects together the brain proper and the optic gan- 
glion. It is worthy of notice that all these strands of fibres 
do not stain like good nerve-fibres, but take on a peculiar tint 
with various staining reagents as pointed out by Ahlborn. 
Further, we see, from the series of horizontal sections that in 
the more superficial sections in which the left ganglion habenulze 
is only slightly or not at all involved, the internal medullary 
part is divided into two portions, of which the one is composed 
of the more superficial fibres of the commissura tenuissima 
and tenia thalami, while the other forms the medulla of the 
ganglion habenule itself. In fig. 27, Pl. XXVIII, I give one 
of a series of horizontal sections through the pineal eye and 
ganglion habenule of an Ammoccetes 66 mm, long, which was 
stained whole in boro-picro-carmine. The peculiar shape of the 
arrangement of the nerve-fibres which form the medulla of the 
ganglion with respect to the nerve-cells is shown in the figure. 

In every respect, then, it appears to me the right ganglion 
habenule -proves itself to be the optic ganglion of the dorsal 
pineal eye. 

6. The Left or Ventral Eye.—There is, however, a left 
ganglion habenule and a left Meynert’s bundle, very much 
smaller than the right ; if this also is an optic ganglion the 
eye belonging to it ought to be visible, though we should expect 
to find it more degenerated, and its structure less easy to define 
than the dorsal eye connected with such a vigorous optic ganglion 
as is represented by the right ganglion habenule. This second 
eye is the epiphysis iii of Ahlborn, i.e. the lower vesicle which he 
describes as possessing a structure similar to that of the upper 
vesicle, or epiphysis ii, which is now recognised as the pineal 
eye.! In this cavity, however, Ahlborn finds the remains of 

' He has evidently made a mistake in supposing that the lower vesicle 


communicates with the upper; all my sections show clearly that they are two 
separate structures. 


VERTEBRATES FROM A GRUSTACEAN-LIKE ANCESTOR. 439 


strands of tissue similar to what he has noticed in the cavity of 
the dorsal eye. A description of it is given in Owsjannikow’s 
paper (7), and he gives it the name of ventral eye, in contra- 
distinction to the larger, more perfect organ which is the dorsal 
eye. This second eye is not only perfectly plain in all my 
sections, especially the horizontal ones, but it is clearly, as 
Ahlborn has pointed out, connected in a peculiar way with the 
left ganglion habenule: it is built up of similar elements to the 
dorsal eye, except that it is never pigmented, as far as I have 
seen, and it is not connected with the cuticular walls of the 
brain-case. As yet I have not distinguished any rhabdites in 
it, and the terminations of its nerve-end cells have broken down 
to such an extent as to give the whole organ the appearance of 
a tube in connection with the left ganglion habenule. 
Owsjannikow (7) describes its posterior retinal wall as being 
connected by means of nerve-fibres with a group of nerve-cells, 
to which he gives the name of the ganglion of the eye. This 
group of nerve-cells seen in fig. 21, Pl. XX VII (g/l;.), with 
the strands of fibres which proceed from them towards the long 
nerve-end cells of this eye, is called by Ahlborn the “ Zirbel- 
polster,” and is recognised by him as a part of the left ganglion 
habenule (gh/,. in his figures). This outlying part of the left 
ganglion habenule is united with the rest of that ganglion (ghd. 
of Ahlborn) by a nervous stalk (gh/,. of Ahlborn), which accord- 
ing to Ahlborn is short in Ammoceetes, but longer in the adult 
Petromyzon; this connection can be easily followed in a 
series of transverse sections ; it cannot be shown in any one 
horizontal section, and therefore is not visible in fig. 21. In 
Ammoceetes it is so thick that undoubtedly Ahlborn’s ghi,. 
and ghl,. are simply parts of the same left ganglion habenule. 
Ahlborn further describes how a fold of pia mater entirely 
separates this outlying part of the left ganglion habenule 
from the lower vesicle, i.e. from the ventral eye, except at 
one place where the continuity of the membrane is broken by 
the passage of a bundle of fibres connecting the two struc- 
tures. My sections show clearly the same appearances as 
those of Ahlborn, and in fig. 21 we see how the fold of pia 


436 W. H. GASKELL. 


mater is interrupted by the passage of the fibres (f.) from 
ghl;. to the ventral pineal eye (pz,.). I feel, therefore, inclined 
to look upon the group of cells (gl;.) which form the ganglion 
of Owsjannikow, and the cortex of the “ Zirbel-polster”’ of 
Ahlborn, as the cortical layer of cells of the optic ganglion of 
the ventral eye, while the nerve-fibres (f.) which pass from 
them into the eye itself form the medulla and nerve of that 
eye ; the main part (g//,.) of the left ganglion habenule being 
formed of the cortical cells and medullary fibres which connect 
this optic ganglion with its neighbour, and with the supra- and 
infra-cesophageal ganglia. 

We see, then, that the original Crustacean-like ancestor 
had a pair of median eyes each with its optic ganglion, and 
its connections with both supra- and infra-cesophageal gan- 
glia; the right eye of the two remained functional longer 
than the left, with the result of producing in the lamprey the 
noteworthy asymmetery of the ganglia habenule, and of the 
two Meynert’s bundles. 

I have no doubt but that farther observation in the light 
of the facts narrated in this paper will suffice to prove that the 
pineal eye of lizards, &c., is also of the same kind as in the 
Ammoceetes ; throughout the mistake of previous observers, 
with the exception of Ahlborn, has been to rely too much 
upon sections cut in the wrong direction, and to omit the 
most important contents of the eye, under the delusion that 
they were a coagulated albuminous fluid, and therefore due to 
the method of preparation, and not essential elements in the 
eye. 


Srctr.—9. The Structure of the Supra- and Infra- 
esophageal Ganglia. 


So close, indeed, is the comparison of the central nervous 
system of the Ammoccetes with that of an Arthropod, that I 
feel sure further investigation will bring out a complete coinci- 
dence, not only in the topographical arrangement of the several 
parts, but also in histological structure. Such an investiga- 
tion is now being conducted in the Cambridge Physiological 


VERTEBRATES FROM. A CRUSTACEAN-LIKE ANCESTOR. 437 


Laboratory under my direction, and I will at present only 
draw attention to one or two points. In the first place, we see 
that the nerve-cells of the central nervous system of such an 
animal as a crayfish vary in size, being divisible into giant- 
cells, large cells, and small cells ; so also in the Ammoceetes we 
find the same three classes of cells. In the Ammoceetes the 
giant-cells are connected with the large Miillerian fibres which 
constitute a system of longitudinal paths in the nervous 
system itself; the large cells give origin to the series of seg- 
mental spinal and cranial nerves ; and the small cells form the 
closely set masses of ‘‘ beerenformigen ” cells which constitute 
the chief part of the spongy portion of the grey matter in 
the upper portions of the central nervous system. 

In the crayfish the large cells are in connection with the 
outgoing segmental nerves; the small cells form closely set 
masses of cells which bear a strong resemblance to the cells 
of the spongy portion of the grey matter, and are connected 
with the reticulated substance (Punctsubstanz) rather than 
directly with outgoing nerves. Ahlborn’s description of the 
arrangement in the Petromyzon of these small cells, in rows 
like bunches of grapes on a stalk, and the appearance of them 
as I myself have seen them, is strikingly illustrated in Patten’s 
fig. 7, on the plate illustrating his paper (25) on the develop- 
ment of the eyes of Vespa, &c. Whether the giant-cells in 
the Arthropod nervous system resemble those in the Ammo- 
coetes, and are connected with a system of longitudinal fibres 
in the central nervous chain itself, I do not as yet know, 
although I think it very probable that such will be found to 
be the case. 

Again, the similarity in the extent of this small-celled 
group is very striking. In the spinal cord of the Ammo- 
coetes, and indeed of all Vertebrates, the small cells of the 
grey matter are but few, and confined mainly to the region 
of the posterior horn; when, however, we reach the higher 
regions of the central nervous system we see how greatly this 
particular class of cell increases in number, forming the main 
feature of the grey substance ; we see how masses of these 


438 W. H. GASKELL. 


cells are grouped round the diverticula of the central canal in 
the region of the cerebral and olfactory lobes; how the com- 
mencing optic lobes are formed of a great lateral bank of 
these cells, spreading dorsally on each side of the aqueduct of 
Sylvius; and how, too, the commencing cerebellum is largely 
formed of an increase of these same kind of cells; further, we 
see, as already mentioned, how the cortex of the large 
ganglion habenule is formed of similar cells, arranged in a 
strikingly similar manner to those of the optic ganglion of an 
Arthropod. 

Similarly, in the Crustacean the most striking difference 
between the higher parts of the nervous system—the supra- 
cesophageal and infra-cesophageal ganglia — and the lower 
gangliaof the ventral chain is the great increase of these small- 
celled masses, which are in connection with the reticulated 
substance (Punctsubstanz), and do not possess well-defined axis 
cylinder processes asin the case of the large pear-shaped nerve- 
cells. 

In fact, the central nervous system of the Ammocetes, and 
therefore of all other Vertebrates, is the direct descendant of 
the Arthropod nervous system in all respects; and it is for this 
reason, and not because similarity of function requires similarity 
of structure, as suggested by Bellonci (27), that the remarkable 
resemblance exists between the structure and connections of 
the olfactory lobes in the Vertebrates and in the higher 
Arthropods which he has pointed out in his paper (27), in the 
‘ Archives Italiennes de Biologie.” In his conclusions he says: 
“ La structure et les rapports des lobes olfactifs présentent 
chez les arthropods supérieurs et les Vertébrés le méme plan 
fondamental. Dans les uns comme dans les autres les fibres olfac- 
tives et les fibres de connexion des lobes olfactifs se résolvent 
en un fin réticule, qui, se groupant par places, forme ce que 
jai nommé le glomérule olfactif.”’ In the Ammoceetes the 
olfactory glomeruli resemble exactly in appearance, and in their 
reaction to staining with osmic, the reticulated substance (Punct- 
substanz) of the Arthropod nervous system ; here more clearly 
than anywhere else in the Vertebrates, we see the characteristic 


TEB VERRATES FROM A CRUSTACEAN-LIKE ANCESTOR. 439 


appearance of this peculiar Invertebrate tissue. I will not, 
however, dwell further here upon the question of the olfactory 
nerves, as I shall deal with them and the rest of the cranial 
nerves in my next paper. 

Although I have dealt with the nervous system, I have 
purposely omitted the consideration of the pituitary body 
because it does not in reality belong to the central nervous 
system; it represents, in my opinion, the Crustacean green 
glands, and will be considered when I come to deal with the 
formation of the Vertebrate skeletal tissues, and of the 
excretory organs. 

In the next chapter I propose to point out how the present 
alimentary canal arose; and as it is perhaps advisable to make 
a preliminary communication in this paper, I will say now 
that the view, the truth of which I intend to discuss, is as 
follows : 

1. The alimentary canal is formed by the prolongation of a 
respiratory chamber. 

2. The respiratory chamber contains the gill-bearing legs of 
the Crustacean-like ancestor, which are still present in the 
Ammoceetes as the bars bearing branchie, and still retain their 
Invertebrate type of muscles. 

3. The first pair of such appendages do not bear branchie ; 
they form the so-called velum, and are simply the foremost 
pair of legs of the Crustacean-like ancestor. 

4. The segmental cranial nerves are the nerves which arise 
from the infra-cesophageal and thoracic ganglia, and supply the 
foremost appendages up to and inclusive of the gill-bearing legs. 

5. The first two cranial nerves are the nerves of special 
sense arising from the supra-cesophageal ganglia. 


LITERATURE REFERRED TO. 


1. Gasket.‘ On the Relation between the Structure, Function, Distribu- 
tion, and Origin of the Cranial Nerves; together with a Theory 
of the Origin of the Nervous System of Vertebrata,” ‘Journ. of 
Physiol.,’ vol. x, p. 153. 


440 W. H. GASKELL. 


14. 


15. 


16. 


if. 


18. 


19. 


. GaskELL.—“ On the Origin of the Central Nervous System of Verte- 


brates,” ‘Brain,’ vol. xii, p. 1. 


. AHLBORN.— Untersuchungen iiber das Gehirn der Petromyzonten,” 


‘Zeitschr. f. Wissens. Zool.,’ vol. xxxix, p. 191. 


. AHLBORN.—“ Ueber die Bedeutung der Zirbeldriise,”’ ‘Zeitschr. f. 


Wissens. Zool.,’ vol. xl, p. 331. 


. Miter, W.—“ Ueber Entwicklung und Bau der Hypophysis und des 


Processus infundibuli cerebri,” ‘ Jenaisch. Zeitschrift,’ vol. vi, 1871. 


. Ronon.—“ Ueber den Ursprung des Nervus Acusticus bei Petromyzonten,” 


‘Sitzung. Bericht d. k. Akad. d. Wissens.,’ vol. Ixxxv, 1882, Wien. 


. Owssannikow.— Ueber das dritte Auge bei Petromyzon Fluviatilis,” 


‘Mémoires de |’Acad. Impér. d. Science de St. Pétersbourg,’ vol. 
xxxvi, 1888. 


. WicuTman.—‘On the Ventricular Epithelium of the Frog’s Brain,” 


‘Studies from Biological Laboratory of the Johns Hopkins Uni- 
versity,’ vol. iv, p. 261. 


. RetssnER.—“ Beitrage zur Kenntniss vom Bau des Riickenmarks, &c.,” 


‘ Miiller’s Archiv,’ 1860. 


. SHIPLEY.— On some Points in the Development of Petromyzon Fluvia- 


tilis,”’” ‘ Quart. Journ. Micr. Sci.,’ 1887. 


. Scort.—‘* Notes on the Development of Petromyzon,” ‘Journ. of 


Morphol.,’ vol. i, p. 253. 


. Corninec.— Ueber die Entwicklung der Substantia Gelatinosa Rolando 


beim Kaninchen,” ‘ Archiv f. mikr. Anat.,’ vol. xxxi, p. 598. 


. His.—‘* Zur Geschichte des Menschlichen Riickenmarks und der Nerven- 


wurzeln,” ‘ Abhandl. d. math.-phys. Classe d. Kénigl. Sachs. Gesell. 
d. Wissens.,’ vol. xiii. 

His.—* Die Neuroblasten und deren Entstehung im embryonalen Mark,” 
‘Archiv f. Anat. u. Physiol. Anat, Abth.,’ 1889, p. 249. 

SAGEMEHL.—“ Beitrage zur vergleichenden Anatomie der Fische.” 
II. “ Kinige Bemerkungen tiber die Gehirnhiute der Knochenfische,” 
‘Morphol. Jahrbuch,’ vol. ix, p. 457. 

Hunter.— Arris and Gale Lectures on Transfusion,” ‘ Brit. Med. 
Journ.,’ 1889; also paper in forthcoming number of ‘Journ. of 
Physiol.’ 

E1st¢.—* Capitelliden,” ‘Faun. u. Flor. d. Golfes vy. Neapel,’ vol. xvi, 
1887. 

Epineer.—“ Vergleichende Anatomie des Gehirns.” I. ‘Das Vorder- 
hirn,’ Frankfort-on-M., 1888. 


Wiepersueim.— Das Gehirn von Ammocoetes und Petromyzon Planeri,” 
‘Jen. Zeitschr.,’ vol. xiv, 1880. 


VERTEBRATES FROM A ORUSTACEAN-LIKE ANCESTOR. 441 


20. Bearp.—“ The Parietal Eye of the Cyclostome Fishes,” ‘ Quart. Journ. 
Micr. Sci.,’ 1888. 

21. LanxesteR & Bourne.—‘The Minute Structure of the Lateral and 
Central Eyes of Scorpio and Limulus,” this Journal, vol. xxiii, p. 177. 


22. GrenacHEeR.—‘ Sehorgan der Arthropoden,’ Gottingen, 1879. 

23. SpencER.—‘“‘On the Presence and Structure of the Pineal Eye in Lacer- 
tilia,”’ this Journal, 1886. 

24, De Graar.— Bouw en de Ontwikkeling der Epiphyse bij Amphibien 
en Reptilien,’ Leiden, 1886. 

25. Patren.— Studies on the Eyes of Arthropods,” ‘ Journ. of Morphol.,’ 
vol. i, p. 193. 

26. Parren.— Studies on the Eyes of Arthropods,” ‘Journ. of Morphol.,’ 
vol. ii, p. 97. 

27. Batitoncr.—“ Sur la structure et les rapports des Lobes Olfactifs dans les 
Arthropods supérieurs et les Vertébrés,” ‘ Archives Ital. de Biologie,’ 
vol. iii, p. 191. 


CamBripGE; March, 1890. 


EXPLANATION OF PLATES XXV, XXVI, XXVII, 
XXVIII, 


Illustrating Mr. W. H. Gaskell’s paper “On the Origin of 
Vertebrates from a Crustacean-like Ancestor.” 


Examination of Reference Letters. 


6. v. Blood-vessel. &.s. Blood-sinus. cd. Cerebellum. cer. Cerebral and 
olfactory lobes (supra-cesophageal ganglia). ch,, ch, chs. Choroid plexuses 
(free walls of cephalic stomach). com. ten. Commissura tenuissima. er. 
Cranial wall. g. 4.7. Right ganglion habenule (right optic ganglion). g. &. 7. 
Left ganglion habenule (left optic ganglion). /. Arachnoidal tissue (cephalic 
liver). 7.¢. Lens of pineal eye. 7. d. Ganglion interpedunculare (duct of 
cephalic liver). 7.2. Lobus infundibuli (old esophagus). 7. ¢. Lamina 
terminalis (anterior end of cephalic stomach). MM. d.7. Right Meynert’s 
bundle. M.4./. Left Meynert’s bundle. med. Medulla of pineal eye. IL/ 
Miillerian fibres. 2. Nerve of pineal eye. xc. Notochord. x, g. Nerve- 
cells of grey matter. 0.c. Central cavity of optic vesicle.  o/. Olfactory 


442 W. H. GASKELI. 


nerve. op. ¢h. Optic thalamus. 4. w. Anterior wall of optic vesicle. pz. 
Dorsal pineal eye. pz . Ventral pineal eye. yp. com. Post-commissure. 2. 
Retina of pineal eye (layer of nerve-end cells). 7. Layer of rhabdites. ra. 
Raphe. 7. ag. Raphe of aqueduct. s¢. Lining epithelium of central cavity 
of brain (epithelium of stomach wall). ¢7. Trabecule. 7. ¢/. tenia thalami. 
v. Vitreous layer or corneagen. V. ag. Ventricle of aqueduct. V.c. Ven- 
triculus communis. /V/. da¢. Lateral ventricle. V;. Third ventricle. Vj. 
Fourth ventricle. 


Fic. 1.—Dorsal view of brain of Ammoccetes. 


Fic. 2.—Transverse section through the infundibular region of the brain 
of an Ammoccetes, immediately after its metamorphosis. Picro-carmine and 
eosine preparation. The section cuts through the very commencement of 
the right ganglion habenule. 


Fic. 8.—From the same series as Fig. 2. The section cuts through the 
posterior commissure and the very commencement of the notochord. Figs. 2 
and 3 correspond closely to Ahlborn’s Figs. 27 and 26 respectively. 

Fic. 4.—Transverse section through the commencement of the formation 
of the fourth ventricle. Hematoxylin preparation. 


Fic. 5.—Magnified drawing of the ventral portion of one side of the nerve- 
substance of the cord in Fig. 4. 

Fic. 6.—Horizontal section through the brain of a full-grown Ammoceetes. 
Osmic preparation. 

Fic. 7.—Magnified drawing of the part of the brain marked x in Fig. 6, 
taken from a hematoxylin preparation of the brain of a full-grown Ammoceetes. 

Fic. 8.—From the same series as Fig. 6, to show the optic thalamus, third 
ventricle, ventricle of the aqueduct, and the fourth ventricle. 

Fie. 9.—From same series as Fig. 2 and Fig. 3, to show substantia centralis 
gelatinosa and arrangement of groups of nerve-cells in the epichordal portions 
of the brain. 

Fie. 10.—Magnified drawing of a portion of Fig. 9. 

Fic. 11.—Transverse section through the conus post-commissuralis of 
nearly full-grown Ammoceetes. Osmic preparation. To show how the limits 
of the lumen are marked out by fat-globules, 

Fie. 12.—From the same series as Fig. 11. Magnified drawing of the grey 
matter of the epichordal part of the brain, to show the formation of the raphe 
and the limiting line of the layer of fat-globules. 

Fic. 13 a, , ¢, d, e.—Selections out of a series of transverse sections 
through the region of the conus post-commissuralis, to show the presence of 
an occluded ventral duct in that region, Osmic preparation of the head of a 
half-grown Ammoceetes, 


VERTEBRATES FROM A CRUSTACEAN-LIKE ANCESTOR. 443 


Fie. 14.—Transverse section through the region of the cerebellum, to show 
the arrangement of the cells of the arachnoidal tissue and the ventral and 
dorsal blood-sinuses. Boro-picro-carmine preparation. 

Fig. 15.—Magnified drawing of the cells of the arachnoidal tissue. Carmine 
preparation. 

Fie. 16.—Cells of arachnoidal tissue. Osmiec preparation from the same 
series as Fig. 20, 

Fie. 17.—Transverse section through the epichordal portion of the brain, 
to show the increase of the pigment round the brain. Carmine preparation 
of Ammoccetes 28 mm. in length which had been kept alive in the laboratory 
for a year. a. Pigment bearing blood-corpuscles in the ventral sinus. 

Fie. 1$.—Magnified drawing of red blood-corpuscles, to show formation of 
pigment in them. a. Normal blood-corpuscle. 4. Commencing pigment 
formation round the nucleus. 

Fie. 19.—Transverse section through the spinal cord of a full-grown 
Ammoceetes. 

Fic. 20 a, b, c, d—From the same series as Fig. 6 and Fig. 8, to show the 
relation of the dorsal pineal eye and its nerve to the surrounding tissue, and 
also the shape of the cavity of the eye. ez. Place of entrance of nerve. 
ec. nas. Nasal cartilage. 

Fie. 21.—Horizontal section through the ventral pineal eye of a full-grown 
Ammoceetes. Boro-picro-carmine preparation. ghl,. Main portion of left 
ganglion habenule. g//;. Optic portion of left ganglion habenule. pz. Fold 
of pia mater which separates the ventral eye from the left ganglion habenule, 
and is perforated by the bundle of fibres (/) in one place. 

Fie. 22.—Magnified drawing of a portion of the pineal eye represented in 
Figs. 20a—d. The drawing is taken from the section immediately following 
204 as we pass in the direction of 20c, to show the nature of the contents of 
the central cavity, and the arrangement of the lines of pigment. 

Fic. 23,—Ends of the nerve-end cells of a lateral eye of Euscorpius. 
From Lankester and Bourne, fig. 5. 

Fic. 24.—Section across the nerve-end cells of a lateral eye of Euscorpius, 
showing the cuticular rods. From Lankester and Bourne, fig. 6, 

Fie. 25.—Termination of nerve-end cells bearing rhabdites of an Acilius 
larva. From Grenacher. 

Fre. 26.—Horizontal section through the dorsal pineal eye of a full-grown 
Ammoceetes, to show the relation of the eye to the walls of the cranial cavity. 
Carmine preparation. as. Nasal cavity. 

Fic. 27.—From same series as Fig. 21, to show the structure of the dorsal 
pineal eye and of the right ganglion habenule. The dotted lines represent 


444 W. H. GASKELL. 


the course of the nerve as seen in the whole series of sections. The part of 
nerve shaded was the only part visible in this particular section, 


Fic. 28.—Diagram of the pineal eye of Ammoceetes as restored, to show 
its Arthropod characteristics. For the sake of clearness most of the pigment 
surrounding the nerve-end cells is not shown in the diagram, but its limits 
only are indicated by the lines of black dots. 


Fic. 29.—Diagram to show the transformation of the walls of the Crustacean 
cephalic stomach into the lining epithelium of the cavities of the brain of 
Ammoceetes. ed represents the contour of the cephalic stomach. Yellow, 
the contour of the lining epithelium of the cavities in the brain of Ammoccetes 
if the choroid plexuses were unfolded [cf. Ahlborn (8), fig. 41, Pl. xvi]. Black 
indicates the position of the principal nervous structures in relation to the 
Crustacean cephalic stomach, and to the brain-cavities respectively. 


DEVELOPMENT OF ATRIAL CHAMBER OF AMPHIOXUS. 445 


The Development of the Atrial Chamber of 
Amphioxus. 


By 
E. Ray Lankester, M.A., LL.D., F.R.S., 


and 


Arthur Willey, 
Student of University College. 


With Plates XXIX, XXX, XXXI, XXXII. 


Last year one of us (Mr. Willey) collected during the months 
of May, July, and August many hundreds of embryos and 
larve of Amphioxus in Sicily. 

The material so obtained has been worked out in the labora- 
tory of University College. The period of the development, 
to which we determined first of all to give attention, was that 
before which Hatschek’s well-known work stops short. Series 
of sections were prepared in order to ascertain the mode in 
which the atrial chamber takes its origin and the subsequent 
history of the gill-slits, viz. as to how the slits on the left side 
of the pharynx originate. The relation of the larval to the 
adult mouth and the details of the curious process of move- 
ment of the mouth from a unilateral to a median position were 
included in the scope of our inquiries. 


Amphioxus occurs in great numbers in a comparatively 
small lake, or pantano, which is situated behind, and sepa- 
rated from the sea by, the village of Faro, near Messina. It 


VOL. XXXI, PART IJI.—NEW SER. GG 


AAG E. RAY LANKESTER AND ARTHUR WILLEY. 


is connected with the Straits of Messina by a narrow canal, 
some two or three hundred yards in length. 

The bottom of the pantano, in contrast to that of the 
Straits, consists of foul mud ; and it may be mentioned in this 
connection, as stated by Professor Kleinenberg,' that Amphi- 
oxus is only occasionally met with in the Straits, and is 
entirely absent from another larger pantano which lies 
behind the neighbouring village of Ganzirri, and is joined 
by a short canal to the one at Faro. 

The embryos float on the surface, and are to be had by 
dredging on the surface at sunrise; but the readiest method of 
obtaining them in quantity is to take the adults in glasses 
and allow them to spawn there, if they will. Spawning takes 
place about an hour after sundown. 

The ova, if fertilised, must be very carefully distributed 
among several glasses containing clean, but unfiltered, water 
from the pantano. If the water is filtered, or if sea-water is 
employed, or if too many ova are placed in one glass, they will 
certainly either die or develop abnormally. 

The first outward and visible sign of fertilisation is the 
separation from the egg-cell of the yolk-membrane (Dotter- 
membran). 

Most, if not all, of the ova obtained were discharged through 
the atriopore. 

If Kowalevsky? had not seen them issuing from the mouth, 
it would not have been supposed that they could pass into 
the pharynx in opposition to the constant outflow of water 
between the gill-bars. 

Segmentation always commences at dusk—between the 


1 J desire to express my sincere thanks to Professor Kleinenberg, of 
Messina, for his kindness and for the invaluable assistance which he gave to 
Mr. Willey in accomplishing the object of the latter’s visit to Sicily. I have 
also to thank the Government Grant Committee of the Royal Society for 
the funds which enabled me to obtain Mr. Willey’s services in this inquiry.— 
1g oi EH 

2 « Wntwick. des Amph. lance.” (‘Mém. Acad. Impér. des Sciences de St. 
Pétersbourg,’ series vii, vol. xvi, 1866). 


DEVELOPMENT OF ATRIAL CHAMBER OF AMPHIOXUS. 447 


hours of seven and eight—and goes on very rapidly through 
the night. 

The early stages have been so fully described by Hatschek ! 
that it will only be necessary to refer to them in the briefest 
manner. 

At 8 p.m. segmentation commences; at 11 p.m. invagination 
commences; at 1 a.m. the gastrulais complete; at 3 a.m. the 
gastrula begins to revolve by cilia within the yolk-membrane ; 
and at 5 a.m. two pairs of myoccelomic pouches have been 
formed, and the embryo ruptures the egg-membrane and 
becomes free-swimming. 

During the first day the embryo grows in length and adds 
several] pairs of somites. By about eight o’clock on the 
second morning—that is, thirty-six hours after the commence- 
ment of segmentation—the embryo has acquired a mouth on 
the left side of the body, and a gill-slit, which arises at first in 
the median ventral line, and subsequently comes to lie on the 
right side of the body. 

The anus is formed soon after the appearance of the mouth 
and first gill-slit. 

The embryonic period is now at an end, and the larval 
period begins. As Hatschek states, the only way of obtaining 
the larval stages is by pelagic fishing. This consists in 
dredging at depths varying from fifteen to twenty fathoms. 
At this depth the Amphioxus larve float in the midst of 
countless thousands of Sagitta larve. 

A long, but not yet clearly ascertained interval (probably 
about a fortnight) elapses between the formation of the first 
and second gill-clefts. 

In the period during which it is free-swimming the larva 
acquires from twelve to fifteen consecutive unpaired gill-slits, 
each one arising in the mid-ventral line, and then growing in 
such a manner as to lie on the right side of the body. This 
applies to the anterior two-thirds of the pharynx, but it is not 
quite clear yet as to whether the last two or three median slits 
ever move up to the right side. Meanwhile, longitudinal 


1 Claus’s ‘ Arbeiten,’ 1881. 


44.8 E. RAY LANKESTER AND ARTHUR WILLEY. 


ridges, which are subsequently concerned in the formation of 
the atrium, have appeared (see Fig. 6). In this stage the larva 
rests habitually on one side at the bottom of the vessel in which 
it is kept, and does not bury itself in sand or mud. 

At the time of the completion of the atrium, which occurs at 
the close of the larval period, some remarkable changes in the 
relative position of parts of the body in the anterior region 
take place, by which the mouth becomes median, and the gill- 
slits are arranged in two series, a right and a left. The larva 
emerges from this critical phase in its development as a 
symmetrical animal, but the details of the process of “ symme- 
trisation”—the strongly marked character of which justifies the 
use of an otherwise undesirable term—are still rather obscure. 
The larva, now really a young Amphioxus, with atrium and 
paired gill-slits, ceases to lead a pelagic life, and takes to the 
sand, where it passes the rest of its life. In this condition it 
does not rest on one side on the sand, but buries itself upright 
tail downwards with the oral hood alone projecting from the 
sand (Willey obs.). Hence in the adult condition there are not 
one-sided relations of the Amphioxus to its environment. 

Spawning occurs at least from April to September inclusive. 
The best month, however, in which to obtain the embryos is 
June, while all the larval stages, up to the passage into the 
adult form, are to be found during July and August. 


Previous View as to the Formation of the Atrium. 


The hitherto accepted method of formation of the atrial 
chamber of Amphioxus is that described by Kowalevsky,! 
and more fully by Rolph.? 

Kowalevsky says that after a certain number of gill-slits 
have been formed, two longitudinal folds appear on oppo- 
site sides of the body, which grow round and meet, and 
finally fuse together in the median ventral line, leaving a wide 
aperture at one end—the atriopore. His figures, two of which 


1 © Archiv fiir Mikrosk. Anat.,’ vol. xiii, 1877. 
2 *Morphol. Jahrbuch,’ vol. ii, 1876. 


DEVELOPMENT OF ATRIAL CHAMBER OF AMPHIOXUS. 449 


are here reproduced (Figs. 1 and 2), bear this description out, 


Inne, IL ites Ye 


Copy of Kowalevsky’s figures of transverse sections through a larva of Am- 
phioxus with fully formed atrium. Fig. 1 represents a section taken 
between pharynx and atriopore; and Fig. 2, one taken just behind the 
atriopore of the same larva. Int. Intestine. a¢. Atrium. /.2. Colom. 


more or less, while Rolph’s schematic figures bear it out 
entirely. The latter are reproduced in Figs. 8, 4, and 5. 


Copy of Rolph’s theoretical section through the pharyngeal region of a larva 
before the formation of the so-called epipleural folds. MW. Nerve-cord. 
M. Muscles. D. Intestine. 4. Epidermis. Ch. Notochord. ZL’. Celom. 
a. Intestinal epithelium. 


The most serious error in Kowalevsky’s view lies in the fact 
that he makes the space in the lateral outgrowths continuous 


4.50 


EK. RAY LANKESTER AND ARTHUR WILLEY. 


with the body-cavity, and consequently calls it “ Leibeshohle,”’ 
or celom. 


= saanel 


Copy of Rolph’s theoretical section through an older larva, showing the com- 
mencing longitudinal downgrowths. #. Epidermis. #,. Visceral epi- 
thelium of the (future) atrial cavity. #,. Parietal epithelium of same. 
U. Subcutaneous tissue. 


Other letters as in Fig. 3. 


— 


—— 


— 


= 


wd 


————— 


Z AL 


> 


> 


Copy of Rolph’s theoretical section, showing the meeting together of the 
“epipleura” in the ventral middle line. 4. Atrium. &. Raphe. 
Other letters as in Figs. 3 and 4. 


DEVELOPMENT OF ATRIAL CHAMBER OF AMPHIOXUS. 451 


There is no evidence that this space is an offshoot of the 
original myoccelomic pouches: it arises apparently as an inter- 
cellular space in the midst of the connective tissue; in fact, it 
would seem to belong to that category of spaces to which the 
term “ pseudocel” has been applied. If this should prove to 
be its history it would stand in contrast to the spaces in con- 
nection with the dorsal and ventral fins, which have been 
shown by Hatschek to be derived directly from the myo- 
coelomic pouches. 

Rolph’s figures (Figs. 3, 4, 5) do not profess to be more than 
diagrams. They show the epipleur originating as a depending 
ridge on each side of the pharynx (Fig. 4). Into this ridge the 
ceelom is extended. The epipleura meet finally in the middle 
line below the pharynx according to this theory (Fig. 5). It is 
no doubt true that the scheme of growth thus sketched by 
Rolph, and based upon Kowalevsky’s erroneous figures, would 
account satisfactorily for the condition of the atrial chamber 
and its epipleural walls, as observed inthe adult. It also gives 
a basis for the suggestion made by Kowalevsky that the 
epipleura are comparable to the opercula of Teleostean fish. 

We shall now give an account of our recent observations. 


Formation of the Atrial Chamber as now 
determined. 


The first indication of the commencing formation of the 
atrial chamber is to be found in larve with nine or ten gill-slits 
on the right side. Behind the region of the pharynx we find 
that the mid-line of the body has become marked with a narrow 
groove, so that in section it is bifid (Fig. 6). The short up- 
standing ridges which limit the groove are the metapleura of 
the adult. Though at first solid, the connective tissue within 
the ridge soon becomes hollowed and forms a lymph-space, 
which we have not traced into connection with the cclom. 
These ridges can be traced from about the middle of the larva’s 
body forward towards the pharyngeal region, where they 
diverge considerably from one another (Pl. XXX, figs. A, B, C). 


A452 E. RAY LANKESTER AND ARTHUR WILLEY. 


That belonging to the animal’s left side keeps a more or less 
median position, and can be traced (though but small in eleva- 
tion) when twelve gill-slits are present as a ridge situated at 
the lower or ventral margin of the gill-slits, and dying out in 


iGs "6s re. 7. 


i) 


SRS SN ON TFb eer ta ee 


Fic. 6.—Transverse section through a larva with eleven or twelve unpaired 
gill-slits, showing the minute sub-atrial ridges. d.m. Dorsal division 
of myoccel in which the fin-ray will lie when it is developed. xe. 
Nerve-cord. chk. Notochord. m. Muscle-plate. my. Cavity of myoccel. 
d.a. Dorsal aorta. nt. Intestine. d./.m. Double-layered membrane 
separating the myoccel from the splanchnoccl. sp. Primitive splanch- 
nocel. v.a. Ventral vessel. me¢. Metapleur. s.a.7. Sub- atrial ridges. 


Fic. 7.—Transverse section through a slightly older larva. The sub-atrial 
ridges (s.a.7.) have fused for a short distance between atriopore and 
pharynx; but in the pharyngeal region the atrium is unclosed, and conse- 
quently the gill-slits still open directly to the exterior. a. Atrium. 


the anterior region of the pharynx (Pl. XXIX, fig. 6). The 
right-hand ridge, or metapleur, takes a course to the right of 
the gill-slits (which, it will be remembered, are on the right 
side of the body), and overhangs the upper limit of the slits to 


DEVELOPMENT OF ATRIAL CHAMBER OF AMPHIOXUS. 453 


a small extent. It dies out in front of the first gill-slit, where 
it bends towards the middle line. 

The atrium is formed by asmall horizontal growth (s.a.7. in 
Fig. 6), which starts from the inner face of each metapleur and 
floors in the deeper half of the groove or area between the two 
metapleura (Fig. 7, at.). 

These horizontal growths may be called the  sub-atrial 
folds. 

They are at first extremely small, and the atrial space 
floored in is a mere canal. Later the width of the atrial space 
increases greatly, and the sub-atrial folds consequently widen 


1ikes, tey 


Transverse section through an advanced larva with fully-closed atrium. The 
latter has begun to encroach on the ccelom (splanchnoceel) (sp.). Letters 
as in Figs, 6 and 7. 


also, becoming that pleated expansible floor of the atrial 
chamber, with its transverse muscular layer, which all observers 
of Amphioxus know so well (Fig. 9, s.a.r.). 


454 E. RAY LANKESTER AND ARTHUR WILLEY. 


The atrial groove becomes floored in first in the region of the 
atriopore. The growth of the sub-atrial folds extends gradually 
forwards, and the closure proceeds along one side (the right) of 


- 
\ 

—S —a' 
ES = 
— S 

q fi ji 
= SS 
\ ES SS 4 


\. 
s 


N 


Ag, 


Transverse section through an adult Amphioxus. The atrium has grown 
up so as to divide the primitive splanchnoccel into two portions—an inner 
or splanchnic, and an outer or pleural, portion (sp'. and sp"’.). sp. The 
portion of the primitive splanchnoccel which is not so affected by the 
atrium, and which persists as the dorsal ccelom. sp”. Pleural ccelom. 
sp'". Its perigonadial dilatation. fr, Fin-ray. Other letters as in pre- 
ceding figures. 


the pharynx. The whole atrium thus formed is a very small 
tube-like space. The closure by means of the small horizontal 
sub-atrial outgrowths in the region of the large gill-slits is 
somewhat difficult to explain. The small left metapleur 


DEVELOPMENT OF ATRIAL CHAMBER OF AMPHIOXUS. 455 


actually moves in course of growth from the mid-line, and 
rises on to the right side somewhat (Pls. XXX and XXXI, 
figs. 6, 7, 14, and 14a, J. met.). At the same time the much 
larger right metapleur is deepened, and overhangs the slits. 
Then the little horizontal junction is effected, and we get 
actually a nearly tubular atrium receiving the openings of suc- 
cessive gill-slits. With subsequent growth the narrow atrial 
tube widens and pushes itself right and left, so as to encroach 
on the space hitherto occupied by the ccelom, and finally it 
extends so far dorsalwards as nearly to surround the alimentary 
canal (see Figs. 8 and 9). 

The evidence of this history, in the form of careful drawings 
of various sections, at various stages in the closure of the 
atrium, together with drawings of whole larve in two stages 
of development, is given in the plates (Pls. XXIX, XXX, 
XXXI, and XXXII) accompanying this paper. It is important 
to point out that the mode of formation of the atrium as a 
narrow groove, which closes and sinks (as it were) into the 
body of the Amphioxus, is really different in important 
respects from the enclosure of a space by downgrowth of large 
folds, though ultimately no doubt the two contrasted modes of 
formation come to the same thing so far as the more obvious 
morphological relations are concerned. The mode of for- 
mation which really occurs in Amphioxus is readily har- 
monised with the existence of the post-atrioporal extension of 
the atrium which gradually tapers to a fine cecal canal. It 
also gives us an essentially different view of the region called 
“epipleur ” by Lankester, and generally so designated, from 
that which Rolph’s theory necessitated. That portion of the 
epipleur into which the myotomes of the body-wall extend is 
seen now to be no downgrowth, no extension or fold. It is 
the original unchanged body-wall which bounds the sides of 
the animal’s body in front of the atriopore, just as much as it 
does behind. The only new growth in the atrial region 
which takes part in the limitation of the surface is the sub- 
atriai growth formed by the two little horizontal folds which 
floor in the atrium when it is a mere canal. These in the 


456 E. RAY LANKESTER AND ARTHUR WILLEY. 


adult are represented by the region of longitudinally pleated 
ventral wall between the two metapleura. 

The formation of the atrium as a narrow groove which 
closes, sinks into, and expands within the body of Amphi- 
oxus, is much more readily comparable to what is known 
of the formation of the atrial chamber in the Ascidians than is 
the Kowalevsky-Rolph scheme. In the Ascidian a pair of in- 
pushings are formed, each with a circular orifice of invagina- 
tion; they expand within the body, fuse with one another to 
form one cavity, and one of the circular orifices disappears. 
In Amphioxus we have a single in-pushing with a longitu- 
dinal orifice of invagination, which closes as the invagination 
forms, excepting at its hindermost border, and then expands 
to a greatly increased volume. 

The comparison of the so-called epipleura of Amphioxus 
with the opercula of fishes has only a remote morphological 
basis, and probably no genetic relationship exists between 
these two structures. On the other hand, it is very probable 
that whilst the median fin-rays and fin including the ventral 
fin with its double rays represent the median fins of fishes—the 
metapleura represent morphologically the primitively con- 
tinuous lateral fins. The duplication of the fin-rays in the 
median ventral series of adult Amphioxus appears to be only 
a complete carrying out of a tendency to bifid structure 
which is found in the young dorsal fin-ray (see Lankester— 
Amphioxus, ‘Quart. Journ. Micr. Sci.,’ vol. xxix, Pl. 
XXXVI, s. fig. 11) ; and though in both dorsal and ventral 
median fins the fin-ray lymph-space is single, yet the floor of 
this space has a bilateral origin according to Hatschek. 

The figures which are givenin Pl. XXTX represent two stages 
of the larve of Amphioxus, an earlier with three gill-slits 
and the rudiment of a fourth (figs. 1, 2,3), and a later with 
twelve gill-slits and the rudiments of two more (figs. 4, 5, 6). 

Though the older larva is considerably larger than the 
younger, the two are, for the sake of comparison, represented 
of the same size. 


DEVELOPMENT OF ATRIAL CHAMBER OF AMPHIOXUS. 407 


Each larva is illustrated by three views: one a surface- 
view of the left side complete, one a surface-view of the right 
side of the anterior end, and one a deep focus of the anterior 
end. 

The drawings were made from carefully preserved specimens 
(killed with corrosive sublimate), stained with carmine and 
mounted in balsam. They are diagrammatic in the sense that 
they represent the results of observation rather than an actual 
view as obtained by one focussing. 

The most striking feature in both larve is the large mouth 
on the left side. In the younger larva the form of the tail, 
with its peculiar larval fin-rays, is noteworthy. In many 
respects this larval tail-fin recalls that of young Teleostean 
fishes. It is also closely similar to that of some Ascidian 
tadpoles (e. g. Styela). The small number and large size of 
the myotomes (indicated by numbers in the drawings) in the 
anterior region of the body are also remarkable. No evidence 
could be obtained by us of the intercalation of new myotomes, 
nor of the multiplication of anterior myotomes by division. 

The new myotomes appear to form exclusively at the caudal 
extremity. 

In the larger larva the full number of adult myotomes has 
been attained, and the larval tail-fin has become greatly modi- 
fied, giving place to the mesoblastic expansion which forms the 
tail-fin of the adult. 

When we remember that in the adult the oral sphincter lies 
in the vertical line of the apex of the tenth myotome, it is not 
a little astonishing to note the position of that myotome rela- 
tively to the alimentary eanal in the younger larva, and even in 
that which has attained the full complement of myotomes. The 
independence of the metamerism of the body-wall from that 
of the gill-slits and alimentary canal is thus very sharply 
indicated. 

In the cephalic region of both the older and the younger 
larva we see two remarkable larval structures, which lie in 
front of the mouth—the one in front of the buccal cavity, and 
the other within its area. These are the preoral pit and the 


458 E. RAY LANKESTER AND ARTHUR WILLEY. 


club-shaped gland. They have been figured by Hatschek, who 
has described the przoral pit as consisting of a ciliated de- 
pression and a short glandular tube, and has traced to this 
structure the thickened ciliated epithelium which is found on 
the inner face of the oral hood of the adult forming there—the 
so-called ‘‘ Rader-organ.” 

Hatschek, in his important memoir in the ‘ Arbeiten a. d. 
Zool. Institute d. Univ. Wien,’ vol. iv, 1881, does not figure 
any larva later than one with a single gill-slit. In one of the 
wall-plates of Leuckart and Nitsche, however, received by us 
during the progress of this work, there are a number of figures 
of later stages, which have to some extent assisted us in arriving 
at an understanding of the later unillustrated note by Hatschek 
(‘ Zoolog. Anzeiger,’ 1884, p. 517). None of the published 
figures exactly coincide with our younger larva as to age, and 
our later larva is even less closely represented in the diagrams 
above mentioned, so that the figures we are able to publish are 
new, and will probably be of service to naturalists. The club- 
shaped gland, though figured by Hatschek and earlier observers, 
has not been described. It is remarkable for its early develop- 
ment (observed by Hatschek), and for the fact that it seems to 
entirely disappear in the adult without leaving any trace. The 
gland is a sac with a large lumen. It lies obliquely on the 
right wall of the buccal cavity, and, bending round below, 
tapers to a narrow canal as it rises on the left wall of the 
buccal cavity, where it opens just below and external to the 
margin of the mouth. In the younger of the two larve 
figured the club-shaped gland has no internal opening ; it ends 
blindly just below the notochord. But in the later stage 
(drawn in figs. 4, 5, 6) the gland has acquired an opening into 
the cavity of the mouth. This orifice is placed at the opposite 
end of the glandular sac to its external opening (Pl. XXIX, 
fig. 5; and Pl. XXX, fig. 5, int. a., and fig. 2, ext. a.). 

By the side of and anterior to the club-shaped gland is a 
tract of modified epithelium of the buccal cavity of about twice 
the breadth of the gland itself, and divided by a median clearer 
space into two parallel tracts. This strangely-placed group of 


DEVELOPMENT OF ATRIAL CHAMBER OF AMPHIOXUS. 409 


cells has sometimes the appearance in published drawings of 
larve—of being a shadow cast by the gland, or in some cases 
looks like a duplication of it. It can be traced in the section 
drawn in fig. 2, Pl. XXX, where it is marked me. 

In the deep-focus drawings, Pl. XXIX, figs. 2 and 5, another 
tubular structure is figured, which is also seen in the transverse 
sections (Pl. XXX, figs. 2 and 3, neph., and fig. 4, neph. a.; and 
Pl. XXX1, fig. 18, neph.). This, so far as we can judge from 
the drawings given in Leuckart and Nitsche’s diagram, is the 
structure which Hatschek has described as a nephridium in the 
‘Zoolog. Anzeiger, 1884, p. 517, without a figure. In the 
condition in which we have observed this structure (viz. in 
larve ranging from the stage with three gill-slits up to closure 
of the atrial cavity) there does not seem to be any special 
reason for regarding it as a nephridium. We should prefer to 
call it the subchordal tube. It appears to end blindly ante- 
riorly, and to open into the buccal cavity near the recurved 
extremity of the glandular tract which accompanies the club- 
shaped gland. The tube les below, and to the left of, the 
notochord. 

The drawings of the larger larva (Pl. XXIX, figs. 4 and 6) 
show some interesting features as to the disposition of the 
gill-slits and the metapleura. In this larva the atrial tube 
has formed from behind (the atriopore) forwards as far as the 
hindermost still very small gill-shts (gs. 9). It is a remark- 
able fact that all the gill-slits up to this stage originate 
in the median ventral line. This is true of the first and of 
all that fellow up to the fourteenth, and possibly some few 
more. It is, however, not true of the formation of new 
gill-slits after the right and the left lateral series of gill-slits 
have become established. The figures in our plate show that, 
whilst gill-shit No. 1 occupies an entirely lateral area on 
the animal’s right side—not reaching below to the median line 
—this position is gradually receded from by the hinder slits, 
which from No. 6 onwards are seen to encroach more and 
more on the left side. When we remember that gill-slit 
No. 1 as well as all that follow it originated in the 


4.60 E. RAY LANKESTER AND ARTHUR WILLEY. 


median line, it is clear that the anterior slits must undergo a 
translation in growth which moves them up the right side. 
Now, if we look at the slits following No. 6, it appears as 
though a translation of these hinder slits were in progress, 
tending to bring them into position on the left side when 
fully formed. We do not, however, consider it likely that 
such a movemeut of the hinder slits to the animal’s left side 
takes place, but believe that they also in due time move up 
firstiy to the right side, alongside of those in front of them. 
We have found it impossible with our present material to trace 
the immediately subsequent history (subsequent to the stage 
drawn in figs. 4 and 6) of the gill-slits. We are of opinion 
that Kowalevsky’s very definite statement and figures given in 
the ‘ Mémoires de l’Acad. Imp. de St. Pétersbourg,’ 7th series, 
vol. xvi, No. 12, 1866, must be accepted. According to that 
account, after some dozen gill-slits have taken up their position 
on the animal’s right side—having moved into that position 
from the median line—a new and startling change occurs. 
The whole series moves downwards across the median line and 
up the left side of the pharynx, so that the primitive right-side 
gill-slits become the left-side series; and in the meanwhile a 
new series corresponding in number make their appearance 
not one by one, but all together, in the right side of the 
pharynx, occupying, as it were, the position deserted by the 
rotated primitive series. This movement of growth appears to 
be a general one affecting the whole pharynx, for, simulta- 
neously with the translation of the primitive gill-slts from 
right to left, the great larval mouth moves from its extra- 
ordinary position on the animal’s left side, and, becoming 
relatively very much smaller, takes up its permanent position 
as an anterior median orifice whilst its hood and tentacles 
appear. We have not, we regret to say, at present been able 
to study any larvee in which these remarkable changes are in 
progress. We have, however, many larve in which they are 
completed. It is noteworthy that these larve are scarcely, if 
at all, larger than that of Pl. XXIX, fig. 6; and yet they have 
the mouth reduced in size and nearly median in position, the 


DEVELOPMENT OF ATRIAL CHAMBER OF AMPHIOXUS. 461 


anterior closure of the atrium completed, and a symmetrically 
placed right and left series of gill-slits. 

We have taken steps to obtain the critical stages in the 
living condition during the present summer, and propose to 
ascertain whether the second row of gill-slits originates by any 
kind of fission from the first. If not, it is a curious fact that 
the morphologically median plane of the pharynx of the young 
larva becomes the left side of the adult, whilst the relations of 
the mouth to median plane, in adult and larva respectively, are 
even more curiously divergent. It is probable enough that in 
these differences the larva does not present the more archaic 
condition, but an adaptational arrangement. We do not at 
present know what are the conditions of life which render its 
excessive asymmetry advantageous to the larva. 

The closure of the atrium by the growth of the little hori- 
zontal sub-atrial ridges from the median face of each metapleur 
is shown in the sections of various larve given in Pls, XXX, 
XXXI, XXXII. 

In the drawings, figs. 4 and 6 of Pl. XXIX, we can trace 
the two metapleura in the still unenclosed region of the 
pharynx. The right-side metapleur is seen to have its free 
edge somewhat high on the animal’s side, whilst the left 
metapleur in the perforated pharyngeal region is almost 
coincident with the median ventral line. (The reference line 
in fig. 4, Pl. XXIX, lettered “edge of left metapleur,” has 
been by oversight carried up to the right metapleur. It 
should stop at the ventral line.) The right metapleur is 
larger and deeper than the left, which is barely traceable as 
a thickening of connective tissue, when its fellow of the 
opposite side is large and provided already with the character- 
istic lymph-space (see fig. 7, 7. met. and J. met., Pl. XXX). 

The figures A, B, C, in Pl. XXX, represent diagrammatic- 
ally three stages in the closure of the atrial tube, showing in 
A the metapleurs or metapleural ridges without any hori- 
zontal sub-atrial floor; in B the formation of this floor in the 
hinder region, where there are no gill-slits; and in C its 
continued formation so as to enclose the perforations of the 


VOL. XXXI, PART III.—NEW SER. HH 


462 E. RAY LANKESTER AND ARTHUR WILLEY. 


pharynx. It must be pointed out that the sections are com- 
plicated and rendered a little difficult of interpretation at first, 
by the fact that the margins of the gill-slits are irregularly 
curved and folded, so that they cross the plane of section, and 
(as in fig. 7, Pl. XXX) the slit itself becomes divided in the 
section by a part of the projecting margin. A further modi- 
fication in appearances is due to the greater or less opening of 
the gill-slits, which can be varied by muscular action during 
the life of the animal. The atrial tube or cavity is also found 
to vary in size and dimensions as soon as it is formed, owing 
to the varying extension or contraction of its muscular floor 
formed by the union of the sub-atrial ridges (compare fig. 12, 
Pl. XXX, and figs. 18 and 20, Pl. XXXII). As was pointed 
out by one of us in the case of adult Amphioxus distended 
with genital products (see Lankester, ‘ Quart. Journ. Micr. 
Sci.,’ vol. xxix, Pl. XXXV, fig. 4), so here in the larva the 
atrium can be distended to such an extent as to practically 
obliterate the metapleural ridges and their lymphatic canals, 
which reappear when the distension ceases. 

The description of the individual figures seriatim will be 
found, it is hoped, sufficiently explanatory of points which 
have not been specially mentioned in the general body of the 
memoir. 


DEVELOPMENT OF ATRIAL CHAMBER OF AMPHIOXUS. 463 


EXPLANATION OF PLATES XXIX—XXXII, 


Illustrating Professor Lankester’s and Mr. Willey’s memoir 
on the “ Development of the Atrial Chamber of 
Amphioxus.” 


PLATE XXIX. 


Figs. ] and 2.—Right and left surface-views of larva, with four gill-slits. 
Length 1°496 mm. 

Fig. 8.—Head of latter, seen with a deeper focus. Club-shaped gland 
open to exterior only. 

Fies. 4 and 5.—Right and left surface-views of larva, with fourteen gill- 
slits. Length 3°485 mm. 


N.B.—In Fig. 4 the reference line belonging to the words “edge of left 
metapleur ” has been carried too far, and touches the right metapleur. It 
should stop at the ventral line of the larva. 


Fic. 6.—Head of latter, seen with a deeper focus. Club-shaped gland, 
open at both lower and upper extremities, externally and internally respec- 
tively. 


PLATES XXX, XXXI, anp XXXII. 


The ¢éalics in Plates XXX, XXXI, and XXXII have the significance given 
below. 


ant. at. Anterior opening of atrium. aé. Atrium. at. p. Atriopore. Jr. e. 
Modified intestinal epithelium bordering the gill-slits. d. a. Dorsal artery. 
d.l.m. Double-layered membrane, separating myoccel from splanchnoccel. 
d. m. Dorsal division of myoccel in connection with dorsal fin. d. w. Dorsal 
wall of atrium. ea?. a. External aperture of club-shaped gland. yg. s. Gill- 
slit. Int. Intestine. Jnt. a. Internal aperture of club-shaped gland. fd. 
Club-shaped gland. /. a. Left dorsal artery (unpaired). /. m. Lower lip of 
mouth. 7. met. Left metapleur. m. Mouth. m.e. Modified epithelium on 
wall of mouth-cavity. mus. Muscle-plates. my. Primary myoceel. my’. 
Secondary upgrowth of myoccel, between the muscle-plates and notochord and 
nerve-cord. .c. Nerve-cord. «ch. Notochord.  xeph. Nephridium of 
Hatschek. xeph.a. Opening of so-called nephridium into mouth-cavity. 
o. h. Commencing oral hood. 7. d. Right embryonic diverticulum from the 
intestine. 7. c. So-called “renal” cells of W. Miiller. 7. met, Right meta- 


464 E. RAY LANKESTER AND ARTHUR WILLEY. 


pleur. s. a. 7. Sub-atrial ridges or floor. s. 0. Sense-organ (part of preoral 
pit). som. Somatopleur. sy. Splanchnocel. sp. py. Splanchnopleur. ». a. 
Ventral vessel. 7. o. Ciliated organ (of preoral pit). 


GENERAL REMARKS. 


The intestinal epithelium is ciliated throughout. The epithelium bordering 
the gill-slits is much modified, being divided up into innumerable small cells, 
the cell-divisions between which cannot be seen under ordinary circumstances. 

The nerve-cord consists of a nucleated portion surrounding the central 
canal and a peripheral fibrous portion. 

Nuclei are to be seen in the notochord, and in the superior and inferior 
canals of the notochord. 

There are nuclei in the muscle-plates, but, as Hatschek points out, there is 
no epithelium on the outer wall of the muscle-plates. The nuclei on the 
inner wall are sufficiently scanty. 

The sense-organ and ciliated organ of the preoral pit are derived together 
from the left anterior diverticulum of the archenteron of the embryo, while 
the right diverticulum becomes simply the space occupying the anterior end 
of the body. It is included in Fig. 1, but not in Fig. 13. 

A reference to the drawings of the whole animal in Pl. XXIX will show 
approximately through what regions the sections have been taken. 


Fies. A, B, C.—Three diagrams of larvee, seen from ventral aspect, to 
illustrate the origin and relation of the metapleural ridges to one another, 
and the gradual closure of the atrium from behind forwards. 


Fig. A. No atrium. 
Fig. B. Atrium behind pharynx. 


Fig. C. First two gill-slits open to exterior, all the rest now open into 
the atrium. 


Fic. 1.—Transverse section through the region of the ciliated organ and 
sense-organ of the preoral pit, just in front of the opening of the latter into 
the former. The anterior commencement of the splanchnoccel, and the pos- 
terior portion of the right embryonic diverticulum are shown. The epithelium 
of the preoral pit is of hypoblastic origin (Hatschek). This larva had twelve 
gill-slits, and no closed atrium. Preparation: osmic acid, borax car., fol- 
lowed by Meyer’s carmine. 

Fic. 2.—Transverse section through the commencement of the mouth- 
opening, showing the external aperture of the club-shaped or tubular gland. 
It also passes through the tract of modified epithelium. The very thin piece 
of epithelium, two thirds of the way up, is the cause of the clear space or line 
which gives a double appearance to the tract. The thickening of the right 
metapleur is tending to the right side. This larva had eleven slits. Prepara- 
tion: sublimate and acetic; hematoxylin. 


DEVELOPMENT OF ATRIAL CHAMBER OF AMPHIOXUS. 465 


Figs. 3 and 4.—Portions of sections through another larva of same age, 
and prepared in same way as the last, taken just posterior to the region 
represented in Fig. 2, to show the opening of the nephridium of Hatschek 
into the mouth-cavity. Notice that the nephridium lies immediately below 
the left dorsal artery. 

N.B.—In the larva there are not two dorsal arteries—right and left—in the 
pharyngeal region, as there are in the adult ; but only one, and that on the 
left side of the notochord. 


Fie. 5.—Section through first gill-slit of same larva, showing the club- 
shaped gland opening into the mouth-cavity at its upper extremity. The 
ventral vessel lies on the right wall of the intestine in the pharyngeal region. 
No cavity yet in the right metapleur. 

Fic. 6.—Section through same larva as Fig. 1 (twelve gill-slits), through 
the same region as preceding, to be compared with Fig. 5 (with eleven 
gill-slits) where the mouth is half shut. In this case the mouth is wide open, 
and the appearance of the section is considerably altered owing to the expan- 
sion of the ventral portion of the celom. The right metapleur is more 
advanced, but still has no cavity in this region. 


Fic. 7.—Section through the sixth gill-slit of the same larva. The double 
appearance of the slit is due to a fold in the wall of the slit. The right meta- 
pleur has a cavity here. The left metapleur has commenced as a thickening. 

Fic. 8.—Section through twelfth and last gill-slit of same larva. The 
metapleural folds are nearly equal. There is a very small cavity in the right 
and none in the left fold. 

Fie. 9.—Section through the post-pharyngeal region of a larva preserved 
with osmic acid vapour, rather older than Fig. 8, but with no part of the 
atrium floored in. The various divisions of the mycccel will be understood by 
a reference to Hatschek’s figures, reproduced in Professor Lankester’s paper 
in this Journal, vol. xxix, Pl. XXXVIA, figs. 6 and 7. 


Fic. 10.—Section through the twelfth (last but one) slit of a larva of the 
age of that represented in Pl. XXIX, fig. 4. Preparation: concentrated 
sublimate ; borax carmine. 

Fig. 11.—Section through the post-pharyngeal region of the same larva 
(ef. Fig. 9), showing the fusion of the sub-atrial ridges. The character of 
the latter as ridges on the inner faces of the metapleura is not so well seen 
here as in other sections. 

Fic. 12.—Section through the same region of another larva of the same 
age, showing the method of fusion of the sub-atrial ridges as described in the 
letterpress. Preparation: osmic acid and picro-carmine. 

Fie. 13.—Section through the compound sense-organ (= preoral pit) of a 
larva in which all the gill-slits, except the first two, opened into a floored-in 


466 E. RAY LANKESTER AND ARTHUR WILLEY. 


atrium. It shows the sense-organ (s. 0.) opening into the ciliated organ (7. 0.), 
and the latter opening widely to the exterior. It also shows the independent 


origin of the oral hood (0. 4.). Preparation: concentrated sublimate ; borax 
carmine. 


Fic. 14.—Section through the anterior opening of the atrium, in the same 
larva. It shows very well the position of the atrium anteriorly on the right 
side, also the sub-atrial ridges. Note the relatively huge size of the right 
metapleur, and the almost entire absence of any indication of the left meta- 
pleur; thus showing that the sub-atrial ridges are distinct structures from, 
and only secondarily dependent on, the metapleural folds; and that the latter 
serve a function (probably vascular) other than that of merely contributing 
to the formation of the atrium. This section is between the second and third 
gill-slits. The second slit opens to the exterior, the third opens into the 
atrium. The large cells at the bottom of the right metapleur are still in the 
epidermis ; whereas, in fig. 7, Pl. XXX, they have migrated inwards. 


Fic. 14a@,—Section through the same larva as the preceding, two or three 
sections farther back, showing a gill-slit (the third) opening into the laterally 
placed atrium. 


Fic. 15.—Section between the tenth and eleventh slits of the same larva, 
showing an older condition of the atrium than that represented in Figs. 11 
and 12, with “renal”? cells on the dorsal wall. Note also the large size of the 
metapleura. 


Fic. 16.—Section through the post-pharyngeal region of the same larva, 
showing a still more advanced condition of the atrium. The gelatinous sub- 
cutaneous tissue has disappeared from the dorsal wall of the atrium, leaving 
a thin double membrane, consisting of ccelomic and atrial epithelium 
(= somatopleur). 

Fic. 17.—Section through the atriopore of same larva. 


Fic. 18.—Section through the last gill-slit but two of a larva with the 
atrium floored in over three slits, showing expansion of atrium and temporary 
obliteration of metapleural spaces. Preparation: osmic acid and picro- 
carmine. 


Fics. 19 and 20.—Sections through a larva in which the atrium had closed 
over two slits (the fourteenth and fifteenth), showing a narrow condition of 


the atrium in front (Fig. 19), followed by a more expanded condition behind 
(Fig. 20). 


A NEW GENUS OF OLIGOCHATA. 467 


On the Structure of a New Genus of Oligocheta 
(Deodrilus), and on the Presence of Anal 
Nephridia in Acanthodrilus. 


By 


Frank E. Beddard, M.A.. 
Prosector of the Zoological Society of London. 


With Plates XX XIII and XXXIIIa. 


I. On the Structure of a New Genus of Oligocheta. 


THE present paper is based upon the study of only a single 
example of the worm. It was collected some years ago by 
Prof. Moseley in Ceylon, and was kindly entrusted to me for 
description by Prof. W. Hatchett Jackson. 

The specimen measures thirteen inches in length by nearly 
half an inch in diameter at the broadest part (at the end of 
the 8th or 9th segment). Its intermediate characters lead 
me to suggest the generic name Deodrilus; the specific name 
I propose to associate with Mr. Jackson. 


§ External Characters. 


The prostomium is entirely absent, as it is in some other 
genera allied to the present. 

The first or peristomial segment is traversed by longitu- 
dinally running grooves, which give it a characteristic appear- 
ance, often seen in worms when there is, as in this genus, no 
prostomium.! 

The three following segments are of about equal antero- 
posterior diameter, though increasing rapidly in their breadth 


from side to side. 
1 Cf., for example, Beddard (1). 
VOL. XXXI, PART 1V.—NEW SER. 101i 


468 FRANK BE. BEDDARD. 


The 5th segment is the first which shows the commencing 
formation of annuli. A slight furrow crossing this segment 
partially divides it into two rings ; the setz are implanted just 
in front of this furrow. 

The 6th ring is similar to the fifth, only that it is slightly 
longer. 

The 7th ring is broader still, and is divided by two furrows 
into three annuli, the sete being conspicuous upon the middle 
one of the three. 

The 8th to the 14th segments are very broad, and each is 
divided by four furrows into five annuli, of which the middle 
one carries the sete. 

The 14th segment, though of equal diameter to those 
preceding it, has only three furrows; but there is an indica- 
tion of the fourth. 

The 15th and 16th have three annuli each. 

After this point I cannot give accurate details, as the ridges 
which carry the male generative pores have introduced altera- 
tions into the annuli of the neighbouring segments. 

Clitellum.—I am unable to map exactly the boundaries of 
the clitellum, as it did not appear to be fully developed. 

On dissection, only two segments (15 and 16) showed a 
very marked difference in the structure of the body-wall from 
the others ; here the deep yellow colour was very apparent, and 
was fully as well developed on the ventral as on the dorsal side, 

The genital orifices, as already stated, are borne on two 
longitudinally running ridges, coinciding in position with the 
ventral series of sete on each side. 

These ridges, particularly in the immediate neighbourhood 
of the male pores, hada very glandular appearance. It is quite 
possible that the clitellum when fully developed extends as far 
as these ridges do, 

In that case it may be stated provisionally that the clitellum 
extends over four segments (viz. 15—18). 

It will be noted, however, that the clitellum differs in its 
anterior and posterior regions. The first two segments are 
entirely invaded by glandular substance, while the three posterior 


A NEW GENUS OF OLIGOCHATA. 469 


segments will in all probability be found to have a median area 
upon which there is no great modification of the epidermis. 
This area is of course bounded laterally by the genital ridges. 
It seems, therefore, that the clitellum of Deodrilus is consti- 
tuted upon the same plan as that of Acanthodrilus. 

Dr. Rosa has made some use, in his scheme of classification 
of earthworms, of the form of clitellum, which he terms saddle- 
shaped (“ clitello a sello”’), or complete (‘‘ cingulo completo ”), 
admitting that Acauthodrilus offers an intermediate con- 
dition. 

The fact is that it is not possible to classify the various modi- 
fications of the clitellum in this way. 

There is a considerable series of gradations which renders it 
impossible to make a fixed demarcation between the different 
forms of clitellum. To commence at one extreme, we have 
species of Lumbricus and Allolobophora with a distinc- 
tively saddle-shaped clitellum: in these forms the glandular 
modified epidermis is only to be found on the dorsal and lateral 
regions of the clitellum ; ventrally there is a wide space of equal 
diameter throughout, which has no trace of glandular tissue. 

In such a form as Rhinodrilus Gulielmi the clitellum 
is divisible into two regions: in the last six segments of which 
it is composed the glandular substance is arranged quite as in 
Lumbricus; but in the first four the ventral area is en- 
croached upon by the glandular tissue, though it is not 
completely invaded. 

In Urocheta the clitellum is constituted in a way quite 
resembling that of Rhinodrilus, but the anterior bare ventral 
space appears if anything to be somewhat narrower. 

In Deodrilus (which, as will be shown later, has points of 
affinity with Rhinodrilus) the anterior part of the clitellum 
is completely developed—extends all round the body—but the 
greater part is still only laterally developed. 

Acanthodrilus has aclitellum inwhich the anterior portion 
is at least as great as, and may be greater than, the hinder 
portion, which still retains its saddle-shaped character. 

Finally, we have such a form as Perionyx, in which the 


470 FRANK E. BEDDARD. 


six or seven segments of the clitellum are completely occupied 
by the modified epidermis.' 

Hence it appears to me to be inadvisable to use the 
characters of the clitellum to help in associating together, as 
Rosa has done, his families Lumbricide and Geoscolecide. 
The structure of Deodrilus shows that this cannot be done; 
and there are, among the Geoscolecide of Rosa, intermediate 
conditions leading to Deodrilus. 

Genital Papillew.—Most earthworms are furnished with 
genital papillae, which are often very characteristic of the 
species in which they are found. 

Deodrilus is not an exception to this rule, and has two 
sets of genital papille. 

The first set consists of a single pair of large flattened 
papille, which are fused together to form a dumb-bell shaped 
area extremely conspicuous. The outer convex border on each 
side reaches to the level of the inner sete of the outer pair. 

The papille (see Pl. XX XIII, fig. 12,) are situated between 
the 11th and 12th segments; they occupy the last annulus of the 
former segment and the first two annuli of the latter: the furrows 
dividing the annuli and the two segments from each other can be 
seen as faint lines traversing the papille. The furrow separating 
the last annulus of Segment 11 from the penultimate forms the 
anterior boundary of the papillz, while its posterior boundary 
is formed by the furrow separating the second from the third 
annulus of Segment 12. The second set of genital papille are 
very much less conspicuous ; they are to be found near the male 
genital pores, and no doubt correspond to the tubercula puber- 
tatis: in front of each of the male orifices are two papille—one 
in front of the other; and behind each is a single papilla. So 
far as I can ascertain, they belong to Segments 17, 18, and 
19; but, as I have already said, it is difficult to be certain 
about the limits of the segments in this region of the body. 

Setz.—The sete are entirely restricted to the ventral 
surface of the body, where they are implanted in pairs. 


1 T do not for the present consider how far the form of the clitellum may 
be influenced by the position of the generative pores, 


A NEW GENUS OF OLIGOCHATA. 471 


On the 8th segment I found that the distance separating 
the two ventral pairs from one another was 2} mm.; the 
lateral pairs were separated by a dorsal area measuring 14 mm. 
The shape of the sete is very remarkable, and is illustrated 
in fig. 18. Their general outline is similar to that of the 
sete of other earthworms, but instead of terminating in a 
hooked extremity they present a truncated appearance, which 
will be understood by a reference to the figure cited. It 
occurred to me, when first observing the sete attached to 
fragments of stripped-off cuticle, that they might be of the 
normal form, but with their extremities broken off. It fre- 
quently happens, as a result of rough usage, that the majority 
of the sete, or at any rate a large number, are broken off 
short ; but it is evident that that is not what has happened in 
the present case. The free extremity of the sete showed no 
signs of having undergone any fracture; and, moreover, the 
shape of the freely projecting extremity is not such as would 
be produced by a fracture, or wear and tear. 

The sete illustrated in fig. 18 are drawn as seen on a 
lateral view; a@ and & represent the free extremities of such 
sete more highly magnified. 

Another peculiarity in the structure of the sete is the fact 
that their distal region, i.e. that part which lies external to the 
slight swelling in the middle of the setz, is ornamented by 
minute pointed processes. 

The description just given, and the figures which illustrate 
it, refer to setz from the first ten segments or so; but I have 
ascertained that the sete in the posterior segments are abso- 
lutely identical in size and structure with those from the 
anterior segments. 

The first five segments of the body are entirely deprived of 
sete ; a microscopic examination of these segments did not 
enable me to find any trace of the presence of setz. 

The disappearance of the sete from the first few segments 
of Deodrilus, and of the species of Diachzta which I de- 
scribed recently in this Journal (1), is a remarkable fact. 
No other instances are at present known among earthworms, 


4,72 FRANK E. BEDDARD. 


though possibly Microcheta Rappi may prove to be one; 
at any rate, it is excessively difficult to recognise the sete 
upon the first few segments; I have not myself been able 
to find them at all. They are, however, figured by Benham 
(11, pl. xv, fig. 1); but he makes no definite statement as to 
their presence on the first two or three segments. Leaving 
Microcheta aside, the absence of setz on the anterior seg- 
ments is correlated with the entire absence of a prostomium. 
Not a vestige of this structure could be recognised in either of 
the two forms mentioned. This correlation, however, is not 
universal, for Urocheta has no prostomium, and yet the 
sete are visible, as usual, from Segment 2 onwards. It is 
furthermore noticeable that in Diacheta Windlei the 
anterior non-setigerous segments show another modification in 
the presence of the specialised bundle of transversely running 
muscular fibres, which I have figured and described as occur- 
ring in the segments beginning with the 6th. 

In most earthworms the first or peristomial segment is so 
far unlike the rest that it is grooved longitudinally, and that 
its epidermis is not clearly distinguishable into two classes of 
cells, or at least is not so clearly distinguishable as is the epi- 
dermis of the following segments. Sometimes this modifica- 
tion appears to affect the 2nd as well as the 1st segment. 

In connection with this modification the varying position 
of the prostomium may be pointed out. Sometimes the 
prostomium is attached to the anterior border of Segment 1; 
in other species it encroaches upon this segment, and finally 
it often completely divides the lst segment, and reaches the 
anterior border of the 2nd. 

Earthworms, in moving along, use the mouth as a kind of 
sucker, even protruding a portion of the buccal cavity; this 
is remarkably the case with Pericheta indica (8), which 
everts what appears to be the whole of the buccal cavity 
at each movement. Conversely, there is often a temporary 
withdrawal of the peristomial segment into the mouth-cavity. 
These two phenomena appear to have led to the different enu- 
meration of the segments of Urochzta adopted by Perrier 


A NEW GENUS OF OLIGOCHATA. 473 


(18), Horst (12), and Rosa (10). As the latter has pointed 
out, Perrier appears to have described a specimen in which the 
buccal cavity was partly everted; while Horst, in stating that 
the mucous gland opened on to the first segment of the body, 
was deceived by the introversion of the peristomial segment. 

Among the half-dozen series of longitudinal sections of 
Urocheta corethrura which I possess there is one in which 
the 1st and a portion of the 2nd segment are introverted, 
and the mucous gland appears, therefore, to open into the 
buccal cavity ; in fact, it actually does open into a temporary 
extension of the buccal cavity. 

It is, in my opinion, possible to believe that a temporary 
introversion,suchas that to which I havejustreferred, 
may become permanent. In this case what will happen 
are two events of importance. In the first place the body will 
be shortened by one segment ; in the second place the ‘‘ mucous 
gland” will come to open into the anterior section of the 
alimentary tract. 

As to the second point, I may call attention to the remark- 
able condition of the anterior nephridia in Acanthodrilus 
multiporus (4). In that worm the anterior segments are 
occupied by a mass of glandular tubes, clearly of nephridial 
nature, on each side of the pharynx. Each mass communicates 
with a long duct, which opens into the buccal cavity. It 
seems impossible to doubt that the nephridial masses of this 
Acanthodrilus originally opened on to the exterior, and that 
their connection with the buccal cavity is only secondary. That 
this “‘ secondary”? connection may be really the original point 
of opening, masked by the partial or entire introversion of the 
lst segment, is surely not incredible. 

With regard to the first point, the possible shortening of the 
body in this way involves really no serious difficulty, though it 
seems, of course, rather ridiculous to gravely assert that a worm 
becomes shorter by swallowing its own head. The structure of 
the epidermis of the 1st segment is more like that of the buccal 
cavity than it is like that of the succeeding segments. 

These remarks, however, apply more particularly to such 


A474 FRANK E. BEDDARD. 


worms as Diacheta and Deodrilus, in which there is no 
prostomium, and in which, therefore, such an inversion will 
cause no external change of importance. 

In worms which have a prostomium this structure may be 
prolonged backwards, so as to commence as an outgrowth of 
the 2nd segment; if, in such a case, the peristomial segment 
were permanently invaginated the prostomium would be left 
attached to the lst segment of the body, i.e. in the more usual 
position; on the other hand, a permanent eversion of the com- 
mencement of the buccal cavity in a worm in which the 
prostomium arises from the anterior margin of the peristomial 
segment would lead to the apparent prolongation of the pro- 
stomium back to the 2nd segment. 


§ Reproductive Organs. 

I have not found the testes nor the ovaries and oviducts. 
Two pairs of sperm-sacs were to be seen attached to the an- 
terior wall of the 10th and 11th segments; these organs are 
racemose in form, as in so many genera. 

The vas deferens funnels appear also to be limited to a 
single pair, which open into the 11th segment. 

The atrium, or prostate gland, is a compact flattened body 
on each side of the body connected by a short muscular duct with 
the male pore; it lies in the 18th segment. The atrium appears 
to be branched, and to resemble the same organ in Pericheta. 

Connected also with the male reproductive apertures is on 
either side a thin-walled sac filled with penial sete. ‘Two of 
these sete are shown in figs. 15, 16; it will be seen from 
an inspection of those figures that the form of the two setz 
selected for illustration differs very considerably. In one the 
distal extremity is covered with numerous minute points like 
those which cover the distal half of the ordinary seta. In the 
other seta these points are entirely absent, and fine wavy lines 
are found on the distal part of the seta, ceasing, however, some 
little way in front of the extremity. 

I call particular attention to the fact that two such different 
forms of penial sete are met with in the same individual, inas- 


A NEW GENUS OF OLIGOCHATA. A475 


much as these setze have been made use of by myself and others 
as specific characters. I may recall the fact that in Acantho- 
drilus Georgianus (5) there is an analogous dimorphism 
of the penial setz. 


§ The Intersegmental Septa. 


As is so constantly the case among earthworms, certain of 
the intersegmental septa are specially thickened, as well as con- 
nected with each other and with the parietes by muscular 
bands. 

In the present species the septa between Segments 6 and 
13 are thus strengthened, there being, therefore, seven. As 
in other cases, these septa are very concave forwards, the 
middle region lying much behind the peripheral attached 
margin ; the septa present, therefore, have the appearance of 
a series of cups, each fitting within the one which follows it. 


§ Alimentary Canal. 


The gizzard lies in Segment 6 (cf. fig. 14). 

The esophagus extends as far back as Segment 18; it 
does not, however, abruptly widen into the intestine, which 
only commences (fig. 13) in the 20th segment. 

In Segments 15, 16, and 17 are three pairs of calciferous 
glands. As shown in the figure (fig. 13), each of these glands 
is divided into two by a transverse furrow. 


§ Nephridia. 

The worm was not in a sufficiently good state of preserva- 
tion to allow of any observations upon the minute structure of 
the nephridia. 

So far as could be ascertained by dissection, the nephridia 
throughout the body appear to be of the “ diffuse” type. 

Lying alongside of the pharynx on each side was a con- 
spicuous glandular body, which is doubtless similar to the 
salivary gland of Acanthodrilus multiporus, and to the 
“ mucous” gland of Urocheta, Diacheta, &e. 


4.76 FRANK E. BEDDARD. 


§ Affinities of Deodrilus. 


The question of the systematic position of this Annelid 
necessitates some review of recent attempts to classify the 
group. As, however, I intend to publish an attempt at the 
classification of these Annelids with a criticism of existing 
schemes, I shall make my references as brief as possible. 

Deodrilus evidently belongs to the “ Intraclitellian” 
group of Perrier (13) ; but as Perrier’s scheme—undoubtedly 
a great advance upon what had gone before—is not now gene- 
rally accepted, I shall not urge any reasons why Deodrilus 
does not fall in with that classificatory attempt. 

The most recent schemes are those of Rosa (9) and Vail- 
lant (14).1_ Rosa has divided earthworms into six families— 
Lumbricide, Geoscolecide, ? Moniligastride, Acan- 
thodrilide, Eudrilide, Perichetide. 

It is only with the second and fifth of these families that 
Deodrilus can have any connection. 

These families are defined by Rosa as follows. 


GEOSCOLECID. 


(1) Male pores within the clitellum between the dorsal and 
ventral sete, occupying segments or intersegmental grooves 
which are very variable. 

(2) Clitellum generally saddle-shaped ; length and position 
variable. 

(3) Sete eight per segment, in pairs or singly, or diversely 
arranged in the anterior and posterior segments. 

(4) Copulatory sete longer than the others and of a dif- 
ferent form. 

(5) Gizzard (or gizzards) placed anteriorly. 

(6) Sperm-sacs one or two pairs. 

(7) No prostates or penial setz. 

1 Since writing the above I have received through the kindness of the 


author Mr. Benham’s ‘‘ An Attempt to Classify Earthworms,”’ ‘ Quart. Journ, 
Mier. Sci.,’ vol. xxxi, pp. 201—815. 


A NEW GENUS OF OLIGOCHATA. 477 


Evuprizip#.! 

(1) Male pores one pair, on Segment 17 or 18, within or 
behind the clitellum, corresponding to the ventral sete. 

(2) Clitellum complete, occupying generally Segments 13 
(14) —16 (18) = 38—6. 

(3) Sete eight, paired or singly, but always parallel. 

(5) Gizzard (or gizzards*) anterior in position. 

(6) Sperm-sacs generally two pairs. 

(7) Prostate and penial setz present. 

The only characters, therefore, which are decisive are (1), 
(2), and (7). 

Deodrilus agrees with the Eudrilide in (1) and (7), and 
it is intermediate between the two in (2). 

But there are other facts in its structure which signify that 
it combines the characteristics of genera which have been 
included in the Geoscolecide and in the Eudrilida. The 
absence of a prostomium is characteristic of certain genera 
of Geoscolecide—Urocheta and Diacheta. It is true 
that Typhzeus, which Rosa refers to the Eudrilide, has no 
prostomium ;? but it must be remembered that this genus 
agrees with Urocheta, Geoscolex, and Diacheta in 
having a single pair of long tongue-shaped sperm-sacs, and 
only a single pair of sperm-ducts. 

In the absence of sete from the first few segments Deo- 
drilus resembles a species of Diacheta, of which a descrip- 
tion has appeared in a recent number of this Journal (1). 

Typheus appears to have no sete upon the first two 
segments, but, as stated in my paper upon that worm (6), I 
am not absolutely certain of the fact. 

The presence of ornamented sete affines Deodrilus to 
Rhinodrilus; nothing of the kind is met with in any of 
Rosa’s Eudrilide. 


1 [ may remark that Audrilus itself does not agree with this definition in 
all characters. 

2 Tadd “gizzards” myself, so as to include Perissogaster, &. 

® Loe. cit., p. 111. This statement requires confirmation, since Bourne 
has lately described a prostomium in Typheus Masoni. 


478 FRANK E. BEDDARD. 


The diffuse nephridial system has not yet been met with in 
any of the Geoscolecide; in this particular, therefore, as in 
the presence of prostates and penial sete, Deodrilus re- 
sembles certain genera (e.g. Cryptodrilus) of the Eudri- 
lide. Finally, the absence of diverticula to the spermathece 
brings Deodrilus into relations with the Geoscolecide. 
It seems to me that the relationship of Deodrilus to some 
other forms is best expressed by the following diagram : 


Deodrilus 
“‘ Geoscolecide ” 


Pontodrilusandmany Typheus een 
“Hudrilide” } 


Cryptodrilus 


Perichetide 


Genus—Deodrilus. 


Sete arranged in pairs upon the ventral surface, peculiar in 
shape, and ornamented. Absent upon the first four seg- 
ments. 

Prostomium absent. 

Clitellum occupying Segments 15—18 (or thereabouts), 
complete anteriorly, saddle-shaped posteriorly. 

Gizzard in Segment 6. 


1 T do not for the present particularise the exact limits which I apply to 
these families. 


A NEW GENUS OF OLIGOCHAYA. 479 


Sperm-sacs, two pairs, in 10, 11, racemose. 

Nephridia diffuse, a mass of tubules in the neighbourhood 
of the pharynx aggregated into a compact gland. 

Atria lobate, each with a sac of penial sete opening on 
to Segment 18. 


Species—D. Jacksoni.! 


Large worm, measuring 13 inches. 

Copulatory papille forming a dumb-bell shaped area between 
Segments 1] and 12. 

Penial setz of two kinds, one ornamented by minute pro- 
tuberances, the other transversely striate. 

Two pairs of spermathecz, without diverticula, in Segments 
8 and 9 (see fig. 19). 


II. Anal Nephridia in Acanthodrilus. 


I have aiready (7) described some of the anatomical and 
histological characters of the nephridia in Acanthodrilus; 
their organs are in some species (e. g. in A. multiporus) repre- 
sented by irregular tufts, which are furnished with numerous 
funnels, and numerous external orifices in each segment. 

I have now to describe a connection of the nephridia with 
the terminal region of the intestine, which occurs in a species 
referable, I believe, to my Acanthodrilus multiporus. 

The material I owe to the kindness of Mr. W. W. Smith, of 
Ashburton, New Zealand. 

In fig. 1 of Pl. XX XIII, is represented half of a portion of the 
posterior end of the body of one of these worms, comprising 
about thirty segments. The body was divided by a cut at 
right angles to the dorso-ventral axis ; the intestinal canal is 
Jaid open, aud the typhlosole is seen to occupy the dorsal line 
of the intestine. The figure is twice the size of nature. 

The typhlosole ends abruptly about an inch in front of the 
anus; a faint streak is, however, recognisable, extending for 
perhaps a quarter of an inch beyond the end of the typhlosole, 


1 Named after Mr. W. Hatchett Jackson. 


480 FRANK E. BEDDARD. 


The part of the alimentary tract lying behind the typhlosole 
is thus sharply marked off, and the distinction between it and 
the terminal section is possibly an important one; I should 
be inclined to regard the “ rectum” as being proctodeum. 
The diameter of the rectum, as shown in the figure, gradually 
narrows to the anus; its walls are marked by longitudinally 
running furrows. 

In transverse sections the gut is seen to be lined by an 
epithelium of tall columnar cells, broader towards their ex- 
tremities and narrower at their attachment. The folds 
observable (fig. 5) in such sections are, of course, due to the 
longitudinally running furrows. Outside the epithelium 
is a circular coat of muscular fibres, and outside this again 
longitudinal fibres; these latter do not form at all a thick 
layer, and they are partly interspersed among the meshes of 
a peculiar form of connective tissue which extends beyond 
them, and forms the outermost wall of the intestine. This 
connective-tissue layer is also found beneath the epithelium ; 
in parts it consists of a meshwork of fine fibres with nuclei 
present, chiefly at the nodal points ; in other parts the mesh- 
work becomes very wide, and the tissue presents the appearance 
of a fenestrated membrane; the fenestre are, relatively speak- 
ing, small, and the tissue lying between them is somewhat 
gelatinous in appearance, with fine fibrils passing through it 
(fig. 8); nuclei are present, which are frequently attached 
closely to the fenestre, bulging out into these latter as depicted 
in fig. 3. 

At the extreme end of the body these layers are not so con- 
spicuous, owing to the crowding together of the last two or 
three of the intersegmental septa, and the continuity of these 
with the intestine. 

The coelomic space, on the ventral side of the body at any 
rate, is almost filled with the nephridia, which form two 
principal masses, one on each side of the nerve-cord. In a 
single section several funnels can be seen connected with the 
nephridia, and their ducts can be observed to perforate the 
body-walls, and to open on to the exterior by many pores. 


A NEW GENUS OF OLIGOCH ATA. 481 


In such sections the outer connective-tissue coat of the 
intestine may be observed to include numerous tubules cut 
across in various directions (fig. 5,2), indicating therefore a 
somewhat tortuous course ; these tubules appear, in sections 
taken between two successive septa, to have no relation what- 
ever with the nephridial tufts that have been already men- 
tioned as occupying the celom. And yet they are clearly 
nephridial in their nature. 

Fig. 5 represents a slightly magnified section through a 
portion of the intestine, showing the general appearance of 
these tubes. 

Fig. 4 is a portion of the same more highly magnified, and 
drawn with the help of the camera lucida. 

From this drawing it may be seen that the structure of the 
tubes is precisely that of the nephridia, although they are for 
the most part considerably wider. Their walls are granular, 
with large nuclei interspersed here and there, showing the 
lumen of the tube to be intra-cellular. 

In the section figured a smaller tubule is seen to project into 
the lumen of the larger tubule, and another small tubule seen 
in transverse section lies entirely within the larger tubule. 

This telescoping of one tube within another at once recalls 
the peculiar structure of the leech’s nephridium, made known 
by the investigations of Bourne (15) and others. The dif- 
ference is that in the leech the inner tube is closely invested 
by the outer, while in the earthworm there is a wide space 
between the two. 

These nephridial tubes, which appear to be so curiously cut 
off from the general nephridial system, are in reality not so 
cut off. A series of sections shows that they become con- 
tinuous with the general nephridial network at the septa; at 
any rate their branches pass along the septa, and can be traced 
into the nephridial tufts : these branches are of the same calibre 
as those of the general nephridial system. 

Traced in the other direction, these nephridial tubes 
may be followed through the lining epithelium of the 
gut, into the lumen, which they open, 


482 FRANK E. BEDDARD. 


The wide tubes with an intra-cellular lumen become gradually 
crowded with nuclei,! though the boundaries between the 
individual cells are not to be discerned in my preparations; the 
lumen, however, is here clearly intercellular, the individual 
cells being apparently more or less cubicalin form. The nuclei 
are quite similar to those of the portion of the nephridium 
with an intra-cellular lumen. The tube then bends up towards 
the epithelium of the gut, and its lumen becomes much con- 
tracted, owing to the great increase in the size of the cells 
which form its walls. In this region of the nephridium the 
cells are quite indistinguishable from those which form the 
lining membrane of the rectum. 

It seems, therefore, probable that the diverticula of the 
intestine were developed as diverticula, and that the nephridia 
afterwards acquired a connection with them, just as the external 
portion of nephridia opening on to the surface of the body is 
developed from a separate epiblastic involution.* 

In the absence of embryological data, I cannot do more 
than regard as highly probable the suggestion made above con- 
cerning the morphological nature of the rectum. The rectum 
is much more likely to be proctodeum than hypoblastic in 
origin. If this is not the case, then the facts recorded in this 
paper have an obvious bearing upon Lang’s views (17) of the 
hypoblastic origin of the nephridia. I should, however, prefer 
for the present to consider the terminal section of the gut, 
into which the nephridia open, to be epiblastic in origin. 

So far as I am aware, there has been no description of 
nephridia connected with the rectum in any other Cheetopod. 

The nearest group in which anything of the kind occurs is 
the Gephyrea; in Bonellia, and other forms belonging to the 


1 Such a fact as this appears to me to show that the morphological distine- 
tion which some have attempted to draw between nephridia with intra-cellular 
lumen and intercellular lumen, e.g. between those of Oligocheta and 
Polycheeta, cannot be maintained. 

2 Bergh (16) has, however, denied that this is the case in Criodrilus; 
according to him the nephridium is entirely mesoblastic, and bores its way to 
the exterior. 


A NEW GENUS OF OLIGOCHATA. 483 


Gephyrea Cheetifera, we have the branched “ respiratory trees,” 
whose structure appears to show that they are nephridial in 
nature; even in Sipunculus, two rudimentary diverticula 
attached to the intestine close to the anus may be homologous 
structures. It is true that some observers, such as Greef and 
Spengel, have denied that the anal glands of the Gephyrea are to 
be compared to nephridia; but the view of the former was based, 
partly at least, upon a misunderstanding of these organs. 

There is no regularity that I could detect in the position of 
the apertures of these tubes ; sometimes two could be observed 
quite close together, at other times an interval separated two 
adjacent apertures. One point of importance with regard to 
the position of the apertures is their limitation to three seg- 
ments; whether the most anterior of these three segments 
marks the commencement of the proctodeum or not I am 
unable to say. Behind the last of the three segments in which 
they occur the limitations of the individual segments were 
obscured ; they begin, therefore, in the last properly developed 
segment. Itis well known that earthworms increase in length 
by the formation of new segments at the posterior end of the 
body ; it is possible that, as the body increases in length, more 
of these proctodzeal nephridia are developed.! In fig. 6 is 
illustrated the greater part of a single nephridial tube; but it 
by no means always happens that a tube has so long a course 
without branching. Very often I have noticed such a tube to 
divide into two shortly before its opening. 

It has been already mentioned that the nephridia which 
open into the proctodeum communicate with the general 
nephridial system only at the septa. As may be seen in suit- 
able sections, the surfaces of the septa are covered by innu- 
merable tubes, which anastomose in every direction to form 
most complicated networks. I have never yet been able to 
show by a satisfactory preparation the actual form of the net- 
work in Pericheta or Acanthodrilus, or in any other 
form where a network must exist. Spencer (18) has figured 

1 Tn another specimen I found these nephridia opening into the gut, in the 
last seven segments at least. 

VOL, XXXI, PART IV.—NEW SER, KK 


484, FRANK E. BEDDARD. 


the network of Megascolides, but in quite a diagrammatic 
way. I think, however, that the figure which I now give (fig. 
7) of the network in Acanthodrilus is sufficient to con- 
vince anyone of the reality of its existence. Usually the net- 
work exhibited the characters shown in that figure; that is to 
say, the individuality of the several tubes was quite distinct in 
spite of their anastomoses in every direction and at such fre- 
quent intervals. Very often, however, the network exhibited 
the appearance shown in fig. 2; here it will be observed that 
the tubes are in very close approximation, so much so that the 
mass formed by their fusion presents the appearance 
of an irregular system of lacune enclosed within a 
definite wall. This character, it is perhaps worth remark- 
ing, belongs to the nephridia of several invertebrate groups. 

In describing the nephridia of Acanthodrilus I stated 
that the tufts of successive segments were isolated from one 
another ; this is, however, as shown in fig. 9, not always the 
case. In that figure, which represents a longitudinal section, it 
will be seen that the nephridia which pass up the septa do not 
always open at once into the lumen of the gut, but apparently 
become connected with nephridial tubes derived from other 
segments, and course along the walls of the gut; a communi- 
cation is thus established between the nephridial 
networks of a number of segments. In fig. 8 is illus- 
trated a portion of one of the septa which is invaded by the 
nephridial tubes on their way to the intestinal walls: it will 
be seen that in such places hardly any traces are left of the 
muscles of the septum; the septum appears to be entirely 
built up of a mass of frequently anastomosing nephridial 
tubes. 

I have not yet examined a large series of earthworms with 
diffuse nephridia, with a view of finding out whether anal 
nephridia are present in other species. I believe, however, 
that they are not present in A. antarcticus—a near relative 
of A. multiporus. 

It is of importance to find undoubted nephridia in a com- 
paratively low type of Cheetopod opening into the anal section 


A NEW GENUS OF OLIGOCHATA. 485 


of the intestine. Such a discovery strengthens the current 
opinion that the respiratory trees of Bonellia are really of 
nephridial nature. 

A remarkable fact about these nephridia in Acantho- 
drilus is the swollen terminal portion, which is lined by cells, 
and which has the characters (fig. 10) of a diverticulum of the 
gut. Ihave distinguished between the terminal section, which 
opens into the intestine and is lined with an epithelium, identical, 
so far as I can see, with the epithelium of the intestine, and the 
next section, the cells of which are rather different. If the 
nephridial tubes connecting this with the general nephridial 
system were to disappear, we should have the gut furnished 
with a series of tubular outgrowths ending bluntly. This is 
exactly what we meet with in the Tracheata, and even in some 
Crustacea—the Malpighian tubes; these structures have been 
compared with nephridia, particularly with the anal nephridia 
of the Gephyrea. It appears that the Malpighian tubes of 
Arthropods are formed by outgrowths of the proctodeum,} 
though in the Amphipods Spencer is inclined to regard them 
as appendages of the mid-gut. It is at least possible that we 
may trace the Malpighian tubes of Arthropods to these gut 
diverticula. Seeing how widely spread the Malpighian tubes 
are among Arthropods, it is not unreasonable to seek for their 
homologues in lower groups; and the Chetopods are the 
nearest group to which we can trace the Arthropods. It is 
true that Peripatus, which appears to stand somewhere near 
the base of the Tracheate series, has no Malpighian tubes; 
this seems to imply a great break between the anal nephridia 
of worms and the Malpighian tubes. It must be remembered, 
however, that Peripatus is furnished with paired nephridia, 
which structures are wanting in the Tracheata. Hence, there 
might possibly be no need for the extra nephridia opening into 
the extremity of the gut; the ordinary nephridia being absent 


1 Recently Mr. Wheeler (19) has shown that in Doryphora decemlineata 
the Malpighian tubes, when first formed, open on to the exterior; they are 
subsequently drawn in along with the proctodwal invagination of the ecto- 
derm. 


486 FRANK E, BEDDARD. 


in the Tracheata, they have been functionally replaced by the 
somewhat metamorphosed equivalent of the anal nephridia. 

It must be admitted, however, that more facts are required 
before the above remarks can be considered as anything more 
than a suggestion. 


List oF LITERATURE REFERRED TO. 


(1) Bepparp, F. E,—“ On the Anatomy of a Species of Diacheta,” ‘ Quart. 
Journ. Micr. Sci.,’ vol. xxxi, p. 159. 

(2) Bepparp, F, E.— On the Structure of a New Genus of Lumbricide 
(Thamnodrilus Gulielmi),” ‘ Proc. Zool. Soc.,’ 1887, p. 154. 

(8) Bepparp, F, E.— On a Living Species of Pericheeta, with Remarks on 
other Species of the Genus,” ‘ Proc. Zool. Soc.,’ 1890, p. 52. 

(4) Bepparp, F. E.—On the Specific Characters and Structure of certain 
New Zealand Earthworms,” ‘ Proc. Zool. Soc.,’ 1885, p. 810. 

(5) Bepparp, F, E,—“ Contributions to the Anatomy of Earthworms, with 
Descriptions of some New Species,” ‘ Quart. Journ. Micr. Sci.,’ 
vol. xxx, p, 421, 

(6) Bepparp, F, E.—‘ On the Structure of Three New Species of Earth- 
worms, &c.,” loc. cit., vol. xxix, p. 101. 


(7) Bepparp, F, E.—‘On the Occurrence of Numerous Nephridia,” &c., 
loc. cit., vol. xxviii, p. 397. 


(8) Rosa, D.—‘ I Lumbricidi del Piemonte.’ 
(9) Rosa, D.— Nuova Classificazione dei Terricoli,” ‘ Boll. Mus. Comp, 
Zool. Torino,’ vol. iii, No. 41. 
(10) Rosa, D.—“I Lombrichi raccolti nell’ Isola Nias,” &., ‘Ann. Mus. 
civ. Genova,’ vol. vii, 1889. 
(11) Bennam, W. B.—“Studies on Harthworms,” No. I, ‘ Quart. Journ, 
Micr. Sci.,’ vol. xxvi, p. 213. 
(12) Horst, R.— Midden Sumatra Expeditie,” ‘ Natuul. Hist. xii Afdeeling, 
Vermes.’ 
(18) Perrier, E.—‘ Mémoires pour servir 4 l’histoire des Lombriciens 
terrestres,’ ‘Nouv. Arch. Mus,,’ t. viii, p. 142. 
(14) Vartuant, L.—‘ Histoire naturelle des Annelés marins et d’eau douce, 
t. iii, Paris, 1889. 
(15) Bournz, A. G.— Contributions to the Anatomy of the Hirudinea,” 
‘Quart. Journ, Micr. Sci,,’ vol, xxiv, p. 419. 


A NEW GENUS OF OLIGOCHATA. 487 


(16) Bercu, R. S.—‘‘ Zur Bildungsgeschichte der Exkretionsorgane bei 
Criodrilus,” ‘ Arbeiten Zool.-Zoot. Inst. Wirzburg,’ Bd. viii, p. 228. 


(17) Lane, 8.—“ Die Polycladen,” in ‘ Naples Monographs.’ 


(18) Spencer, W. B.—“ The Anatomy of Megascolides australis,” 
‘Trans. Roy. Soc. Victoria,’ vol. i, pt. i. 


(19) Wuerter, W. W.—“ The Embryology of Blatta germanica and 
Doryphora decemlineata,” ‘J. Morph.,’ vol. iii, p. 291. 


DESCRIPTION OF PLATES XXXIII anp XXXIIIa, 


Illustrating Mr. Frank E. Beddard’s paper “ On the Structure 
of a New Genus of Oligocheta (Deodrilus), and on the 
Presence of Anal Nephridia in Acanthodrilus.” 


Figs. 1—11.—Acanthodrilus multiporus and nephridia. 

Fig. 1. Dorsal half of the posterior twenty or thirty segments of Acan- 
thodrilus multiporus, enlarged to twice the natural size. 

Fig. 2. Part of a network of nephridia. The fusion of the tubules is so 
complete as to produce the impression of a system of lacunar spaces. 
Fig. 3. Connective tissue from outer coat of rectum. x. Nuclei. ». 

Spaces. 

Fig. 4. A portion of connective-tissue sheath of rectum, with contained 
nephridial tubes. m. Muscular fibres. v. Spaces. 47. Blood-vessels. 
z. Small tubule apparently contained in a larger one. 

Fig. 5. Part of a transverse section through rectum. ep. Lining epi- 
thelium. . Nephridial tubules in connective-tissue sheath. 

Fig. 6. Part of Fig. 5, more highly magnified to illustrate structure of 
different regions of anal nephridia, V., V'z. 

Fig. 7. Network of tubules lying in and upon septum. a. Muscular 
fibre. 2. Walls of nephridium. 

Fig. 8. Mass of tubules passing up septa, the relations of which are 
shown in Fig.9. 47. Blood-vessels. a, b,c, d. Nephridial tubes. 

Fig. 9. Longitudinal section through a portion of rectal wall. 0. Orifice 
of anal nephridium. sp. Intersegmental septum. 

Fig. 10. Section through point quite close to opening of nephridium into 
rectum. WV. Nephridium. #. Epithelium of rectum continuous with 
nephridium. 


488 FRANK E. BEDDARD. 


Fig. 11. Diagrammatic transverse section of a portion of posterior ex- 
tremity of body, to show relations of anal nephridia. 2. Their rectal 
orifices. o. External orifices. Funnels. 


Fies. 12—14.—Deodrilus. 

Fig. 12. Ventral view of anterior segments. VI. Sixth segment, the first 
setigerous segment. jp. Genital papilla between Segments 11/12. 
cli. Clitellum. ¢ sperm-duct orifice on 18th segment. 

Fig. 13. Dissection of a portion of the body, to show the position and 
appearance of the calciferous glands, ca. d. Dorsal vessel. @. 
(sophagus. 

Fig. 14. Diagrammatic longitudinal section, to show position of various 
parts of the alimentary tract. yg. Gizzard. ca. Calciferous glands. 
The roman numerals refer to the segments. 


PLATE XXXIIIa. 
Deodrilus. 
Figs. 15, 16. Penial seta, with chitinous sheath. 


Fig. 17. One of the pairs of hearts (4.), with their connection with the 
dorsal (d.) and ventral (v.) vessel. qs, @sophagus. 

Fig. 18. Sete of body. 

Fig. 19. Two segments dissected to show spermatheca (Sp.), Sp.2). 2. 
Nerve-cord. v. Ventral blood-vessel. J/. Lateral blood-vessel. 0. 
Orifice of spermatheca. m. Intersegmental septum. 


EXORETORY TUBULES IN AMPHIOXUS LANCEOLATUS. 489 


Excretory Tubules in Amphioxus 
lanceolatus. 


By 


F. Ernest Weiss, B.Sc., F.L.S., 
University College, London. 


With Plates XXXIV and XXXV. 


In the spring of last year, through the kind permission of the 
British Association for the Advancement of Science, I had the 
' privilege of occupying the table which the Association supports 
at the Zoological Station in Naples. By the admirable arrange- 
ments made at this station I was able to have a constant and 
unlimited supply of living specimens of Amphioxus lanceo- 
latus, and thought this a good opportunity to undertake 
some experiments with a view to ascertaining whether the 
curious patches of modified epithelial cells on the ventral wall 
of the atrium of Amphioxus had any excretory function, as 
Johannes Miiller (1) had held probable; and whether also 
the atrio-ceelomic funnels first described by Professor Ray 
Lankester (2) in 1874; and again more recently (8), had any 
such function. 

Feeding experiments were made with Indian ink, carmine, 
and Bismarck brown, the carmine alone leading to good results, 
as the Indian ink is not able to be distinguished in the pig- 
mented cells of the atrial cavity ; and the Bismarck brown, 
though colouring the excreting cells very readily and deeply, 
penetrated also into many other cells. 

Carmine is only very slightly soluble in sea water, but when 
well ground up in a mortar it remains suspended in granules 
sufficiently small to be taken up by the intestinal cells of the 


490 F. ERNEST WEISS. 


Amphioxus. I found it most expedient to leave the Am- 
phioxi in water thickly clouded with carmine, through which 
a constant current of air was passed. As a further precaution 
I changed the whole of the water every day, adding fresh car- 
mine as before. In this way I was able to keep the Amphioxi 
for weeks together, and they seemed to remain as healthy as 
those kept in running water. The current of air which was 
passed through the water had the secondary but useful action 
of preventing the carmine granules from settling down, and 
this ensured a constant inhalation by the Amphioxi of carmine- 
laden water. Such a current seems to haye no irritating 
action on these animals. I have been able by the aid of a 
syringe to pass sea water very highly charged with carmine 
through individuals, which were at the time in quite clear 
water, without their seeming to notice it, certainly without 
their closing their mouth for an instant, which they do imme- 
diately anything obnoxious to them is brought into their 
neighbourhood. 

After a day or two the Amphioxi will have taken up a con- 
siderable amount of carmine in its very finest granules into 
the cells of the intestine, and their feeces are made up almost 
entirely of the coarser granules which could not be incepted 
by the cells. 

From the intestinal epithelium the carmine is passed into 
the intestinal blood-vessels, which seem charged with corpus- 
cles (lymph-cells ?). 

Amphioxi which are kept longer still in carmine take up a 
considerable amount of it into their vascular system, so that I 
was enabled to follow out some of the blood-vessels, which are 
otherwise very difficult to make out. 

The intestinal vessels join anteriorly to form a single vessel 
(Pfortader of Miiller [1]; Darmvene of Schneider [4] ), which, 
as Professor Lankester rightly states, is continued into the 
endostylar or cardiac subpharyngeal trunk, though Miller 
asserted that it ended at the commencement of the cecum. 

In each gill bar, whether primary or secondary (tongue bar), 
there are two blood-vessels, one running on the inner side of 


EXORETORY TUBULES IN AMPHIOXUS LANCEOLATUS. 491 


the bar, and communicating, as is seen from fig. 4, with the 
dorsal aorta. This vessel then separates either entirely into 
two parallel trunks, or is only constricted medianly, so that 
we find one vessel running below the median inner epithelial 
band, and the other one on the inner face of the chitinoid rod, 
as figured by Professor Lankester. In the case of the secon- 
dary or tongue bars carmine is found also within the hollow 
of the chitinoid rod, so that I think Schneider’s conception of 
this space as a blood-vessel connected with the one running 
along the inner surface of the rod is probably correct. A 
very small vessel seems to run on the outer edge of the 
chitinoid rod. It would seem to be connected with the blood- 
vessel running beneath the excreting tubule in fig. 1, and 
being like those in the suspensory folds connected with the 
excretory function of the atrial epithelium. The two dorsal 
aorte are either connected at intervals by very fine vessels, or 
else each gives off branches which run beneath the epithelium 
of the dorsal groove of the pharynx. 

Besides these branches of the aorte I was able to confirm 
Schneider’s statement that branches are given off to the 
muscles of the body, passing up by the side of the notochord, 
and other branches to the inner face of the body-wall. These 
latter branches run beneath the cceelomic epithelium, and more 
ventrally beneath the atrial epithelium; they connect the 
dorsal aorta with a longitudinal vessel described by Miller as 
running on the inner body-wall above the gonads. 

Indeed, the whole inner surface of the atrium seems very 
well supplied with blood-vessels running beneath the atrial 
epithelium, and in the case of these modified epithelial cells, 
described by Miller as possibly renal organs, diverticula are 
formed by the blood-vessel between the cells (fig. 6). 

Along the inner lamella of cutis, too, of the ventral atrial 
wall I found a considerable amount of carmine in what I con- 
clude must be a blood-vessel or a vascular space. I have not 
here given a general account of the vascular system of Amphi- 
oxus, but have confirmed statements by various observers, 
statements which Professor Lankester in his last memoir has 


492 F, ERNEST WEISS. 


cited as wanting confirmation. I feel sure, however, that 
much that is new might be made out by careful investigation 
of Amphioxi fed with carmine as above described. 

The coelomic cavities were singularly free of carmine, so 
that it would seem as if the vascular system were more dis- 
tinctly separated from the ccelomic system than has hitherto 
been supposed. 

In the metapleural lymph-spaces, however, I constantly 
came across cells containing carmine, though not to any very 
considerable extent. 

My object, however, being to trace the carmine to the 
excretory organs, I found it best to examine specimens in 
which the carmine was already disappearing from the vascular 
system. After a week or a fortnight Amphioxi kept in the 
-carmine-containing water would have assumed a quite definite 
pink coloration, and I then transferred them into a tank with 
running water, where they gradually became paler. Amphioxi 
thus treated gave the best results for the purposes I had in 
view. 

As already mentioned, the patches of modified epithelial 
cells on the ventral wall of the atrium have a very definite 
blood-supply, of the nature of a subepithelial blood-space with 
short blood-vessels running up between the cells (fig. 6), and 
when the vessels were well coloured I found also carmine 
granules in these cells. In many cases they were not readily 
distinguished on account of the deep-coloured granules con- 
tained in the cells, and only in specimens which were very 
slightly pigmented could these carmine granules be unmis- 
takably seen. This pigmentation of the atrial epithelium pre- 
vented my ascertaining whether excretion takes place to any 
great extent over its entire surface, or whether it is confined 
to the specialised portion described by Miller. It was this 
same circumstance which prevented me from obtaining any 
positive results from examination of the atrio-ccelomic funnels 
described by Professor Lankester. They are applied, as shown 
in Professor Lankester’s drawings, to the wall of the dorso- 
pharyngeal ccelom, along which wall, as I have stated before, 


EXORETORY TUBULES IN AMPHIOXUS LANCEOLATUS. 493 


we find well-marked blood-vessels. This fact lends further 
support to the very justifiable view that they may have an 
excretory function. 

A further set of modified atrial epithelial cells are those 
which cover the outer portion of the gill bars. These are 
columnar in shape, with a large nucleus near the base, and 
generally a considerable amount of granules at the opposite 
end. Definite pigment granules are found only in one or two 
cells where the epithelium of the alimentary tract borders on 
the atrial epithelium. 

In the specimens fed with carmine the atrial epithelial cells 
of the secondary or tongue bars all showed carmine granules. 
In the primary bars I was unable to distinguish any carmine. 
This fact is, I think, to be explained by the circumstance that 
in the secondary bars the blood-vessel, which I described as 
running along the outside of the chitinoid rod, hes imme- 
diately beneath the atrial epithelium, while in the primary 
bars of course it lies beneath the ccelomic epithelium, and 
only very fine ramifications, if any, would pass round to the 
atrial epithelium. 

On staining sections, and also in feeding living individuals 
with Bismarck brown, the same fact will be observed, the atrial 
epithelium covering the secondary bars becoming much more 
intensely stained than the similar epithelium of the primary 
bars. I made use of this colouring matter on the recom- 
mendation of Dr. Paul Mayer, of the Zoological Station at 
Naples, who invariably used it to colour excreting cells, chiefly 
in Crustacea. 

Similar to the atrial epithelium of these gill bars behaves 
the epithelium of the dorsalward extension of the atrium 
lining the so-called suspensory or pharyngo-pleural folds. 
The cells of this portion of the epithelium are even larger than 
those of the gill bars in many parts, and stain deeply in their 
more granular portion with Bismarck brown. 

That so large an amount of the atrial epithelium should be 
excretory does not seem strange or improbable to me, as Dr. 
Eisig, in experimenting with Capitellide, found these worms 


494, F. ERNEST WEISS. 


to excrete over the greater portion of their epithelial struc- 
tures. 

But the excretion of this atrial surface is not comparable in 
quantity, at least in the excretion of carmine, with the amount 
excreted by some small tubules which have hitherto remained 
unnoticed, but become very evident in the specimens fed with 
carmine. The course of one of these tubules is figured in the 
drawings of the successive sections (figs. 1 to 5). They lie on 
the outer side of the topmost of the pharyngo-pleural folds 
which connects the uppermost primary bar with the lateral 
ridge, from which all the gill bars start. Along this ridge we 
find a continuous chitinoid rod, which is connected in turn 
with each rod of the series of bars. These tubules occur thus 
serially, the last one being in connection with the last primary 
bar, and lying therefore much more ventrally than the fore- 
most ones. They seem to project into the celomic cavity, but 
at the same time are covered with the thin celomic epithelium 
which lines that cavity. 

In fig. 1 the tubule is cut longitudinally, and on its inner 
side will be seen a blood-vessel which is a branch from the 
dorsal aorta, and passes over the longitudinal chitinoid rod of 
the lateral ridge to join the aortic trunk. This branch of the 
aorta is seen passing upwards in figs. 3,4, and 7. It is from 
this branch, too, that the blood-supply to the pharyngo-pleural 
folds and the vessel along the outer edge of the chitinoid rod 
of the gill bar is derived. 

From the subsequent sections it will be seen that the tubule 
runs upwards and backwards, and then bending downwards 
opens (figs. 4 and 8) into the upward extension of the atrium 
at the highest point, which the latter reaches between the 
pharyngo-pleural folds. The opening is always placed at the 
origin of a secondary gill bar, and is just the place at which 
an excreting organ might profitably open to ensure its excreta 
being carried away. The chief difficulty with regard to the 
making out of these points was the obtaining of a suitable 
stain which would not hide the carmine granules. Carmine 
and hematoxylin proved quite unsuitable, and after trying 


EXORETORY TUBULES IN AMPHIOXUS LANCEOLATUS. 495 


most varied aniline colours, I finally adopted a colouring 
agent recommended me by Dr. Hugo Hisig, of the Naples 
Zoological Station, which answered the purpose better than 
any other stain. This was a solution of picric acid in turpen- 
tine, which has the advantage that after mounting the sections 
they need not be passed through weak alcoholic solutions. 
Though this stain, however, shows up the carmine well, it 
does not stain the nucleus or cell-wall differentially, so that in 
making the drawings I had to fill up most of the details from 
specimens stained with different colours, such as Bismarck 
brown, carmine, or hematoxylin. In sections coloured in 
this way I was able to make out these excreting tubules, but 
in the case of Amphioxus a very large number of individuals 
must be used to determine any point definitely, owing to 
the different appearance almost every individual presents, due 
to differences of contraction of the gill bars, distention by the 
genital products, and shrinkage during the killing, hardening, 
and embedding. 

I have thus not been able to settle clearly whether these 
tubules just described have any internal opening to the cceelom, 
a point which is of very great interest. Still, to be true 
nephridial tubules they need not retain permanently a com- 
munication with the colom, and they may subsequently be 
found to possess such continuity at some stage in their develop- 
ment. 

They may, on the other hand, be simply upward extensions 
of the atrium, though to this view I should not be very ready to 
give my support, owing to their bend downwards and their rela- 
tion to the blood-vessel, which in all parts of its course seems 
to run outside the atrial and inside the ccelomic epithelium. 

The cells, too, are devoid of one marked characteristic of the 
atrial epithelium, namely, the pigmentation, though this might 
be due to specialisation. 

If these tubules are coelomic in origin they would come very 
near true nephridial structures. Indeed, they would be very 
typical segmental organs devoid of any connecting duct, and 
their only difference from the most primitive form of nephridia 


496 F. ERNEST WEISS. 


would be the fact that they had become closed off from the 
remaining portion of the celom. Yet even this I do not con- 
sider quite settled, as the tubules in many instances seem very 
imperfectly closed off at their lower end, but I could not satisfy 
myself that this was not due to faulty preparation. Still there 
seemed to be no definite nephridial funnel, so that I must 
await further proof to consider them as closed off from the 
celom. 

They are in this particular similar to the atrio-ccelomic 
funnels described by Professor Lankester, but that they are 
homologous structures with these belonging to a series of 
primitively equal dorsalward extensions of the atrium I am 
not inclined to believe, as the excreting tubules I have just 
described occur regularly in connection with the secondary 
gill bars of the region in which the atrio-ccelomic funnels have 
their opening. If the excreting tubules are simply extensions 
of the atrium, they must be looked upon as analogous to and 
not homologous with the atrio-ccelomic funnels. 

Before concluding I should like to express my thanks for 
help and suggestions I received during my work to my teacher, 
Professor E. Ray Lankester, and to Dr. Paul Mayer and Dr. 
Hugo Eisig, of the Zoological Station at Naples. 

April, 1890. 


MEMOIRS REFERRED TO. 


1. Jonannes Mijttnr.—“ Ueber den Bau und die Lebenserscheinungen des 
Branchiostoma lubricum, Costa, Amphioxus lanceolatus, 
Yarrell,’’ ‘Abhandlungen der Konigl. Akad. der Wissenschaften,’ 
Berlin, 1841. 

2. KE. Ray Lanxester.—“‘Some New Points in the Anatomy of Am- 
phioxus lanceolatus,” ‘ Quart. Journ. Micr. Sci.,’ vol. xv, 1875. 

3. EB. Ray Lanxester.—* Contributions to the Knowledge of Amphioxus 
lanceolatus, Yarrell,” ‘ Quart. Journ. Mier. Sci.,’ vol. xxix, 1889. 

4. Anton Scunuiper.—‘Beitrige zur Anatomie und Entwicklung der 
Wirbelthiere,’ Berlin, 1879. 


EXCRETORY TUBULES IN AMPHIOXUS LANCEOLATUS. 497 


EXPLANATION OF PLATES XXXIV & XXXV, 


Illustrating Mr. F. Ernest Weiss’s paper on the ‘‘ Excretory 
Tubules in Amphioxus lanceolatus.” 


PLATE XXXIV. 


Fie. 1.—Portion of a section through the pharyngeal region of Amphioxus, 
showing the right dorso-pharyngeal ccelom and an excretory tubule cut 
longitudinally. A blood-vessel coloured by the carmine runs on the inside of 
the tubule. 

Fie. 2.—The next section, showing the beginning of the upward bending 
of the tubule. In these and the following sections some carmine is seen in 
the dorsal aorta. 

Fic. 3.—This section shows the origin of the blood-vessel lying beneath 
the subexcretory tubule in Fig. 1 from the dorsal aorta. The atrial epithe- 
lium of the secondary gill bar also shows excreted carmine, though to a 
slighter extent than the tubule. 


Fig. 4 shows the opening of the excretory tubule into the upward exten- 
sion of the atrium at the commencement of the secondary gill bar. It shows 
also the branching from the dorsal aorta of the blood-vessel of the secondary 
gill bar, and of the blood-vessel to the excretory tubule. 

In Figs. 3 and 4 can be seen the communication between the cclomic 
cavity of the primary gill bar and the dorso-pharyngeal ccelom. 


PLATE XXXV. 


Fic. 5 shows the portion of the excretory tubule behind its opening into 
the atrium. 

Fic. 6 shows the thickened patches of atrial epithelium in the hind region 
of the atrium described by Johannes Miiller. Their cells contain excreted 
carmine granules. Below them is a blood-filled sinus, or possibly a network 
of closed vessels, with processes running up into the thickened portions. 

Fies. 7 and 8 show an excretory tubule on the left side of the pharyngeal 
region, and the opening of the tubule at the origin of the secondary gill bar. 


Pi i . ai LY fl Aa 


Peart ie eld LE FUP BET OWL OR giG Bee! lia Lary, ull 
. . 


b | i 4 t aA a 
arch J » ¥ . } t « om yy 
* ‘ 
. i ° 
. j a ii a ih » 
‘ i : 
‘ 
= - a _ be 
§ 
) if 4 a : 4 M3 
' f ’ 
] 4 a - 
ag 
rl hy y if ~ ‘ ee 
t? e i 3 ot G4) td toot 
are 
a ‘¥ - . j 
ba al ad 
A ’ 


; a 
hialiv Ath St i inh iF 64 ‘ 1j jw ue 2a] os(tek 


are i 4% Lb he SMA Se 14 (De aeu roe 


afta yf bila is Gb weil ol lh AR 


eat ili ) F ul Tt wile bi a» he Laeteg ba wels 6 ie 


STUDIES IN MAMMALIAN EMBRYOLOGY. 499 


Studies in Mammalian Embryology. 


II.—_The Development of the Germinal Layers 
of Sorex vulgaris. 


By 


A. A. W. Hubrecht, LL.D., C.M.Z.S., 
Professor of Zoology in the University of Utrecht. 


With Plates XXXVI—XLII. 


INTRODUCTION. 


Or all embryological problems those concerning the gastru- 
lation process of the Amniota and the formation of their meso- 
blast and notochord may certainly be said to be at the present 
moment one of the most burning questions ; not only because 
of the number of important investigations that have of late 
years been directed towards their solution, nor of the high 
authority of the names of many of those who have been en- 
gaged in these investigations, but more especially because of 
the very conflicting results and conclusions to which these 
different researches have conducted their authors. 

If I venture to add new fuel to this fire it is only because 
I am not without hopes that the manner in which the facts 
present themselves in the as yet uninvestigated species of 
primitive Mammalia (Insectivora) which have been the object 
of my own researches may serve to reconcile varied and oppo- 
site interpretations to which the rabbit (Kolliker, van Beneden, 
Rabl, a. o.), the guinea-pig (Carius, Keibel, Strahl), the cat 
(Fleischmann), the opossum (Selenka), the bat (van Bene- 
den), the mole (Lieberkiihn, Heape), and the sheep (Bonnet) 
have led those respective authors. 

VOL. XXXI, PART IV.—NEW SER. LL 


500 A. A. W. HUBRECHT. 


Before entering upon a description of the embryological 
data, such as we notice them in the shrew, a short sum- 
mary of the principal points upon which the above-named 
authors agree, and of those upon which they differ, may serve 
to clear the ground, to refresh the memory, and to facilitate 
the interpretation of the phenomena which are here brought 
to bear upon the points in contest. 

In this brief summary I will aim at the utmost conciseness, 
referring the reader who wishes to make a full study of the 
intricate subject to the original treatises, of which I will give 
a list in an appendix to this article. 

Omnium consensu, the primitive streak and the primi- 
tive groove are looked upon in the light in which Balfour first 
taught us to see them, 1. e. as the region that corresponds to 
the blastopore of Amphioxus, of the Cyclostomata, and of the 
Amphibia. 

The lips of the blastopore are stretched, and of the circular 
opening a longitudinal groove (Primitiv-rinne) is the repre- 
sentative, which may henceforth be designated, with Bonnet 
and Fleischmann, as the gastrula-groove. 

The gastrularidge (Gastrula-leiste, Fleischmann) is a very 
adequate name for the tissue which proliferates inwards from 
the epiblast in this region. ‘This proliferating ridge was 
hitherto designated as the primitive streak. 

From its cell-strata mesoblastic tissue spreads between the 
two primary germinal layers. 

The inevitable consequence of our comparison with lower 
types is that we have to look upon this median streak 
of tissue that has proliferated inwards— although it un- 
deniably originates out of the epiblast—as essentially a 
formation of hypoblastic material (Urentoderm, Kupffer, 
Bonnet, a. o.).! 


1 This allows of considerable simplification of one point, about which much 
unnecessary controversy has been going on, viz. in how far the mesoblast of 
the Amniota is partly of epiblastic, partly of hypoblastic origin. As soon as 
the gastrula ridge is looked upon as representing archaic hypoblast, the 
question as above formulated can advantageously be allowed to drop. 


STUDIES IN MAMMALIAN EMBRYOLOGY. 501 


The chief questions about which the numerous authors 
above cited disagree are— 

1. In what relation stands this archaic hypoblast of the 
gastrula ridge to the remaining hypoblast of the didermic 
blastocyst, which forms a closed layer, and in most mammals 
even a closed sac beneath it ? 

2. Is the production of mesoblast restricted to the tissue of 
the gastrula ridge, or does the hypoblast proper take part in 
such production ? 

3. Does the notochord originate solely out of a median 
forward growth (Kopffortsatz) of the gastrula ridge (in which 
a central canal may arise to asmaller or greater extent), which 
pushes forwards between the primary layers, and which finally 
fuses with the hypoblast (Hinschaltungsstadien, Keibel), after 
which this notochordal tissue again separates from the hypo- 
blast .(Ausschaltungsstadien, id.), which then coalesces below 
it to form the dorsal wall of the gut ? 

4. Or does a flat portion of the lower layer of the didermic 
blastocyst participate in the formation of the anterior region 
of the notochord in addition to this median rod of tissue 
(Kopffortsatz) which has grown forwards out of the gastrula 
ridge ? 

5. Do both these median structures, which we shall henceforth 
designate as the protochordal wedge (median anterior pro- 
longation of the gastrula ridge, often enclosing a protochordal 
canal or rudiments of it) and as the protochordal plate 
(developed in situ as part of the hypoblast), take part in the 
formation of lateral paired plates of mesoblast ? 

6. Does the gastrula ridge and its anterior prolongation 
(the protochordal wedge), togethtr with the longitudinal 
cavity therein contained, represent the whole hypoblast and 
the archenteric cavity of Amphioxus, so that the inner layer 
of the didermic blastocyst is not the homologue of one of 
the germinal layers, but simply a yolk-membrane (para- 
derm, Kupffer; lecitophor, van Beneden), peculiar to the 
Amniota ? 

It will be clear from the enumeration of these points of 


502 A. A. W. HUBRECHT. 


disagreement that a satisfactory solution of the question of 
gastrulation and mesoblastogenesis in the Amniota is stilla 
desideratum. More than one of the points formally exclude 
each other (e. g.6 and 4). Still, it is impossible to arrange 
them in two or more groups into which the different investi- 
gators could be respectively brought together. To prove this 
it may be sufficient to say—restricting ourselves for the present 
to researches on mammalian embryology—that point 2 is now 
answered in the affirmative, as far as its first proposition goes, 
by Kdélliker (18, 19), van Beneden (8), and Fleischmann (8) ; 
in the negative (i.e. affirmative as far as its second proposition 
goes) by Heape (9), Bonnet (5), and myself (14). 

Point 3 is affirmatively answered by Carius (7), Keibel (17), 
and van Beneden. 

Point 4 by Heape and Bonnet, against whom Keibel argues 
at great length. 

Point 5 is very diversely answered, certain authors, as Strahl, 
considering the paired lateral plates of mesoblast alongside of 
the median notochord to be lateral forward outgrowths of the 
posterior mesoblast in the region of the gastrula ridge, whereas 
others (amongst whom Rabl indulges in more general and com- 
parative speculation) consider these plates as lateral wing-like 
growths of the median streak of tissue. 

On point 6 van Beneden has committed himself to very far- 
reaching speculations, which are viewed with favour and 
adopted by Rabl (24), doubted by Bonnet (5, 11) and myself 
(15), combated by Keibel (17). 

This being a very general outline of some of the chief points 
in contest, I do not think it desirable to enter at any greater 
length into historical retrespects, or to go into more detail 
concerning the argumentation of the different authors. Many 
of the papers contain careful comparisons of the results therein 
brought forward with those of the other authors; many are 
ballasted by polemical remarks directed against the authors of 
such researches as would appear to lead to divergent con- 
clusions. 

As my object is in the first place a rearrangement of the 


STUDIES IN MAMMALIAN EMBRYOLOGY. 503 


known and published facts, for which we are indebted to those 
different authors, and professes to be an attempt (based on 
independent observations) towards harmonious interpretation, 
rather than a contribution towards polemics and criticism, I 
am all the more inclined to refrain from further “ weighing 
of evidence.” 

And so now a full description of the chief points which I 
have noticed in the early developmental stages of the shrew 
may here follow. 


Cuap. I.—EHarzty DEVELOPMENTAL STAGES OF 
SorEx VULGARIS. 


1. The Blastula and the Didermic Blastocyst. 


The earliest ovum of which I possess transverse sections is 
in a stage just following upon the phase of the well-known 
rabbit’s ovum, which was described and figured by van 
Beneden, and has since passed into every text-book of embryo- 
logy as a specimen of the earliest mammalian blastocyst. 

There is a very distinct zona pellucida, internally clothed by 
a laver of flattened cells, of which I count six to eight in the 
largest circumference, and at one spot there is an agglomera- 
tion of larger and more bulky cells of somewhat different size. 
The first-named layer is the trophoblast (vide Hubrecht, “The 
Placentation of Erinaceus,” ‘ Quart. Journ. of Micr. Sci.,’ vol. 
xxx, p. 298). The accumulation of cells contains the material 
for the embryonic epiblast and for the hypoblast (figs. 5—7). 
Counting the nuclei in this and the remaining sections, the 
embryo here figured is found to consist of from fifty to sixty cells. 
A cavity which is clearly marked though not yet over-spacious 
extends beneath the polar thickening. It is, properly speak- 
ing, the segmentation cavity; after the development of the 
ccenogenetic hypoblast the greater part of it is enclosed by 
hypoblast-cells, and becomes the cavity of the yolk-sac. The 
trophoblast forms its outer wall, the thick zona is again out- 
side this. The embryos here described were found freely 
suspended in the lumen of the uterus, without any attachment 


504 A. A. W. HUBRECHT. 


to or enclosure by the uterine wall, as was described for the 
hedgehog (16). 

Of the stages following upon the one here noticed numerous 
specimens are at my disposal, and as long as the blastocyst 
remains in the monodermic, and even in the didermic stage, it 
is either quite free in the lumen, or it shows the very first 
signs of bulging out, and adhering with its outer layer against 
the uterine epithelium. However, it has first to undergo a 
considerable increase in size, and during this growth there is 
at first an evident stretching of the existing cells, much more 
than an active subdivision and increase in number. Simulta- 
neously with the stretching of the widening blastocyst the 
zona pellucida diminishes in thickness (figs. 6—11, 22—24). 
In the last phases of the didermic stage, just before the first 
traces of a mesoblast and of a gastrula ridge appear, the zona 
has reached its limit of tenuity (fig. 26). After that it dis- 
appears, and the definite attachment of the trophoblast to the 
uterine tissue is brought about. 

In the preserved specimens which I examined, this increase 
in bulk of the blastocyst, coupled with the decrease in thick- 
ness of the zona, causes the blastocyst to lose its spherical 
shape, and to fall into folds of very varied form and dimension 
(figs. 8—12, 22—27). This folding—which is no doubt an 
artefact, as we may suppose the blastocysts to be spherical 
and elastic in life—is only counteracted when the trophoblast 
begins to adhere against the uterine walls (fig. 12). The 
blastocyst is then seen to reassume, in sections of uteri that 
were preserved in toto, its spherical aspect. For our appre- 
ciation of the developmental processes that go on in the 
trophoblast and in the inner cell-mass—which might also be 
designated as the embryonic knob—these folds are but of very 
little consequence, and do not interfere with an accurate 
interpretation of the phenomena. 

To prove that indeed the increase in size of the blastocyst 
to its double diameter (i. e. an increase in bulk of eight times 
the original cubic span, and a surface increase of one to four) 
is more a phenomenon of stretching than of cell-division, we 


STUDIES IN MAMMALIAN EMBRYOLOGY. 505 


have only to count the number of cells that now compose the 
blastocyst, which we find at about fifty-six for the trophoblast, 
twenty-six for the embryonic knob; whereas these numbers 
were in the earliest phase thirty-six and nineteen for one and 
thirty and seventeen for another embryo. 

In the sections of other embryos of about the same size the 
embryonic knob has the appearance of being more bulky (figs. 
10 and 11, Pl. XXXVI). In the one case it, however, contains 
twenty-two, in the other twenty-three cells; the apparent dis- 
crepancy is thus evidently occasioned by obliquity of the 
section plane. Another somewhat older embryo, which 
immediately precedes the stage in which the hypoblast is 
going to separate from the inner cell-mass, shows a further 
increase in the trophoblast, with a stationary number of inner 
embryonic cells. The stage alluded to (fig. 9) contains twenty- 
‘two cells in its embryonic knob; the trophoblast is formed by 
eighty cells, and as yet no coating of hypoblast-cells is detected 
in any part of this blastocyst. 

In the following stage the hypoblast-cells are seen to spread 
through the blastocyst, and at the same time the embryonic 
knob is more flattened, and projects somewhat less into the 
central cavity (figs. 22—24). There can be no doubt that 
these large hypoblast-cells which gradually form a con- 
tinuous layer clothing the trophoblast, and which then 
constitute with it the didermic blastocyst, take their origin 
from the embryonic knob. The cells of the latter are 
less flattened and more bulky, the nuclei larger than those 
of the trophoblast ; and by the latter peculiarity it is easy, 
even before the hypoblast-cells form a continuous layer, to dis- 
tinguish them from the trophoblast-cells against which they are 
being applied (fig. 23). 

With regard to the embryonic knob, two questions are not 
without a certain importance: (1) Is there an indication that 
the trophoblastic cell layer stretches above the embryonic 
knob, or does it merge into this all along the border of the 
latter ? 

The significance of this question will be understood when 


506 A. A. W. HUBRECHT. 


we bear in mind that in the rabbit (Rauber, a. 0.), in the mole 
(Heape), in the hedgehog (Hubrecht), perhaps also in the bat 
and in other mammals, either isolated cells (Deckzellen) ora 
continuous layer of cells are present outside of the layer that 
is going to become the embryonic epiblast. These outer 
layer cells in some cases disappear or fuse with the embryonic 
epiblast (rabbit) ; in others they are separated more fully from 
it in later developmental stages (hedgehog). 

In the shrew similar cells have been found by me; they do 
not, however, form a continuous nor a substantial layer, as in 
the mole and hedgehog. They are detected in the didermic stage, 
and figured on Pl. XX XVII, fig. 26, #7’. Their presence here 
can hardly leave a doubt that alsoin the earlier stages the tro- 
phoblast-cells with their smaller nuclei stretch above the em- 
bryonic knob. If we keep in mind that the nuclei of the 
trophoblast are in the beginning very wide apart, we cannot 
wonder that in sections through the earliest embryonic knob 
the outer trophoblastic covering cannot always be indubitably 
traced. 

Certain of the sections figured give, however, indications 
that also in this earlier stage the embryonic knob may be said 
to be situated (figs. 6, 7, and 10) within the trophoblastic 
layer, and still to be continuous with it at the border. The 
latter connection distinctly prevails later on (figs. 28—381, 
&c.), when the sparse “ Deckzellen” have wholly disappeared. 

(2) Is the difference in size of certain cells composing the 
embryonic knob which is noticed in one of the specimens of 
the earliest phase (although I have not as yet specially men- 
tioned it) a distinction by which at that early age the mother- 
cells of the later epiblast-cells and of the later hypoblast-cells 
are distinguished ? Or isit an individual and fortuitous occur- 
rence peculiar to one of those early specimens? (fig. 5). 

I cannot definitely answer the question, but will merely 
call attention to it, remarking at the same time that in the two 
other specimens of the same age (figs. 6 and 7), as well as in 
those that follow (figs. 8—11), I could not recognise any differ- 
ence in size between the cells composing the embryonic knob ; 


STUDIES IN MAMMALIAN EMBRYOLOGY. 507 


so that, even if the question were to be answered in the 
affirmative, we should have to add that the possibility of distin- 
guishing potential epiblast- from potential hypoblast-cells is 
only limited to that very early stage. 

When once hypoblast-cells have begun to emigrate from the 
embryonic knob towards the periphery of the blastocyst a stage 
is soon reached in which, in the region of the embryonic knob, an 
outer layer—more than one cell thick—of epiblast can be dis- 
tinguished from a subjacent stratum, of which the cells have 
assumed a more flattened aspect (fig. 24), and are continuous 
with the hypoblast-cells beyond the embryonic knob. From 
this moment onwards we shall do well to drop the term of em- 
bryonic knob, and to cali the thickened circular or oval patch 
of embryonic epiblast, in accordance with the name chosen by 
Bonnet and other authors, the embryonic shield. This is thus 
a purely epiblastic structure, whereas the embryonic knob con- 
tained both epiblastic and hypoblastic elements. 

As the blastocyst gradually enlarges, and the zona still 
further attenuates, the embryonic shield increases both in size 
and in thickness (figs. 12 and 25; 26 and31). The hypoblast 
now forms a complete and closed sac, clothing the entire 
inner surface of the trophoblast. This completion of the 
hypoblast into a closed and independent sac (nowhere coales- 
cent or fused with the epiblast) is thus attained before the first 
trace of the origin of a middle layer has become apparent. The 
actual diameter of the blastocyst is now about 0°8 to 1 mm. 
Reconstructions of the entire surface of the embryonic shield 
out of a continuous series of sections show that in this stage 
the shield is hardly ever quite circular, but has generally an 
unmistakable oblong, sometimes an ovoid, shape, the thinner 
end of which corresponds to what will afterwards be the ante- 
rior, the thicker to the posterior end of the embryo.! 


1 In some specimens (fig. 21) it is the anterior end that is the broader. 
Even in the later stages of Pl. XXX VIII the oblong shield is sometimes wider 
posteriorly, though generally anteriorly. At the same time it must not 
be forgotten that the surface views of figs. 16—21, 32—35, 62—64, and 
79—S1 were not drawn from the fresh specimens, but are careful recon- 


508 A. A. W. HUBRECHY. 


A most remarkable fact, to which I must now call attention, 
is this, that it is not in the posterior region of the epiblastic 
shield that the formation of the middle layer and its earliest 
representatives—notochord and lateral mesoblast plates—is 
first inaugurated. Ir 1s In THE HyYPoBLAST that the first 
differentiation occurs which ultimately leads to the 
formation of the above-mentioned structures, In the 
dozen or more of embryos of this stage which are in my posses- 
sion I invariably find that the hypoblast undergoes an 
important modification always in the selfsame spot, 1. e. just 
below the anterior end of the embryonic shield. Here the 
hypoblast, which from the beginning was never more than 
one cell thick, thickens over an area which soon counts 
some five or six dozen of cells. It is not at the outset a 
splitting process, but the cells simply become thicker and more 
massive (fig. 30), and lose their flattened fusiform shape, 
whereas the nuclei become much more closely packed. 

In fig. 20 the hypoblast-nuclei underlying the epiblastic 
shield are drawn in situ. This will give a general apprecia- 
tion of the phenomenon, and at the same time prevent that the 
dark patches in figs. 18—21 and 32—385, by which the region 
here mentioned is indicated, should be regarded as very sharply 
separated from the rest of the hypoblast. 

This patch of modified hypoblast-cells has at the beginning 
an oval shape, with the iong axis perpendicular to that of the 
embryonic shield. Part of this patch will develop into the 
anterior portion of the notochord; for this reason I will call 
it the protochordal plate. 

In the stages immediately consecutive upon this it retains 
its isolated position in the rest of the hypoblast ; but now an 
increase in thickness is noted by the cells dividing in a plane 


structions from an unbroken series of sections. If the plane of section is 
not quite perpendicular to that of the embryonic shield, the reconstruction 
will not exactly correspond to the shape of the latter, but rather to its pro- 
jection on another plane. Still, with due allowance for similar small devia- 
tions, the figures here given may be said to give an accurate representation of 
the developmental phases of the embryonic shield. 


STUDIES IN MAMMALIAN EMBRYOLOGY. 509 


parallel to the hypoblast, so that the protochordal plate 
becomes—at least the central part of it—more than one cell 
thick (figs. 28, 40, and 41). 

While this process is going on, the gastrula ridge has, how- 
ever, made its appearance in the posterior part of the em- 
bryonic shield, and its shape, extension, superficial groove, 
mode of development and of production of a peripheral sheet 
of mesoblast correspond to what has so often been described 
for other mammals. Still I will enter into a few details con- 
cerning these processes, yet not without once more insisting 
on the important fact that the formation of a protochordal 
plate has preceded the appearance of the very first indications 
of a gastrula ridge (primitive streak). 


2. The Development of the Mesoblast. 


Towards the origin and further development of the middle 
layer in Sorex vulgaris three distinct sources contribute. 

It is only for a short period that they and the mesoblast 
which they produce remain distinct. Soon the mesoblast 
becomes a confluent plate of cells, more than one cell thick, 
stretching further and further between the two primary layers, 
and separated in the median axis of the embryo by the noto- 
chord or by the tissues that give rise to it. There is every 
reason to suppose that after a short time the process by which 
mesoblast originates from the three different sources above 
mentioned ceases, and that from this period onward the growth 
of the mesoblast is due to self-division of pre-existent meso- 
blast-cells. It follows from this that only by examination of 
certain favorable early stages we can obtain reliable data 
concerning the actual origin of the mesoblast. 

The three sources above alluded to are— 

(1) The protochordal plate noticed on p. 508. 

(2) The gastrula ridge and its median prolongation forwards, 
which advances between epi- and hypo-blast, and which I 
have proposed to designate by the name of the protochordal 
wedge (Kopffortsatz, auct.). 

(3) An annular zone of hypoblast situated just outside the 


510 A. A. W. HUBRECHT. 


limits of the embryonic shield, and thus enclosing—but at the 
outset independent from—the protochordal plate. 

This annular zone makes its first appearance when the gas- 
trula ridge has already passed through its first stages of 
development. Its formation is thus a later phenomenon of 
differentiation in the hypoblast than was that of the proto- 
chordal plate. 

Still it is of the same order, being essentially in the first 
instance a thickening and approximation of a number of hypo- 
blast-cells that are contained in the annular zone below, but 
just outside the border of the epiblastic shield. This change in 
the hypoblast-cells appears to take place almost simultaneously. 

Embryo 73f (fig. 32) was taken from the same uterus and 
preserved in the same way as the embryos 78a, b, and d (figs. 
33—35). In it there was not yet a trace of the zone here 
alluded to, and also in other respects this embryo is somewhat 
behind the others in development. In the three others the 
closed ring can be discerned which is formed of the hypo- 
blast-cells here alluded to; the formation of the ring may 
hence be concluded to take place rather suddenly. 

There will be some difficulty in giving a concise description 
of the origin of the mesoblast from the three different sources 
here stated, because of the very early tendency of their pro- 
ducts to become mixed up into one continuous layer. Only at 
the very earliest stages is the distinction a clear one. And 
so it is more especially these which will have to be described 
somewhat more fully. The embryos No. 73, above alluded to, 
are for this purpose all the more valuable, as they differ in 
development just enough to elucidate the actual course of the 
phenomena we are here investigating. 

It must here be remarked that there is an apparent disad- 
vantage in the fact that some of the sections are neither strictly 
transverse nor strictly longitudinal. This could not always be 
attained, since it had become obvious that the best way of 
cutting the blastocyst was to imbed it in situ in the uterus, 
even though in the early stages there is no strict con- 
cordance between the long axis of the uterus and that of the 


STUDIES IN MAMMALIAN EMBRYOLOGY. 511 


embryo. Still it was soon found that this disadvantage was 
imaginary; that, the series being complete, entire accuracy of 
reconstruction could be obtained, and that, more careful perusal 
and comparison being necessary under these circumstances, 
certain points came to light which otherwise might have re- 
mained unnoticed, whereas certain other points are much less 
evident in either transverse or longitudinal section series than 
in oblique ones. 

The utmost care was bestowed on the surface views, which 
have been traced by means of direct reconstruction (in a given 
plane of projection) of the camera lucida drawings that were 
made of the actual sections. And so we commence our descrip- 
tion with what occurs in the anterior source of mesoblast—the 
protochordal plate. 

This plate of thickened hypoblast in embryo 73f/ (fig. 32) 
can hardly be said already to contribute to the formation of 
independent cells between hypoblast and epiblast (figs. 40 and 
41). But in 736 (fig. 33) it does thus contribute. In the 
centre its original character as a thickened patch of hypoblast 
is retained, and even more marked by still more considerable 
local proliferation (figs. 53 and 54). At the borders of this 
patch fusiform cells are seen (in sections not here figured) to 
radiate from it, and to spread between the two primary layers 
in a direction more or less perpendicular to the long axis of 
the embryonic shield. 

Thus the first trace of lateral wings of cells (that have 
originally a decidedly fusiform ‘‘ mesenchymatic” aspect) de- 
veloping from the protochordal plate becomes marked. 

In embryo 73d (fig. 34) the mesoblast-cells in this anterior 
region have already attained a more considerable numerical 
development (figs. 43 and 44), it being at the same time very 
worthy of note that they do not to any extent spread back- 
wards, 1.c. in the direction of the front end of the gastrula 
ridge. 

In the following stage, represented by the embryos No. 45 
(figs. 62—64), the mesoblast is already one continuous plate. 
But in the anterior region of the embryonic shield we easily 


512 A. A. W. HUBRECHT. 


recognise (though I have not, as in figs. 31—34, marked this by 
a darker shading) our protochordal plate, broader and more 
massive, and laterally merging into plates of mesoblast (figs. 66, 
67, 70, and 71). The cells of this mesoblast have no longera 
fusiform aspect, but are more rounded and similar to the meso- 
blast-cells in the middle and posterior regions of the embryonic 
shield. A comparison of fig. 66 with 67, and of 70 with 71, 
will show that in 67 and 71 it is more difficult to distinguish 
between protochordal plate-cells, mesoblast-cells, and hypoblast- 
cells than in 66 and 70, although these latter sections are 
situated further forwards. It is in this stage that the proto- 
chordal plate may be said to have reached its maximal exten- 
sion. In the following stage (embryos 42, figs. 79—81) it is 
hardly more than one or two cells thick; the mesoblast has be- 
come much more distinctly separate from it. In still later stages 
the protochordal plate-cells are fairly on the way of becoming 
indistinguishable for a time from the adjoining enteric hypo- 
blast. With this they extend as a continuous plate (only one 
cell thick) below the medullary groove or canal, and only later 
the front portion of the notochord folds off, this being the last 
phase in the developmental phenomena of what we have called 
the protochordal plate. This has been figured by Heape for 
the mole, and is not further entered upon in this paper. 

It is important to distinguish this primary excalation of noto- 
chordal tissue in the anterior protochordal plate region from the 
secondary excalation which takes place further backwards in 
the region where the protochordal wedge has first fused with 
the enteric hypoblast. This latter phenomenon has been de- 
scribed by other authors (7, 24, 29). 

Passing to the second source of mesoblast-cells, the gastrula 
ridge, we find it to be indeed a very considerable contributor 
to the increase of the middle layer. 

We have already noticed that it appears after the proto- 
chordal plate has become very distinctly differentiated. 

When the gastrula ridge for the first time becomes apparent 
(embryo 73/, figs. 32 and 37—389) we notice a fusion which 
then obtains between hypoblast-cells (that have hitherto been 


STUDIES IN MAMMALIAN EMBRYOLOGY. 513 


subjacent to but independent from the epiblastic shield) and a 
posterior swollen knob of this epiblastic shield. This fusion is 
at the early stage so superficial that it is as yet easy to distinguish 
between the proliferated epiblastic cells of the gastrula ridge 
and the flattened hypoblast-cells that adhere to them (figs. 37 
and 38). Moreover, the fusion at first only affects half a dozen 
hypoblast-cells. Certain important data should here be borne in 
mind as following from stage 73/, viz. (a) that the first gas- 
trula-ridge proliferation arises at the posterior end of the 
epiblastic shield, and not anywhere towards its middle; (4) that 
the first indication of a forward growth (protochordal wedge) of 
this proliferation (fig. 39, p. w.) is unmistakably present; (c) 
that the epiblast in front of the gastrula ridge is henceforth 
quite distinct from that which belongs to the ridge ; (d) that a 
perforation of the epiblast in the anterior end of the gastrula 
ridge can be faintly noticed (figs. 88 and 39 p), it being very 
questionable whether it actually opens out at both ends. Also 
later on the traces of a lumen (canal) in the protochordal 
wedge (Kopffortsatz) are so sporadic (figs. 50, 69, 73, 78, 79, 
81, 85, 86, 90) that I would not venture to apply the diagram 
which Heape has given for the mole’s neureuteric canal without 
restriction to the shrew. 

The point where this fusion between the originally inde- 
pendent sheets of epiblast and hypoblast first comes about is no 
doubt the spot that also passes by the name of Hensen’s knob. 
Forward from it there is a growth that gives rise to notochord 
(pro parte) and gastral mesoblast (Rabl); posteriorly we 
find the region of the peristomal mesoblast. Such peristomal 
mesoblast is as yet absent in the embryo 73 /; to its formation 
the further backward growth of the proliferation is pre- 
liminary. The gastrula ridge gradually stretches backwards. 
Posteriorly it again dilates into a caudal swelling (Schwanz- 
knoten, Endwulst; figs. 34, 35, 56, 59, 62, 63). There is 
here a very sudden passage from the epiblast of the embryonic 
shield, which is involved in the formation of this caudal swell- 
ing of the gastrula ridge into the epiblast outside of the 
embryonic shield (ef. fig. 56 and fig. 59). 


514. A. A. W. HUBRECHT. 


The gastrula ridge gives rise to mesoblast in a way that was 
often described for other mammals and Sauropsida. Cells, 
evidently in direct connection with the superficial epiblast, 
proliferate so as to shove in between the primary layers as the - 
lateral plates of peristomal mesoblast. At the posterior end 
of the gastrula ridge which, as above remarked, is so sharply 
defined in longitudinal sections, mesoblast is similarly seen to 
proliferate backwards in the direction of the long axis of the 
embryonic shield, so that the sheet of peristomal mesoblast 
had better not be compared to two lateral sheets fusing 
together posteriorly,| but to a fan-like extension of a cell 
layer proceeding from a linear source of proliferation, the 
gastrula ridge. It should be noted that the posterior and 
superficial portion of this proliferating sheet is in all my pre- 
parations marked by a somewhat different histological aspect, 
distinguishing the component cells from the subjacent strata 
of mesoblast. For this I refer to the figs. 56—60, and 83, 87, 
91, but do not wish for the present to enter upon speculations 
concerning the eventual significance of this histological dis- 
tinction (cf. p. 549). 

That the peristomal mesoblast can indeed be compared to a 
fan that is opened to its maximum width, can be noticed in 
figs. 33---35 of stage 73 a, 6, and d. The dotted line marks 
off the outer edge of mesoblast; it is more or less circular, 
with the gastrula ridge for a radius.’ 


1 In similar sense Kolliker expresses himself for the rabbit when he says 
(‘Festschrift Wiirzburg,’ Bd. i, 1882, p. 34): “ Angesicht gewisser neuerer 
Erfahrungen tiber die Entstehung des Mesoderms aus paarigen Anlagen 
betone ich dass beim Kaninchen Axenplatte und Mesoderm bei ihrem ersten 
Auftreten eine zusammenhangende Lage darstellen und dass auch das 
Mesoderm bei seinem Weiterwucheren wenigstens nach der einen hinteren 
Seite eine unpaare Bildung darstelt.” 

2 It must be observed that in these figures no dotted line is used to indicate 
the boundary of the less sharply defined mesoblast that originates from the 
protochordal plate, whereas the dotted line of figs. 62—64 and 79—81 does 
include such mesoblast, which has then become fused both with the gastrula- 
ridge mesoblast of figs. 33—385 and with the peripheral mesoblast that 
originates out of the annular hypoblastic zone. 


STUDIES IN MAMMALIAN EMBRYOLOGY. 515 


Still I must now introduce a new factor into our considera- 
tions, viz. that not all the mesoblast-cells enclosed within the 
circle here indicated have. arisen from the proliferation along 
the gastrula ridge just described, but that many of them can 
be seen to take their origin from what I have termed the third 
source from which mesoblast arises, viz. the annular ring of 
hypoblast round and outside the edge of the embryonic shield 
(figs. 33—85, 53—55,hy.az.). Figures are given on P]. XXXIX 
(figs. 57—61) in which the participation of this hypoblastic 
zone towards the formation of mesoblast is put beyond all 
doubt. They are taken from three different embryos, and in 
all of them the plane according to which the nuclei divide, as 
indicated by the well-preserved karyolytic figures, is such as 
to leave no doubt that in each of these cases a cell divisioh 
was just on the point of taking place, one of the cells of which 
was going to number amongst the mesoblast-cells, whereas 
the other was going to remain in the underlying layer of 
hypoblast. 

That, moreover, this latter layer also increases by multiplica- 
tion of its cells can be demonstrated in these same prepara- 
tions ; for such an increase the plane of the nuclear division is, 
however, seen to be not coincident with that of the hypoblast, 
but, on the contrary, perpendicular to it. 

Fig. 58 gives the most strongly marked karyolytic stage ; 
fig. 59, z, a stage in which the nuclei have just regenerated 
after division. This latter figure would offer no convincing 
proof of the process here advocated ; it is, however, a welcome 
corroboration of what is noticed in earlier phases in the other 
figures. The reconstructions on P]. XXX VIII allow us to point 
at the exact spots, when seen from above, where the karyolytic 
stages of the foregoing figures are situated. We can then see 
that it is not only the posterior, but also the lateral parts of 
the annular ring that thus contribute to the formation of 
mesoblast-cells. 

Also in the anterior portions of the hypoblastic ring, the 
participation of its constituent cells towards the formation of 
mesoblast can be proved by karyolytic figures. It would, 

VOL. XXXI, PART 1V.—NEW SER. MM 


516 A. A. W. HUBRECHT. 


however, seem that the process is most active in the region 
posterior to the gastrula ridge. Even after mesoblast has 
been derived from it, the annular zone of hypoblast has a dif- 
ferent aspect from the circular patch of hypoblast which it 
encloses, as a glance at fig. 68, Pl. XL, will reveal. 

There can hardly be a doubt that the annular zone of meso- 
blast-formation here described is homologous with the annular 
mesoblast which Bonnet has described and figured in the 
early developmental phases of the sheep. 

Whereas the embryos 73 have allowed us to analyse the 
first origin of the mesoblast, the embryos 45 reveal a further 
step in the development of the middle germinal layer, which 
is very instructive. As far as the protochordal plate is con- 
cerned, we have already above (p. 512) noticed in what respect 
its features are changed when compared with stage 73. It is 
not less important to observe that the forward growth from 
the front end of the gastrula ridge which is often indicated as 
the ‘‘ Kopffortsatz ” of the gastrula ridge, but which has here 
been designated as the protochordal wedge, has evidently, 
during its forward growth, carried forward with it the wings 
of mesoblast. In stage 73 the protochordal wedge was as yet 
only in its very first phases, and the phenomenon here alluded 
to not yet very distinct. In the embryos 45 (figs. 68, 72 —74) 
lateral plates of mesoblast are seen to be confluent with the 
median thickened string of tissue, the protochordal wedge. 
The latter fuses with the hypoblastic protochordal plate, thus 
constituting a continuous band of tissue, from which in the 
further stages the notochord will develop. At the same time 
the mesoblastic wings of protochordal wedge and gastrula 
ridge fuse with the lateral mesoblast that has sprung from the 
protochordal plate, and thus one continuous sheet is formed, 
which in the reconstructed surface views of Pl. XL, figs. 62— 
64, is again indicated by a dotted line. ‘This dotted line marks 
off the outer boundary line of the sheet of mesoblast; it is 
again seen to be more or less circular, though, as we have seen, 
it is not simply the increase of the mesoblast within the circle 
of figs. 83—35 that has given rise to it, but also the anterior 


STUDIES IN MAMMALIAN EMBRYOLOGY. Bled 


mesoblast connected with and originating from the proto- 
chordal plate, and that from the annular zone of hypoblast. 

Thus the region bounded by the dotted line in stages 45 
and 42 (figs. 62—64, and 79—81) contains mesoblast that has 
arisen from the three distinct sources above indicated. His- 
tologically the mesoblast offers no salient points by which, 
from this stage onward, this different mode of origin could 
rigorously be detected. Still it deserves special attention, 
that in the oblique sections through the stage 42 (figs. 79— 
81) the difference between protochordal-plate-mesoblast, pro- 
tochordal-wedge-mesoblast, and gastrula-ridge-mesoblast could 
yet be distinguished with some precision. Fig. 87 is the 
most striking example of this, and at the same time teaches 
us (as do also the figs. 84, 89, and 90) that the protochordal- 
wedge-mesoblast gives ample evidence of being from the 
beginning a double plate of cells. 

The circular patch of hypoblast enclosed by the aunular 
ring was yet more distinct in the embryos 73 than it is in the 
embryos 45. As far as it underlies the gastrula ridge and proto- 
chordal wedge it undergoes a retrogressive metamorphosis, 
and can no longer be distinguished as a separate layer. 

It will be of better aid towards the compreheusion of the 
description above given to compare figs. 66 to 78, which repre- 
sent sections through stage 45, with those of Pl. XXXIX, 
than to enter into any further detailed statements. The pre- 
parations have been very carefully reproduced in the tigures 
that were first sketched with the camera, and for criticism of 
the views here advanced the original preparations are at the 
disposal of such investigators as should wish to convince 
themselves by personal inquiry. 

The passage from stage 45 to stage 42 is a more gradual 
one than was that from 73 to 45. The mesoblast covers a 
wider area. Protochordal plate and wedge now definitely con- 
stitute an axial strip of tissue, from which the notochord will 
take its origin when it has become isolated out of the enteric 
hypoblast, with which it is as yet continuous. The mesoblast, 
which was originally in direct lateral continuity with these 


518 A. A. W. HUBRECHT. 


two, is more distinctly, though not as yet entirely, separated 
from them along this axial line, the two lateral plates of meso- 
blast being, however, continuous with each other, both in 
front of the protochordal plate and behind the embryonic shield. 

Moreover, the extra-embryonic celom has become apparent 
in the stage 42 all along a curved line, just behind the em- 
bryonic shield (figs. 79, 80, 83, 87, 91, cwl.). The cavity is 
widest and most spacious just behind the gastrula ridge, in the 
region where the development of the amnion will begin, 
and where the allantois will make its appearance. In front 
this celom is not yet continued into the mesoblast underlying 
the epiblastic shield. In one of the embryos 42 (see figs. 80 
and 88) I find, however, two small spaces symmetrically 
situated in this mesoblast. I have as yet no definite suggestion 
to make, although presumably these two spaces may be looked 
upon as the first indications of the pericardial cavity. The 
further exposition of these phenomena and the participation 
of the protochordal plate in the formation of the wall of the 
fore-gut and pharyngeal membrane (primitive Rachenhaut, 
Carius), as well as in that of the heart, I wish to reserve for 
a future publication. 


Cuap. I].—THEOoRETICAL CONSIDERATIONS ON THE GASTRULA- 
TION OF THE MAMMALIA. 


In the preceding chapter a detailed description has been 
given of the mode of development of epiblast, hypoblast, and 
mesoblast in the shrew. We have seen that the last-named 
germinal layer has a multiple origin, and that the exact data 
concerning this latter fact can only be gathered from a study 
of certain particular developmental stages which are rapidly 
passed through. After this the mesoblast continues as a 
separate layer to increase in extension without revealing 
anything about its primary origin or the multiple foci of its 
formation. The notochord, the mesoblastic somites, and the 
lateral plates of mesoblast are the representatives of the middle 
layer in these later phases; whereas the formative foci of the 


STUDIES IN MAMMALIAN EMBRYOLOGY. 519 


earlier stages just alluded to are—(a) the gastrula ridge with its 
forward prolongation—the protochordal wedge; (0) the proto- 
chordal plate; (c) the peripheral annular zone of thickened 
hypoblast. 

As late as 1889 a very careful investigator of the develop- 
ment of mesoblast in the Mammalia (R. Bonnet, in ‘ Archiv 
f. Anat. u. Phys.,’ Anat. Abth., 1889) has written about 
this last-named peripheral zone (Il. c., p. 60): “‘ Was die ento- 
blastogene Entstehung des peripheren Mesenchyms anlangt, 
so gebe ich von vornherein gerne zu, dass diese Frage mit zu 
den schwierigsten Aufgaben der Embryologie gehért ;” and 
about the connection between protochordal wedge and proto- 
chordal plate (1. c., p. 84) as follows :—‘‘ Was die ganze, der 
definitiven Abschniirung der Chorda vorausgehende Canali- 
sirung und Hinlagerung des Kopffortsatzes in den Entoblast 
bedeutet ist, mir wenigstens, zur Stunde noch absolut unklar.” 
Rabl expresses himself in a similar sense when in the same 
year (1889) he writes (24, p. 140) concerning the Mammalia, 
“ Die von mir bisher untersuchten Stadien lassen eine Zuriick- 
fihrung auf einander und eine klare Darlegung der Meso- 
dermeutwickelung noch nicht moglich erscheinen.” 

The latest author who has grappled with the problem of 
mammalian gastrulation is Keibel (‘Archiv f. Anatomie u. 
Physiologie,’ 1889, Anat. Abth.), who in the conclusion of his 
essay writes: ‘ Ferner hat die Arbeit gezeigt dass die bis jetzt 
aufgestellten Theorien der Gastrulation fiir das Saugethierei 
nicht durchzufiihren sind und dass dies insbesondere auch von 
den neuesten Versuchen auf diesem Gebiete, von den Theorien 
von Rabl und van Beneden, gilt. Leider ist der Autor 
[Keibel] nicht im Stande diesem Verlangen [nach einer Neuord- 
nung des wieder auf einen Haufen zusammengeworfenen 
Thatsachenmaterials] nachzukommen.” 

Under these circumstances a renewed attempt to deduce 
general conclusions from the facts now known to us will 
certainly not be looked upon as superfluous—general conclu- 
sions in which it is sought to correlate the phenomena of 
gastrulation, mesoblast-formation, and chordatogenesis, as we 


520 A. A. W. HUBRECHT. 


find them in Mammalia, with what we notice in the anamniotic 
Amphibia, Cyclostomata, and Amphioxus. 

It appears to me that the principle of precocious se- 
gregation, which plays such an important part in ontogeny, 
will serve to explain many of the riddles of which even such an 
acute and eminently painstaking observer as Bonnet complains. 
In our case this principle must be applied to part of the 
hypoblast. 

I assume—and we actually observe the fact in the 
opossum (27), the mole (9), the hedgehog (15), the shrew, 
the rabbit (1, 19), the bat (2, 3)—that in the monodermic 
stage of the blastocyst certain cells forming part of its wall 
separate from this and arrange themselves into a layer, which 
either at a very early (hedgehog) or at a later stage (rabbit) 
forms a closed sac inside of the outer layer. A didermic 
stage of the blastocyst is thus inaugurated before the 
actual process of gastrulation has set in. It seems 
to me that we have here an eminent example of precocious 
segregation, the determining cause of which I will discuss 
later on (pp. 529—533). 

As I have supposed that only a portion of the hypoblast 
has partaken in this precocious segregation, another portion of 
it can be expected to arise in a more palingenetic fashion.* 
This we actually notice in the gastrula ridge, and I have 
already above (p. 500) referred to the results of Balfour, Rabl, 


1 Van Beneden has pretended that in the rabbit the hypoblast remains 
absent at the pole of the blastocyst opposite the embryo. Hensen, on the 
contrary, has figured (I. ¢., pl. viii, fig. 18) hypoblast-cells at the incriminated 
spot in an early didermic stage. Keibel (I. c., pl. xxiv, fig. 46) again agrees 
with van Beneden, at least for the early stages. 

2 In the ‘ Anatomischer Anzeiger, Band iii, 1888, p. 911, I have already 
hinted at the probability of the existence of two separate phases in which the 
process of gastrulation of the Mammalia has become subdivided. Keibel has 
taken up this suggestion in his essay above referred to (‘Archiv fiir Anat. 
u. Phys.,’ 1889, Anat. Abth.), but has not worked it out any further, which 
is for the first time done in this paper. Keibel’s paper contains a severe but 
well-founded criticism of van Beneden’s theory of the blastophore and 
lecithophore. 


STUDIES IN MAMMALIAN EMBRYOLOGY. 521 


Bonnet, Fleischmann, a. o., who all agree in looking upon 
this as the homologue of the blastopore, and who must conse- 
quently regard the tissue which from here proliferates inwards 
as in the first instance hypoblast.' 

If for a moment we leave out of consideration the participa- 
tion of the coalescing lips of the blastopore in the formation of 
mesoblast, then the first question that presents itself to us is 
this: How does the palingenetic hypoblast arising in the 
region of the gastrula ridge reunite with the precociously se- 
gregated portion of the hypoblast which is already spread out 
below it as a closed sac, at all events as a continuous layer ? 
If we suppose that instead of the solid gastrula ridge and 
the shallow gastrula groove of mammals an ingrowth along the 
lips of a wide-open circular or elongated blastopore could be 
noticed, even then the direct fusion of this palingenetic hypo- 
blast with that which had been cenogenetically developed as 
a closed sac could not come about, and could not give rise to 
an arrangement resembling the gastrula of Amphioxus or of the 
Amphibia unless a circular patch of the cenogenetic hypoblastic 
tissue were to disappear by which—on the supposition here 
made—the palingenetic archenteron would be shut off from 
the cenogenetic cavity of the umbilical sac. Only after dis- 
appearance of the cell layers separating them the two cavities 
combined would correspond to that of the archenteron of 
Amphioxus. 

And yet not strictly because of the additional space which 
the increase of food-yolk in the Hypotheria and Prototheria 
(Huxley), as represented by the fluid contents of the umbilical 
sac in the Eutheria, has called forth within the hypoblastic 
dominions. : 

We find an analogy to this circular patch of tissue separating 
as a thin cell layer the two cavities just mentioned in the 
membranes by which mouth and anus are primarily closed in 
early developmental stages. These membranes undergo a 

1 Possibly it is the presence of a continuous sheet of hypoblast below the 


gastrula ridge which has hitherto been so much in the way of the recognition 
of the real sequence in the phenomena, 


522 A. A. W. HUBRECHT. 


process of resorption; they leave no trace, and they can 
certainly not be looked upon as palingenetic structures having 
phylogenetic significance ! 

In the same way I am inclined to look upon a definite 
rounded patch of hypoblast below the embryonic shield and 
gastrula ridge as of cenogenetic significance, and I will now 
point out certain morphological, histological, and develop- 
mental peculiarities which substantiate this view, at least for 
certain mammals, viz. the shrew here described and the sheep 
(Bonnet). The patch to which I allude forms part of the 
space which is enclosed inside the annular ring of modified 
hypoblast, with the anterior part of which the protochordal 
plate (which arises independently of it) afterwards fuses (see 
p- 515). In figs. 33—35 the extension of the oblong patch 
here alluded to is distinctly marked.! In the sections the 
hypoblast-cells of this region are seen to be uncommonly flat, 
the nuclei very wide apart. 

It is with these cells that the lower layer of the gastrula ridge 
fuses ; it is here that the so-called connection between epiblast, 
mesoblast, and hypoblast comes about (cf. fig. 38), a connection 
which in Mammalia is undoubtedly secondary. If my view is 
correct the connection is one between primary (or palin- 
genetic) and secondary (or cenogenetic) hypoblast, and we 
can well understand on this hypothesis that the fusion becomes 
so intimate that soon it is impossible to notice any boundary 
line. Actual resorption of the flattened patch of secondary 
hypoblast by the much more massive, bulky, and active cells 
of the primary hypoblast (gastrula ridge) may, for aught I 
know, take place. At all events, I have not noticed the 
slightest fact in support of what Heape has brought forward 
for the mole, and Lieberkiihn and Hensen for other mammals, 
that at the point of fusion between gastrula ridge and under- 
lying hypoblast (it is the anterior, not the posterior region 


1 Bonnet figures early stages of the sheep’s blastocyst (‘ Archiv. f. Anat. u. 
Phys.,’ 1884, Anat. Abth., pl. ix, figs. 8, 14, 15, 36, and 38). The light 
space in the centre of his annular “ Mesoblast-hof” corresponds to the 
region which is figured for the shrew, and which is here alluded to. 


STUDIES IN MAMMALIAN EMBRYOLOGY. 523 


of the gastrula ridge that is here understood !) the latter takes 
any part inthe production of new cell material for the gastrula 
ridge or the mesoblast. Karyokinesis of the nuclei forming 
part of this thin stratum of hypoblast in a plane coinciding 
with it was never noticed by me, and I am here in entire 
accordance with Bonnet.! However, if we carefully consider 
the wording of Heape’s conclusions (‘ Quart. Journ. Micr. 
Sci.,”” 1883, p. 47), we find that he writes, “ Immediately 
below the primitive groove there is no layer of hypoblast to be 
distinguished, and here mesoblast is produced from hypoblast- 
cells (sic). Laterally all three layers are distinct, but in the 
middle line they may be said to combine with one another, 
and in this region, therefore, the middle layer is formed from 
both epiblast and hypoblast.”’ 

This proves that he deduces the participation of the under- 
lying hypoblast in the front region of the primitive streak, 
towards the formation of mesoblast, not from any karyolytic 
indication, but from the mere fact that the distinction of the 
three layers being in this region no longer possible, a partici- 
pation of both the primary ones in the formation of the third 
one cannot be strictly denied. 

Since then the progress made in the study of karyokinesis 
has rendered us more careful, but at the same time more 
decided. So that I hold that I may safely deny any active 
proliferation, in the sense here alluded to, of the hypoblastic 
layer that first underlies and afterwards fuses with the front 
end of the gastrula ridge (primitive streak); and may, on the 
contrary, assume that a process of resorption of this patch of 
tissue is more in accordance with what we observe in sections. 
This resorption also applies to the short stretch of hypoblast 
underlying the protochordal wedge (see figs. 49, 55, 68, 74, 


1 Bonnet writes (I. c., 1889, p. 41), “ Dafiir, dass dem Knoten oder der 
Gastrulaleiste vom Entoblast her durch Theilung oder Ausschaltung der Ento- 
blast-Zellen Zuwachs geliefert werde, finde ich, auch nach erneuter Revision 
meiner Schnittserien, ebensowenig Anhaltspunkte wie v. Kolliker beim 
Kaninchen. Die miihsame Controle..... ergab stets dass die Thei- 
lungsebene senkrecht auf der Entoblastflache stand.” 


524 A. A. W. HUBRECHT, 


78), and I have no doubt that this small strip of tissue has 
actually been eliminated, when once the protochordal wedge has 
become fused with the protochordal plate, and when the meso- 
blast extends as a continuous sheet (though of multiple origin) 
between epiblast and hypoblast. The gastrula ridge and its 
forward prolongation—both of them palingenetic hypoblast— 
thus contribute material to the formation of the dorsal wall of 
the intestine, in a region where the underlying cenogenetic 
hypoblast-cells soon become indistinguishable—a fact in which 
nearly all observers concur, although they may offer different 
interpretations. 

When this phenomenon is accomplished the effects of the 
precocious segregation above sketched have been eliminated, 
and comparison with lower vertebrates is mucheasier. Those 
portions of hypoblast that are enclosed inside the annular 
“© Mesoblast-hof,’’ and that have not been sacrificed to the 
fusion here alluded to, are in their turn utilised in the forma- 
tion of the dorsal enteric wall: they lose their flattened aspect 
of the early stages, and become indistinguishable from the rest. 

The fusion between palingenetic and cenogenetic hypoblast 
having thus finally come about, we can commence our com- 
parison of the lowest vertebrates and of the mammals with 
respect to the further processes. 

To begin with the notochord, we may say that its formation 
out of a continuons strip of mediodorsal hypoblast is also met 
with in mammals, but that here the strip has been divided 
into an anterior and a posterior portion, the former (proto- 
chordal plate) belonging to the group of hypoblast-cells that 
have undergone cenogenetic displacement, the latter (proto- 
chordal wedge) forming part of the palingenetic hypoblast and 
actually growing forwards from the blastopore, and connected 
with the epiblast in the front wall of the neurenteric canal, as 
we see in Amphioxus. The mesoblast in Amphioxus arises to 
the right and left of this strip of protochordal hypoblast in the 
form of separate diverticula, the walls of which are directly 
continuous with it. 

In such mammals as the shrew, notwithstanding the very 


STUDIES IN MAMMALIAN EMBRYOLOGY. 525 


considerable cenogenetic changes, we find mesoblast to arise 
from and be connected with the selfsame strips of tissue from 
which we have seen the notochord to originate. 

Both the protochordal plate and the protochordal wedge 
give rise to lateral wings of mesoblast, which posteriorly pass 
into the mesoblast that is segregated from the coalesced lips 
of the blastopore along the gastrula ridge. The duplicity in 
the origin of the notochord, here ascribed to precocious segre- 
gation of a portion of the hypoblast, very naturally also 
applies to the mesoblast that originates in the corresponding 
regions. At the same time it appears equally natural that 
when the fusion of the two protonotochordal halves has come 
about, the two pairs of mesoblast wings should also reunite. 

Direct comparison between Amphioxus and Mammalia is 
less easy when we come to consider the mesoblast that arises 
from the gastrula ridge and that which takes its origin from 
the peripheral hypoblast. When the embryo folds off from 
the yolk-sac this latter hypoblast constitutes the greater part of 
the wall of the gut (Darmentoblast). As far as we can judge 
from Hatschek’s and Kowalevsky’s observations, nearly all the 
mesoblast of Amphioxus is derived from the archenteric 
diverticula above alluded to. Hatschek further mentions and 
figures two larger polar cells of mesoblast at the hind end of the 
embryo, right and left of the blastopore, but we only learn 
in very general terms that they participate in the formation 
of mesoblast for the caudal region (1. ¢., p. 35).! 

Here, however, the Cyclostomata and the Amphibia furnish 
us with data that are valuable for the explanation of the phe- 
nomena in the Mammalia. 

To Calberla, Scott (254), Nuel (22), and Shipley (28) we 
are indebted for the facts as they present themselves in Cyclo- 
stomata; to Scott and Osborn (25), O. Hertwig (12), O. Schultze 
(26), and many others to those which are noticed in Amphibia. 
Still there is divergence of opinion amongst these authors on 


1 «« Wir werden sehen dass diese Zellen die stets den hinteren Kérperpol 


bezeichnen bei der Bildung des Mesoderms den hinteren Abschluss desselben 
bilden.”’ 


526 A. A. W. HUBRECHT. 


some important points. One of these has more particular 
bearing on the question which we are here considering. It is— 
does the hypoblast (constituted in these groups of much more 
bulky cells than the epiblast) contribute towards the formation 
of mesoblast in the ventral and caudal regions by actual de- 
lamination from the surface ? 

Nuel, Scott and Osborn answer in the affirmative ; Shipley 
and O. Hertwig in the negative. As the latter, however, very 
distinctly admit and figure a participation of hypoblast-cells to- 
wards the formation of mesoblast in the immediate vicinity 
of the blastopore the question is more one of quantity than 
of quality. Whether new cells are added from the hypoblast 
to the mesoblast along circles of increasing radius, or whether 
this increasing radius of the mesoblast is due to growth of the 
free edge of the mesoblast, which was primarily derived from 
hypoblastic cell matter, is in itself a phenomenon which may 
be modified either one way or the other by more or less preco- 
city of the segregation process.1 

The importance of the phenomenon lies in the fact that a 
special secondary portion of mesoblast (peristomales Meso- 
blast, Rabl) originating from the hypoblast can also be traced 
in the Amphibia, aud that this portion must be homologous 
to that which in Amphioxus is derived from the (similarly 
hypoblastic) terminal “ pole-cells of the mesoblast.”” We have 
seen above that the anterior part of the mesoblast in front of 
the blastopore (gastrales mesoblast, Rabl), as it originates in 
Amphibia, allows of very close comparison with the similarly 
situated mesoblastic diverticula of Amphioxus, a comparison 
which was firmly established by the most valuable investigations 
of O. Hertwig. 

Turning to the Mammalia, we have seen that the com- 
parison of the medio-dorsal mesoblast-wings (gastrales meso- 


1 Ose. Schultze regards the phenomena as they present themselves in 
Amphibia in a very different light from Hertwig’s. I will not here enter 
upon the points of dispute between them, but merely remark that the 
figures in both his papers would allow of an interpretation of the formation of 
ventral and posterior mesoblast in the sense of Scott (for Cyclostomata), and 
of Scott and Osborn (for Amphibia). 


STUDIES IN MAMMALIAN EMBRYOLOGY. 527 


blast, Rabl) with the similar formations in Amphibia is easy 
enough, 

And as for the so-called peristomal mesoblast, we shall have 
to look for that in the first place in the immediate vicinity 
along the sides and at the posterior end (Endwulst) of the 
gastrula ridge. The identification of the lateral wings of meso- 
blast that spring from the gastrula ridge with the Amphibian 
peristomal mesoblast has already been effectuated by Rabl 
himself. Still it appears to me that the woodcut which he 
gives on p. 173 of his essay (24) is not complete, but ought to 
show a posterior loop connecting the two parallel dotted lines, 
thereby expressing that from the gastrula-ridge mesoblast does 
not originate in the shape of two separated halves which after- 
wards coalesce posteriorly, but that the plate of mesoblast 
could better be compared to a fan which was brought to its 
maximum of expansion (see above, p. 514). It is the exten- 
sion backwards of this continuous mesoblast plate that can of 
course be directly compared to what takes place in Amphibia. 
For the shrew, I have above demonstrated that at the posterior 
end of the gastrula ridge new cells are added to this plate of 
mesoblast, which directly spring from the underlying hypo- ° 
blast belonging to the modified annular zone. This pheno- 
menon is again comparable to what was noticed in Amphibia 
concerning the participation of a certain number of yolk-cells 
towards the formation of the peristomal mesoblast. I have 
also noticed above that laterally numerous indications were 
found of the actual participation of hypoblast-cells to a 
similar end; whereas for the sheep, Bonnet contends that to 
a no less considerable extent the hypoblast participates in 
the formation of peripheral mesoblast along the whole ex- 
tension of an annular region slightly larger than the embryonic 
shield. To this annular zone, which I have above alluded 
to more fully, he has given the name of ‘‘ Mesoblast-hof.” I 
have above described an exactly similar ring of tissue in the 
shrew, which is at all events peculiarly modified peripheral hypo- 
blast, even if we cannot fix for the present the exact extent 
to which either the whoie or only a part of it actually produces 


528 A. A. W. HUBRECHT. 


mesoblast-cells by karyolytic cell-division. This latter point, 
however, is, as we shall see, to a certain extent secondary, in 
the same way as we have judged it secondary, whether in 
Amphibia mesoblast was produced from a larger or from a 
smaller extent of surface belonging to the hypoblast-cells that 
will finally constitute the lower and posterior wall of the larval 
intestine. 

The difficulty that remains is this: is there any possibility 
of comparing that hemispheral surface belonging to the lower 
and posterior portion of the larval amphibian hypoblast with 
the annular zone observedin mammals? Ithinkthis comparison 
will offer no difficulties if for a moment we were to suppose 
the larval amphibian when it was in the stage of fig. 92 to 
increase by the addition of food-yolk. It might then be ex- 
pected to expand ventrally, the actual cells which were after- 
wards to partake in the formation of the ventral wall of the 
gut being pushed aside, whereas at the same time a further 
inferior expansion of both epiblast and hypoblast furnished a 
sac in which this increased yolk might be expected to find its 
place. Of the state of things here described I have given an 
outline sketch in fig. 93, to the details of which I will presently 
return. It requires no straining of the imagination to picture 
to ourselves fig. 93 here alluded to still further expanding into 
a spherical sac, on the top of which the future embryonic tissue 
was flattened out, and we then immediately see that an annular 
zone of hypoblast would thus make its appearance (fig. 95) under- 
lying the free borders of the embryonic epiblast, and coutribut- 
ing, when once the folding off of the embryo might have set 
in, towards the formation of the ventral and posterior wall of 
the gut. This assumption of a very considerable increase of 
food-yolk indeed serves to explain the change of shape and 
size, the origin of a vascular area on the yolk-sac, &c, 

We have reason to expect that between the Amphibia and 
the Hypotheria a phylogenetic link has once existed in 
which actual food-yolk formed a very considerable addition to 
the early blastocyst. The case of the Ornithodelphia is most 
important in this respect. There is little or no food-yolk in 


STUDIES IN MAMMALIAN EMBRYOLOGY. 529 


the Didelphia, none in the Monodelphia; and the very large 
size (when compared to the embryonic area) to which the 
blastocyst of the higher Mammalia increases has generally 
been looked upon as a repetition, called forth by heredity, of 
these ancestral lecithophorous arrangements. To me it has 
always appeared that this explanation is rather strained. A 
yolk-sac without yolk would be an encumbrance to a mammal 
that completed its development inside the maternal genital 
ducts, and would long since have been reduced or even eliminated 
by natural selection—unless under the changed circumstances 
a new and important function has come to be fulfilled by it, 
which is of equally vital importance to the continuation of the 
species. 

This has, I hold, been the case in Mammalia. When the 
nutritive contents of the yolk-sac were no longer of primary 
importance, and a considerable reduction in size of the blasto- 
cyst might have gone hand in hand with the change from 
mesoblastic to holoblastic segmentation, this was not effectuated 
because another factor came into play. 

The vascular area which heredity called forth on the surface 
of the yolk-sac, and by the aid of which the nutritive contents 
of that sac were elaborated and absorbed, must have rendered 
eminent service for the establishment of a different mode of 
nutrition as soon as the embryo underwent a considerable 
part of its development inside the maternal generative ducts, 
The beautiful figures which Selenka has given for the opossum 
(27) demonstrate this most forcibly, and the temporary abdi- 
cation of the allantois in this particular case is also most 
instructive. Now, for a satisfactory working of the new 
arrangement it is undoubtedly of the utmost importance that 
the surface of the area vasculosa should be stretched to its 
maximum extent, and at the same time should be elastic 
against pressure tending to throw it into folds. The change 
required would thus be the substitution of liquid contents, 
serving the purpose just alluded to, instead of the nutritive 
contents characteristic of the Hypotherian ancestors. With 
the absorption and retention of this liquid, under a certain 


5380 A. A. W. HUBRECHT. 


pressure, the outer layer of cells of the mammalian embryo— 
the trophoblast—has no doubt been specially entrusted. For 
this purpose it is undeniably the most favorably situated. 

It is certainly significant that in all Mammalia it forms a 
closed sac at the very earliest period after segmentation of the 
ovum has commenced. More significant yet is the fact which 
I have noticed in Sorex that a considerable increase in size of 
the early blastocyst is brought about (cf. figs. 5 and 6 with 
figs. 8—11) without any adequate increase of the number of 
cells composing it. This is the actual demonstration of the 
fact that the increase in size is due to an increased tension, 
which, in the case of the spherical blastocyst, can only be 
brought about by the accumulation of liquid contents that 
are under a higher pressure inside the blastocyst than is the 
surrounding medium, This in its turn has to be ascribed to 
inherent properties of the protoplasm of the trophoblast-cells— 
properties which may either be of a more secretive or of a more 
osmotic nature, as will some day have to be more carefully 
determined. The actual high elasticity of a mammalian blasto- 
cyst has often been observed and been commented upon. 

The utility of this arrangement has probably contributed 
more towards the retention of what I would call the pseudo- 
meroblastic condition of the blastocysts of the higher Mammalia 
than has the hereditary tendency towards the production of 
this condition. Moreover, other factors came into play to 
increase the significance of this elastic and spacious blastocyst. 
It offers a very safe lodging for the developing head of the 
embryo, which already in Reptilia is seen to be enclosed in a 
proamnion that bends downwards into the yolk. Such a 
protection is all the more effective for the mammalian embryos 
that are no longer protected by a hard shell, but enclosed in 
moveable and contractile maternal tissue.’ 


1 Another reason which might apparently be given for the elasticity and 
the increase in bulk of the mammalian blastocyst has here been intentionally 
left in the background, viz. the reason that thereby the walls of the uterus are 
bulged out, in consequence of which nutritory facilities are obtained. I am 
not inclined to attach any value to this argument, which appears to me to be 


STUDIES IN- MAMMALIAN EMBRYOLOGY. 531 


The different size which the blastocyst of the same degree 
of development attains in different mammals (extremes being 
represented e.g. by the rabbit on one side and by the hedgehog 
on the other) may partly be influenced by the more or less favor- 
able conditions of nutrition under which the vascular area finds 
itself placed. In a former publication (‘ Quart. Journ. Micr. 
Sci.,’ 1889) I have described these as particularly favorable 
in the case of the hedgehog. In short, I have here touched 
upon several points which have all contributed to an important 
change in the function and also in the development of the 
outer wall of the early blastocyst. Now this change will, I 
presume, have been equally momentous for the development of 
the inner layer of the didermic blastocyst—the hypoblast. 

And there can hardly be a doubt that the earliest function of 
the trophoblast, as above hypothetieally described, can certainly 
be rendered more effectual if at the same time the hypoblast 
follows suit, and constitutes at the earliest possible moment 
an inner lining to the trophoblastic sac. 

The area vasculosa spreads out between these two mem- 
branous cell layers. The danger of a slight defect in a mono- 
dermic expanded blastocyst might be reduced by 50 per cent. 
if the blastocyst is not monodermic, but didermic. The latter 
consideration may still further help us to understand a pecu- 
larity in the gastrulation of the Mammalia, as compared to 
that of the Reptilia (lizards [Weldon, Strahl, Hoffmann] ; 
tortoises [Kupffer, Mitsukuri and Ishikawa], a. 0.). The 
palingenetic phenomenon of infolding at the lip of the blasto- 
pore, which in the latter is so clear and considerable, and so 
intimately linked with the formation of a neurenteric canal, is 
ever so much more obscured in mammals. Now, if the elastic 
tension of the mammalian blastocyst is a distinctive charac- 
teristic which has developed in the way that has been above 
hypothetically sketched, then we can very well understand that 
an open-mouthed blastopore has come to be more or less oblite- 
too mechanical. Selection will probably not have operated in such a direct 


way, and the swellings of the maternal tissues are parallel to the increase in 
size of the ovum; they are certainly not occasioned by it (cf. fig. 12). 


VOL, XXXI, PART IV.—NEW SER. NN 


532 A. A. W. HUBREOHT. 


rated and retarded. Supposing for a moment the palingenetic 
phenomena in the region of the primitive streak were to follow 
the type of the Reptilia, a wide-open blastopore ensuing, and 
now this palingenetic hypoblast to fuse with the pre-existent 
cenogenetic hypoblast underlying it, and to turn into a gas- 
trula in the way that was indicated above (p. 521), the fatal 
effect would inevitably be that the fluid contents would escape, 
and that the blastocyst would collapse. A first safeguard 
against this danger is the fact that the hypoblast is a closed 
sac; a second that the canal in the protochordal wedge—a 
palingenetic remnant of the posterior medio-dorsal region of 
the archenteron—is (1) often absent, and that (2) when present 
it is exceedingly fine, capillary resistance thus counteracting 
the tendency of the enclosed fluid to escape by that canal. 

Moreover the canal is, firstly, very much bent forwards 
under nearly right angles to the radius of the blastosphere, 
which is another physical impediment towards the escape of 
fluid ; secondly, only in later stages, when the so-called ‘‘in- 
tercalation ’’ in the hypoblast has come about, it is in a more 
or less extensive communication with the cavity of the yolk- 
sac; thirdly, the attachment of the blastocyst within the 
uterine cavity has by that time become more definite, and 
thereby the pressure above the germinal area more or less equal 
to that inside the blastocyst.1 

1 There is a figure in Selenka’s treatise on the Opossum (27, pl. xviii, 
fig. 3) which at first sight would seem to go dead against the hypothesis here 
developed, because it shows a small pore in a blastocyst at an early though 
already considerably expanded stage. It should be noted—l. That the 
possible escape of fluid contents may in this case be most effectually coun- 
teracted by the thick albuminiferous layer enveloping the blastocyst. 2. That 
in other similar stages (l.c., figs. 4 and 10) there is no trace of a similar 
opening. So that I think my suggestion also holds good for the opossum. 
The phenomenon of stretching of the blastocyst wall, without increase in the 
number of cells, is very marked in Selenka’s figures (cf. |. ¢., pl. xviii, figs. 2 
and 3). 

A ae directly comparable to the one just noticed is figured by Keibel (17, 
1889, pl. xxiv, figs. 464 and 47) for the rabbit, and by Heape (9, pl. viii, 
fig. 31) for the mole. In both cases it cannot be said to be an actual perfora- 
tion. In all the three cases it is in the region just behind the embryonic shield 


STUDIES IN MAMMALIAN EMBRYOLOGY. 533 


At all events, these various considerations allow us to catch a 
glimpse—however hypothetical—of the causes that may have 
brought about the necessity of the precocious segregation of part 
of the hypoblast in mammals. And wecan very well understand 
that once the double closed sac having been constituted, the 
further processes should again offer close analogies to what is 
observed in Sauropsida. The precocious segregation has not 
necessarily had any altering influence on the hereditary 
tendencies of the different portions of the hypoblast from 
which mesoblast originates. As might be expected, the palin- 
genetic phenomena are more closely comparable; the cenoge- 
netic changes do not, however, in any way escape the possi- 
bility of comparative analysis. 

I will now attempt to further elucidate the argumentation 
contained in the foregoing pages by the discussion of four 
diagrams given on Pl. XLII. 

Fig. 92 is a diagram of a developmental stage in either 
Cyclostomata or Amphibia after the infolding has commenced, 
and when from the medio-dorsal wall of the archenteron both 
the notochord and the lateral plates of gastral mesoblast have 
developed, whereas at the lower lip of the blastopore peri- 
stomal mesoblast is originating. If for a moment we give no 
attention to the different colours in this diagram, nor to the 
fact that the solid mass of hypoblastic yolk-cells is here only 
represented by a few polygonal outlines, we know that the 
continuation of the process just commenced leads both in 


that the incriminated spot is found. Judging from Selenka’s figures, it would 
seem to be the starting-point from whence the inwandering of cenogenetic 
hypoblast commences, and he on purpose applies the name blastopore both to 
it and to a yet earlier stage in which he noticed an actual displacement 
inwards (I. ¢., pl. xvii, fig. 8). It is all the more suggestive that this spot is at 
all events in the immediate vicinity of what will become the first trace of the 
gastrula ridge, i.e. of the point of origin of the palingenetic hypoblast 
appearing so much later, when the cenogenetic is already a closed sac. 

That in Sorex the case lies somewhat differently, and that here no corre- 
sponding thinner spot is noticed, must no doubt be ascribed to the thickness 
of the epiblastic embryonic shield, which is comparatively considerable even in 
the very earliest stages. 


534 A. A. W. HUBRECHT. 


Triton and Petromyzon—(1) to a further excavation of an 
archenteric cavity in the direction which in this figure is 
marked by five blue dots ; and (2) to the simultaneous growth 
of notochord and paired mesoblast plates (gastrales mesoblast, 
Rabl) in that advancing anterior region. 

It was above noticed that authors do not agree as to the 
extent of mesoblastic cell material yet produced by delamina- 
tion from the hemispheric lower surface of the hypoblastic 
cell-mass, nor does it matter for the argumentation here put 
forward. Still I notice this point because by-and-bye we will 
have to reconsider this possibility, and will then have to 
picture to ourselves the hemispheric surface here alluded to, 
which in this diagram has received a uniform blue tint, and 
forms the lower layer of the hypoblastic tissue. 

Recapitulating, we thus find that in this figure the round 
dots, both the white and the blue, are meant to designate the 
zone where notochord and gastral mesoblast originate ; the white 
stripes, the zone where the peristomal mesoblast arises; and 
part of the uniform blue hemispherical region, the zone of what 
I will designate as the peripheral mesoblast. 

The passage from this stage to one in which the yolk has 
very considerably increased, as is the case in Sauropsida 
embryos, has already been described by Rabl; his figure of a 
diagrammatic longitudinal section is reproduced with a very 
slight modification and without colours in fig. 94. The blas- 
topore, the anterior lip of which is at the same time the 
anterior surface of the neurenteric duct, is easily identified in 
both figures. In front of this anterior lip (i.e. to the left in 
Rabl’s figure) is the embryonic region; to the right is the 
region of the primitive streak, where the lips of the blastopore 
may be said to coalesce. The regions indicated by white dots 
and by white stripes in fig. 92 (notochord and peristomal 
mesoblast) are here indicated by the letters pw. and pst. ; 
below them is a sheet of cells, the hypoblast, which is always 
distinct, although its significance has lately been differently 
interpreted (paraderm, Kupffer; lecithophore, van Beneden), 
and which I have here somewhat more distinctly indicated 


STUDIES IN MAMMAIIAN EMBRYOLOGY. 535 


than is done in Rabl’s original figure. We can identify this 
layer with the light blue layer overcapping the yolk-cells in 
fig. 92. 

In how far also in Sauropsida regions might be distin- 
guished that could be identified with protochordal plate and 
protochordal wedge in a stricter sense, and in how far an 
annular mesoblast-producing zone of hypoblast can also here be 
distinguished, are questions that will have to be reinvestigated 
very fully. We have not as yet enough data for any definite or 
exhaustive answer.! The chief point is that there is no difficulty 
in comparing the diagrammatic stage (fig. 92), which we start 
from in attempting to interpret the gastrulation of the Mam- 
malia, with the diagrams for the Sauropsida which Rabl has 
given. This is all the more important because with respect to 
the Mammalia, this author, while acknowledging the insuffi- 
ciency of the data at his disposal, yet inclines the other way, 
and has expressed himself in favour of van Beneden’s views, 
which I must dissent from. 

In Sauropsida (1) the great bulk of the yolk ; (2) the parti- 
cipation of the upper layers in the phenomenon of retarded 
cleavage (Nachfurchung), by which new cell-material is added 
to the embryonic tissues ; and (3) the simultaneous appearance 
(at least in the chick) of hypoblast and mesoblast, are pheno- 
mena which obscure the early points in contest by which the 
formation and homology of the layers can be judged. 

In discussing the Mammalia, where, on the contrary, a well- 
defined didermic stage is indisputably present, we can there- 
fore not derive much benefit from the diagram that applies to 
the Sauropsida, and we shall have to fall back upon another 
hypothetical intermediate stage. However, before doing this, 
the diagram here given of the phenomena as we actually find 
them in the Mammalia must first be more closely looked at. 
It is fig. 95 which represents a diagrammatic longitudinal 
section of a mammalian blastocyst (Sorex, Talpa, Ovis), 

1 Figs. 53, 56, and 60 in Duval’s ‘ Atlas d’Embryologie’ (1889) are very 
suggestive as far as the protochordal plate is concerned; but I will refrain 
from any further discussion at present. 


536 A. A. W. HUBRECHT. 


when the formation of the notochord and mesoblast has 
definitely commenced; ». p.is the front end of the gastrula 
ridge, and marks the dorsal lip of the blastopore. When a 
neurenteric duct is present it is here that its dorsal opening is 
situated. The epiblast of the embryonic shield is dark black, 
which in the pre-blastoporian region (e.) is applied as a uniform 
tint, in the region of the gastrula ridge interrupted by hori- 
zontal white stripes. 

From the front end of the gastrula ridge a forward growth 
inserts itself between epiblast and hypoblast ; it is our proto- 
chordal wedge (Kopffortsatz, auct.), and is here marked by a 
black band with round white dots. 

The lateral border of the epiblastic embryonic shield is indi- 
cated in semi-perspective by a black boundary line, its general 
surface by a light grey tint, which also marks the trophoblast 
that forms the outer wall of the blastocyst. 

Turning to the hypoblast, we find it represented by a blue 
tint clothing the inner surface of the blastocyst, forming a 
continuous layer beneath the embryonic shield. Phenomena 
of fusion between the blue hypoblastic lining of the blastocyst 
and the palingenetic hypoblast of which the protochordal 
wedge and the gastrula ridge consist, are in this phase already 
apparent, but not marked in the diagram. A blue semi-perspec- 
tive annular band (hy. az.) indicates the annular region of the 
hypoblast that was fully described above, and that takes part 
in the formation of mesoblast. A thickened patch of hypo- 
blast, enclosed within the anterior border of this ring (our 
protochordal plate), is represented in this longitudinal section 
by a blue band marked by darker blue round dots and con- 
tiguous to the black band of the protochordal wedge. 

The mammalian diagram having thus been explained in its 
general outlines, we shall have to consider how we can bridge 
the considerable gulf that separates it from the diagram 92 of 
the lower Ichthyopsida from which we have started in our 

1 For better interpretation of the diagrammatic section, comparison with 


the figs. 33—35, 62 and 64, in which the blastocyst is seen from above, will 
be useful. 


STUDIES IN MAMMALIAN EMBRYOLOGY. 537 


attempts at harmonious interpretation of the gastrulation 
phenomena in the Amniota. 

A hypothetical intermediate stage may serve this purpose. 
It is given in the diagram of fig. 93. It may be derived from 
fig. 92 by supposing the ventral wall of this latter stage to 
bulge out into a more capacious reservoir for the retention 
either of nutritive yolk or of elastic fluid contents, as was 
already noted above (p. 528). If this process of development 
takes place at the ventral pole of the spherical blastocyst, then 
the ventral hemispherical cap of hypoblast of fig. 92 which in 
the Cyclostomata and Amphibia gives rise to the lateral and 
ventral walls of the intestine, will open out in trumpet fashion, 
as indicated in fig. 93. The additional sac-like reservoir will 
be a medio-ventral appendage to this intestine, and the hemi- 
spherical zone will have become annular, as is indicated in semi- 
perspective in fig. 93. 

The regions which in fig. 92 were marked by the dark blue 
dots will retain their position, and so will the layer hy’. that 
temporarily forms the floor of the archenteric cavity of invagina- 
tion. It should be borne in mind that also in Amphibia and 
Sauropsida this floor is only a temporary one, that it dwindles 
away as the so-called Dotterpropf is being resorbed, and that it 
is then finally replaced by the definite floor of the intestine 
which has developed out of the hypoblast-cells that ab origine 
occupied the lower surface of the hemispherical cap (= the 
annular zone of fig. 93). 

If the transitory character of this layer hy’., which is indis- 
putable in the lower Ichthyopsida, is retained in the hypo- 
thetical stage of fig. 93 and further in the mammalian (and 
sauropsidan) development, we shall have to look for it just below 
the protochordal wedge and the front end of the gastrula ridge, 
these being the coinciding regions that are stretched out above 
it. Now in Mammalia such a portion of hypoblast does exist, 
and coalesces with the cells of gastrula-ridge and protochordal 
wedge. Ceuntrifugally it merges into those hypoblastic surfaces 
which actively contribute to the formation of what will ulti- 
mately be the lateral and ventral wall of the intestine. So in 


538 A. A. W. HUBRECHT. 


this respect the comparison of fig. 92 with fig. 95 holds good. 
It also does this if we compare the lower hemispherical hypo- 
blastic surface of fig. 92, which we find back in the diagram of 
fig. 95 as a flat annular band of hypoblast. This shape it 
roust necessarily have taken if we suppose the transitional 
phase of fig. 93 to have yet further become bulged out, so that 
finally, as was already discussed on p. 528, the formative 
blastoderm became spread out flat on the upper surface of a 
much larger spherical blastocyst, as is indicated in fig. 95. 

The identity of the regions marked by the blue and white 
dots and by the thin white stripes in the figures is, moreover, 
self-evident. 

And so now we have to turn to our hypothesis of precocious 
segregation of part of the hypoblast, and see how it can be 
applied to the diagrams here given. For this I have made use 
of the different colours, and will first discuss the mammalian 
diagram. The epiblast, whether of the embryonic shield or of 
the trophoblast, is black or grey. So is the palingenetic hypo- 
blast, which arises in the gastrula ridge and extends forwards 
as the protochordal wedge. The blue sphere is the closed sac 
of cenogenetic hypoblast, the constituent cells of which have 
wandered inwards before the actual gastrulation process com- 
mences. On this cenogenetic hypoblast certain modified por- 
tions—the protochordal plate and the annular ring—appear 
before the definite fusion between the palingenetic and ceno- 
genetic elements has been accomplished. 

“Those portions of the amphibian hypoblastic cell-mass, 
which I consider to be homologous to the mammalian ceno- 
genetic hypoblast, are in fig. 92 artificially distinguished from 
the remaining part of the hypoblastic invagination by a similar 
blue colour. The way in which I picture to myself that from 
stage 92 the mammalian stage 95 has been arrived at was fully 
discussed above. I have here only to add that in all these 
diagrams a prominent part is allowed to the hypoblast in the 
formation of the notochord. This is indicated by the dark 
blue dots in the dorso-median region contiguous with the proto- 
chordal wedge. Suppose for a moment that the confirmatory 


STUDIES IN MAMMALIAN EMBRYOLOGY. 539 


results to which Bonnet, Heape and myself have arrived in 
respect to this question had not been obtained, but that in all 
mammals it was only the protochordal wedge from which the 
notochord developed, as Carius, v. Beneden and others will 
have it for the guinea-pig, rabbit, bat, a.o.; even then the 
hypothesis that has here been developed would to me seem 
more acceptable than van Beneden’s hypothesis of the blasto- 
phore and lecithophore. It would at any rate not have to 
undergo any important modification. 

Should future researches bring to light that indeed the 
formation of notochord and mesoblast in different mammals 
takes place according to divergent modes, as would appear to 
follow from the researches here cited, even then the theory of 
mammalian gastrulation here developed would remain appli- 
cable, whereas that of van Beneden collapses as soon as a 
definite participation of cenogenetic hypoblast (his lecithophore) 
in the formation of these tissues has been demonstrated. And 
this is now the case at any rate for the sheep and for the 
shrew. 

As, moreover, van Beneden has to deny direct homology 
between the mammalian inner layer and the hypoblast of 
Amphioxus, whereas in my hypothesis these two remain per- 
fectly comparable and homologous germinal layers, I think 
the latter hypothesis will also have a priori grounds in its 
favour. At all events, the supporters of van Beneden’s view 
will have to bring Bonnet’s and my own results in accordance 
with their own hypothetical solution. I myself do not see my 
way to effect this. 


CuHap. I1].—Pornts or Comparison IN EaARruier INVESTI- 
GATIONS BY OTHER AUTHORS. 


I have not in the preceding pages very fully and repeatedly 
referred to the literature of the subject. Whenever references 
were indispensable they have been inserted. Still a large 
amount of embryological research which runs along parallel 


540 A. A. W. HUBRECHT. 


lines remains unnoticed. From this I have extracted such 
points as seem to have special bearing on the subject here 
treated, either as directly confirming or as in apparent contra- 
diction to what has above been described for Sorex. 


1. The Protochordal Plate of Rabbit and Cavia. 


Carius’ dissertation (p. 26) purposely refrains from deciding 
whether in the rabbit there might not possibly be a participa- 
tion of the entoblast towards the formation of anterior portions 
of the notochord. A comparison of his fig. 8 with our figs. 
44, 67, 70, and 71 undoubtedly suggests the presumption of 
their homology, i.e. of the presence also in the rabbit of a 
distinct hypoblastic protochordal plate which certainly is less 
distinct at the outset in the rabbit than we have found 
it in the shrew. In that case we shall also be allowed to 
institute more direct comparisons between the figures which 
he gives of the changes in the blind fore-gut with what 
is found in the shrew, but has not been entered upon in this 
paper. 

Concerning Cavia Cariusis much more emphatic in denying 
any participation of hypoblast in the formation of the noto- 
chord. Here, too, a renewed comparison with what the 
shrew has taught us will allow us to decide whether the proto- 
chordal plate is perhaps so much reduced that it most 
naturally escapes detection, or whether it is wholly absent, 
the protochordal wedge supplying the whole of the notochord. 
At all events, the extreme inversion of the layers, as we 
find it in Cavia, must necessarily induce us not to look 
upon Cavia as a fit representative of the normal mammalian 
development. 

Keibel’s publication (17) concerning the formation of the 
notochord is posterior to that of Carius, and on p. 27 of the 
reprint of his article he holds himself justified to exclude 
for the rabbit any participation of the hypoblast towards the 
formation of the notochord. He recognises that at the fore- 
most extremity the decision in the sense he advocates is 


STUDIES IN MAMMALIAN EMBRYOLOGY. 541 


rendered extremely difficult. He has noticed a hypoblastic 
thickening in this anterior region, but only holds it to be of 
significance for the formation of the primitive “ Rachenhaut ” 
(pharyngeal membrane), whereas a direct comparison with the 
shrew’s protochordal plate and with the chorda-entoblast 
which Bonnet describes and figures in the sheep would perhaps 
lead to different results. That Keibel feels all the importance 
of the contradiction between his own views and those of 
Bonnet (which are so fully supported by the facts observed 
in the shrew) may be concluded from the following passage 
which I translate from p. 345 of his article. He says, “I 
will not here further refer to Bonnet’s doctrine of the meso- 
blast derived from the entoblast. Nothing like it is found 
either in the rabbit or in the guinea-pig, and even Bonnet’s 
own figures have not convinced me. . . . Bonnet’s observa- 
tions are, at all events, not available in support of Rabl’s or 
van Beneden’s hypothesis. If they were confirmed this would 
mean a further difficult complication of our problem, and for 
this reason I believe I may be relieved from further entering 
upon Bonnet’s data.” 

If we refer to van Beneden’s early article on the rabbit’s 
development (‘ Archives de Biologie,’ vol. 1, 1880), we find 
that he confounds for the earlier stage (v1) epiblast and meso- 
blast (pl. vi, fig. 2), regarding the trophoblast (Rauber’s 
Deckzellen) as the definite epiblast. This confusion has been 
refuted by Lieberkiihn, and seems to have since been recog- 
nised as such by the author, who on the same plate (figs. 
11—18) gives correct interpretations of the three layers in a 
later stage (1x), a stage of which he definitely affirms that 
there was as yet no trace of Hensen’s node (i.e. of the gas- 
trula ridge). These three figures should be somewhat more 
carefully considered by us. They show that mesoblast is 
present in the rabbit before there is any trace of the gastrula 
ridge. This van Beneden emphatically states on p. 220, sub. 
13. This mesoblast appears in crescent shape in the posterior 
region of the embryonic shield. Van Beneden does not men- 
tion how he has made out that this was indeed the posterior 


(542 A. A. W. HUBRECHT. 


region, nor does his surface view in fig. 5, pl. vi (l.c.), throw 
any light on this question. I presume that he may have thus 
concluded on a priori grounds, and feel inclined to suggest 
that the region where sections 12 and 13 were taken was 
actually the anterior region of the embryonic shield. In 
that case the crescent-shaped mesoblast might be interpreted 
as mesoblast derived from a hypoblastic protochordal plate 
(not further mentioned, however, by van Beneden), and its 
presence before the appearance of any trace of the gastrula 
ridge would be very well in harmony with facts which we have 
observed in the shrew. However, it is only tentatively that I 
advance this proposition, which only renewed researches of 
very numerous early stages of the rabbit’s blastocyst can 
bring to a definite test. 

That the veteran leader in embryology, von Kélliker, retains 
theoretical objections against any participation of the hypo- 
blast towards the formation of mesoblast is well known, as 
also that these views are all the more emphatically brought 
forward in his later publication. He has experienced the 
gratifying sensation that van Beneden, who in the publication 
above cited (p. 142) most vehemently attacked Kolliker’s 
interpretations in terms which Kolliker resented, though he 
referred to them very magnanimously (‘ Die Entwickelung 
der Keimblatter des Kaninchens, Historische Vorbemerkungen,’ 
p. 5, Festschrift Wiirzburg, 1882), has since turned over an 
entirely new leaf. In his latest, though as yet only prelimi- 
nary communications on the rabbit and the bat (‘ Tageblatt 
der Naturforschervers., Berlin, 1886, and ‘Anat. Anzeiger, 
ili, 1887) van Beneden not only wholly accepts Kélliker’s 
views, but draws very full and far-going conclusions from 
them. As such we may consider his theory of the gastrulation 
of the Mammalia, their blastophore and lecithophore. 

If we consider the plate by which Kolliker’s essay just 
cited is illustrated, we find in the surface views a crescentic 
“ vorderer Randbogen.” This seems to correspond with van 
Beneden’s crescentic mesoblast above alluded to. In the 
text (p. 9) Kolliker compares these stages with those of van 


STUDIES IN MAMMALIAN EMBRYOLOGY. 543 


Beneden, and being apparently also doubtful whether van 
Beneden’s definition of what is the anterior and posterior 
region of the embryonic shield was accurate, he places a ? 
behind the word “anterior” with which he indicated this border 
region, and which I am inclined to homologise with the 
region of the protochordal plate in stages 52 and 73 of the 
shrew. 

Moreover, comparing Kolliker’s pl. i, fig. 1, on which an 
embryonic shield is figured which very strongly resembles our 
fig. 32, pl. c, of the shrew’s stage 73f, we in the first place 
notice the presence of the first trace of the gastrula ridge at 
the lower border of the embryonic shield. It seems impro- 
bable that from here the primitive streak should grow forwards 
as Kolliker interprets it. Fig. 3 would no doubt seem to 
militate in favour of such an interpretation, but then we must 
not forget that it is not drawn on the same scale of enlarge- 
ment. Were we to draw it on the same scale as fig. 1 the 
embryonic region in front of the gastrula ridge would be of 
about equal size to that of fig. 1, and we might look upon the 
gastrula ridge as having arisen by a backward extension of 
the original posterior proliferation hw. of fig. 1, together with 
a general growth of the epiblastic shield much in the same 
way as we have been led to interpret the phenomena in the 
shrew. 

A comparison of Kolliker’s figs. 4 and 5 with our figs. 
33—53 of embryos 73, a, 6, d, and of his figs. 7, 8, with our 
figs. 62—64 of embryos 45, and figs. 79—81 of embryos 42, 
will also prove instructive. It must be borne in mind that 
Kolliker’s drawings are made from surface views, whereas 
mine are reconstructions, and as such—though somewhat less 
reliable as far as the outline goes—more exact where they in- 
dicate histological differences. Kolliker himself says (1. ¢., 
p. 23), “I have made no special study of the differentiation 
of the cells of the germinal layers, nor of the karyokinetic 
processes in them.” We further notice that in Kolliker’s 
fig. 29 there is a central portion of flattened hypoblast, right 
and left of which a histological modification is apparent, 


544 A. A. W. HUBREOHT. 


which offers points of comparison with the modification by 
which the annular zone (Ay. ar., figs. 53, 54, 68) of the shrew 
is characterised. Moreover, fig. 47 of Kolliker raises the 
doubt whether also in the rabbit certain points of the hypo- 
blast might not finally be detected, where its participation in 
the formation of mesoblast could not be as decidedly demon- 
strated as it has been done by Bonnet for the sheep and by 
myself for the shrew. 

It does seem that the rabbit is in this respect a very recalci- 
trant subject, if we notice how also Rabl (“‘ Theorie des Meso- 
derms,” ‘ Morphol.-Jahrbuch,’ 1889, p. 113, pl. ix) confirms 
the views as advocated by Kélliker. From Rabl’s essay I will 
notice certain details which should be compared with what I 
have above advanced for the shrew. Rabl states that in an 
early stage of the rabbit’s blastocyst the two layers are at 
first in no way directly connected, which corresponds with 
what was observed in the shrew. 

Contrary to what we have noticed in Sorex, he states (l.c., 
p. 144) that in the rabbit this mutual independence of the two 
layers is maintained even after the gastrula ridge and proto- 
chordal wedge of the rabbit have made their appearance, and 
the formation of the mesoblast is already considerably ad- 
vanced. Somewhat later, however (l. c., p. 149, pl. ix, 
fig. 9), the fusion of the three layers in the region of Hen- 
sen’s knob comes about, although as yet to no considerable 
extent. 

As to Rabl’s figures (1. c., pl. ix, figs. 5—9), I must again 
invite comparison with what was figured for the shrew on 
Pl. XL, figs. 66—78. These figures do not exclude the possi- 
bility of seeing in the anterior part of his “ Chordaplatte” a 
portion homologous to our hypoblastic protochordal plate, 
although, of course, Rabl’s text goes the other way, and he 
wishes to interpret them as confirmative evidence for van 
Beneden’s gastrulation hypothesis. 


STUDIES IN MAMMALIAN EMBRYOLOGY. 545 


2. The Protochordal Plate, Protochordal Wedge, and Annular 
Zone of Modified Hypoblast in other Mammals, according 
to Bonnet, van Beneden, Selenka, Fleischmann, a. o. 


In Bonnet’s description of the developmental phenomena of 
the sheep very numerous points of coincidence with what has 
here been described for the shrew have already been repeatedly 
pointed out. Bonnet emphatically insists on the participation 
of an anterior (ab origine hypoblastic) chorda-entoblast 
towards the formation of notochord and mesoblast, in which 
further backwards the “ Kopffortsatz” (our protochordal 
wedge) and gastrula ridge participate. His statements are 
all the more valuable as in his later paper (I. c., 1889) he 
recognises (p. 75) that the different results at which other 
authors have arrived have led him to most careful and repeated 
reperusal of his series of sections, and have changed his mind 
in this sense, that in a former publication (1. c., 1884) he gave 
too prominent a place to the anterior hypoblastic plate in the 
formation of the notochord. He is now willing to accord a 
much more considerable part to the “‘ Kopffortsatz,’”’ of which 
he had formerly underrated the length ; but he holds on as 
strongly as ever to the participation of a purely hypoblastic 
anterior portion. It will be best to quote his own words, 
which are nearly verbally applicable to the shrew. He says 
(1.c., 1889, p. 84): 

“The definite formation of the notochord is only brought 
about when the gutter-like or flat ‘ Kopffortsatz’ that has become 
intercalated in the entoblast is again pinched off longitudi- 
nally ; only now we may, rigidly speaking, apply the terms 
‘notochord’ and ‘ notochordal lumen,’ in so far as by this 
latter name one would wish to designate remains of the folded- 
off enteric cavity in the segregated notochord. In the genesis 
of the notochord we must thus very strictly keep apart the ori- 
ginally solid but subsequently canalised Kopffortsatz, its ventral 
fusion by which it opens and becomes gutter- or plate-shaped, 
and intimately connected with the enteric hypoblast. Poste- 
riorly the ‘Kopffortsatz’ passes into the gastrula ridge, 


546 A. A. W. HUBRECHT. 


anteriorly into the ‘chorda-entoblast’ without any strict 
boundary ; all this together represents the materia] from which 
the notochord will originate. But only when the chorda- 
entoblast and the intercalated ‘ Kopffortsatz’ (intercalated in 
the hypoblast) have again become pinched off—only then the 
definite notochord has originated. It is only the chorda- 
entoblast at the cranial extremity and the gastrula ridge con- 
tribution at the caudal extremity which, as follows from this 
description, is directly converted into notochord. By further 
differentiations in gastrula ridge and terminal knob the noto- 
chord extends further backwards; by further processes of 
growth in the split-off chorda-entoblast it extends further 
forwards.” 

The points of comparison between shrew and sheep will be 
self-evident for whomsoever compares certain figures from 
Bonnet’s earlier paper (‘ Arch. f. Anat. u. Physiol.,’ 1884, Anat. 
Abth., pls. ix—xi) with those here given for the shrew,! more 
especially— 


Bonnet’s fig. 30 ; 3 : with our figs. 45, 46 
sy a rolae 5 : 5 » 43, 44 
oe) » 29 C : : ” » 48 
aie PRRs gic oe noattheeelae » 9 49, 50 
a) Leteetb- ad. belane Fo eee 
Se eoROO Ol cal die > soit Hanktt Os Me 
POAT ADS eget te 5. dh Ons 
” 9 59, 60 . ° . 23 ”? 74, 76 


Finally I may draw attention to a fact which I am inclined 
to attach importance to, viz. that Bonnet describes and figures 
in the anterior portion of the sheep’s protochordal plate 
numerous downward proliferations of this hypoblastic tissue, 


1 [hardly think Bonnet is justified in expressing doubt (p. 66, 1. ¢., 1889) 
as to whether the median, anterior thickening of the hypoblast, noticed in 
early stages in front of the Kopffortsatz, stands in any relation to the forma- 
tion of the notochord. JI must acknowledge that I cannot succeed in 
connecting this statement with Bonnet’s description of his ‘‘ chorda-ento- 
blast,” which somewhat later occupies the same position (our protochordal 
plate), and which would naturally be looked upon as a further thickening of 
the region already similarly recognisable in the earlier stages. 


STUDIES IN MAMMALIAN EMBRYOLOGY. 547 


protruding into the cavity of the fore-gut and marked sp 
(Entoblast sprossen) in his figs. 53 and 54. It is with these 
figures that I am inclined to compare the woodcuts which 
van Beneden has given on pp. 712 and 7138 of vol. iii of 
the ‘Anatomischer Anzeiger,’ for Vespertilio murinus. 
In these he figures cellular matter (marked B on p. 718), 
which at the foremost extremity of these early embryos is 
meant by him to stand for the bottom of the “‘ Chorda-kanal,” 
i.e. belonging to the forward growth from the node of Hensen, 
which we have termed the protochordal wedge (Kopffortsatz, 
v. Beneden, a. 0.). 

In this forward growth the ‘ Chorda-kanal”? makes its 
appearance, and as the Kopftortsatz is intercalated in the 
hypoblast, the “canal” fuses (beginning in the middle, and 
from thence both backwards and forwards) with the lumen of 
the yolk-sac. This intercalating process can be noticed, 
according to van Beneden, even under the originating pros- 
encephalon. It is with respect to this point that I must 
again refer to what I have noticed above when discussing 
Carius’ and Keibel’s description of the formation of “ Kopf- 
fortsatz’’ and notochord in Cavia and rabbit. 

Here, too, a very rigid inquiry will have to test the facts as 
stated by van Beneden. The possibility of a pre-existent 
hypoblastic protochordal plate must be rigorously excluded 
experimentally before we are justified in giving up the possi- 
bility that to a certain extent, however much reduced, there 
might be coincidence between the mammals which Bonnet, 
Heape, and myself have investigated, and those which have 
served for van Beneden’s, Carius’, Keibel’s, a. 0. researches. 
As stated above, the woodcuts furnish a starting-point for such 
comparison, which, however, I only bring forward in the very 
tentative manner here explained. 

In the opossum the earliest developmental phenomena of 
the notochord were studied by Selenka, and although in his 
text (‘ Das Opossum,’ p. 152) he declares himself an adherent 
of Kolliker’s theoretical views, his figures certainly admit of a 
different interpretation, which if verified would bring the 

VOL. XXXI, PART IV.—NEW SER. 00 


548 A. A. W. HUBRECHT. 


facts in the opossum on a level with Bonnet’s and my own 
results in placental mammals. 

Thus on Selenka’s fig. 2, pl. xxi, the front end of the 
notochord in its very first stages is figured. A thin layer 
with nuclei beneath it should be erased in this figure, as we 
are told on p. 151 of the text. When this erasion has been 
brought about, the figure very strongly suggests the identity 
of the massive plate of 7—10 thick entodermal cells, with the 
protochordal plate of our figs. 66 and 67. It seems to me 
hardly possible here to adopt van Beneden’s or Carius’ views, 
and look upon this portion as the widened anterior part— 
intercalated in the hypoblast—of the “ Kopffortsatz.” The 
front end of the latter is perhaps situated in Selenka’s fig. 4 
(or 3). If we now consult his surface view (fig. 1, pl. xxi), 
we see that just in the level of section fig. 4 the notochord is 
very considerably constricted. Later researches will have to 
make out whether the two regions, anterior and posterior to 
this spot, may be identified respectively as protochordal 
plate and protochordal wedge. Of the very earliest phases 
of the gastrula ridge Selenka does not give any figures, 
nor do his figs. 8—1], pl. xviii, furnish material for a 
profitable comparison with those very earliest stages in 
Sorex. 

A very emphatic opponent to some of Bonnet’s views, par- 
ticularly those which refer to the formation of an annular zone 
of hypoblast, from which mesoblast (on purpose I do not follow 
Bonnet in the use of the term mesenchyme) is originated, is 
found in Fleischmann, who has more especially occupied himself 
with the development of Carnivora. As we have been able to 
show that in the shrew Bonnet’s results are fully substantiated, 
and to bring forward additional evidence in karyolitic figures 
which are not figured by Bonnet, we necessarily find ourselves 
in conflict with Fleischmann. This author (‘ Unters. iiber 
- einh. Raubthiere, 1889, p. 17) does acknowledge the possi- 
bility that in this respect fundamental differences between the 
Carnivora and the mammals which Bonnet and myself have 
studied exist. Still his a priori argumentation against the 


STUDIES IN MAMMALIAN EMBRYOLOGY. 549 


existence of an annular peripheral mesoblast, directly deriving 
from hypoblast-cells, is very sweeping. 

Another detail in Fleischmann’s paper (p. 11) to which I 
will direct attention concerns a difference which he has 
noticed in the staining properties of splanchnic and of somatic 
mesoblast-cells in the region of the primitive streak. <A 
similar phenomenon in the cat was observed by Strahl 
(‘ Arch. f. Anat. u. Phys.,? Anat. Abth., 1886, p. 160) for the 
rabbit, and has been referred to above by myself (p. 514), and 
figured on Pl. XLI, figs. 83 and 91. Certain preparations 
(a. 0., figs. 56 and 59) would lead me to believe that there is 
a more direct connection between this portion of the somatic 
mesoblast and the superficial layers of the gastrula ridge. 
Whether the phenomenon has a higher theoretical significance 
can for the present not be answered ; attention should, how- 
ever, be directed towards it. In Selenka’s sections of the 
opossum a somewhat similar phenomenon—not, however, 
noticed in the text as being marked by different staining pro- 
perties—is figured on Pl. X XI, fig. 7, sm. and splm. 

One other point in Fleischmann’s paper should be reflected 
upon, viz. whether his fig. 10, pl. i, in which three layers are 
figured in a section just in front of the medullary plate, and in 
which the hypoblast is uncommonly thick and proliferating, 
might not after all be interpreted as belonging to a region 
homologous to that of the protochordal plate in the shrew, or 
of Bonnet’s chorda-entoblast. 


3. Sundry Observations by Different Authors in Support of the 
Hypothesis of Precocious Segregation of Part of the 
Hypoblast in Mammals. 


The views here brought forward with respect to the gastru- 
lation of the Mammalia, which have been fully discussed 
above, receive independent corroboration from certain facts 
already passingly alluded to, which were noticed by other 
observers and by myself in other mammals. I wish to insist 
somewhat more fully on these observations. 


550 A. A. W. HUBRECHT. 


In the opossum, Selenka (1. c., p. 114) describes and figures 
avery early blastula stage in which one or a few cells get 
separated from the outer blastula wall, wander into the archi- 
coel (segmentation cavity), and form the starting-point for 
the development of the hypoblastic sac. In the beginning 
(Selenka’s figs. 3, 8, and 11, pl. xvii; 2, 3, and 4, pl. xviii) 
these rapidly multiplying cells adhere to the outer blastula wall 
in a spot which Selenka calls the blastopore. When the for- 
mation of the hypoblastic sac is completed there is a stage in 
the development of the opossum’s didermic blastocyst when 
the two layers are no longer thus primarily con- 
nected, but are separate though contiguous spherical 
sacs, the hypoblast consisting of flattened, the epiblast of more 
massive cells (Selenka’s figs. 2, 3, 6, pl. xix). 

Only in a further stage—immediately following on the fore- 
going—the palingenetic hypoblast (gastrula ridge and proto- 
chordal wedge) makes its appearance; the fusion which 
in. consequence of this is established between epiblast and 
hypoblast stands in no direct relation to the earlier con- 
nection above alluded to, and which is noticed in the blastula 
stage. 

There can thus be no doubt that the wall of the monodermic 
blastocyst of the opossum produces cell-material, which arranges 
itself into a second closed sac inside the primary one. The 
spot from whence this cell production emanates is distinct 
and circumscribed, it disappears without leaving any trace 
when the didermic blastocyst is fully established. Shortly 
afterwards there appears in about the same region of the 
outer layer a renewed cell proliferation which leads to the 
formation of the gastrula ridge. 

The latter is what I have proposed to call the palingenetic, 
the former the cenogenetic hypoblast. In the opossum the 
clearness of the phenomenon leaves nothing to be desired. It 
is not a priori necessary that the spot from whence the ceno- 
genetic hypoblast originates should be the same as that where 
afterwards the palingenetic blastopore (gastrula ridge, primi- 
tive streak) makes its appearance. That it appears to do so 


STUDIES IN MAMMALIAN EMBRYOLOGY. Sore 


more or less in the opossum is nevertheless in favour of the 
hypothesis of precocious segregation here brought forward. 

It remains doubtful whether it will be desirable to designate 
the first-named spot as Selenka does by the name of “ blasto- 
pore.’ An additional adjective will render good service, and 
the name “ cenogenetic blastopore ” might recommend itself. 

Of the eight embryos in this stage which Selenka has 
examined, five show this cenogenetic blastopore as a closed 
proliferative spot; in three of them small openings (Zellen- 
liicken) were noticed! in the wall of the blastocyst, which soon 
closed up. In the hedgehog I have myself described (‘ Anat. 
Anzeiger,’ ili, pp. 511 and 907; ‘ Quart. Journ. Mier. Sci.,’ 
1889, p. 286) a monodermic blastocyst, with hypoblast-cells 
adhering to the wall at a spot which I think we are justified 
in comparing to the opossum’s cenogenetic blastopore. The 
comparison of still earlier stages of the hedgehog is as yet 
a desideratum in order to definitely sanction this comparison. 
Here, too, the faint trace of the spot where the cenogenetic 
hypoblast probably originates out of the wall of the mono- 
dermic blastocyst is obliterated ; here, too, a didermic stage 
follows with the two layers separate; and here, too, the 
connection between epi- and hypoblast which results from the 
formation of the palingenetic blastopore is only a third phase 
in this developmental process. 

For the mole we have Heape’s observations, according to 
which (Il. c., pl. ix, figs. 17—27) the hypoblast arises much 
in the same way as we have noticed in the shrew. It is hardly 
possible to indicate a spot from whence the cenogenetic hypo- 
blast can more particularly be seen to take its origin. And 
the perforation which Heape indicates in his fig. 31 would 
seem to me not to be homologous to the spot above noticed 


1 Selenka’s figures of the blastocysts, in this and the following stages, are 
drawn on very varied magnifying scales. This should be borne in mind in 
comparing his figures with each other. It is apt to create some confusion, 
from which in this paper I have endeavoured to keep clear by adhering very 
strictly to the use of very few and always identical scales of enlargement for 
the different_figures. 


552 A. A. W. HUBRECHT. 


for the opossum, nor for that observed by Keibel in the rabbit, 
and to which I will refer lower down, but rather to the less 
markedly perforated hinder region of the embryonic shield of 
Sorex figured in Pl. XX XVIII, fig. 39. It should be especially 
kept in mind that Heape’s fig. 31 has been preceded by another 
stage (his figs. 10 and 30), in which the hypoblast was already 
completed and an independent closed sac inside the epiblastic 
vesicle, so that their fusion and the perforation represented in 
fig. 31 are no doubt already a secondary connection, as is the 
fusion of the figs. 37 and38o0f Sorex. This early perforation 
before the protochordal wedge has as yet very far advanced is 
no doubt a peculiar feature which should be submitted to 
repeated observation, but it is in no way an obstacle to the 
theoretic views here given concerning the fusion of a palin- 
genetic and cenogenetic hypoblast. Moreover another dif- 
ference between Talpa and Sorex runs parallel to it, viz. the 
greater distinctness and width of the neurenteric perforation 
(see about this p. 532). 

Finally, an observation of Keibel in the rabbit deserves 
more particular consideration with respect to this matter. 
He represents (17) in his section, figs. 46a, 460, and 47, a 
stage which very nearly approaches what Selenka figures in the 
opossum. Keibel draws attention to this correspondence, as 
also to that between these stages and that of the mole according 
to Heape. I have above given my reasons for which in this 
latter respect I am inclined to differ from his interpretation ; 
but I think we are fully justified in looking upon the spot BP 
in Keibel’s figures as the true cenogenetic blastopore of the 
rabbit, homologous to that of the opossum. 

From Keibel’s text I gather with satisfaction, as I have 
already had occasion to notice above (p. 520, foot-note), that 
he is not adverse to the view which I first put forward two 
years ago, and have more fully advocated in this paper, viz. 
that the formation of the mammalian hypoblast (i.e. the _ 
phenomenon of gastrulation) is accomplished in two stages 
separated by a short interval. I cannot, however, acquiesce 
to his suggestion (I. ¢., p. 876) that in the first of these two 


STUDIES IN MAMMALIAN EMBRYOLOGY. 553 


phases the enteric hypoblast (Gdtte’s enteroderm) is formed, 
whereas in the second the formation of notochord and mesoderm 
is accomplished. I have shown in the foregoing pages that the 
hypoblast which belongs to the first phase (cenogenetic hypo- 
blast) has undoubtedly its share in the formation of both 
notochord and mesoblast, and I have at the same time 
attempted more definitely to determine the exact amount of 
this share, and also that of the palingenetic hypoblast. 


12. 


List or AUTHORS REFERRED TO. 


. BENEDEN, H. vay.— Recherches sur l’embryologie des Mammiferes. 


La formation des feuillets chez le lapin,’ ‘Archives de Biologie,’ 
1880, t. i. 


. BenEpEN, HE. van.—‘ Erste Entw. Stad. von Saugethieren,” ‘ Tageblatt 


der 59 Versamml. deutscher Naturf. u. Aertzte, Berlin,’ 1886, p. 374. 


. Benepen, HE. van.—‘‘ Unters. iiber die Blatterbildung, den Chordakanal 


u. die Gastrulation bei den Saugethieren,” ‘Anat. Anzeiger,’ vol. iil, 
p. 709. 


. BENEDEN, E. vAN, ET JuLIN, C.— Observations sur la maturation, &c., 


de l’ceuf chez les Cheiroptéres,’”’ ‘ Arch. de Biologie,’ 1880, t. i. 


. Bonnet, R.—“ Beitrige zur Embryologie der Wiederkéuer gewonnen 


am Schafei.” I. ‘Arch. f. Anat. u. Physiol.,’? 1884, Anat. Abth. 
II. ‘ Arch. f. Anat. u. Physiol.,’ 1889, Anat. Abth. 


. Bonnet, R.—‘‘ Ueber die Entw. des Allantois und die Bildung des 


Afters, &.,” ‘ Anat. Anzeiger,’ 1888, S. 105. 


. Cantus, F.—‘ Ueber die Entwickelung der Chorda und der primitiven 


Rachenhaut bei Meerschweinchen und Kaninchen.’ Dissertation. 
Marburg, 1888. 


. Fierscumann, A.—‘ Embryologische Untersuchungen.’ 1 Heft. “ Unter- 


suchungen iiber einheimische Raubthiere,” mit 5 Taf., Wiesbaden, 
1889. 


. Heare,W.— The Development of the Mole (Talpa europea),” ‘Quart. 


Journ. Mier, Sci.,’? 1883. 


. Hears, W.—‘ The Development of the Mole,” ibid., 1886. 
LL 


HeEnsen.—* Beobachtungen iiber die Befruchtung und Entw. des Kanin- 
chens und Meerschweinchens,” ‘Arch. fiir Anat. u. Phys.,’ 1876, 
Anat. Abth. 

Hertwic, O.— Die Entw. des mittleren Keimblattes der Wirbelthiere,’ 
Jena, 1883. 


554 A. A. W. HUBRECHT. 


13, Horrmayy, C. K.— Die Bildung des Mesoderms, die Anlage der Chorda 
dorsalis, und die Entwickelung des Canalis neurentericus bei Vogel- 
embryonen,” ‘Verhandl. d. kgl. Akad. d. Wissensch. Amsterdam,’ 
1883. 

14. Husrecut, A, A. W.—“ Keimblatter-bildung und Placentation des 
Tgels,” ‘ Anatomischer Anzeiger,’ 1888, S. 510. 

15. Husrecut, A. A. W.—“ Die erste Anlage des Hypoblastes bei den 
Saugethieren,” ‘ Anat. Anzeiger,’ 1888, S. 906. 

16. Husrecut, A. A. W.—“ Contributions to the Embryology of the Mam- 
malia.” I. “The Placentation of Erinaceus europeus,” ‘Quart, 
Journ. Mier. Sci.,’ December, 1889. 

17. Kerset, F.—“ Zur Entwickelungsgeschichte der Chorda bei den Siu- 
gern,” mit 4 Tafeln, ‘Arch. f. Anat. u. Physiol.” 1889, Anat. 
Abth. 

18. Kontixer, A. v.—‘‘ Entwickelungsgeschichte des Menschen und der 
hoheren Thiere,” Leipzig, 1879. 

19. Kouiiker, A. v.— Die Entwickelung der Keimblatter des Kaninchens,” 
‘Festschrift fiir Wiirzburg,’ Leipzig, 1882. 

20. Kuprrer, C.— Die Gastrulation an den meroblastischen Hiern der Wir- 
belthiere und die Bedeutung des Primitivstreifs,” ‘ Archiv. f. Anat. u. 
Physiol.,’ 1882 u. 1884. 

21. Linserkt'un.— Ueber die Keimblatter der Saugethiere,” ‘ Gratulations- 
schrift an Nasse,’ 1879. 

22. Nurt, P. J.— Recherches sur le développement du Petromyzon 
planeri,” ‘ Archives de Biologie,’ ii, p. 403. 

23. Rast, C.—“ Ueber die Bildung des Mesoderms,” ‘Anat. Anz.,’ 1888, 
S. 654. 

24, Raspi, C.—“ Theorie des Mesoderms,’? ‘Morphologisches Jahrbuch,’ 
vol. xv, p. 113. 

25. Scort anp Osporn.—‘On some Points in the Early Development of 
the Common Newt,”’ ‘ Studies from the Morpholog. Laboratory in the 
University of Cambridge,’ 1880. 

25a. Scorr, W. B.—Beitrage zur Entwickelungsgeschichte der Petromy- 
zonten,’’ ‘ Morphologisches Jahrbuch,’ Bd. vii, p. 101. 

26. ScHuttzr, O.— Die Entw. der Keimblitter und der Chorda dorsalis 
von Rana fusca,” ‘ Zeitschr. f, wissensch. Zool.,’? Bd. xlvii, 8. 325. 

26a. Scuutrzn, O.—* Zur ersten Entwickelung des braunen Grasfrosches,” 
‘ Festschrift fiir A. vy. Kolliker,’ 1887. 

27, SELENKA.—‘ Studien iiber Entwickelungsgeschichte d. Thiere,’ Wies- 
badeu. Heft, iv and v, ‘ Das Opossum.” 

28. SHipLEy.—* On some Points in the Development of Petromyzon 
fluviatilis,” ‘Quart. Journ. Micr. Sci.,’ vol. xxvii, 1887, p. 325. 

29. STRAHL UND Canius.—“ Untersuchungen iiber den Kopffortsatz des 
Kaninchens,” ‘ Marburger Sitzungsberichte,’ 1887. 


STUDIES IN MAMMALIAN EMBRYOLOGY. 555 


EXPLANATION OF PLATES XXXVI—XLII, 


Illustrating Prof. A. A. W. Hubrecht’s paper “Studies in 
Mammalian Embryology. II.—The Development of the 
Germinal Layers of Sorex vulgaris.” 


PLATE XXXVI. 


Figs. 1—4.—(The numbers in brackets refer to the catalogue number of 
the uterus from which the embryos were taken.) Four uteri of Sorex con- 
taining the earliest embryonic stages that were investigated. The swellings 
visible in 16 are not caused by the enclosed embryos, but are the last visible 
remnants of a preceding pregnancy, and mark the are of placentation in 
their latest retrogressive phase. 

These figures are given in natural size. The unpaired median portion is 
the vaginal part in which embryos are never found. 

Fies. 5—7.—Transverse sections of three young blastocysts of No. 16. 
Z. Zona pellucida. Zr. Trophoblast. #. The inner cell-mass which will 
give rise to epiblast and hypoblast of the embryo. 

In Fig. 5 there is a difference in size between certain cells of the inner 
mass, which was not noticed in Figs. 6 and 7. In all there is sufficient 
evidence of trophoblast-cells between the inner mass and the zona 
pellucida, The total number of cells constituting these blastocysts 
varies from fifty to sixty. The figures were drawn (as are all the 
other figures of sections on this plate) with Zeiss’ apochr., oc. 4, obj. 4. 

Fig. 5.—Mus. Utr. Cat. n* Sorex 16 a and 4, 37. 98.1 

Fig. 6.— 55 is 16 cg, 27.1458. 

Fig. 7.— a x 16 cg, 27.145. 

Fie. 8.—Section through an embryo of No. 44. Lettering as in the pre- 
ceding figures. There is an artificial bulging in on one side of the blastocyst, 
which was absent in other embryos of the same age, and which must un- 
doubtedly be ascribed to the thinning out of the zona simultaneously with the 
increase in size of the blastocyst. It must be ascribed to the reagents used 
in hardening the tissues. 

It will be seen from the subsequent figures on this and the following plates 
that this phenomenon is very much on the increase as the size still further 
augments, and is only counteracted when once the blastocyst has become 
attached to the walls of the uterus (see Fig. 12). 

Mus. Utr. Cat. n” Sorex 44a, 6,47.11s. 


' See this Journal, vol. xxx, p. 393. 


556 A. A. W. HUBRECHT. 


Fre. 9.—A similar section through an embryo of No. 105, the zona being 
much folded. 
Mus. Utr. Cat. n* Sorex 105,27. 6s. 


Fies. 10 and 11.—Section through two other blastocysts belonging to 
No. 110. Lettering as in the foregoing. 
Mus. Utr. Cat. n®” 5, Sorex 110,27. 8 and 10s. 


PLATE XXXVII. 


Fic. 12.—An early blastocyst of Sorex in situ in the expanded portion 
of the uterine lumen. The section is transverse to the axis of the uterus. A 
small swelling was externally visible (see Fig. 20, which represents the same 
uterus, natural size, as drawn from the spirit specimen). w. e. The uterine 
epithelium. 4/. Cavity of the blastocyst, already adhering by a small portion 
of its walls against the surfaces of the mucosa. The embryonic shield is seen 
to occupy the topmost surface, and to be turned towards the anti-mesometrical 
concavity of the uterine swelling. In this stage the hypoblast forms already 
a completely closed sac, constituting the inner layer of the didermic vesicular 
blastocyst. 


Fies. 13—15.—Three outline sketches in natural size of the uteri (spirit 
specimens) from which the embryos were taken, of which sections are figured 
on this plate. 

Cat. n°® 2, 65, 52. 


Fries. 16—21.—Surface views of the embryonic area reconstructed from 
the series of sections through the early stages that were obtained from the 
uteri of the preceding figures. 

These surface views, and similarly all those that are represented on the 
following plates, were obtained from camera lucida outlines that had been 
drawn with Zeiss’ apochromatic objective 16 mm., oc. 4. They are thus 
about sixty-two times enlarged. 

The outline given is not that of the blastocyst, but of the embryonic shield 
on the top of it. 

The cross arrows indicate the exact extent and direction of the sections 
that are figured on this plate ; the small numbers beside the arrows refer to 
the figure which represents the section. 

In Figs. 18, 19, and 21 a distinct protochordal plate was present in the 
hypoblast ; its outline, as underlying the anterior part of the embryonic 
shield, is indicated by a grey space. 

In Fig. 20 all the nuclei that are present in the hypoblast below the 
epiblastic shield have been figured. It is here seen how they are much 
more closely packed at the anterior end, i.e. in the region which has 
here been termed the protochordal plate. 


STUDIES IN MAMMALIAN EMBRYOLOGY. 557 


Fic, 22.—Trausverse section of a blastocyst (cf. Fig. 16) in which the 
hypoblast is just being formed by cells detaching themselves from the polar 
thickening (cf. Figs. 10 and 11), which has become more flattened. 

The zona is much thinner and more considerably folded. 

Lettering and magnifying scale as in Figs. 5—11. 

Mus. Utr. Cat. n® Sorex 122,2,17.11s. 


Fic. 23.—Part of the folded zona with trophoblast- and hypoblast-cells of a 
similar embryo. The hypoblastic nuclei, Ay., are distinctly larger sized than 
those of the trophoblast, ¢7. 

Mus. Utr. Cat. n° Sorex 122, 2,1 7.17s. 

Fic. 24.—The embryonic shield of another blastocyst in the same stage. 

Distinct flattened hypoblast-cells are present below the epiblastic shield, Z. 
Mus. Utr. Cat. n° Sorex 122,27. 7s. 


Fic. 25.—Blastocyst of a later stage, having become didermic in its whole 
extent. At the spots ¢r’. the trophoblast is thickened preliminary to the 
attachment of the blastocyst to the uterine mucosa (Zeiss’ apochr., oc. 4, 
obj. 16). 

Mus. Utr. Cat. n° Sorex 24,17. 30s. 

Fic. 26.—The epiblastic shield of the same embryo (cf. Fig. 17) more 
considerably enlarged (Zeiss’ apochr., obj. 4, oc. 4). The zona, z., has become 
extremely attenuated. ¢7’. Trophoblast-cells above the epiblastic embryonic 
shield (cf. Figs. 5 and 7). In later phases these have entirely disappeared. 

Mus. Utr. Cat. n° Sorex 24,17, 29 s. 


Fie. 27.—The modified trophoblastic zone, ¢7”., of Fig. 25, more consider- 
ably enlarged (Zeiss’ apochr., oc. 4, obj. 4). 
Mus. Utr. Cat. n° Sorex 24,17. 30s. 


Fics. 28 and 29.—Two sections through another didermic blastocyst, of 
which the former passes through the hypoblastic protochordal plate, pp. (cf. 
Fig. 18). 

Mus. Utr. Cat. n° Sorex 527,47.11s. 
47.3 5S. 


Ee) 22 ” 


Fics. 30 and 31.—Two sections through yet another didermic blastocyst 
(cf. Fig. 19). The protochordal plate is cut along its longest diameter. 
Mus. Utr. Cat. n° Sorex 52 ¢,37. 26s, 
4r.8 5. 


bed 3 bP) 


558 A. A. W. HUBRECHT. 


PLATE XXXVIII. 


Fies. 32—35.—The embryonic shield of four blastocysts that were taken 
from the same mother (Cat. n* 73). These tracings were obtained by the 
reconstruction of the details of structure in the different regions of the 
embryonic shield from the unbroken series of sections. The epiblastic shield is 
marked by a dark outline. The annular ring of modified hypoblast has a 
uniform grey tint, and so has the protochordal plate, which was also already 
indicated on the foregoing plate. 

The gastrula-ridge makes its first appearance in Fig. 32; in the other 
figures it is much farther advanced, and traces of a gastrula-groove are found. 
The regions right and left of the gastrula-ridge, as well as behind it, contain 
the continuous plate of mesoblast, of which the outer boundary is indicated 
by a dotted line. The arrows have reference’ to the different sections as 
figured on this and the next plates; the numbers attached to them are the 
same as the respective numbers of the figures in which those sections are 
represented. If the sections do not extend through the whole length of the 
embryonic region, the arrows are correspondingly shortened. 

In Fig. 34 the front part of the annular ring of modified hypoblast is 
omitted, because this portion being bent in the plane of the section 
could not allow its outlines to be rigorously traced in the reconstruc- 
tion process, 

In Fig. 32, although obtained from the same uterus, no annular zone of 
modified hypoblast was as yet present. This zone may consequently 
be said to appear as a simultaneous modification of the region in 
question. 


Fies. 86—39.—Four sections through the posterior region and first origin 
of the gastrula-ridge of the embryonic shield of Fig. 32. In four sections, 
two of which are figured, there is adhesion between the proliferating epiblast 
(palingenetic hypoblast) and the subjacent layer of flat cells (ccenogenetic hypo- 
blast). At the front end of the incipient gastrula-ridge there are faint traces 
of a perforation indicated by p. in Figs. 38 and 39. pw. Cell-mass in front of 
this: incipient protochordal wedge (cf. Figs. 46—48, and 55, pw., of slightly 
older stages). gv. Proliferating cell-mass behind it: incipient gastrula-ridge. 

Fig. 36.—Mus. Utr. Cat. n° Sorex 73 f,27. 15s. 


Fig. 37.— By af 5 17 s. 
Fig. 38.— . 5 ss 19 s. 
Fig. 39.— “5 es = 20 s. 


Figs. 40 and 41.—Two sections through the anterior region of the embry- 
onic shield of Fig. 32. pp, The hypoblastie protochordal plate, already more 
than one cell thick. 

Fig. 40.—Mus. Utr. Cat. n* Sorex 73 f, 27.19 s. 
Fig. 41.— e ore: 


STUDIES IN MAMMALIAN EMBRYOLOGY. 559 


PLATE XXXIX. 


Fie. 42.—The uterus (Cat. n* Sorex 73) from which the blastocysts of 
this and the foregoing plate were taken. Natural size; drawn from the 
spirit specimen. 

Fies, 43—52.—Ten sections (Zeiss’ apochr., oc. 4, obj. 4) through dif- 
ferent regions of the embryonic shield of Fig. 34. The exact situation and 
extent of the sections are marked in that figure by numbered arrows. 

The protochordal plate, pp., is cut in Figs. 43 and 44, and is giving origin 
to mesoblast-cells. 

Mus. Utr. Cat. n® Sorex 73 d,27.19s, 

a op A op PS 

In Figs. 45 and 46 there is as yet no mesoblast laterally. The median 
cells, pw., belong to the foremost extension of the protochordal wedge, 
which is not yet confluent with the protochordal plate. A thickened 
keel of epiblast is present in the median line, 

Mus. Utr. Cat. n® Sorex 73 d, 27. 305. 

55 - op 3r1s, 

In Fig. 47 the protochordal wedge, pw., is connected with lateral meso- 
blast plates ; the mesoblast is also present in the annular zone of modi- 
fied hypoblast, hy. az. 

Mus. Utr. Cat. n* Sorex 73. d,3 7.4. 


In Figs. 48—50 the unfolding of the lips of the palingenetic blastopore 
(front end of the gastrula-ridge) is especially distinct and expressed 
even in the position of the cells. 

In Fig. 50 a faint trace of a partial perforation is observed in the median 


line. 
Fig. 48.—Mus. Utr. Cat. n° 73d, 37.6 s. 
Fig. 49.— A . re 
Fig. 50.— 3 ee eet ara 


Figs. 51 and 52 are sections further backwards through the gastrula- 
ridge and its lateral plates of mesoblast. 
Mus. Utr. Cat. n* 73 d, 3 7. 15 s. 
is op » 2058, 

Figs. 53—55.—Three sections through the whole extent of embryonic 
shield and annular zone of modified hypoblast (Ay. az.) of the blastocyst of 
Fig. 33 (Zeiss’ apochr., oc. 4, obj. 8). gp. Protochordal plate. Z. Hpiblastic 
shield. Ay. Unmodified flattened hypoblast both inside and outside the ring 
above mentioned. ¢7. Trophoblast. pw. Protochordal wedge. gr, Gastrula-ridge. 


Fig. 58.—Mus. Utr. Cat. n* Sorex 73 07,2 7. 20s. 
Fig. 54,— x 3 - 3 LAs: 
Fig. 55.— 3 ne ee tal) ce 


Fie. 56.—Section through the posterior knob of the gastrula-ridge with 


560 A. A. W. HUBRECHT. 


vhe mesoblast stretching backwards. The future somatic mesoblast-cells 
(also in Figs. 57—60) are distinguished by their more flattened shape and 
stronger absorption of staining reagents. 

Mus, Utr. Cat. n° Sorex 75 62,1 7.14, 15 s. 


Fies. 57 and 58.—Two sections through mesoblast that has originated 
partly from the gastrula-ridge, partly from the annular zone of hypoblast ; 
karyolytic figures determining the details of the latter process are very 
distinct. 

Fig. 57.—Mus. Utr. Cat. n* Sorex 73 0}, 47, 21s, 
Fig. 58.— > 9 + 17s. 

Fies. 59—61.—Three sections through other blastocysts in which further 
confirmation of these karyolytic processes is distinct. 

In Fig. 59, @ indicates a spot where the dividing nuclei have just sepa- 
rated and rearranged themselves. 

Fig. 59.—Mus. Utr. Cat. n* Sorex 73 al, 37. 23 s. 

Fig. 60.— BS of = 18 s. 

Fig. 61.— ie 3 de Sai Ded oee, 

(This latter section belongs to a series of which no surface reconstruction 
is given.) 


PLATE XL. 


Fies. 62—64,—Three surface views (obtained by reconstruction from an 
unbroken series of sections drawn with camera) of three embryos from the 
same uterus (No. 45). The mesoblast, of which the outer boundary is indi- 
cated by a dotted line, stretches some distance beyond the epiblastic shield. 
It is in this stage already a continuous plate, only interrupted in the median 
line below the epiblastic shield by the gastrula-ridge (gr.), the protochordal 
wedge (pw.), and the protochordal plate (pp.). 

The arrows indicate the exact situations, and also the extent of the sections 
figured under the corresponding numbers. 


Fie. 65.—The uterus, No. 45, natural size, drawn from the spirit specimen. 


Fries, 66—68.—Three sections through the embryonic region of Fig, 62. 
pp. Protochordal plate. pw. Protochordal wedge. Ay.az. Remnants as yet 
fairly distinct of the annular zone of modified hypoblast of Figs. 33—35. 

Fig. 66.—Mus. Utr. Cat. n°’ Sorex 45 c, 47. 5s. 
Fig. 67.— ” 5B) » 2s, 
Fig. 68.— és Mi » 3718s. 

Fic. 69.—An oblique section through another embryonic shield of the 
same stage. Letters as above. C. Rudiment of a canal in the protochordal 
wedge. 

Mus. Utr. Cat. n* Sorex 45e,47.10s. 


STUDIES IN MAMMALIAN EMBRYOLOGY. 561 


Fies. 70—77.—Kight sections through part of the embryonic shield of 
Fig. 63. Lettering as above. 
Fig. 70.—Mus. Utr. Cat. n* Sorex 45 0,27. 38s. 


Fig. 71.— ra 3 re 25 s. 
Fig. 72.— Be 36 So Mose 
Fig. 73.— a5 53 3 5s. 
Fig. 74,— ” ” 2 7 Ss. 
Fig. 75.— 35 * . 9s. 
Fig. 76.— 33 A A alley 
Fig. 77.— 55 5 17s. 


Fie. 78.—Part of a section through an embryonic shield not here figured. 
Here, again, there is an unmistakable rudiment (c.) of a canal in the proto- 
chordal wedge. 

Mus. Utr. Cat. n* Sorex 45 a, 47. 22 s. 


PLATE XLI. 


Fies. 79—81.—Three surface views of three embryos from the same uterus 
(No. 42), They were obtained in the same way as those of the foregoing 
plates. pp., pw. gr. as in Figs, 62—64. Cel. Regions of the celom. 
(per.: in Fig. 80 first indication of pericardial ccelom, in transverse sections 
in Fig. 88.) 

Fig. 82.—The uterus, No. 42, drawn from the spirit specimens (natural 
size). 

Fies. 83—87.—Five sections through the embryonic shield of Fig. 79. 
The sections are oblique, i.e. asymmetrical. This brings out all the more 
clearly (especially in Fig. 87) the slight but still marked differences between 
the mesoblast that more particularly belongs to the protochordal plate (mes. 
pp.), that which is continuous with the protochordal wedge (mes. pw.), and 
that which belongs to the region of the gastrula-ridge (mes. gr.). Lettering 
as before. gr. and gg. The front end of gastrula-ridge and gastrula-groove 
as seen in oblique sections. (In Figs. 85 and 86, and also in Fig. 90, there 
are rudiments of a protochordal canal inside the protochordal wedge.) 

Fig. 83.—Mus. Utr. Cat. n° Sorex 42 ce, 3 7. 6 d. 


Fig. 84.— x A x Ode 
Fig. 85.— i: ¥ a ea 
Fig. 86.— ws a ulate 
Fig. 87.— < se ain Ss 


Fics. 88 and 91.—Two sections as indicated in Fig. 80. 


Fig. 88.—Mus. Utr. Cat. n* Sorex 42¢,376 5. 
Fig. 91.— ry i - MIS e: 


Fics. 89 and 90.—Idem for Fig. 81. Lettering as above. 
Fig. 89.—Mus. Utr. Cat. n° Sorex 42 6,47. 27s. 
Fig. 90.— + 3 4 22s. 


562 A. A. W. HUBRECHT. 


PLATE XLII. 


Figs. 92—95.—Four diagrammatical figures to illustrate certain theoretical 
speculations developed in the text (p. 533), 

Fig. 92.—An amphibian or cyclostomatous gastrula-stage. 

Fig. 93.—A hypothetical transition between the foregoing and the 
mammalian diagram, fig. 95 (partial perspective). 

Fig. 94.—A sauropsidan gastrula-stage (copied with very slight modifi- 
cations from Rabl). 

Fig. 95.—A mammalian stage corresponding to the phase which is also 
represented in the surface views 62—64 (partial perspective). Grey 
and black: trophoblast, embryonic epiblast. Black with white dots: 
palingenetic hypoblast of the protochordal wedge. Black with parallel 
stripes: palingenetic hypoblast of gastrula-ridge. Blue: ccenogenetic 
hypoblast with protochordal plate (pp.) and modified annular zone 
(hy. az.). 

In Figs, 92 and 93 the distinction between palingenetic and ccno- 
genetic hypoblast is not actually existent, but arbitrarily introduced 
in order to elucidate the facts as presented by 95. 


TERMINATIONS OF NERVES IN TORTOISE-SHELL. 063 


Terminations of Nerves in the Nuclei of the 
Epithelial Cells of Tortoise-shell, 


By 


John Berry Haycraft, M.D., D.Sc., 
From the Physiological Laboratory of the University of Edinburgh. 


With Plate XLIII. 


Tue land tortoise (Testudo greca), so commonly imported 
into England from the south of Europe, appears to be a very 
sluggish animal. This is not really the case, and its move- 
ments on a hot summer day are the reverse of phlegmatic. 
In this condition its carapace is sensitive to the slightest 
impact. Ifthe carapace or plastron be very gently tapped, 
the nearest leg is alone withdrawn, a heavier tap causing a 
withdrawal of its whole body. We have here, therefore, a 
structure which is a true sensitive surface, and like the soft 
skin of a frog or of a man, it is brought into relation- 
ship with the central nervous system. Like the soft skin of 
other animals it may be mapped out into areas, from which 
the nerve-fibres passing to the spinal cord are all especially 
connected with outgoing motor nerves, so that the definite 
reflex movements of limbs as already described may come 
about. 

The above experiment naturally suggested that the sensory 
nerves passed right through the thick bone of the carapace and 
plastron, and ended near the outer surface, either in the 
epithelial tissue of the tortoise-shell itself or in the layer of 
connective tissue which unites it to the subjacent bone. 

After removal of a scute of tortoise-shell the connective 

VOL. XXXI, PART IV.—NEW SER. PP 


564 JOHN BERRY HAYORAFT. 


tissue outside the bone was found, in confirmation of this 
surmise, to contain sensory nerve-fibres, for the application of 
acetic acid, or of an interrupted galvanic stimulus, caused 
definite reflex defensive movements, similar to those which follow 
the application of acid to the frog’s skin. 

Inasmuch, therefore, as sensory nerves evidently end quite 
superficially, it became an interesting question to determine 
their exact mode of termination in the curiously modified tissues 
of the carapace. Portions of the carapace, generally taken 
from the region of the costal plates, were softened in chromic 
acid and nitric acid fluid, frozen, and cut with a thick-bladed 
razor. In this way one can obtain fairly thin sections even of the 
tough tortoise-shell. The sections were treated in various 
ways, with a view of demonstrating nerves or nerve termina- 
tions, and in no case was I able to discover any nervous struc- 
ture in the tortoise-shell itself. 

In the subjacent connective-tissue layer, however, were 
bodies which I, at first sight, thought were end organs (Pl. 
XLITI, fig. 1). They turned out to be the transverse sec- 
tions of curiously modified nerve-fibres. These nerve-fibres 
are easily distinguished from the blood-vessels (which in this 
situation are devoid of a muscular coat) by their solid appear- 
ance (fig. 2), strong connective-tissue covering, and by their 
occasional transverse section, which is very characteristic 
in appearance. 

Fig. 1 represents, in transverse section, two of these fibres 
bound together by a common sheath of connective tissue (H). 
Each fibre consists of an external layer of concentrically 
arranged connective tissue, consisting of laminz of colloid 
granular material with intervening connective-tissue corpuscles 
(m). Within this is acolloid-looking core (Kk), devoid of nuclei, 
and also staining pink with picro-carmine. 

In the centre of the core is generally to be found a small 
spot, probably an axis-cylinder, somewhat differentiated from 
the rest of the core (Gc). 

No trace of medullary matter is to be found in connection 
with any of these nerves, nor are ordinary medullated fibres 


TERMINATIONS OF NERVES IN TORTOISE-SHELL. 565 


to be found in this region; and one is forced to conclude that 
owing to their peculiar situation—perhaps on account of the 
pressure of the hard scutes placed immediately above them— 
the medullated nerves are replaced by axis-cylinders, enclosed 
and protected by the sheaths of modified connective tissue 
just described. 

These modified nerve-fibres can be traced back a little way 
into the bone, and no doubt ultimately pass into the ordinary 
medullated nerves found so plentifully on the inner surface 
of the carapace. Under the scutes they freely branch, 
becoming smaller and smaller, and ultimately terminate in the 
lower epithelial cells of the tortoise-shell. We have, therefore, 
medullated fibres passing from the central nervous system to 
the bone of the carapace and plastron, these then pass into the 
medullary nerve-fibres seen under the scutes, from which, as 
we shall presently see, fine terminal naked axis-cylinders run 
into the scutes. 

The final intra-epidermic termination of the nerves was 
never seen in any of the sections, for the softening of the 
tissue previously to its cutting prevented their subsequent . 
demonstration by staining agents. The nerve endings may, 
however, be demonstrated by another very simple method. 

The scutes from a recently killed tortoise are removed in 
pieces with a sharp scalpel, care being taken to keep attached 
to their under surfaces as much as possible of the subjacent 
connective tissue; and it will be found advisable before doing 
so to remove as much as possible of the dense outer part 
of the scute. In this way one can obtain thin and fairly 
transparent pieces of tissue, consisting of the lower layers of 
the tortoise-shell and the tissue connected with it. These are 
placed in absolute alcohol 2 parts, and distilled water | part, 
and after twelve hours are thoroughly steeped in distilled water 
until every trace of alcohol is removed. The tissues are then 
placed in a solution of hematoxylin until they are sufficiently 
stained ; they may then be mounted in balsam, the connective 
tissue or deeper layer being above the epithelium and next 
the cover-glass. 


566 JOHN BERRY HAYCRAFT. 


Hematoxylin Solution. 
Ammonia alum, 3 grammes ; Pure hematoxylin, 3 grammes ; 
 UDistilled water, 100 c.c. * C Absolute alcohol, 16 c.c. 


Mix A and B, keep in diffuse daylight for two weeks, and dilute with 20 
volumes of distilled water. 


On looking down into the connective tissue with a power of 
three or four hundred diameters, the modified nerve-fibres are 
seen branching in all directions. On deeper focussing the 
lower cells of the epithelium are seen from below. Their out- 
lines are in most situations fairly well seen, and their nuclei 
should be stained with the hematoxylin. 

In these preparations the nuclei frequently shrink within 
the nuclear cavities, appearing as dark blue granular masses 
(B, fig. 4) ; but in most cases they fill the nuclear cavity, and 
their chromatin filaments can clearly be made out. The 
greater number of cells are devoid of any nerves, but here 
and there nerve-fibres may be seen branching again and again 
in the connective tissue, and sending their finest ramifications 
to the nuclei of the epidermic cells. These are what appear to 
be definite sensitive spots where alone the nerves terminate. 
These spots are of variable size, so small as to correspond to 
a space occupied by only a dozen cells, or so large as to occupy 
two or three fields of the microscope. I should say that some 
twenty or thirty of these “spots” might be found on one 
square inch of a costal scute. Between these spots the epithe- 
lium presents, as already observed, nothing very remarkable, 
but within the spot the appearance is very striking. 

The non-medullated fibres deeply stained with the logwood 
divide again and again, sending, in many cases, hundreds of 
fibres to the epithelial cells. Fig. 3 represents a sector of one 
of these spots carefully drawn from a specimen. At the cireum- 
ference (B) the fibres terminate in only a few of the epithelial 
cells, but towards the centre all or nearly all of the cells 
receive fibres. The outlines of the cells are not well marked, 
the fibres at first sight appearing to terminate in little round 
blue masses, which are in reality the nuclei of the cells. 


TERMINATIONS OF NERVES IN TORTOISE-SHELL. 567 


In fig. 4 a very small portion at the outer part of a sensi- 
tive spot is more highly amplified. At the upper part of the 
figure the epithelium outside the spot is seen. Below this the 
terminations of the nerves can readily be made out. They 
certainly pass into the nuclear cavity. Whether they end in 
little flat plates within the nuclear cavity and closely applied 
to the outside of the nucleus, or whether they are prolonged 
into the chromatin of the nucleus, I should not like dog- 
matically to state. I am inclined to believe in the latter view, 
and think it probable that they are continued into true nuclear 
substance. The appearances seen at x HB, fig. 4, are pro- 
bably due to shrinkage as a result of treatment with alcohol ; 
but in c ¢, fig. 4, the nuclear cavity is completely filled by the 
nucleus, all the chromatin substance having apparently gone 
to form the knob or cup at the end of the nerve, leaving the 
rest of the nucleus almost devoid of granular matter, and very 
faintly tinted by the hematoxylin. 

The nerves end in the cells of the rete alone, for it is impossi- 
ble to trace them beyond the deeper layer of the epithelium. 
This is what might be expected, for in the adult tortoise-shell 
the rete consists of one, two, or perhaps three layers of nucle- 
ated rounded cells, and above these, with hardly any transi- 
tional tissue, there are the dense lamin of the horny layer, 
made up of flattened keratinised scales with unstainable nuclei. 

It follows from the foregoing remarks that the scutes of the 
tortoise, in spite of their hard, dense nature, form a very typical 
epidermic sensory covering for the animal. As in the soft 
skin of mammals, the nerves end in localised sensitive spots 
in the epidermis, and before penetrating this tissue they form 
a horizontal plexus in the upper part of the connective tissue. 

The final terminations of nerves in epithelium has received 
much attention from histologists, who have studied this sub- 
ject perhaps most fully in the tadpole’s tail. 

In some situations the nerves appear to run entirely between 
the cells—indeed, this appears to be generally the case (Ran- 
vier, 1; Klein, 2; Eberth, 3; Leboucq, 4). They either end in a 
simple plexus, or terminate in very small knobs or plates, which, 


568 JOHN BERRY HAYORAFT. 


judging from the drawings and preparations I have seen, are 
for the most part much smaller than the chromatin knobs of 
the tortoise. 

But some authors have traced nerves into the epithelial cells 
themselves, where they appear to end in little knobs embedded 
in the cell protoplasm, near but never in the nucleus. Thus 
Pfitzner (5), working with the Amphibia, finds this to be the 
case; and more recently Macallum (6) describes nerves termi- 
nating both between and within the epidermic cells of the 
tadpole’s tail. 

In the tortoise-shell the nerves certainly pass right into the 
nuclear cavity, within which the only structures deeply stained 
by hematoxylin are the club- or cup-shaped masses into which 
the nerves pass. A very remarkable fact is the ease with which 
these preparations are obtained. I have made over twenty, and 
in all cases good demonstrations were obtained. I have tried 
several gold methods, but they were vastly inferior to the log- 
wood, and, as usual, chiefly characterised by want of uniformity 
in the results obtained. In other situations the non-medul- 
lated nerves of the tortoise do not stain at all readily with 
logwood. 


PAPERS QUOTED IN TExtT. 


1. Ranvier.—‘ Traité technique d’histologie,’ p. 900. 

2. Kierm.—‘ Atlas of Histology.’ 

3. EpertH.—‘ Archiv f. mikr. Anat.,’ Bd. ii, p. 490, 1866. 
4. Lrsovce.— Bull. de Acad. Roy. de Belgique,’ 1876. 
5. Prirzner.— Morph. Jahrbuch,’ Bad. vii, p. 726. 

6. Macauium, A. B.—‘ Quart. Journ. Micr. Sci.,’ vol. xvi. 


TERMINATIONS OF NERVES IN TORTOISE-SHELL. 569 


DESCRIPTION OF PLATE XLII, 


Illustrating Dr. John Berry Haycraft’s paper on “ Termina- 
tions of Nerves in the Nuclei of the Epithelial Cells of 
Tortoise-shell.” 


Fig, 1.—Section through the lower part of the tortoise-shell, the sub- 
jacent connective tissue, and part of the bone of a costal plate. 4. Lower 
lamine of horny layer of tortoise-shell. B. Cells of rete, with big nuclei 
and deep ridges passing into subjacent connective tissue. (c) Process— 
profile view of a ridge—of connective tissue running into epithelial cell. 
D. Pigment-cell in tissue around nerve-fibres. H. Sheath common to two 
nerve-fibres. m. Outer covering of granular modified connective tissue. 
K. Inner part of nerve-sheath, consisting of granular non-nucleated connec- 
tive tissue. G. Spot in centre, probably an axis-cylinder. #, Bone. 

Fic. 2.—Longitudinal view of a nerve considerably smaller than the one 
represented in transverse section (Fig. 1). 

Fic. 8.—Part of a sensitive spot; the tortoise-shell is viewed from below. 
x 250. A. Centre of spot. Branching nerve-fibres seen ending in nuclei of 
epithelium. Border of cells not seen because of low power used. B. At 
periphery of spot, where most of the epithelial cells are unconnected with 
nerve-fibres. 

Fic. 4.—Part of the same highly magnified. x 800. .4. Nucleus outside 
sensory spot. It contains chromatin filaments. B. Nucleus that has shrunk 
within nuclear cavity. cc. Nerve-fibre passes into nuclear cavity, and 
apparently ending in the chromatin of the cell, the rest of the nuclear cavity 
being filled with clear, almost colourless material. sxx. Nothing is to be 
seen inside the nuclear cavity except the knobs terminating the nerves. The 
rest of the nucleus has probably shrunk around this. #. Towards centre of 
spot the outlines of the epithelial cells are very indistinct. 


rar 


Ma etn at) 


ales Vie 


PND x On WO ln xo. 


NEW SERIES. 


Acanthodrilus, anal nepridia of, 467 


Amphioxus, development of its atrial _ 


chamber, by Lankester and Willey, 
445 
# excretory tubules of, by 

Weiss, 489 

Arachnids, the origin of vertebrates 
from, by Patten, 317 

Atrial chamber of Amphioxus, its 
development, by Lankester and 
Willey, 445 


Beddard on Deodrilus and on anal 

nephridia in Acanthodrilus, 467 
= on the structure of an earth- 

worm belonging to the genus 
Diacheta, 159 

Benham on the classification of Harth- 
worms, 201 

Bourne, A. G., on Cheetobranchus, a 
new genus of oligochetous Cheto- 
poda, 83 

Buchanan, F., on Hekaterobranchus, 
a new genus of Spionide, 175 

Biitschli’s imitation of protoplasmic 
movement, 99 


Cerata of Nudibranch Mollusca, by 
Herdman, 41 

Chetobranchus, a new genus of 
oligochetous Chetopoda, by A. G. 
Bourne, 83 


| 


Deodrilus, Beddard on, 467 
Diacheta, a species of, by Beddard, 
159 


Earthworm of the genus Diachata, 
by Beddard, 159 


_ Earthworms, an attempt to classify, 


by W. B. Benham, 201 
Embryology of Mammalia, by Hu- 
brecht, 499 
5 of Scorpion, by Laurie, 
105 
Eyes of Arthropoda, morphology of, 
by Watase, 143 


Gaskell on the origin of Vertebrates 
froma Crustacean-like ancestor, 379 


Haycraft on terminations of nerves in 
tortoise-shell, 563 

Hekaterobranchus Shrubsolii, 
anew genus and species of Spionide, 
by F. Buchanan, 175 

Herdman on the cerata of Nudibranch 
Mollusea, 41 

Hubrecht, studies in mammalian em- 
bryology, No. II, 499 


Lankester and Willey on the develop- 
ment of the atrial chamber of 
Amphioxus, 445 

Laurie on the embryology of a Scor- 
pion, 105 


572 


Marshall, C. F., on the histology of | Sorex, development of germinal layers 


striped muscle, 65 
Muscle, histology of 
Marshall, 65 


Nephridia, anal, in Earthworms, 467 

Nerve-terminations in tortoise-shell, 
by Haycraft, 563 

Nudibranch Mollusca, the cerata or 
dorsal papille of, by Herdman, 41 


striped, by 


Oligocheta, a new genus of, by A.G. 
Bourne, 83 


Patten on the origin of Vertebrates 
from Arachnids, 317 

Phymosoma varians, by Shipley, 1 

Porter on the presence of Ranvier’s 
constrictions in the spinal cord of 
vertebrates, 91 

Protoplasmic movement, imitation of, 
by Bitschli, 99 


Ranvier’s constrictions in the spinai 
cord of vertebrates, by Porter, 91 
Scorpion, embryology of, by Laurie,105 

Shipley on Phymosoma, 1 


INDEX. 


of, by Hubrecht, 499 

Spiders, spinning apparatus by, War- 
burton, 29 

Spinal cord, Ranvier’s constrictions 
in, by Porter, 91 

Spionide, a new genus of, by F. 
Buchanan, 175 


Tortoise-shell, nerve-terminations in, 
by Haycraft, 563 


Vertebrates, their origin from a 
Crustacean-like ancestor, by Gas- 
kell, 379 

3 their origin from Arach- 
nids, by William Patten, 317 


Warburton on the spinning apparatus 
of spiders, 29 

Watase on the morphology of the 
compound eyes of Arthropoda, 143 

Weiss on excretory tubules of Am- 
phioxus, 489 

Willey and Lankester on the develop- 
ment of the atrial chamber of 
Amphioxus, 445 


PRINTED BY ADLARD AND SON, BARTHOLOMEW CLOSE. 


6 


wg 


Opes ewes a Sh ein te ee 
mt Sete ON ee re ayers emmy ev 


Aero ee 


Aa raer - 


Soe tk ee 8 ~~ 


PN IN ET LTE ET 


a A NO a ill EN Hy = ee ee et, 
ee a Se ee ee  naemn 


SE Pe eae OPO nce hanna 
nee etree ae OR Gen ae ee ee om 
PPO OTD SO LALLA LORE IES 


a OR ll ey ENED ee AOL ne eR Oe AI I et Oy NS ten hem 
Agee te eee 
a A a a a i AEA A AEA AN AOTC np pn DOE Race ies 
A Ne ah RD eA cel OPE te Oy AE ONG OE OO = 0 AO RO ly eel 8 AO SE MOBS wi et 
Rat Ae hk Oe pers Reh tO Ope Nee tee hey Rg ee 2 he 
: — 
<p eon ay ns te 8 
Re nee h A Ah Ae Ae ES Ree ae ee AOR Ain PEO BEE Ap Ot aon Sons odes 
eer RAs Rw Le oe enh AE eee AO ONO O OE TAY LEG Oe He OY os a pe 
my at: mee