arene iy GI UT ee en a anwar is
nea a te ain ae:
ar
HARVARD UNIVERSITY
coh
LIBRARY
OF THE
Museum of Comparative Zoology
Yoda
Pa PG) pie Via /s,
alt y) Lo as
a.
a rr :
Siren hes rit
‘
+
?
{ J
* os ‘
y
: F
¢ 7
'
4
‘ :
ri
s ;
4
«
‘
Yount)
.
} /
:
‘
.* ‘
phe
1
»
; ’ ’
P
‘ y
‘ ; ’
aA Shy
Ps
a |
s. ;
« .
AS
]
i Som
,
af
‘ ; ‘
he re
ae f
§
fe bs
>
‘
oa
4 '
/ v4 ¥ 7a
é i 4
en ;
’ {
*
ve
mie Y p
‘ j
; ,
en ‘
‘
\ ’
S i
feet
f OAg® 4
;
ar '
id ius 4,
aby ‘ e CL t
ay
; y ’
ot, ‘ Meld Vd h
my? y Lies!
Solna its
oP ace hye |
re se A
}
QUARTERLY JOURNAL
MICROSCOPICAL SCIENCE:
EDITED BY
E. RAY LANKESTER, M.A., LL.D., F.R.S.,
Linacre Professor of Comparative Anatomy, Fellow of Merton College, and
Honorary Fellow of Exeter College, Oxford.
WITH THE CO-OPERATION OF
ADAM SEDGWICK, M.A., F.RS.,
Fellow and Lecturer of Trinity College, Cambridge ;
AND
W. F. R. WELDON, M.A., F.R.S.,
Jodrell Professor of Zoology and Comparative Anatomy in University College, London;
Fellow of St. John’s College, Cambridge.
VOLUME 36.—NeEw Serizs.
With Rithographic Plates and Engrabings on Wood,
LONDON:
J. & A. CHURCHILL, 11, NEW BURLINGTON STREET,
"1894,
ls YET OF
TSS SATA AAD
Olea gh Goat
‘ Ants, ro] Aa ih in
Goo ae spaldo 629 aa ile
-
ce NAGE
> YAOIOOS C00 e0N
a ,
| CCIM BAOIKEN AO tears
“Tot
CONTENTS.
CONTENTS OF No. 141, N.S., APRIL, 1894.
E. RAY LANKESTER (with Portrait)
MEMOIRS:
A Contribution to the Morphology of Bacteria. By E. Kur, M.D.,
F.R.S., Lecturer on General Anatomy and Physiology at St.
Bartholomew’s Hospital Medical School. (With Plate 1)
On Certain Points in the Development and Anatomy of some
Earthworms. By AtFrrep Gisss Bourne, D.Sc., Professor of
Biology in the Presidency College, Madras. (With Plates 2—5)
On the Law of Development commonly known as von Baer’s Law ;
and on the Significance of Ancestral Rudiments in Embryonic
Development. By Apam Senewick, M.A., F.R.S.
A Contribution to our Knowledge of the Annelidaa—On some
Points in the Structure of Huphrosyne. On Certain Young
Stages in Magelona, and on Claparéde’s unknown Larval Spio.
By W. C. McIntosu, Marine gracias St. Andrew’s, Ri a:
Plates 6—8) : 5
Spolia Nemoris. By A. A. W. Soho LL.D., C.M.Z.S., Pro-
fessor of Zoology in the University of Utrecht. (With Plates
9—12) : . ‘ : : :
CONTENTS OF No. 142, N.S., JUNE, 1894,
MEMOIRS :
Studies on the Comparative Anatomy of Sponges. VI.—On the
Anatomy and Relationships of Lelapia australis, a Living
Representative of the Fossil Pharetrones. By ArtHuR Denpy,
D.Sc. (With Plate 13) .
b
11
35
53
17
127
iv CONTENTS.
The Structure of the Bill and Hairs of Ornithorhynchus
paradoxus; witha Discussion of the Homologies and Origin of
Mammalian Hair. By Epwarp B. Poutton, M.A., F.R.S., &.,
Hope Professor of Zoology in the University of Oxford. (With
Plates 14, 15, and 15a) . :
A Contribution to our Knowledge of the See of Tropical
Eastern Africa. By Frank E. Bepparp, M.A., F.R.S., Pro-
sector to the Zoological Society of London. (With Plates 16
and 17)
A Further Contribution rn the nating af lutinanbiele tan-
ganyice. By R. T. Gintuer, B.A., Lecturer of Magdalen
College, Oxford. (With Plates 18 and 19)
Notes on the Minute Structure of Pelomyxa palustris (Greeff).
By Litian J. Govup, Hall Scholar, Somerville Hall, Oxford.
(With Plates 20 and 21).
CONTENTS OF No. 148, N.S., JULY, 1894.
MEMOIRS :
On Moniligaster grandis, A. G. B., from the Nilgiris, S.
India; together with Descriptions of other Species of the Genus
Moniligaster. By Atrrep Gipps Bournez, D.Sc.Lond., Pro-
fessor of Biology in the sean rae Madras, (With
Plates 22—28) . :
A Review of Professor Spengel’s oe on saaidaeiestn
By E. W. MacBrins, B.A., Fellow of St. John’s College, De-
monstrator in Animal Morphology in the University of Cam-
bridge. (With Plates 29 and 30)
Notes on a Gregarine of the Harthworm (Lumbricus herculeus)
By Wm. Ceci, Bosanquet, M.A., Fellow of New College,
Oxford. (With Plate 31) :
.
PAGE
143
201
271
295
307
385
421
CONTENTS.
CONTENTS OF No. 144, N.S., AUGUST, 1894.
MEMOIRS :
Some Abnormal Annelids. By EH. A. AnpREws. (With Plates
32—34) . ; : : : ‘ ;
Studies on the Nervous System of Crustacea. By HEpear J.
ALLEN, B.Sc.Lond. I.—Some Nerve-elements of the Embryonic
Lobster. (With Plates 35 and 36)
Studies on the Nervous System of Crustacea. By Epear J.
ALLEN, B.Se.Lond. Il.—The Stomatogastric System of Astacus
and Homarus. III.—On the Beading of Nerve-fibres and on
End Swellings. (With Plates 37 and 38)
The Sensory Canal System of Fishes. Part I. Ganoidei. By
Watter EHpwarp Co.uiner, Demonstrator of Zoology and
Comparative Anatomy, Mason College, Birmingham. (With
Plates 39 and 40) : : ;
InDEX
PAGE
435
461
483
499
539
Vaan
’
area eng
; > ae
Sic
Pana ae
ay aia
T. Way lith.
L. Stillman del.
EK. RAY LANKESTER.
Ir is now five-and-twenty years since Professor Lankester
first undertook the task of editing the ‘ Quarterly Journal of
Microscopical Science,’ and by issuing the present number his
colleagues desire to mark the occasion, and at the same time
to take the opportunity of offering to him their hearty con-
gratulations on the success which has attended this quarter of
a century of effort on his part.
The Journal was founded in the year 1853 by the publisher,
Mr. 8. Highley, and was edited by Dr. Edwin Lankester and
Mr. George Busk. In 1856 the publisher’s business was
transferred to Mr. John Churchill, with which firm it has
remained ever since. Up to 1868 the Journal published the
‘Transactions of the Royal Microscopical Society of London,’
but in 1869 the Society started its own publication, and a new
editorial arrangement of the Journal was made. Mr. George
Busk retired, and Mr. Ray Lankester, who had lately taken
his degree at Oxford, joined his father in the editorship.
Mr. Ray Lankester’s connection with the Journal began in
1863 with the publication of a paper “On our Present Know-
ledge of the Gregarinz,” followed in 1864-5 by a memoir, in
three parts, on “The Anatomy of the Earthworm.” In 1865
he suggested the publication of a Quarterly Chronicle of the
progress of Histology and Microscopic Investigation, and
joimed Mr. Busk in its preparation. Curiously enough, this
feature has been abandoned since 1872, whilst the Royal
Microscopical Society has taken the task in hand, and pro-
duces an admirable and extensive record.
In 1872 Ray Lankester’s father ceased to take part in
a
ll E. RAY LANKESTER.
editing the Journal, and was succeeded by Dr. J. Frank
Payne. Lankester and Payne added Mr. Thiselton Dyer
(now Director of Kew Gardens) to their editorial body in
1873, and he was succeeded in 1876 by Mr. Archer, of Dublin,
the Secretary of the Dublin Microscopical Club, and the
author of so many interesting discoveries among fresh-water
Rhizopoda. In 1877 Dr. Payne retired, and Dr. Klein joined
the editorial staff.
In 1878 a further change was made. Professor Lankester
became sole editor, with the co-operation of Archer, Francis
Balfour, and E. Klein. This arrangement has continued ever
since, with various changes in the list of those co-operating.
Thiselton Dyer returned for a few years as one of those giving
his co-operation ; and Moseley and Milnes Marshall have in
turn assisted in the conduct of the Journal, and have published
in it many of their most important papers, inducing their pupils
to adopt the same mode of publication.
The number of contributions which this energetic policy
attracted to the Journal soon made it necessary to enlarge it ;
and the term of Lankester’s editorship has been marked by a
continuous increase in the amount of letterpress and in the
number and excellence of the plates. This has of necessity
been accompanied by a rise in price. The original price was
four shillings per number—the numbers being issued quarterly.
At that time the volume consisted of some eight-and-twenty
demy octavo sheets and twenty plates, mostly also octavo. The
last volume contained thirty-six royal octavo sheets and forty-
two plates, many of which were coloured, while the majority
were of quarto size. The change from demy to royal octavo
was effected at the commencement of 1883, and in 1890 the
strict quarterly publication of the Journal was abandoned, so
that more than four numbers could be issued in the year.
During the eleven years which have elapsed since 1883 sixty-
one numbers, divided into fifteen volumes, have been issued ;
so that the increase in size and price has not only affected the
magnitude of each number, but has been accompanied by an
increased rapidity of publication.
E. RAY LANKESTER. lil
Every reader will remember that Professor Lankester’s
energy has by no means been exhausted in merely editing the
Journal, for besides his many writings elsewhere, he has pub-
lished more than sixty memoirs in the pages of this Journal
alone ; and we may, perhaps, be permitted to mention a few of
. the more prominent of these—such as that on ‘‘ The Develop-
ment of the Pond Snail” (1874), which marks the starting-
point of his well-known investigation of the development of
Mollusca; the “Notes on the Embryology and Classification
of the Animal Kingdom” (1877), which exercised so great an
influence upon the whole tendency of morphological specula-
tion; the descriptions of Limnocodium (1880) ; the series of
memoirs on Apus and Limulus (1881—1884), and on Rhabdo-
pleura (1884) ; the first description of the atrio-ccelomic funnels
in Amphioxus (1875), and the subsequent memoir on the ana-
tomy of the same animal, together with the account, commenced
in conjunction with his pupil Mr. Willey, and continued by
Mr. Willey alone, of the later history of its remarkable larva.
It would be useless to enumerate all the naturalists who
have contributed to the Journal since Professor Lankester’s
successful enterprise has made it the chief medium of publica-
tion for English morphological work; but it is interesting to
notice that the contributors have constantly included foreign
naturalists of distinction, including E. van Beneden, Bow-
ditch, Carriére, Claparéde, Dollo, Giard, Hubrecht, Iijima,
Ischikawa, Kingsley, Mitsukuri, H. F. Osborn, Oudemans,
Packard, Patten, Pelseneer, Pouchet, Ranvier, Whitman, and
others. Some of these have taken the opportunity, by contri-
buting to the present number, of joining in the hearty con-
gratulation on his past achievement, and sincere good wishes
for the future, which Professor Lankester’s associates now
offer to their chief.
A. SEDGWICK.
April, 1894. W. F. R. Wepon.
”
j at
~~
¢
’
{
iy ee ke i A
"8 « roe @
ereci:
ag
Le at
. Pe
TA"
-
_
®
© ity
ea
|
ont i Que
12 pees aft
=) 2 ota)
=
~~ Sy -..-
ee
ea
f =
A Contribution to the Morphology of Bacteria.
By
E. Klein, M.D., F.R.S.,
Lecturer on General Anatomy and Physiology at St. Bartholomew’s
Hospital Medical School.
With Plate 1.
Since the classical papers on Bacteria published by Pro-
fessor F. Cohn in the ‘Zeitschrift der Biologie d. Pflanzen,’
vols. i to ili, all bacteriologists have accepted the subdivision
of the Schizomycetes into cocci, bacilli, and vibrios or spirilla,
as representing the main morphological fundamental types. But
by Lankester! and Zopfs’ researches on Cladothrix dicho-
toma, and by Hauser’s well-known and exhaustive work
on ‘ Putrefactive Bacteria’ (‘ Ueber Faulnissbacterien,’ Leipzig,
1885), it has become recognised that the shape under which a
particular bacterial species presents itself depends both on the
medium in which it grows, as also on certain inherent charac-
ters of the organism itself. Thus it has become recognised
that while the elements of one species appear as often in the
form of oval as of cylindrical cells, those of another retain,
under almost all conditions, pre-eminently that of cylindrical
cells. To name a few instances: (a) the Proteus vulgaris
of Hauser. This organism—the organism of putrefaction par
ex cellence—is known to occur in the most varied shapes, as
cocci, oval forms, cylindrical and vibrionic forms; but when
growing in gelatine plates at 20° C. it will be found that in the
1 ¢ Quart. Journ. Micr. Sci.,’ vol. xiii, 1873; vol. xvi, 1876.
VoL. 36, PART 1,—NEW SER, A
2 E. KLEIN.
first twenty-four to forty-eight hours the colonies are made up
entirely of cylindrical bacilli, some of extreme length and
forming very characteristic threads. The “swarming” out and
the development of strands from such a colony are uniformly
due to, and consist of, thread-like bacilli (see my article in
Stevenson and Murphy’s ‘Treatise on Hygiene,’ II, pl. 11);
but later on, say after three or four days, when liquefaction of
the gelatine has become extensive, the forms one meets are
those various kinds described by Hauser as coccus forms, ovals,
cylindrical cells, and vibrionic forms.
(5) The Bacillus filamentosus (which I found in sewage,
and which I described in my article in Stevenson and Mur-
phy’s ‘ Treatise on Hygiene,’ II, fig. 14) is, under all conditions
of culture (gelatine, agar, broth, serum, &c.), always made up
of cylindrical cells, either singly or in pairs, or forming longer
and shorter chains.
The same applies to the Bacillus subtilis, the bacillus of
swine erysipelas, and the bacillus of human typhoid fever.
(c) The Bacillus prodigiosus, on the other hand, remains
under artificial cultivation in the various media pre-eminently
of the spherical or slightly oval shape, while there are always
present a few cylindrical forms, and it is owing to the greatly
prevalent number of coccus forms that in former years this
organism was described (Cohn, Schrotter) as the Micrococcus
prodigiosus,
Another interesting point connected with the unstable shape
under which some species appear is that while some, growing
in one medium, appear under one particular shape, this changes
when growing in a different medium. A group of bacteria are
known, the essential biological character of which is that most
of them produce acute septiceemic infection in one or the other
rodent. To this group belong the bacillus of fowl cholera,
of fowl enteritis, of Frettchenseuche, of Wildseuche, of swine
fever, of the Middlesborough pneumonia, of grouse disease, the
Bacillus coli, and others.
Now all these in their cultural characters in the different
media have many points in common, as also the close resem-
A CONTRIBUTION TO THE MORPHOLOGY OF BACTERIA. 38
blance of the acute disease they are capable of producing in
rodents. But it is noticeable amongst them that while some
preserve the same definite shape when grown in one species of
animals or one kind of medium, this becomes changed under
other conditions. Take, for instance, the bacillus of grouse
disease: in the grouse itself, taken from the liver or cultivated
in gelatine, it appears in a short oval form; but in the guinea-
pig or in the mouse the cells are more commonly of a cylin-
drical form, and so also in the gelatine culture from the blood
of these animals.
Amongst the best known examples of permanency of shape is
that of the Bacillus anthracis; so much so that its cylin-
drical elements, single and in short and long chains in the
blood of an animal dead of anthrax, have become as much the
classical illustrations of typical bacilli as those of the Bacillus
subtilis.
A. In the year 1883 (see this Journal, vol. xxiii) I have
described a peculiar change of the anthrax bacilli in culture,
in the course of which the typical cylindrical cells constituting
the well-known threads become transformed into oval and
spherical corpuscles, some containing vacuoles. This change
was named a torula form, because some of the threads resemble
in a remarkable manner the chains of cylindrical, oval, and
spherical cells which are observed on Saccharomyces my-
coderma of thrush. The direct connection between the
typical cylindrical cells and the spherical and oval corpuscles
(three and more times the diameter), and the division of these
into similar corpuscles, was traced in many filaments through
all intermediate stages (fig. 1).
At that time I insisted on these changes not being due to
involution and degeneration, but belonging to an active phase
of growth in the artificial media. In the first place it was then
shown that this morphological change is observed already in an
early stage of growth, when of degeneration there can be no
question ; besides, in later phases, after two, three, and more
days’ growth, the progress of the growth and the resulting
filaments are again of the characteristic appearances. I have
4 E. KLEIN.
within recent years almost constantly observed a similar change
in early phases of the growth of anthrax bacilli in gelatine
plates (beef bouillon; gelatine 10 per cent.; peptone 1 per
cent.; salt 1 per cent.). Already in very early phases, when
the colonies are only just visible as angular greyish spots, and
_when by their numerous outgrowing filamentous prolongations
they become more and more connected with one another—that
is, between twenty-four to forty-eight or seventy-two hours’ in-
cubation at 20° C.—numbers of these filamentous sproutings,
examined in impression cover-glass specimens, are seen to be
made up entirely or partially, not of the typical filaments com-
posed of the cylindrical typical bacilli, but of large spindle-
shaped spherical or oval elements, the protoplasm of which
showing abundant vacuolation. In figs. 2 and 3 such growing
outrunners of young colonies are accurately represented by pho-
tograms. There can be no question, then, of these forms being
indicative of active growth ; as a matter of fact, later on—that
is after four days and more, as growth proceeds and liquefac-
tion becomes pronounced—such forms do not obtain any more ;
the threads are all uniformly made up of the typical cylindrical
bacilli.
When comparing the colonies of the thrush fungus, Sac-
charomyces mycoderma, or Oidium albicans, growing
on gelatine plates, it will be found that the impression pre-
parations obtained therefrom show in many filaments the very
identical appearances ; and that while some threads or parts of
threads are composed of cylindrical cells, others are made up
of oval, spindle-shaped, and spherical cells; the same local
accumulation of the growing protoplasm as huge spindles or
spheres, and the same vacuolation of the protoplasm are ob-
served in both. I conclude from this that, although the
Bacillus anthracis is a typical bacillus in the blood of
animals infected with, or dead from, anthrax, and also in most
conditions of artificial cultivation, it nevertheless under certain
conditions (early stages of growing colonies on gelatine)
assumes a character by which it closely resembles a Sac-
charomyces mycoderma, or perhaps Oidium, and thereby
A CONTRIBUTION TO THE MORPHOLOGY OF BACTERIA. 5
probably it returns to an atavistic stage in its evolutional
history.
B. The second microbe, by which a similar marked change
is exhibited, is the Bacillus diphtheriz. Loffler (‘ Mitth.
aus d. k. Ges.,’ vol. 11) first drew attention to the fact that the
diphtheria bacillus, discovered by Klebs, shows on cultivation
a curious segregation of its protoplasm, and a knob-like or
club-shaped enlargement of one or both ends. Léffler, and
after him others (Fligge, ‘Mikroorganismen ;’ Baumgarten,
‘Pathologische Mycologie,’ and others), considered these changes
as due to involution. I have already, in the ‘ Report of the
Medical Officer of the Local Government Board, 1889-90,’ and
‘Centralblatt f. Bakt. und Parasitenkunde,’ vol. vii, 1890,
shown that this view cannot be correct, for the following
reasons :
(a) In the diphtheritic membrane, in which the progress of
the disease is still active, an abundance of diphtheria bacilli
occur, which show this change in a marked degree, viz. segre-
gation of the protoplasm into spherical, cubical, or cylindrical
particles, and terminal knobs or clubs, sometimes of great
size and containing vacuoles.
(6) On agar cultures, already after twenty-four to thirty-
six hours, when the growth is in its initial and most active
phase, an abundance of bacilli are seen, which are shorter or
longer threads, in which the segregation of the protoplasm and
the terminal knobs and clubs are already very marked (fig. 4).
(c) In the subcutaneous necrotic tumour of the cow, pro-
duced by subcutaneous injection of virulent culture of the
Bacillus diphtheriz, there occur connected masses and
clumps in which at the growing margin the diphtheria bacilli
appear all in the form of threads, in which the spherical or
oval swellings and terminal knobs are most conspicuous, and
strikingly resemble the ends of growing hyphe; the subjacent
muscular fibres become invaded and gradually destroyed by
the growth of the threads into their substance. This process
of the gradual growth and penetration of the diphtheria
threads with swellings and club-shaped ends into the muscular
6 E. KLEIN.
fibres is very marked, and occurs in a large number of places.
It was described and illustrated in the ‘Report of the Medical
Officer of the Local Government Board, 1889-90,’ pp. 173 and
174, plates xiv, xv, and xvi; and I have to add here that
sections of the tumour stained in a mixture of eosin and
methyl blue show this in beautiful contrast, the growing threads
blue, the muscular substance red; and it seems to me that one has
only to examine such a specimen to at once see that the threads
are actively growing; of an involution there can be no question.
Dr. Abbott, in the ‘ Journal of Pathology and Bacteriology,’
vol. ii, while agreeing that the thread-like bacilli with terminal
swellings are not involution forms, and are present in the
growing artificial cultures (on serum), does not agree to their
being comparable to the growing ends of hyphe. This dif-
ference seems, however, merely a question of words; all that 1
maintained was that the threads with knobbed ends strongly
resemble growing ends of hyphe, and that such a change in
bacilli as I described in artificial cultures, and particularly in
the growing threads in the cow’s tumour, if it is not due to
involution, as it certainly is not, is only explained by its
representing a relationship to a mycelial fungus, perhaps the
Saccharomyces mycoderma or an Oidium form.
(d) From the milk of cows successfully inoculated with
cultures of the diphtheria bacilli (Report, 1889-90 and
1890-91) I have isolated by culture the diphtheria bacillus,
and in the gelatine cultures they showed this change in a con-
spicuous manner ; the colonies in their young state are almost
entirely made up of thread-like forms with terminal knob-like
and club-shaped swellings (figs. 5 and 6) quite unlike the typical
bacilli. Here the bacilli are actively growing, and therefore
it is quite out of place to regard these forms as due to involu-
tion; and if they are not involutions, their similarity in
growth and shape strongly suggests the view that I have put
forward, viz. that they are comparable to the hyphe of a
mycelial fungus, e.g. Saccharomyces mycoderma. In these
threads with local accumulations of their substance, and with
terminal knob-like or club-shaped enlargement of their proto-
A CONTRIBUTION TO THE MORPHOLOGY OF BACTERIA. 7
plasm, we have a condition of things which does not harmonise
with the fundamental characters of a typical bacillus, but
rather suggests that this microbe, though under many con-
ditions conforming with what corresponds to a typical bacillus,
may after all not be one, or at any rate the boundary between
it and a mycelial fungus is not a severe one.
C. The most instructive organism, showing a similar and
perhaps more pronounced morphological change, is the tubercle
bacillus of Koch. As is well known, this microbe presents
itself in the tubercular deposits and in serum and agar
glycerine cultures in a form which has vindicated to itself
the term of a typical cylindrical bacillus. But already in
preparations made of the human pulmonary (tubercular)
sputum forms occur which appear more of the character of
threads composed of unequal—i. e. not uniform—elements ;
some of these threads show not unfrequently a terminal element
of the same knob-shaped or club-shaped character as those
mentioned of the diphtheria bacilli. It was on account of
such forms that Metschnikoff expressed the opinion that per-
haps the tubercular bacillus is not a bacillus at all, but belongs
to the group of mycelial fungi. I have already in 1889-90
shown that in glycerine agar cultures of the tubercle bacilli,
after some weeks’ growth at 37° C., there occur large numbers
of such thread-like forms with club-shaped ends; some short,
others long, some smooth, others made up of unequal elements ;
further, long and short threads occur which show undoubtedly
and markedly branchings, these latter either of great length
or only as just commencing sprouts (see figs. 7 and 8). Thatall
these forms are undoubtedly the tubercle parasite is shown by
the transitional forms between the typical cylindrical tubercle
bacilli and the long-branched threads (homogeneous or seg-
mented), with smaller or longer lateral buddings, and by the
fact that all these forms behave in staining (fuchsin, washing
with nitric acid, 1:3) lke the true tubercle bacilli. I have
seen the same forms already after three to five weeks’ growth
on solidified hydrocele fluid; here most of the organisms were
the typical cylindrical bacilli, but there were some undoubted
8 E. KLEIN.
threads with knobbed ends, some branched, others as yet
unbranched,
There can be no question about involution, because, as I
have pointed out (1. c., and ‘Centralblatt f. Bakteriology,’
vol.vii, No. 25), the branched nature of the threads and
the presence of the small lateral buddings conclusively prove
the active growth. Later, Mafucci (‘Archiv f. Hygiene und
Infect.,’ xi, p. 445) described the same forms in the culture of
the tubercle bacilli of the fowl, and Fischel (‘ Fortschr. d.
Med.,’ Bnd. x, No. 22, p. 908) also of the human tubercle
cultures ; and this latter observer arrived at the same conclusion
as myself, viz. that we are dealing with forms which are com-
parable to a mycelial fungus.
From all these facts I think we are justified in concluding
that the above three species are not so well-marked typical
bacilli as has always been assumed ; that is to say, well-defined
species of desmo-bacteria in the sense of F. Cohn. True, under
many conditions they show morphological characters of the
same kind as the typical bacilli; but under other conditions
they easily revert to or assume forms by which their relation
to the Saccharomyces or Oidium (anthrax, diphtheria), or a
still higher mycelial fungus (tubercle) becomes evident.
- A CONTRIBUTION TO THE MORPHOLOGY OF BACTERIA. 9
EXPLANATION OF PLATE 1,
Illustrating Dr. E. Klein’s paper, “A Contribution to the
Morphology of Bacteria.”
All figures are reproductions from photograms magnified 1000.
Fic. 1.—Cover-glass specimen of Bacillus anthracis growing in gela-
tine plate, two days old. Many of the bacilli are changed into spherical or
oval masses, containing vacuoles.
Fic. 2.—Cover-glass specimen from an impression of a gelatine-plate
cultivation of Bacillus anthracis, two days’ growth. Instead of the
typical threads of cylindrical bacilli, there are threads made up of thick spindles,
the protoplasm in many of these vacuolated.
Fic. 3.—From the same plate cultivation as the previous figure. Copious
vacuolation.
Fie. 4.—From an agar culture of the Bacillus diphtheria, grown at
37° C. for two days, showing typical club-shaped filamentous bacilli.
Fies. 5 and 6.—Diphtheria bacilli, derived from the milk of a cow infected
with diphtheria. The bacilli had been growing on gelatine. Typical club-
shaped filamentous forms.
Fies. 7 and 8.—From glycerine-agar cultivations of the tubercle bacilli.
Filamentous bacilli with terminal knob-like enlargements, some showing
distinct branching.
why
a’ f
.
.
ere mee
a) Lae
26
DEVELOPMENT AND ANATOMY OF SOME EARTHWORMS. 11
On Certain Points in the Development and
Anatomy of some Earthworms.
By
Alfred Gibbs Bourne, D.Sc.,
Professor of Biology in the Presidency College, Madras.
With Plates 2—5.
Tue following notes are selected from a more extended
series of observations which I am making upon this subject,
and would not have been published separately, but that I
wished to take this opportunity of testifying to my regard for
one who has been for so long, and still is, my teacher and
friend.
Much of the literature of the subject has not yet reached me
(including the later portions of Vejdovsky’s ‘ Entwickelungs-
geschichtliche Untersuchungen’), and I have been obliged to
defer any detailed reference to it until a future occasion. The
organs dealt with are the sete and the nephridia, and I have
added a synoptical description of the two new species of worms
with which the paper chiefly deals.
New SPECIEs.
The two worms, Mahbenus imperatrix and “ Peri-
cheta” pellucida, both belong to the family Perichetide,
but neither of them to the genus Pericheta. For the one I have
founded a new genus, being fairly confident as to what constitute
its generic characters; for the other I have abstained from so
doing, as it is, I think, allied to some of the new species de-
scribed by Fletcher, while it is the only representative of the
12 ALFRED GIBBS BOURNE.
genus which I have come across. I speak of it for the present
as Perichzta.
MAHBENUS, gen. nov.
Set very numerous. Circles of sete almost continuous.
Clitellum extends over more than three segments.
Male pores, one pair, very close together, no sete between
the male pores.
Gizzard occupies not more than one segment.
Intestinal ceca absent.
Septa normal.
Micronephridia present.
Testes, two pairs, freely exposed.
M. imperatrix, sp. nov.
Length 650 mm. Circumference 35 mm.
Segments 200.
Colour rich dark brown, lighter on the ventral surface.
Prostomium encroaches very slightly upon the peri-
stomium.
Seta.—Segment 11, 52; Segment v, 80; Segment 1x, 110;
no modified setze ; setee remain in the clitellum. The circles
of sete are continuous in Segments 11—xvii and xIx, the circle
is interrupted ventrally in xv111 by about eight seta gaps, and
in all the posterior segments there is a tiny gap recognisable
only with care in the median dorsal and ventral lines.
Clitellum not strictly confined to particular segments, may
extend over Segments xiv—xr1x and part of xx.
Genital Apertures.—Male pores very small and quite close
together (with not much more than a seta gap between them) ;
the region immediately around them becomes raised into an
oval papilla after killing in spirit. Two pairs of glands open
in the region of the male pores, one pair on Segment Xvii
behind the circle of set, and one pair between Segments x1x
and xx; their apertures lie about as lateral as seta 6.
The oviducal apertures are paired.
The spermathecal apertures are very small, and placed
DEVELOPMENT AND ANATOMY OF SOME EARTHWORMS. 13
between Segments VI-vII, VII-VIII, and viI-1x; the two pores
of each pair are remarkably close together, so close as to be
only recognisable with care as separate apertures.
Dorsal Pores.—The most anterior dorsal pore lies between
Segments v and v1; they are present in the clitellar region in
the young worm, but become completely obscured when the
clitellum is developed.
Alimentary Tract.—The gizzard is in Segment vii; dilated
portions of the cesophagus serve as calciferous glands in Seg-
ments 1x to xv; they are large in Segments x1 to xiv only.
* The typhlosole is a simple flap, deep, down to about Seg-
ment LxIv; there it narrows, becomes a mere ridge, and goes
on to Segment cxvul, and there ends abruptly, and at the
same place the alimentary epithelium changes and becomes
rectal.
Vascular System.—The dorsal vessel is double in Seg-
ment vir and onwards for a considerable distance.
Nephridia.—Minute micronephridia are present in large
numbers. ;
Generative System.—The testes lie in Segments x and x1,
and are attached to the septa bounding these segments ante-
riorly. The ciliated rosettes lie in the usual position. Pros-
tates are large and rounded, and are provided with a muscular
duct. The seminal reservoirs lie in Segments 1x and x11. The
spermathece are sausage-shaped sacs with a short duct, and a
small cecum lying in the thickness of the body-wall, and,
what is very unusual, increase in size from before backwards.
A spermatheca of Segment 1x is more than three times as long
as one of Segment vil. The ovaries and oviducts lie in the
usual positions.
Pericheta pellucida, sp. nov.
Length 450 mm. Circumference 12 mm.
Colour.—The body-wall is very transparent, so that the
colouring depends on the blood and the contents of the alimen-
tary canal.
14 ALFRED GIBBS BOURNE.
Prostomium small, not dovetailed into the peristomium.
Setze.—Segment 11, 24; v, 44; 1x, 36; xx,36. The number
is subject to variations,and they are not very regularly arranged.
No modified sete are present. The dorsal gap is equivalent to
about ten seta gaps, and the ventral gap to about three or less.
Clitellum not strictly confined to particular segments ; it
extends over Segments x111I—xvi1r and a little into Segment
xu. The dorsal pore x1i1-x111 lies well within the clitellum,
and posteriorly it extends nearly up to the seta ring of Seg-
ment xx. It is not well developed in the ventral region.
Genital Apertures,—The male pores lie in Segment xvu1
and thus in the clitellar region, although the clitellum is not
developed in their neighbourhood. They lie in a small dumb-
bell-shaped pit in preserved specimens. The distance between
them is equivalent to about five seta gaps. There are no
setee between them. There are no other apertures in this
region.
The oviducal apertures are paired, and lie about in the direc-
tion of seta 1, and just in front of the seta ring in Segment xtv.
The spermathecal apertures are placed between Segments
VII-VIII and vVIII-Ix, and in the direction between seta 1 and
seta 2.
Dorsal Pores.—The most anterior dorsal pore lies between
Segments v and vi. They remain obvious between the clitellar
segments.
Alimentary Tract.—Gizzard is in Segment v.
The cesophagus presents dilatations with vascular walls in
Segments vi to x111. There are well-developed calciferous
glands in xiv, xv, and xvi. Saccular intestine commences in
Segment xVIII.
There are no intestinal ceca.
There is no typhlosole.
Nephridia.—There is a pair of complex (see below)
nephridia in each of Segments vir to x1, and a pair of small
simple nephridia in each of the following segments. There are
no micronephridia.
Generative System.—Testes lie in Segments x and x1,
DEVELOPMENT AND ANATOMY OF SOME EARTHWORMS. 15
and are attached to the septa bounding these segments ante-
riorly. The ciliated rosettes lie in the usual position. The
prostates are long tubular-shaped glands, but remain in Seg-
ment XVIII.
The spermathece are elongated pyriform sacs with a small
ceecum.
The ovaries and oviducts lie in the usual positions.
SETA.
The setze which first develop are, in all the worms which I
have studied, replaced by others in either all or in the greater
number of the segments before the embryo leaves the capsule ;
the replacement takes place in regular order from before back-
wards, and if it has not taken place in the hindermost seg-
ments before birth it does so shortly after. I speak of the
setze which first develop as embryonic sete, and of the sete
which replace them as permanent sete, although, of course,
these may drop out and be again replaced later. The groups
of cells from which the embryonic setz of any segment: de-
velop I term secondary setal matrices. The secondary
setal matrices develop from a primary setal matrix in
each segment on each side of the body whether the complete
number of setz in the segment is four on each side or a larger
number, as in the Perichetide.
Origin of the Primary Setal Matrices,
To exemplify this I have taken figs. 1—8 from sections of
an embryo of Mahbenus imperatrix. ‘The embryo was the
same as that drawn in fig. 33, and was 7 mm. long when
removed from the capsule. The sections were cut longitu-
dinally through the tail end after it had been flattened out as
in fig. 83. They are ‘006 mm. thick. They were drawn with
a camera lucida and Zeiss F, oc. 3 (for figs. 1—6; oc. 2 for
figs. 7, 8), and as drawn are, therefore, magnified about 1100
diameters. The indentation marked z in figs. 1—8, due to a
fold caused by the flattening, serves to mark the relative
16 ALFRED GIBBS BOURNE.
antero-posterior position of those three sections. Fig. 1 passes
through the “ primary mesoblast ” of the right side; the next
section on one side (not figured) is precisely median and passes
between the “ mesoblasts,’’ while the section beyond (not
figured) passes through the “ mesoblast” of the left side.
Fig. 2 is the next section to the right of fig. 1, and fig. 3 the
next beyond; they show the row of cells proceeding from the
“‘mesoblast”’ of fig. 1 passing slightly outwards and forwards
(cf. fig. 33). Fig. 4 is a portion of the section next to fig. 3,
and commences about half its own length beyond where fig. 3
leaves off ; it shows the row. of cells proceeding from the neuro-
blast of the right side. The lines 4 B in figs. 4, 5, and 6 are
drawn at the same level to show the relative antero-posterior
positions of the figures. Fig. 5 is the seventh section from
fig. 1, but the cells a, 6, and c are inserted from the eighth and
ninth sections, while fig. 6 is the tenth section and the most
lateral of the series.
Figs. 5 and 6 show the origin of the primary setal matrices ;
they arise on each side from a longitudinal row of cells which
is, I have little doubt, the row arising from Wilson’s lateral
teloblasts.! These cells, at first very superficial, take up as
they pass forwards a deep-lying position within the cclom.
At first there is one cell on each side in each segment lying
close to the nephridial cell (fig. 5). Their further history is
shown in figs. 7 and 8; fig. 8 comes from near the median line
a little in front of fig. 4, and includes a portion of the nerve-cord ;
fig. 7 lies a little behind and to one side of fig. 8. The setal
matrices are now seen to consist of several cells. There is not
the slightest difficulty in tracing these structures forwards
from segment to segment until they are old enough to have
developed sete.
The embryos of Moniligaster (probably M. sapphirinaoides)
which I have in these early stages show also very clearly that
the primary setal matrices take their origin from continuous
longitudinal rows of cells, but neither here nor in Mahbenus,
at any rate in the youngest stages which I have examined, are
1 «Journal of Morphology,’ vol. i, 1887.
DEVELOPMENT AND ANATOMY OF SOME EARTHWORMS. 17
any of the teloblasts except the ‘‘ mesoblasts” as specially
enlarged as they are in Lumbricus.
I gather from Bergh (‘Zeit. f. w. Zool.,’ Bd. 1, p. 523) that
Wilson has, in a paper which I have unfortunately been unable
to see, shown that “the outer setigerous glands arise from the
lateral cell-cord,” and that he recognises a “ setiblast.”” My
observations, which were originally made without any knowledge
of the’previous literature, corroborate Wilson’s, but go further,
as I show that all the setal matrices, however many there may
be, arise on each side of the body as a cell-cord. My observations
on the development of other organs are still very incomplete,
but so far as they go they have led me to the conclusion that
all the organs which express a metameric segmentation arise
from the cords of cells which grow forwards in the germ bands.
The mesoblastic bands give rise, I believe, to the muscles of
the septa and ccelomic epithelium and blood-vessels only, while
the muscles of the body-wall which do not exhibit metamerism
arise in some other way. I find no evidence that they arise
directly from epiblast, but a certain amount that they arise at
an early stage from the primary mesoblasts, and perhaps also
from the mesoblastic cords before these have become segmented,
growing outwards in all directions, and not in that direction
alone which is taken by the mesoblastic cords. Certain it is
that at a time when the germ bands are still in their infancy
muscles are to be found underlying the whole of the epiblast,
Besides the muscles of the body-wall, the epidermis, and the
alimentary tract, the only other organs which do not arise from
the germ bands exhibit no metamerism in the embryo. The
occurrence of such organs as the gizzard in such varying seg-
ments in different worms would be explained if we can show
that what segmentation the alimentary tract possesses has
nothing to do with the metameric repetition of other organs.
I have further, like Bergh, come to the conclusion that the
nerve-cord arises from two distinct matrices, and that the
ganglia are the only structures which arise from the neuro-
blastic cords. The above theory will, I foresee, land us in
great difficulties when we consider organisms other than worms,
VOL. 36, parT 1.—NEW SER. B
18 ALFRED GIBBS BOURNE.
and I put it forward with great diffidence and in a purely tenta-
tive manner.
Development of the Secondary Setal Matrices.
The primary setal matrices grow in each segment laterally,
and also to a lesser degree towards the ventral median line, as
soon as they have taken up their position in the our and
become covered by cceelomic epithelium.
The primary matricial] cells become segregated to form the
secondary matrices. The number of these formed in each seg-
ment varies; in Moniligaster each seta couple arises from a
single matrix, which accounts for the fact that the two sete of
the couple are always so close together. This is also the case
in Lumbricus.
In Acanthodrilus sp. the sete are slightly “ separated,”
and each seta develops from a separate secondary matrix. In
Urocheeta there may be three or two secondary matrices on
either side, according as the sete are “ scattered” or not in a
particular segment on that side (fig. 16). The exact arrange-
nent which obtains in regard to this matter in “Pericheta”
pellucida is described below, but, speaking generally, in this
form and in M. imperatrix and in Perionyx saltans each
seta arises from a separate secondary matrix.
All the secondary matrices on each side of the body remain
connected together for a longer or shorter time by a band of
tissue which stands out freely into the body-cavity, and is com-
posed of the ccelomic epithelium cells which covered the primary
matricial cells as they grew out to form the secondary matrices,
and may be termed the intermatricial band (figs. 20, 24, &c.,
im.; cf. Vejdovsky, ‘ Entwickelungsgeschichtliche Unter-
suchungen,’ Heft 3, Tab. xxiii, fig. 19, 7. f1, and Tab. xxviii,
fig. 7, 7m.).
Fig. 9 is a portion of an embryo of Moniligaster flattened
out, drawn with a camera lucida and Zeiss BB, oc. 3, magnified
about 150 diameters. At p. m. is seen a series of primary
setal matrices, the more anterior ones commencing to grow
outwards ; in the four segments (¢m.) the formation of the two
DEVELOPMENT AND ANATOMY OF SOME EARTHWORMS. 19
secondary matrices is taking place, these being connected by
the intermatricial band. This band disappears very early in
Moniligaster, and in the anterior segments are seen inner and
outer secondary matrices, as’ at s.m.?. and s.m.o. In the
most anterior segment drawn, a seta (se¢.) is appearing in the
inner secondary matrix. The embryo from which this figure
is drawn has about fifty segments in front of those drawn, but
it is not possible to count any segments behind those drawn,
i, e. they are not distinguishable as segments in such a pre-
paration ; longitudinal sections enable one to count them further
back.
Figs. 10—13 show the development of the secondary
matrices in Moniligaster studied in sections. They are taken
from a series cut transversely through a portion of an embryo
of about the same age as fig. 9, and which had been flattened
out in the same manner. They represent the matrices of four
consecutive segments ; three or four sections are omitted between
those figured. The upper part of each figure lies near the
nerve-cord. They are drawn with a camera lucida to the same
scale as figs. 1—7. Figs. 10 and 11 show primary matrices,
that in fig. 11 projecting freely into the ccelom, while that in
fig. 10 does not ; and in these segments there is quite clearly
no other setal matrix on either side of the nerve-cord. In the
next segment in front, the primary matrix has given rise to two
secondary matrices on each side, those of one side being shown
in fig. 12; four inches of the drawing have been cut out in the
middle to place it upon the plate. In the next segment in
front, the inner and outer secondary matrices are quite sepa-
rated from one another; the inner one alone is figured. The
epidermic thickening marked z is one of a series of such struc-
tures with regard to which I can at present give no further
information. I figure it to show that it is not a “ seta follicle.”
Fig. 14 is a surface view of a secondary matrix, showing the
development of the seta couple (set., set.) ; the coelomic epithe-
lium shown allround is omitted from the surface of the matrix.
Fig. 15 shows in a slightly older stage, from a longitudinal
section, one of the setz of a couple.
20 ALFRED GIBBS BOURNE.
Fig. 16 shows in a diagrammatic manner a late stage in the
development of the secondary matrices in Urocheta. In all
the five segments figured, except the most posterior, the primary
matrix has completely separated into two secondary matrices.
In the segments marked a and ¢ the outer secondary matrix is
further dividing into two to place seta 4 in the very dorsal
position it occupies in alternate segments.
The investigation of the development of the secondary setal
matrices from the primary ones in the Perichetes is attended
with great difficulties. To work out the question thoroughly
satisfactorily it would be necessary to obtain a series of very
thin sections, so accurately transverse to the long axis of the
embryo that in each segment one section passed through the
whole of the region of the future seta ring. No amount of care
would with certainty secure such a series, and my efforts in
this direction have not been attended with any special luck.
I have, however, sufficient evidence to prove that the only im-
portant difference in this respect between the Perichztes and
Moniligaster is in the number of secondary matrices which are
produced. In the preparation shown in fig. 33, from the tail
end of which the sections (figs. 1—8) tbat show the origin of
the primary matrices were obtained, these primary matrices
may be traced growing gradually outwards; older embryos
show that they do this until they have grown right round the
segment on each side. The cells meanwhile segregate to form
the secondary matrices, the segregation commencing at the
ventral ends of the matricial bands, and proceeding gradually
towards the dorsal region.
The later stages in the process of production of the secondary
matrices I have most conveniently studied in Pericheta
pellucida, but the small size of the embryo prevented my see-
ing the earlier stages in that form. I think we may fairly
assume, considering the close agreement that obtains between
two such different forms as a Moniligaster and Mahbenus
imperatrix, that the origin of the primary matrices and the
proliferation of their cells to give rise to secondary matrices
takes place as in those forms. The production of the secondary
DEVELOPMENT AND ANATOMY OF SOME EARTHWORMS. 21
matrices in P. pellucida does not, however, take place in
regular succession, beginning at the ventral and ending at the
dorsal end of the band. It is quite clear that in P. pellu-
cida new secondary matrices are produced between existing
ones, either by matricial cells which separate from the existing
matrices and travel along the intermatricial band, or by the
division of any matrix into two. There is obviously no essen-
tial difference between these two methods, nor is it possible to
draw any definite line of demarcation between them, but the
latter method prevails in the earlier and the former in the later
stages of the production of the full number of secondary
matrices.
Fig. 17, from a transverse section of an embryo of P. pel-
lucida, shows a secondary matrix dividing intotwo. Fig. 18,
from the same series, shows two secondary matrices connected
by the intermatricial band; one of these is the most ventral
of the series belonging to one side of a segment, while the
other one shows the band connecting it with the next matrix
beyond.
Figs. 19—25 are views of portions of matricial bands of
embryos of P. pellucida which had been slit open and
flattened out; they are arranged, with the exception of fig. 23,
in the order of their age. Fig. 22 shows the entire matricial
bands on one side of three consecutive segments.
In P. pellucida the two secondary matrices first produced
from a primary matrix come to lie at the two ends of an entire
band, all the other matrices in the band being subsequently
produced. I have deduced this law from the order of appear-
ance of the setee, assuming, as we fairly may, that the matrices
form these setz on arriving at similar ages.
When all the secondary matrices have been developed they
come to lie at nearly equal intervals from one another, and the
intermatricial band gradually disappears (fig. 25).
Development of the Embryonic Sete.
In this, as in all other matters relating to the segmentally
arranged organs, any segment presents during development
22 ALFRED GIBBS BOURNE.
a slightly further advanced condition than the segment be-
hind it.
I have studied the order of development of the embryonic
setee in a very large number of embryos of P. pellucida, and
figs. 26—28 are given as three typical stages ; they are diagram-
matic, but accurate in respect of the number of sete present.
The embryo is supposed to have been slit open and flattened,
and all the sete on one side of the body areshown. V. V. and
D. D. represent the median, ventral, and dorsal lines.
These figures show that the sete develop, asa rule, in couples,
that the most ventral couple is the earliest to develop, that the
couple which become the most dorsal follow next, and that
then with considerable regularity ventral and dorsal couples
appear alternately; further that from a very early stage the
appearance of the sete becomes retarded in Segment 11, a
retardation which afterwards extends to the next two or three
segments (in stages later than those figured).
All the sete in figs. 26—28 are embryonic sete, no perma-
nent having yet appeared even in fig. 28.
In Mahbenus and in Perionyx the order in which the sete
appear is different; the most ventral appear first, then those next
to them, and so on, the most dorsal appearing last. I expect
that this represents a more modified condition; in P. pellucida
each segment passes in the condition of its sete through a
stage which remains as the permanent condition in an octo-
cheetous form.
In Moniligaster, Acanthodrilus, and Lumbricus the earliest
seta to develop on each side in each segment is the seta which
becomes seta ] (i.e. the most ventral), the next to develop is
seta 4, the next seta 2, and then seta 3.
Adult Condition and Formation of the Permanent
Setz.—The sete in the adult P. pellucida in any segment
lie at irregular intervals from one another, and the number of
the sete in a ring varies to a small extent. The actual num-
ber of sete in a particular specimen lying on one side of the
body in Segments xxv—xxvi1 are shown in fig. 29, the relative
distance between one seta and another is accurately shown,
DEVELOPMENT AND ANATOMY OF SOME EARTHWORMS. 23
but the sete themselves are magnified. It is very rare in the
adults of this species to find young sete ready to replace old
ones which drop out, and I believe that the sete shown in
this figure were the “‘ permanent” sete formed in the embryo.
I term them permanent setz in contradistinction to the em-
bryonic setze which precede them. The permanent sete deve-
lop at regular intervals from one another, except for the dorsal
and ventral gaps, and a greater number are formed than occur
in the adult, so that some drop out, leaving those which remain
at irregular intervals from one another. The number of those
that drop out appears to be nearly constant.
The permanent setze begin to appear on either side of the
dorsal and ventral median lines, one in the immediate neigh-
bourhood of each embryonic seta (fig. 30); the last to appear
are the most laterally placed ones. If 50 sete develop in the
ring (the largest number I have found in this species) seta 1
develops first, then seta 25 on each side, then in fairly regular
order, 2, 24, 3, 23, and so on. They always develop first in
the most anterior seta-bearing segment (11), and then very
regularly backwards. They always begin to appear before the
embryonic sete are visible in the most posterior segments. I
give three examples:
1. Embryo 25 mm. long, 160 segments counted, with a
growing tail in which the segments could not be counted when
mounted whole, embryonic setz visible down to Segment Lxv,
but mere dots after xxx, and four pairs only present in the
hindermost of the segments in which setz are already visible
(say Lv to Lxv); the permanent sete are just beginning to
develop, and are visible in Segments 11—V (see fig. 31).
2. Embryo 30 mm. long, embryonic setz visible down to
Segment ci11, the number of embryonic and permanent setz
in the anterior segments are as under :
24. ALFRED GIBBS BOURNE.
Embryonic sete. Permanent sete.
Segment 11 : : a ; : 35
5 III : ; 1D: ee : : 40
Bs IV : : 19 «tC : ; 40
3 Vv “ - Zou ts : : 50
- VI 3 ; Biey ; : 36
Jai By at : ; a9... : ; 27
5) WIIL ; : 40. ; ; 26
55 IX . ‘ 48 ‘ ; 12
ss x 5 : 52a ‘ ; 2
ads XI : . 44, 5 3 0
The seta rings of Segments 11, 111, and 1v of this worm are
shown in fig. 32.
3. Embryo 40 mm., embryonic setz down to CLxXxxvi, the
number of embryonic and permanent sete as under:
Embryonic sete. Permanent sete.
Segment 11 . : 1 ie : : 32
s III ; : YB) oe ; : 42
a IV ‘ 4 40. : : 48
aI Vv : : 30... ‘ : 32
. VI
i bi . 40 t0 50. . few and very small
” Ix
= x and onwards . ; 4 : none
This embryo was much more backward in regard to its seta
development ; it had not lost so many embryonic sete nor deve-
loped permanent sete to such an extent as had the others,
NEPHRIDIA.
Development of the Nephridia in Mahbenu
imperatrix.
I propose to deal here with certain stages only in the de-
velopment of these organs—the stages which have a special
bearing upon the so-called “ plectonephric” condition. I deal
chiefly with Mahbenus imperatrix. All my observations on
the early stages corroborate those of Vejdovsky, and the later
DEVELOPMENT AND ANATOMY OF SOME EARTHWORMS. 25
stages now described resemble very closely those described by
that author in dealing with Megascolides australis
(‘ Archiv fiir mikr. Anat.,’ Bd. xl).
In an embryo such as that from which figs. 5—8 and 33
are taken it is quite clear that the nephridia arise as paired
structures, a pair in every segment except perhaps the first.
Nephridia 7 and 16—19, from the right-hand side of this em-
bryo, are shown enlarged in figs. 34, 35.
Each consists of a preeseptal funnel, a neck connecting the
funnel with the glandular loop, and an excretory duct.
The funnel is at no stage well developed, and is probably
never functional, and afterwards entirely degenerates.
This neck becomes afterwards a very important structure,
and is dealt with below.
The glandular loop arises by budding from the neck region
and rapidly enlarges, ductules develop within it, and it becomes
a very complicated structure, as shown in fig. 35 and in outline
in fig. 39. It certainly corresponds to a macronephridium of
Megascolides, but its further development becomes arrested,
and I have been unable to distinguish it in the adult from the
loops of the micronephridia which subsequently appear.
It is important to note that owing to the imperfect state of
development of the septa these loops are by no means confined
to their own segments.
The excretory duct also arises from the neck region asa solid
outgrowth (fig. 34, neph. 17; and fig. 37, ex. d.) ; it very soon
acquires a lumen and opens to the exterior. It elongates
rapidly, more than keeping pace with the body-wall] in its
growth, and the aperture comes to lie very dorsally; the ex-
cretory pores of all the nephridia in front of nephridium 17 lie
outside the preparation in fig. 33.
The series figs. 836—39, taken from an older embryo, traces
the further development of the nephridia ; the figures represent
the 7th, 55th, 75th, and 86th nephridia respectively. Fig. 39
represents, therefore, a later stage in the development of the
nephridium of fig. 35, but figs. 836—88 represent stages in the
development of nephridia which did not exist in the embryo of
26 ALFRED GIBBS BOURNE.
fig. 33, and show that the mode of development alters in the
nephridia which are late in appearing.
I have been unable, owing to the advanced stage of develop-
ment of all the other structures, to fully trace out the nephri-
dium of fig. 39; but there is sufficient to show that the neck
has undergone great elongation, so that the glandular loop lies
very dorsally. The loop itself has become very complicated,
and the excretory duct extends away beyond the portion
drawn. The great importance of the stage is that it shows
the developments which are taking place in the neck region.
The cells here are giving rise to the secondary loops (z. 1, x.
2, n. 3).
Fig. 88 shows that in a nephridium developing at this late
stage the growth in the neck region commences at a relatively
much earlier stage. The neck is here already much elongated,
while the primary loop is still very undeveloped and the
excretory duct has not yet acquired its proper lumen. It is
clear that the primary loop is the earliest to develop of a
series, and that the whole structure is the homologue of the
nephridium of Lumbricus, and, as these secondary loops form
or give rise to all the scattered nephridia on the one side of
each segment, these are taken as a group, but not individually,
homologous with the Lumbricus nephridium.
Fig. 40 is taken from an old embryo ; 2 2 is a portion of the
original neck region, now much attenuated; a to f are secondary
loops which have arisen from it; of these @, at any rate, has
acquired an excretory pore. These loops show that the same
sort of difference obtains between a secondary loop which de-
velops late and one which develops early as between a ne-
phridium which develops late and one which develops early.
The secondary loops give rise to tertiary loops as outgrowths
from their own neck region. In this way some fifty or more
loops develop which ultimately become separated from one
another, while each develops its own excretory duct and becomes
a micronephridium.
There is no elongation in the neck region and no develop-
ment of secondary loops in the nephridia of the most anterior
DEVELOPMENT AND ANATOMY OF SOME EARTHWORMS. 27
segments ; the whole structure aborts in these segments, so that
in the adult no nephridia occur in them.
I find absolutely no trace of provisional nephridia from a
gastrula stage onwards, though, of course, in a sense all the
loops which first develop are provisional, and I expect that all
so-called provisional nephridia will be ultimately explained by
the fact that the mode of development and ultimate structure,
and even continued existence of a nephridium, depends upon its
time of development; and, further, that the nephridia which
develop early are not confined by septa to their own segments.
This becomes very clear in large embryos like those of Mahbenus.
Fig. 41 shows as much of the anatomy of a micronephridium
of an adult as I have been able to make out; the whole struc-
ture as drawn is 0°25 mm. long.
Development of the Nephridia in some other
Worms.
In P. pellucida the nephridia develop in the same way as
in Mahbenus, certain details excepted, as that the funnel
becomes better developed up to a certain stage, up to such a
stage as fig. 35. The loop has then the same long wandering
character. At this stage all resemblance ends. The loop
becomes broader in proportion to its length, no elongation
occurs in the neck, no secondary loops are formed, and each
segment in the adult never possesses more than a single pair of
nephridia. The five most anterior pairs, after having attained
a well-developed condition, degenerate and entirely disappear.
The most anterior pair of nephridia in the adult belong to
Segment VII.
The nephridia of Segments vii—x1 undergo further modifi-
cation until their structure somewhat resembles that of the
nephridia described by Benham in Microcheta. From a por-
tion of the ordinary loop a number of outgrowths form, into
which the tubules run in a very complicated manner, and this
bunch of outgrowths ultimately form by far the largest portion
of the nephridium.
Fig. 42 is taken from a fortunate preparation obtained by
28 ALFRED GIBBS BOURNE.
macerating a portion of a full-sized embryo in nitric acid. I
found it impossible to make out all the details. At d. is seen
the excretory duct, composed of a very regularly arranged series
of drain-pipe cells; 6. is the apex of the lobe; at ec. the
arrangement of the tubules is very characteristic,—there is a
central tubule and a double set of convoluted tubules, shown in
the figure by single lines ; at a. is the bunch of outgrowths, each
one of which subsequently elongates to a considerable extent.
In a Madras species of Acanthodrilus which has scattered
nephridia in the adult I have found the nephridia to develop
as paired organs, one pair to each segment, which bears out
Beddard’s observations upon this genus (this Journal, vol.
XXxiii, part 4). I have inserted fig. 43 of the 17th nephridium
from a 10 mm. long embryo of this species, as it showed with
absolute clearness the exact course of the greater portion of the -
ductule from the excretory duct d. up to 6.; from thence on-
wards to the funnel duct the ductule was too fine to be traced.
I have not yet obtained stages of this species showing the
development of the micronephridia, but as the process seems
to be so similar in such widely different forms as Mahbenus
imperatrix and Megascolides australis it is probably
the way in which all micronephridia develop.
The So-called “ Plectonephric” Condition.
My own observations, those of Vejdovsky, and in a less
direct way those of Bergh, Wilson, and others who have dealt
with development of the nephridia in meganephric forms,
and even those of Beddard himself (on the development of
A. multiporus), throw grave doubt upon the conclusions
arrived at by Beddard and Spencer with regard to this matter.
Apart from this I have for some years, upon anatomical grounds,
doubted the existence even, of such a condition of the nephri-
dium as was described by Beddard for P. aspergillum. For
one thing, Beddard’s figs. 7 and 10 (this Journal, vol. xxviii,
Pl. XXX) do not seem to me to prove what he would
have them prove; fig. 10 contains an impossible blood-vessel,
branching and returning blood into itself, which looks very
DEVELOPMENT AND ANATOMY OF SOME EARTHWORMS. 29
much as though there had been some confusion between
nephridial tubule and blood-capillary, while fig. 7 hardly
shows a continuity of nephridial tubule from segment to seg-
ment. Further, both Beddard’s and Spencer’s figures, espe-
cially the latter, clearly show that the tubules possess a different
character in different regions, and that they are therefore much
more specialised than the tubules of Pontobdella; and in no
case can one be certain of the continuity of such fine tubules
from an examination of sections alone.
For my own part, (1) in spite of repeated and most careful
search in preparations from so-called plectonephric worms made
in all sorts of ways, I have never seen any connection between
one nephridium and another in the adult. If, however, the
mode of formation of the micronephridia is always such as I
describe for Mahbenus, it is very possible that there are forms
in which all or any of the micronephridia on the same side of
any segment may remain connected together ; but I very much
doubt whether there is ever any continuity from side to side or
from segment to segment.
(2) I have in a large worm like Mahbenus imperatrix
been able to count the micronephridia belonging to a definite
area, and then to mount the cuticle which showed a corre-
sponding number of pores.
(3) The micronephridia have always a complicated structure,
similar to that of a meganephridium. This is the fact which
first led me to doubt the plectonephric condition, and I have
worked out with great trouble as much of the structure of a
micronephridium as is shown in fig. 44, This nephridium
belongs to Pericheta mirabilis. The micronephridia in
this worm are very numerous, and have, or some of them at any
rate have, funnels ; these funnels are not preseptal, and present
eight marginal cells, ciliated on their centrally directed faces
and arranged in a horseshoe fashion, and one central cell, but
no other cells. The funnel is only 0°05 mm. in diameter.
The coils of tubule are shown in the figure. Fig. 41 shows
a similar complicated condition for Mahbenus, and I have
obtained similar results in all the micronephridia which I
30 ALFRED GIBBS BOURNE.
have examined, so that even if they were all connected together
into a network and did not develop, as they appear to do, we
should have a condition very different from that of Pontobdella,
and much more nearly connected with a meganephric condition.
(4) I knew that in Pericheta pellucida and some other
species which, although not to be placed, strictly speaking, in
the same genus as P. aspergillum, P. mirabilis, &c., are
very closely allied forms, one pair only of nephridia were to
be found in any segment, which rendered it, at any rate, un-
likely that anything so fundamentally different from the mega-
nephric condition as the plectonephric condition would occur.
(The existence of Perichetes with meganephridia reopens the
question of the systematic position of Perionyx.)
I think that the condition of the nephridium in Perichetes,
Acanthodrilus, and many other genera must have arisen from
the meganephric condition. That the funnel appears, as in
Mahbenus and Megascolides, only to disappear, and that the
loop which appears earliest in the nephridia which develop
first and attains such great complication, only to be arrested in
its development or even to abort, to set aside all other con-
siderations, is very strong evidence that the development of
micronephridia by budding from it, is a specialised condition.
Taking into account what we know of the development of
the nephridia in other leeches and the presence of funnels (if
M. Bolsius will allow me to call them so), so seemingly out of
keeping with the rest of the nephridium in Pontobdella, I
shall not be surprised to learn from a history of its develop-
ment that that is by no means a primitive structure, and has
possibly no genetic relationship with the tubules of Platy-
helminths.
I am unable at present to bring forward any direct evi-
dence as to the relationship between genital ducts and
nephridia in earthworms, but I cannot refrain upon this
occasion from pointing out that the demonstration of the fact
that a so-called plectonephric condition is not a primitive one
removes the strongest objection to the theory brought forward
by Lankester in one of his earliest contributions to this Journal.
DEVELOPMENT AND ANATOMY OF SOME EARTHWORMS. 31
EXPLANATION OF PLATES 2—5,
Illustrating Professor A. G. Bourne’s paper “On Certain
Points in the Development and Anatomy of some Earth-
worms.”
PLATE 2.
Fies. 1—8.—From Mahbenus imperatrix, fully described in the text.
Fig. 1. W, m. Primary “ mesoblasts.” Zp. Epidermis. al. ep. Alimentary
epithelium (the cells fully shown here and in Fig. 8, partially so only
in Figs. 2—7). Muse. Cells which give rise to the muscle of the
body-wall (their presence here at this stage will be discussed in a
subsequent paper). «. marks a corresponding spot in Figs. 1—3.
Fig. 2, M. Row of cells growing from the primary “ mesoblast.” Other
letters as before.
Fig. 3. Letters as before.
‘Fig. 4. n. Neuroblast. . Cells forming nerve-cord. Sept. Septum.
Cel. ep. Celomic epithelium. The line A B in this and Figs. 5
and 6 marks a corresponding level. Other letters as before.
Fig. 5. a, 4, c. Row of cells giving rise to primary setal matrices, as
s., 8. Neph. Nephridial matrix.
Fig. 6. s. Continuation backwards of the row a, 4, c, of the preceding
figure. Neph. Nephridial matrices; the actual nephroblast does not
come into this section.
Fig. 7. Letters as before.
Fig. 8. Letters as before.
PLATE 3.
Fics. 9—15.—Moniligaster (probably M. sapphirinaoides).
Fig. 9. Portion of an embryo flattened out. Neph. Young nephridia.
a. Portion of the nerve-cord. Sept. Septa. yp. m. Primary setal
matrices. 7. m. Secondary setal matrices in process of formation.
S. m. 4. and §. m. o. Inner and outer secondary setal matrix. Sed.
Rudiment of a seta. wz. Series of large cells of unknown significance.
Figs. 10—13, Transverse sections, described in the text. p.m. Primary
setal matrix. §. m.7. and §. m.o. Inner and outer secondary setal
matrix. NVeph. Nephridium. Zp. Epidermis. 2. Epidermic ingrowth.
Fig. 14. Surface view of an advanced secondary setal matrix. se¢., set.
The two séte of a couple.
Fig. 15. Similar structure from a longitudinal section.
32 ALFRED GIBBS BOURNE.
Fic. 16.—Urocheta sp. Secondary setal matrices of one side of the
five segments a—e. v.v. Ventral median line. Ant. and Post. mark
the anterior and posterior regions.
PLATE 4.
Fies. 17—32.—Pericheta pellucida; Fig. 33. Mahbenus im-
peratrix.
Figs. 17—21. Secondary setal matrices dividing before the appearance
of sete. §. m. Secondary matrices. 7. m. Intermatricial bands.
Cel. ep. Celomic epithelium. ep. HEpidermis.
Fig. 22. Portions of the setal matrices of two consecutive segments with
developing sete. In the lower portion of the figure which is anterior
two new setz are shown developing between an older pair.
Fig, 23. The entire setal matrices on one side of the body in three con-
secutive segments. V. Ventral end. D. Dorsal end. Sete are seen
developing in the most ventral and most dorsal regions, the former
being the elder.
Figs. 24 and 25. Later stages in the development of the secondary
matrices.
Figs. 26—28. Diagrams showing the exact number of sete on one side
of the body in the segments marked 11, 111, Iv, &., in three embryos
of different ages. V.V.and D. D. mark the ventral and dorsal median
lines.
Fig. 29. A similar diagram taken from an adult worm.
Fig. 30. Portion of a matrix in which permanent sete are beginning to
develop alongside the embryonic sete.
Fig. 31. The complete seta ring of Segment 1 of an embryo 25 mm. long.
The long fine lines represent the embryonic sets, the short thick lines
the permanent ones.
Fig. 32. The complete seta rings of Segments 11, 111, and Iv of an older
embryo from which the embryonic sets have nearly all dropped in
these anterior segments; those left are shown by the small fine lines, of
which there are 6, 12, and 19 in the different segments.
Fig. 33. An embryo of Mahbenus imperatrix showing the entire
series of nephridia, The references Weph. 1, Neph. 2, &c., are con-
nected with the excretory ducts; the excretory pores themselves are
within the limits of the preparation in the case of nephridia 7 back-
wards. S. m. Setal matrices. Sep/. Septa torn during the removal
of the alimentary wall.
DEVELOPMENT AND ANATOMY OF SOME EARTHWORMS. 30
PLATE 5.
Figs. 34—41. Mahbenus imperatrix; Fig. 42. Pericheta pel-
lucida; Fig. 48. Acanthodrilus sp.; Fig. 44. Pericheta mirabilis.
Figs 34 and 35. Enlarged views of some of the nephridia from the prepa-
paration, Fig. 338. Letters as before.
Figs. 36—39. Series of nephridia from an older embryo of Mahbenus.
Jun. Funnel. NK. Neck region. 7. Glandular loop. Za. d. Excre-
tory duct. w. 1, 2. 2, 2. 3. Micronephridia beginning to develop.
Fig. 40. From a still older embryo. 2. x. Portion of the neck region of
anephridium such as Fig. 39. a, 0, ¢, d, e, f. Micronephridia develop-
ing from it. g. Micronephridium developing from a. Zz. d. Excretory
duct. Hz. p. Excretory pore. s. m. Setal matrices. 7. m. Inter-
matricial band.
Fig. 41. A micronephridium from an adult Mahbenus.
Fig. 42. A nephridium from one of the Segments vir—x1 of a full-sized
embryo of P. pellucida.
Fig. 43. A nephridium from an embryo of Acanthodrilus sp. B. W.
Body-wall. Sepé. Septum. fuz. Funnel. a. Duct from funnel. From
6 to d the exact course of the duct is shown. d passes downwards to
the excretory pore.
Fig. 44. A micronephridium from an adult Pericheta mirabilis.
fun. Funnel (not preseptal). Av. d. Excretory duct.
VoL. 36, PART 1.—NEW SER. Cc
t
.
. wat . e
e ie
' -
" Pe« : i
s I
ma ‘ . :
a: "a
- ie ”
- iis he Z
bes
* — A : .
; ; C> |
; - ons
a os
- es . . a
» aa : . ‘2 Tia
7 . a a ma o a s
D ree ee rr Ie"
’ ee I a 7?
a 1 oww fy
— Le th
ed gpl’
ee ; i 4: ’
2 7 a ;
pia see
LAW OF DEVELOPMENT KNOWN AS VON BAER’S LAW. oo
On the Law of Development commonly known
as von Baer’s Law; and on the Significance
of Ancestral Rudiments in Embryonic De-
velopment,
By
Adam Sedgwick, M.A., F.R.S.
THe generalisation commonly referred to as v. Baer’s law
is usually stated as follows:—Embryos of different members
of the same group are more alike than the adults, and the
resemblances are greater the younger the embryos examined.
It appears to be pretty clear that v. Baer held some such view
as this,' and there can be no doubt that it is a view which is
largely taught at the present day.” In fact, I think it is safe
to say that all zoologists are brought up with this view as one
of the fundamental postulates of their science.
It will be the object of the following pages to show that this
view is not in accordance with the facts of development.
V. Baer based his view mainly upon the study of Vertebrate
embryos; and it will be convenient for us, in criticising him,
to begin with an examination of this group. We may also,
in the first instance, follow v. Baer in another point, viz. in
limiting ourselves to the consideration of embryos as opposed
to larve. Embryonic development and larval development
take place under entirely different conditions, and in order to
obtain clear ideas they must be considered apart in treating
1 Vide ‘Ueber Entwickelungsgesch. d. Thiere,’ part i, pp. 221, 223,
and 224,
2 Vide Darwin in ‘Origin of Species,’ p. 364, 6th ed.; and Balfour
‘Comp. Embryology,’ vol. i, p. 2.
36 ADAM SEDGWICK.
this subject. They have not as arule been clearly distinguished,
and facts based on a study of larve have sometimes been
assumed to hold true for embryos without further examination;
and this practice has, as I hope to show, given rise to errors
which have prevented our arriving at a clear understanding of
the subject.
It is, of course, impossible to examine fully all the stages of
all Vertebrate embryos. In the comparison necessary for the
criticism I am making it will be convenient to limit ourselves
to typical cases, and I propose to consider (1) the embryos of
animals widely divergent ; (2) the embryos of animals which
are closely allied in the adult state. If it is found that in
neither of these cases is the law of v. Baer followed, then I
think we may reason that there is, to say the least of it, a
strong probability that it will be found not to hold true for
intermediate cases.
(1) Embryos of divergent classes of the Craniate phylum.
The examples I have chosen are the fowl and dog-fish.
The fowl and the dog-fish in the adult state live under
entirely different conditions ; whereas in the embryonic phases
the conditions are very similar, both being developed within
an egg-shell at the expense of ovarian yolk and surrounding
albumen.
According to the law of v. Baer these embryos ought to be
closely similar in the young stage.
Do these embryos, developing under similar conditions, con-
form to the law? Superficially, clearly not. There is no
stage of development in which the unaided eye would fail to
distinguish between them with ease—the green yolk of the
one, the yellow yolk of the other; the embryonic rim and
blastopore of the fish, the absence of these in the chick ; the
six large gill-slits bearing gills on the one hand, the four
rudimentary clefts on the other; the small head, straight
body, and long tail, as opposed to the enormous head, cerebral
curvature, short tail, and so on. A blind man could distin-
guish between them.' These embryos are not closely similar,
1 J do not feel called upon to characterise the accuracy of the drawings of
LAW OF DEVELOPMENT KNOWN AS VON BAER’S LAW. 37
but it is maintained that the law is justified by certain remark-
able features of embryonic similarity which the adults do not
exhibit, and of which the most important are the presence in
the chick of pharyngeal clefts, a tubular piscine heart, and a
similarity in the arrangement of the cardiac arterial system,
a cartilaginous endo-skeleton, oro-nasal grooves, and a noto-
chord. NowlI freely admit that these are striking similarities,
but I question whether they are sufficient to justify the law of
v. Baer. By themselves, no doubt, they would be sufficient to
justify that law ; but are there no differences to set off against
them ? Are there no differences of a morphological value, as far-
reaching and as striking as these similarities? Let us clearly
understand the question at issue. V. Baer’s law, as applied to
the present case, may fairly be held to mean, if it has any
meaning at all, that whereas the differences between the adults
are large and important differences of class value, the differ-
ences between the embryos are slighter and unimportant, and of
less than class value. Now in no single member of the group
Craniata is the mesoderm of the head segmented. According
to our present morphological knowledge, the discovery of an
animal with cranial segments would be a very remarkable one,
and would, we might confidently predict, require the establish-
ment of a class at least separate from all other Craniate
classes—such is our estimation of the importance of this
feature. And if to this character was also added the presence
of a coelomic sac close to the eye, of another in the jaw, and of
a third near the ear; of an aperture of communication between
the neural canal and rectum, of kidney tubules opening into
the muscle-plate coelom as well as into the perivisceral ccelom,
of a Millerian duct opening into the front end of the Wolffian,
I do not think that any anatomist would have any doubt about
the matter. Now it is precisely in these points, amongst
embryos of different classes of the Vertebrata given by Haeckel in his popular
works, and reproduced by Romanes and, for all that I know, other popular
exponents of the evolution theory. As a sample of their accuracy, I may
refer the reader to the varied position of the auditory sac in the drawings
of the younger embryos,
38 ADAM SEDGWICK.
others, that the embryo dogfish differs from the embryo chick.
I am quite aware that arguments tending to reduce the
classificatory value of the embryonic differences I have just
enumerated might with more or less plausibility be adduced.
But one thing at a time. I am at present dealing solely with
the importance of the anatomical resemblances and differences
between the embryos; and I think I have shown, as far as it
can be shown, that they have, if judged by standards used
when comparing adults, at least as great an importance as the
resemblances between the same embryos—the differences, like
the resemblances, relating solely to the embryos, and not
existing in the adults. V. Baer’s law then falls to the ground,
and must be replaced by another law, which is as follows :—
Embryos of different members of the same group often resemble
one another in points in which the adults differ, and differ from
one another in points in which the adults resemble; and it is
difficult, even if possible, to say whether the differences or the
resemblances have the greater zoological value (because we
have no clearly defined standard of zoological value).
It will probably be urged here by my reader—Are you not
beating the air in a vain warfare of words and unessentials of
which we were all aware, and trying to kick up a cloud of dust
by which to obscure the essential point, viz. that embryos
pass through, incompletely if you like, stages of structure
permanent in lower members of the same group? To such a one
I reply, that I am as keenly alive to the importance of the
essential point as he is, but that I differ from him in being
dissatisfied with the explanation which is at present given of it,
and that I am convinced that the form in which this pheno-
menon is referred to in v. Baer’s law has led to the acceptance
of an imperfect explanation of the embryonic phase in animal
development.
(2) But before I come to that point I have to consider
the case of the embryos of closely allied animals. If v.
Baer’s law has any meaning at all, surely it must imply that
animals so closely allied as the fowl and duek would be indis-
tinguishable in the early stages of development; and that in
LAW OF DEVELOPMENT KNOWN AS VON BAER’S LAW. 39
two species so closely similar that I was long in doubt whether
they were distinct species, viz. Peripatus Capensis and
Balfouri, it would be useless to look for embryonic differ-
ences: yet I can distinguish a fowl and a duck embryo on
the second day by the inspection of a single transverse section
through the trunk, and it was the embryonic differences between
the Peripatuses which led me to establish without hesitation the
two separate species. But it is not necessary to emphasise
further these embryonic differences ; every embryologist knows
that they exist and could bring forward innumerable instances
of them. I need only say with regard to them that a species
is distinct and distinguishable from its allies from
the very earliest stages all through the develop-
ment, although these embryonic differences do not
necessarily implicate the same organs as do the
adult differences.
If I have laid great stress—some may think undue stress—
upon the inadequacy of v. Baer’s law, I have done so be-
cause of the importance which is at the present day attached
to this law by teachers of zoology. In support of this, | may
quote the words of three of the greatest teachers of zoology of
this or of any other age—words which show that they at any
rate considered that the law correctly represented the facts.
Darwin, in the ‘Origin’ (p. 387, 6th ed.), says:—‘‘So
again it has been shown that generally the embryos of the
most distinct species belonging to the same class are closely
similar, but become, when fully developed, widely dissimilar.
A better proof of this latter fact cannot be given than the
statement by v. Baer that the embryos of Mammalia, of
birds, of lizards, and snakes, probably also of Chelonia, are in
their earliest states exceedingly like one another, both as a
whole and in the mode of development of their parts ; so much
so, in fact, that we can often distinguish the embryos only by
their size. In my possession are two little embryos in spirit,
whose names I have omitted to attach, and at present I am
quite unable to say to what class they belong. They may be
lizards or small birds, or very young Mammalia, so complete
40 ADAM SEDGWICK.
is the mode of formation of the head and trunk in these
animals.”
This, I think, shows quite clearly Darwin’s view of the
matter,
Huxley, in his ‘Man’s Place in Nature,’ says :—“ The his-
tory of the development of any other Vertebrate animal—lizard,
snake, frog, or fish—tells the same story. There is always, to
begin with, an egg having the same essential structure as that
of the dog; the yolk of that egg undergoes division or seg-
mentation, as it is called, the ultimate products of that seg-
mentation constitute the building materials for the body of the
young animal; and this is built up round a primitive groove,
in the floor of which a notochord is developed. Furthermore,
there is a period in which the young of all these animals
resemble one another, not merely in outward form, but in all
essentials of structure, so closely, that the differences between
them are inconsiderable, while in their subsequent course they
diverge more and more widely from one another.”
Tue SIGNIFICANCE OF ANCESTRAL RUDIMENTS IN
Emsryonic DEVELOPMENT.
The existence of a phase at the beginning of life during
which a young animal acquires its equipment by a process of .
growth of the germ, is of course intelligible enough. We see
such a phase in the formation of buds, and in the sexual repro-
duction of both animals and plants. The remarkable point is
that while in most cases this embryonic growth is a direct and
simple process—e.g. animal and plant buds, embryonic de-
velopment of plant seeds—in some cases—e.g. most cases of
sexual reproduction of animals—it is a circuitous one, and the
embryonic phase shows stages of structure which seem to
possess a meaning other than that of being merely phases of
growth.
As is well known, the explanation which is given of this
circuitous course of embryonic development is that we are
dealing with a special case of the law of heredity— each
LAW OF DEVELOPMENT KNOWN AS VON BAER’S LAW. 4]
organism reproducing the variations inherited from all its an-
cestors at successive stages in its individual ontogeny”
(‘Comp. Emb.,’ vol. i, p. 3).
“These two principles, namely, that slight variations gene-
rally appear at a not very early period of life, and are inherited
at a corresponding not early period, explain, as I believe, all
the above specified leading facts in embryology.” (Darwin,
‘Origin,’ p. 392, ed. vi.)
But this explanation, though good as far as it goes, is not
entirely satisfactory, because it fails to explain (without further
qualifications) the majority of cases (animal and plant buds,
embryonic development of seeds) in which ontogeny presents
no ancestral traces; it is at variance with the fact that in
many cases variations which affect the adult have affected the
whole of embryonic development (see below) ; and it does not
enable us to understand why some organs, e.g. gill slits, have
been retained in embryogeny, whereas other organs which have
much more recently disappeared, e.g. teeth of birds, fore-limbs
of snakes, have been entirely lost. It assumes that the repe-
tition of ancestral characters in embryogeny is the intelligible
rule; and that their omission is the exception which requires
explanation whenever it occurs. This assumption is not
warranted by the fact above indicated that in the vast majority
of ontogenies there are no phylogenetic traces, nor by the
consideration that a number of important organs, such as
teeth and hand-claws in birds, limbs in snakes, gill-clefts in
fishes, have recently disappeared without leaving a trace in
ontogeny.
In fact the balance of evidence appears to me to point most
clearly to the fact that the tendency in embryonic development
is to directness and abbreviation and to the omission of an-
cestral stages of structure, and that variations do not merely
affect the not-early period of life where they are of immediate
functional importance to the animal, but, on the contrary, that
they are inherent in the germ and affect more or less pro-
foundly the whole of development.
I am well aware that in holding this opinion I am running
42 ADAM SEDGWICK.
counter to the great authority of Darwin. In the chapter
from which the above quotation was taken he gives many facts
and arguments in favour of the view that slight variations
generally appear at a not very early period of life, and are
inherited at a corresponding not early period. He admits that
larger variations—monstrosities—do affect the embryo at a
very early period, but he thinks that slight variations do not.
Without considering the difficult question as to where the line
should be drawn between a slight variation and a monstrosity,
I may merely point out that Darwin’s evidence is largely based
upon the experience of breeders that it is impossible to tell
until some time after birth what will be the merits or demerits
of their young animals. In mitigation of the force of this fact
it must be-remembered that a successful breeder is a highly
skilled man—that he possesses powers of observation greater
than the ordinary—that his success depends upon his ability
to see points which escape the eye of other people. If the
points for which the mature animal is selected are thus
ditficult of observation, can we wonder if it is beyond the
power of man to see them when the animal is immature, and
the relative sizes of the parts of the animal, and its whole
appearance, are so different? In support of this way of looking
at the matter I would urge that when the variation is large
and of a nature to be easily observed, it can—in a great
many instances at any rate—be detected all through develop-
ment.
The evidence is of this kind :—(1) Organs which we know
have only recently disappeared are not developed at all in the
embryo. For instance, the teeth of birds, the fore-limbs of
snakes, reduced toes of bird’s foot (and probably of horse’s
foot), the reduced fingers of a bird’s hand. These are instances
which readily occur; I have no doubt that many others might
be quoted by anyone giving attention to the matter. (2) Organs
which have (presumably) recently become reduced or enlarged
in the adult, are also reduced or enlarged in the embryo.
Many examples of this might be given, and it is a most im-
portant point as showing the manner in which variations have
LAW OF DEVELOPMENT KNOWN AS VON BAER’S LAW. 43
affected the whole of embryonic development. As examples
I may mention the small outer toes on the feet of the pig
and probably of other Ungulates, the large digit of the ostrich’s
foot and of the kangaroo’s foot, the spiracle of Elasmobranchii,!
the rudimentary character of the phalanges of the bird’s hand.
I have no doubt that many other instances will occur to my
reader. (3) Organs which have been recently acquired may
appear at the very earliest possible stage; e.g. the double
hallux present in some breeds of fowls makes its appearance
as soon as the other digits; the webbing of the duck’s foot
is not preceded by a stage in which the digits are separate.
In short, the evidence seems to indicate that in a great number
of cases adult variations of any part are accompanied by
precedent similar alteration of the same part in the embryo.
I do not mean to affirm that the alteration of the organ in
' The spiracle of Elasmobranchii is a reduced gill-cleft, and in correspond-
ence with its reduction in the adult it is found to be reduced also in all
stages of its development from its very first appearance, which takes place
after the hyobranchial cleft—not before it, as would be expected from its
position as the anterior member of a series (see Self, “‘ Notes on Elasmobranch
Development,” this Journal, vol. xxxili, p. 572). It would be excessively
interesting in this connection to ascertain whether any trace of the spiracle
is present in the embryos of those Elasmobranchs in which it is absent in the
adult. In fact, an account of the spiracular cleft throughout the Vertebrata
is much needed. Is it present in embryo in Teleosteans and in Lepidosteus ?
Balfour asserts that it is present in the former (‘Comp. Embryology,’ vol. ii,
p. 77, mem. ed.), but I am unable to find his authority for the statement.
He also states that it is present in the embryo Lepidosteus as a double layer
of cells without a cavity (Balfour and Parker on ‘ Anatomy and Development
of Lepidosteus,’ mem. ed., pl. xxxvii, fig. 43), but, so far as I am aware, his
interpretation of this structure has not been confirmed. In Amniota—through-
out which the pharyngeal clefts present a very remarkable constancy—the
spiracular cleft is as large, if not larger, than the succeeding one, and appears
first in development. ‘This is an exceedingly interesting fact, which has not
been sufficiently noted. It tends to show that the Amniota have arisen from
aquatic forms independently of the terrestrial Amphibia, in which group the
spiracular cleft is not formed at all—though a slight rudiment of it does
appear for a short period. In fact, we may take it as a fact of systematic
value that the spiracular cleft is absent or rudimentary in all Ichthyopsida
while it is present in relatively normal development in all Amniota,
44, ADAM SEDGWICK.
the embryonic stage! is the same as it is in the adult. It may
be altered relatively more or it may be altered relatively less ;
the point is that it is altered in the same direction as the
adult organ. And this is surely what we should expect when
we remember that embryonic development is the preparation
of the free form in the most perfect state and at the least
expense. If this view is correct that variations are present in
the embryo—that an organ which is enlarged, diminished, or
suppressed in the adult is correspondingly, or nearly so, en-
larged, diminished, or suppressed in the embryo,—then I ask,
how are we to account for those cases which most undoubtedly
occur in which records of previous states of structure are
present in the embryonic history, e.g. the pharyngeal slits of
Sauropsids, the tubular heart, the vascular arches, the em-
bryonic kidney of the same group, and manysuch. The point
is this: organs which have been recently altered show a
similar alteration in the embryo, whereas some organs, like the
gill-slits, which must have been altered very far back, do not
show a corresponding embryonic alteration, but persist more
or less in their old form without discharging the original
functions or being of any use to the embryo. In other words,
some ancestral organs persist in the embryo in a functionless
rudimentary (vestigial) condition and at the same time without
any reference to adult structures, while other ancestral organs
have disappeared without leaving a trace. The latter arrange-
1 It appears that in some cases, at least, it is less in the embryo. LE. g.
sternal ribs of ostrich are generally five in adult, rarely six; in embryo,
they appear always to be six. In birds the fibula reaches the tarsus in
embryos, but very rarely does so in adults.
A case of this kind which might be investigated is tlis:—In the golden
plover the hallux is entirely absent, whereas in other plovers it is present.
Has the golden plover any trace of it in the embryo?
I am aware that it is often held—Darwin held it—that rudimentary organs
are, relatively to the adjoining parts, larger in the embryo than in the adult.
But unless this fact can be shown to be universal, it has but little value
because it applies to many other organs in the embryo which are not rudimen-
tary, e. g. brain, eye, heart, and kidney. This difference in relative size is
probably simply owing to the fact that the bulk of the skeletal, muscular,
and connective tissues of the embryo is relatively less than in the adult.
LAW OF DEVELOPMENT KNOWN AS VON BAER’S LAW. 45
ment seems to be the rule, the former the exception. How
are we to account for the exception? The key to the position
is, in my opinion, to be found by comparing the conditions of
larval and embryonic development. In larve the organs are
functional and the animal is getting its own living during the
development, whereas in embryos the development takes place
under the protection of egg membranes, the pupal case, or the
uterine wall, and the organs are for the most part functionless,
special arrangements being made for the supply of nutriment.
These two developments have generally not been properly dis-
tinguished by naturalists writing on this subject.
In embryos the organs are for the most part functionless
and without relation to the maintenance of life; consequently
there is nothing to counteract the tendency to the appearance
of a variation at all stages in the life of an organ. In larvae,
on the other hand, the organs are functional and the con-
ditions of life may be different from those of the adult. They
have to maintain themselves during the various phases of their
development ; consequently if a variation of an organ at one
stage is injurious to the same organ at a previous or subsequent
stage, it will be eliminated at the stages at which it is injurious.
In this way, as will be readily seen, natural selection will com-
pel the limitation of variations in an organ to particular stages
in the development of that organ; the power of natural selec-
tion will inevitably prevent a variation useful at one stage
from affecting another stage of the same organ in which its
presence would be injurious to the larva. Thus there must be
in larvee a tendency to the inheritance of variations at corre-
sponding periods, and to the elimination of them at other
periods when they would be harmful to the organism.
Thus it must happen that if variations occur which enable the
adult to change its condition of life, and if at the same time
the old habits of life are retained by the last larval stage, then
the old arrangement of organs will be retained by the larva.
In this way, as the adult form gradually progresses in evolution,
not only one but a whole series of larval stages might become
established, each one being based upon some ancestral stage of
46 ADAM SEDGWICK.
structure and retaining some ancestral habit of life. But of
course these larval stages are liable to vary and are subject
to the natural selection engendered by the struggle for life.
So they may themselves become modified and the ancestral
habits and structure which they have inherited may also become
modified. It thus becomes apparent that larvee will often retain
traces more or less complete of ancestral stages of structure,
and that they will do this in virtue of the operation of the
force of heredity and of natural selection. And the retention
of ancestral features by the larve will be the more complete the
more completely the ancestral habits of life are retained by
them. There is, then, in larve a tendency to the inheritance of
variations at corresponding periods, and in this respect larve
differ from embryos.
To sum up, I would maintain that ancestral stages of struc-
ture are only retained in so far as they are useful to the free-
growing organism, i.e. to the larva in its free development.
Or, to put the matter in another and more recondite form,
modifications appearing in and affecting the adult structures
will similarly affect the same structures all through the
development of the offspring unless the old structural arrange-
ments are called into being in the development of the offspring
by the application of the old stimulus, viz. the same external
conditions of life.
In embryos, on the other hand, the organs are for the most
part functionless, and there appears to be uo reason for the
retention of ancestral conditions of structure. On the con-
trary, as I have shown above, most organs when modified in
the free-living state are similarly modified in the embryo.
And, as I have already insisted, this is what we should expect
when we remember that embryonic development is the pre-
paration of the free form in the most perfect state and at the
least expense. How is it, then, that we do get in embryos in
certain cases a most remarkable preservation of ancestral
organs and conditions of structure which have been lost in the
adult? I think it can be shown that the retention of ancestral
organs by the larve after they have been lost by the adult is
LAW OF DEVELOPMENT KNOWN AS VON BAER’S LAW. 47
due to the absorption of a larval or immature free stage into
embryonic life.
Let us take an example. Let us try to picture to ourselves
the steps by which the tadpole stages of the frog might be lost,
so that the adult frog arose direct from the egg. The larval
organs of a tadpole cannot disappear one by one independently
of one another. If the gill slits disappeared before the heart
had become double and the lungs had developed, the tadpole
would die of asphyxia. In order to completely obliterate the
piscine stage from the tadpole, you require a number of
nicely co-ordinated variations affecting different organs in
very different ways—all tending to the atrophy of those
organs which adapt it to an aquatic life and to the
development of the organs required for terrestrial life.
Such a combination of suitable variations as is here re-
quired—such an inversion of the original evolutionary changes
—is very unlikely to occur,! especially when the same object
can be obtained, namely the obliteration of the piscine phase
in the frog’s life, by a simple single variation—that is to say,
by the mother becoming viviparous and retaining its young
within its uterus or oviduct until the piscine stage of develop-
ment has been passed through; or by the ovarian ovum de-
veloping a greater amount of yolk, so that the whole develop-
ment up to the close of the piscine stage can take place before
hatching at the expense of the yolk. That larval stages do
disappear and embryonic stages arise in this way is shown by
the case of the viviparous salamander (Salamandra atra), in
which the gills, &c., are all developed but never used, the
animal being born without them. Here, therefore, is an actual
case in which the larval phase has disappeared by becoming
embryonic and therefore functionless, and therefore largely
removed from the direct action of natural selection ; once em-
‘It has suggested to me here that this combination of variations must
have taken place in phylogeny, otherwise the terrestrial animal could not
have been evolved; why not then in the larva? To this I reply: there is no
necessity for the long and laboured changes to be gone over again in inverted
order in the case of the tadpole, because the object can be obtained by the
simple inclusion of the tadpole stage within the embryonic period.
48 ADAM SEDGWICK.
bryonic only, the conditions of its existence are totally altered.
Its disappearance is no longer a matter of importance to the
organism, because the embryo being protected from the struggle
for existence the presence of rudimentary functionless organs
is unimportant to it. They therefore persist, and it is this
persistence which has given rise to von Baer’s law. But von
Baer’s law is imperfect, because it omits to take cognisance of
the fact that embryonic features are no more constant than are
the adult characters; that indeed they vary with the adult
characters, so that no adult character is changed without some
precedent alteration of all the previous embryonic phases. ‘The
embryonic life is a connected whole, and it is impossible that
an isolated alteration of one particular stage can have taken
place. All variations must run through the whole develop-
ment; they may come out strongly at one particular stage, but
they must have been led up to and followed by variations in
all other stages.
Embryonic variations are not for the most part acted upon
by natural selection, because they concern rudimentary organs
only ; but when free life is reached, and the organs become
functional, the same variations continued (for continue they
must) are put to the test, and the organism stands or falls by
them. ‘The constancy of development in the same species
proves this point ; for if the embryonic stage could vary without
the free stages being at all affected, then, as natural selection
does not act upon rudimentary embryonic organs, the em-
bryonic organs would run riot, and we should expect to find
the greatest diversity in embryonic development of the same
species, and this we do not find; and this applies not only to
organs which persist into the adult, but also to organs which
disappear before the adult stage is reached. These purely
embryonic structures must have some nexus with structures
which succeed them in development, and a variation in them
must be accompanied by variations in these later appearing
persistent organs. In fact, it seems to me most important to
remember that the various stages in the development of an
animal are just as much correlated as are the different organs
LAW OF DEVELOPMENT KNOWN AS VON BAER’S LAW. 49
of the adult animal with one another. I repeat, the constancy
of development in the same species proves this point, as do the
small but constant differeuces between the embryonic phases of
slightly different but distinct species.
Granting that embryonic rudiments do vary, of which I do
not think there can be any doubt, then it would appear that
the variations must be selected, not with regard to their
intrinsic! merits at the moment as is the case with variations
in functional organs, but with regard to the effect of their de-
scendent or correlated variations in the adult. So it comes
about that the embryonic rudiments in one group of animals,
though resembling generally those of another group of the
same class or phylum (just as the functional adult organs
resemble one another generally), yet differ from them in minor
points, so that the group has its own individual character with
regard to that particular rudimentary organ, just as it has its
own individual character with regard to any adult functional
organ.
The conclusion here reached is that, whereas larval develop-
ment must retain traces (it may be very faint) of ancestral
stages of structure because they are built out of ancestral
stages, embryonic development need not necessarily do so,
and very often does not; that embryonic development, in so far
as it is a record at all, is a record of structural features of pre-
vious larval stages. Characters which disappear during free
life disappear also in the embryo, but characters which though
lost by the adult are retained in the larva may ultimately be
absorbed into the embryonic phase and leave their traces in
embryonic development.
[Throughout the above discussion I have, to avoid complica-
tion, treated all embryonic organs to be functionless; but it
1 By intrinsic merits at the moment, I mean the effect on the organism as a
whole at any particular moment. A variation in a rudimentary functionless
organ of an embryo can have no effect upon the welfare of the embryo
(excluding secondary effects—if any—of interfering with functional organs,
e. g. blood-vessels) ; its utility can only be judged when the free state is
reached.
vou. 36, PART 1.—NEW SER. D
50 ADAM SEDGWICK.
must not be forgotton that some of them are functional and
that these resemble organs of larve in retaining ancestral
features, e. g. the ductus arteriosus, &c. |
To put the matter in another and more general way, the only
functionless ancestral structures which are preserved in develop-
ment are those which at some time or another have been of
use to the organism during its development after they have
ceased to be so in the adult. In this way I should be inclined
to explain the hair of the human fcetus and the teeth of the
foetal whale—that is to say, I should be inclined to suppose that
the possession of the lanugo is due to the fact that there was a
time in the evolution of man when the babe required this
protection against the cold after the necessity for it had dis-
appeared in the adult, and that the young whale in the days
when whalebone was first acquired still retained the ancestral
habits which required teeth. It is, however, possible that
these and other similar cases of the retention of rudimentary
organs in late embryonic life have another explanation, and it
becomes necessary to collect and examine as many cases as
possible of the undoubted retention, as embryonic rudiments, of
organs which we have reason to know have recently disappeared
from the adult stage.
The retention of such organs in the embryo may, as I have
hinted, be due to the fact that they have been retained func-
tionally by the young animal after they have been lost by the
adult; but another explanation is possible, which is that
organs which are becoming functionless, and disappearing at
all stages, may in some cases disappear unevenly; that is to
say, they may remain at one stage after they have totally dis-
appeared at another. In this manner we might get an organ
which had become quite functionless and had quite disap-
peared in the free stage, still persisting, though with a much
reduced development, in the embryo. It is possible that the
lanugo and the teeth of foetal whales may be explicable in this
manner. But that such a retention of organs in the embryo
is not an important or permanent one is shown by the fact of
their comparative scarcity in embryonic histories. This is a
LAW OF DEVELOPMENT KNOWN AS VON BAER’S LAW. 51
most important subject, and I propose in a future paper to
collect and examine as many cases as I can find of the reten-
tion in the embryo of organs which have lately disappeared in
the adult.
There is another aspect of the same question which is sug-
gested by the above considerations, viz. if an organ can disap-
pear unevenly there is no reason in the nature of things, so far
as I can see, why it should not disappear in its devoloping em-
bryonic stages before it does so in the adult, so that there would
still be found in the adult a persistent useless rudiment of it
after all trace had gone in the embryo. And we may even go
further than this, and maintain that if organs can disappear
unevenly it is conceivable that traces of an ancient organ might
appear and disappear more than once in the course of develop-
ment. Of the last-suggested phenomenon I know of more than
one instance, but I know of no instance of an organ disappear-
ing in its embryonic stages while still persisting as a rudiment
in the adult. As an example of the repeated appearance and
disappearance ofa rudimentary organ in embryonic develop-
ment I may mention the neurenteric canal of certain species of
birds as described by Gasser,‘ and quoted by Balfour in the
‘Comparative Embryology’ (vol. ii, p. 162, mem. ed.). The an-
terior neuropore of Ascidians, which appears twice in the
development, is another example of the same phenomenon.
Although I know of no instance of an organ disappearing in
the embryo before it totally disappears in the adult, | do know
of instances of rudimentary embryonic organs which have
disappeared in their earlier stages while still present at a
later stage, e.g. the muscle-plate coelom of Aves, the primitive
streak of Amniote blastoderms, and the neurenteric canal of
Aves; and I have no doubt that many instances of this might
be collected.
From the application of the principles set forth in the pre-
ceding pages it becomes apparent to us why it is that in the
1 Gasser, “Der Primitivstreifen bei Vogelembryonen,” ‘Schriften d.
Gesell. zur Beford. d. gesammten naturwiss.,’ zu Marburg, vol. ii, sup. 1,
1879.
52 ADAM SEDGWIOK.
higher animals it is the early stages of development which
have the greatest interest for us, the later stage having been
added at a time when, as now, the immature stages of free life
were but little marked, and consequently there was but little
chance of the incorporation of any ancestral features in the
embryonic development. It also helps us, I think, to under-
stand why the most interesting of the ancestral embryonic
features were related to the passage from the aquatic to the
terrestrial condition, because when this took place in phylo-
geny there must have been a most pronounced aquatic larval
stage, such as we find to-day in Amphibia.
APPENDIX.
Mr. J. J. Lister has pointed out to me as confirmatory of
the views set forth in the preceding pages that there is at least
one exception to the rule that animals produced by budding
show no ancestral rudiments in their development, viz. the
sexually mature medusoid spore-sacs. These organisms present
in their development traces, as is well known, of many organs
which they must formerly have possessed in a functional con-
dition, e.g. the umbrella cavity, the marginal tentacles, the
circular canal, &c.; but, as Mr. Lister points out, these spore-
sacs differ from other buds in this important fact that they
have most undoubtedly had quite recently a free life during
the maturation of the generative products; and it may be that
it is the impress of this ancestral free life which has given rise
to the ancestral features in the development.
A CONTRIBUTION TO THE ANNELIDA. 53
A Contribution to our Knowledge of the
Annelida.
On some Points in the Structure of Euphrosyne.
On Certain Young Stages of Magelona and on Clapareéde’s
unknown Larval Spio.
By
Ww. C. McIntosh,
St. Andrews Marine Laboratory.
With Plates 6—8.
On some Points IN THE STRUCTURE OF EUPHROSYNE.
In the circumstances under which the present paper is con-
tributed, it is perhaps fitting that a form should be selected
(viz. Euphrosyne) which first came under my notice by the
kindness of Professor Ray Lankester, who sent preparations
procured in Herm in 1865. When an opportunity occurred a
few years later (1868) of becoming familiar with living exam-
ples on the same ground, unfortunately time did not permit a
detailed investigation of internal structure, and accordingly I
have had to content myself with the examination of rather
imperfect spirit-specimens.
While the structure of the genus Spinther, Johnston, has
been more or Jess fully elucidated by the labours of R, von
Drasch (‘ Anatomie von Spinther miniaceus,’ Grube, 1885),
and especially by those of Ludwig von Graff (‘ Die Anneliden-
gattung Spinther,’ 1887), the present genus, so far as I can
ascertain, has received attention only from Schmarda! and
E. Ehlers;? but at the date of their treatises the modern
1 ¢Wirbellosen Thiere,’ 1, ii, p. 136, pl. 33, figs. 264, 287.
2 «Die Borstenwiirme,’ i, p. 67, Taf. i and Taf. ii, 1864.
54 W. ©. MCINTOSH.
methods of microscopic investigation had not been introduced.
The sections were made from examples of Euphrosyne
foliosa, Aud. and Ed., and a few from E. cirrata, Sars, a
northern species, kindly sent by Canon Norman; and I have
to thank Mr. C. H. Williamson, M.A., B.Sc., for preparing
and mounting the slides ; and also for aid in this respect from
Mr. A. T. Masterman, B.A. (Cantab.).
Body-wall.—The cuticle is of considerable thickness, and
in the dorsal region and some other parts has externally in
the preparations a striated granular coat, which may be con-
nected with the presence of cilia. It is, on the whole, some-
what thicker than in Spinther. The hypoderm beneath has
the usual structure, and is very thin in the mid-dorsal line,
but increases laterally, and again ventrally. The inner edge
is sharply defined, so as to separate the circular muscular coat
distinctly, but a definite basement-layer does not appear to be
formed. Proportionately this layer is somewhat less developed
than in Spinther. The circular muscular coat lies imme-
diately beneath the hypoderm in the form of a continuous
sheath, though in the dorso-lateral regions it is modified;
the main mass, however, bounding the perivisceral chamber
and forming a thick ventral layer, pierced here and there
by the fibres of the vertical and oblique muscles. In the
lateral regions a radiate arrangement of powerful muscles
takes place in connection with the bristles; indeed, a double
radiate arrangement is present in many sections where the
upper and lower series of bristles come in the way of the
knife.
The longitudinal muscular layer forms dorsally a series of fine
fasciculi beneath the circular, and it follows the latter in its
course outside the perivisceral chamber, attaining its maximum
thickness after the splitting of the circular coat. Ventrally
the powerful longitudinal muscles are grouped in large fasciculi
separated by the vertical or oblique strands which pass to the
circular layer. The fibres from the oblique muscles decussate
beneath the nerve-cords, many mingling with the circular
(Pl. 7, fig. 4). The latter feature is much less evident in
A CONTRIBUTION TO THE ANNELIDA. 55
Spinther; indeed, none of von Graff’s figures indicate it, and
no allusion thereto occurs in his description.
Caruncle.—In horizontal section this highly sensitive
tongue-shaped organ presents externally a thin layer of cuticle,
the greater part of the area internally being formed of hypo-.
dermic cells and granules. Posteriorly, in the middle line,
however, is a double fibrous band, with at intervals various
lateral fibres, which extend not only into the lateral regions
but also into the narrow spaces between the bands. About
the middle of the organ one band becomes prominent and its
fibres spread more and more outward on each side until the
eyes are reached. These fibres are probably contractile. In
such a section as in Pl. 1, fig. 1, the pennate arrangement of
the median fibres is conspicuous, and they merge laterally into
the hypoderm. Moreover, in this section the complex mus-
cular fibres which intercross at the margins, and which appa-
rently are connected with the flattening of the organ, and other
changes in its outline are clearly shown.
In vertical section (PI. 6, fig. 2) the caruncle has a more or
less radiate appearance, due to the fibres which diverge from
the inferior area into the surrounding cellular stroma of the
hypoderm. The area referred to increases in size from behind .
forward, and in many views has a definite boundary or arch of
decussating fibres superiorly. It (that is, the space below the
arch) is occupied by a series of vertical fibres which radiate
superiorly into the substance of the caruncle, and cause the
somewhat arborescent appearance of the organ in section.
While, therefore, the lower fibres in the area just mentioned
are vertical, the upper are less definitely so. On the other
hand, the fibres are almost wholly vertical in the small or un-
developed condition of the area posteriorly, since they pass
straight upward to diverge into the tissue of the caruncle.
Below the latter the muscular fibres form a powerful band,
which by-and-by are interlaced with longitudinal fibres in con-
spicuous meshes (Pl. 6, figs. 2 and 3). The origin of these
fibres would appear to be the raphe above the folds of the
mouth. The area diminishes anteriorly, and seems to be occu-
56 WwW. OG. MCINTOSH.
pied chiefly by the cut ends of longitudinal fibres, while the
fibres bounding it form a dense fillet. In vertical section (PI.
6, fig. 2) the radiate arrangement of the fibres from the arch
or fillet is well shown, and they constitute a series of loops or
-meshes between the arch and the hypoderm. The area itself
is occupied by vertical and oblique fibres which come from the
dorsum in powerful bands, and which have longitudinal fibres
here and there in their meshes.
In Euphrosyne cirrata, Sars, from Norway, a conspicuous
strand of fibres passes from the nerve-masses (now at the ventral
surface) to the caruncle (Pl. 7, fig. 1), while other fibres con-
verge from the neighbouring region to the same organ. Im-
mediately over the cephalic ganglia in this species the area
presents only fine fibres at the sides of the median space, and
a large amount of opaque granular matter occurs at the sides
of the organ, partly mixed with the hypodermic tissue and
partly in a special capsule at each side, and apparently repre-
sents a pair of glandular organs. The granular substance has
not been stained. Posteriorly (before the termination of the
cephalic ganglia) well-marked median and radiate fibres appear
in the caruncle; then bands pass from the lateral regions both
into the latter directly and into the median longitudinal band
in front of the descending nerve-mass. A very complete raphe
is formed, and the distribution of the fibres is so arranged that
the organ and the region below it can be directly pulled. The
caruncle can thus be elevated or depressed. After the nerve-
mass has become wholly ventral, strong muscular fasciculi pass
from the ventral wall to the base of the caruncle, and spread
out on each side of it; then the band is chiefly median and
attached to the base and to the small knob representing the
organ posteriorly. The caruncle in this species differs consi-
derably in external form from that of Euphrosyne foliosa,
being more or less elongate and free, and the tentacle is like-
wise filiform. The latter merges into the base of the former
anteriorly (Pl. 7, fig. 1).
In Euphrosyne foliosa a pair of closely approximated
eyes lie at the anterior part of the caruncle (Pl. 6, fig. 3, ocd.),
A CONTRIBUTION TO THE ANNELIDA. 57
and another even more closely approximated pair on the ventral
surface of the snout in front of the two prominent pads of the
mouth (ibid., ocv.), and thus separated from the dorsal pair by
a considerable interval. Functionally thus the one pair serve
as organs of vision dorsally, the other for use ventrally. The
eyes have a distinct capsule with a broad margin of pale co-
lumuar cells, within which is the dense black pigment.
In Spinther the tentacle is apparently homologous with the
caruncle in the present genus. It lies over the cephalic nerve-
mass, is supplied with two large nerves, four eyes, and con-
siderably developed hypoderm, but the latter and the muscular
strands are much more largely developed in Euphrosyne,
and the organ is more complex. Vertical muscles pass from
the nerve-cords inferiorly in Spinther, and other muscles
from the lateral regions outside the pharynx, but they are less
developed than in Euphrosyne. The situation of the eyes in
Spinther, however, diverges, since both pairs are dorsal in
position, being located at the anterior and posterior margins
of the tentacle. The figures of Drasche! and von Graff,’ as
well as my own sections of Spinther miniaceus, Grube,’
show that the minute structure of the eyes is the same in both
genera.
In Euphrosyne cirrata the dorsal eyes lie on each side
of the anterior region formed by the fusion of the tentacle
with the caruncle. They have a somewhat radiate arrange-
ment of the clear vesicles with a dense ring of pigment, and lie
in the hypoderm—with the cuticle externally. The ventral pair
are more widely separated than the dorsal, but have the same
structure and relations to the hypoderm and cuticle.
Branchiwe.—Externally are the cuticle and its cilia—with
the thick hypodermic layer beneath—both layers being con-
tinuous with those of the body-wall. Moreover, vertical bands
of muscular fibres pass through the circular and other layers
to the bases of the branchial stems, enter the latter, and form
1 «Anat. von 8. miniaceus, Grube,’ Wien, 1885, p. 10, Taf. ii.
2 Op. cit.
* I am indebted to Professor L. von Graff for various examples of this
species,
58 W. 0. MOINTOSH.
the central longitudinal fibres. The shortening and elongation
of the organs is thus explained. In vertical sections (Pl. 6,
fig. 6) many transverse fibres occur between the longitudinal,
often crossing at right angles to the latter. Though in such
sections the fibres appear to be characteristically transverse,
yet in transverse sections of the basal region (PI. 6, fig. 4) they
present a radiate aspect, passing in front and behind the lumen
of the blood-vessel on each side.
These radiate fibres will readily alter the calibre of the stems,
expansion again occurring probably by the elongation of the
longitudinal fibres and the distension caused by the blood.
The inner margin of each blood-vessel is well defined, appa-
rently by a special coat, while the coagulated blood occupies
the central space. Externally the connective tissue and other
fibres of the branchial stem are closely united with the vessel,
so that no separation other than what has been mentioned
exists. In longitudinal sections in which the two vessels are
slit symmetrically, the lateral regions are occupied by the
latter (vessels), the median by transverse and longitudinal
fibres. The constricted region below the tip is circular in
transverse section, with the hypodermic cells radially arranged
around a central point, for the blood-vessels are now absent
(reaching only to the commencement of this narrow region, as
in Pl. 6, fig. 5), and the fibres pass into the centre of the di-
lated terminal region almost to the tip. In the terminal part
the large cells of the hypoderm (which give it a vesicular ap-
pearance) are also somewhat radially arranged—often sloping
from the central axis outward and upward in vertical section, or
placed regularly around the central axis in horizontal sections.
The tissue in these dilated regions is therefore much more lax
than in the constricted region beneath, as Ehlers observed.
He also describes a slender, circular muscular coat exter-
nally (that is, within the hypoderm), but the preparations did
not satisfy me on this point, though in Euphrosyne cirrata
certain trausverse wrinkles were seen at the base in longitu-
dinal sections. Besides, the arrangement of the radial fibres
would indicate that the functions of such a coat are fulfilled by
A CONTRIBUTION TO THE ANNELIDA. 59
other means. Shortening and elongation of the organs may
occur without the presence of circular fibres.
The absence of specially developed branchie in Spinther is
interesting, but the perivisceral fluid as well as the blood-
vessels are in this form nearer the surface, and the great hypo-
dermic membranous flaps on the dorsum may in some respects
also subserve this function. The intimate connection of the
lateral lamelle with the bristles may also prove of importance,
since the muscles of the bristles must cause extension of the
membranous lamelle.
Dorsal Cirri.—Externally is the somewhat granular layer
covering the cuticle, a condition probably due to the cilia,
which Ehlers describes as being largely developed. A thick
layer of hypoderm occurs beneath, with large cells here and
there. Both layers are continuous with those on the body.
The hypoderm at the extremity of the cirrus is finely granular
and longitudinally streaked, and the cuticle of this region is
very thin. Vertical fibres from the body pass upward into the
base of the cirrus and thence to the tip of the organ. Only a
few circular muscular fibres were observed at the base of the
cirrus (under the hypoderm), the elasticity of the cuticle pro-
bably sufficing to restore the shape of the organ on relaxation
of the longitudinal fibres. These organs are not represented
in § pinther.
Alimentary Canal.—On viewing the alimentary canal
from the dorsum in a spirit preparation of Euphrosyne
foliosa (Pl. 7, fig. 2) the proboscidian region (Schlundkopf,
Schmarda) has a somewhat ovoid outline, and is slightly
narrower in front than behind. It is divisible into two regions,
for anteriorly a glistening whitish layer (a) envelops like a
sheath rather more than the anterior third, splitting in the
middle line and curving outward on each side, so that its out-
line resembles that of a bivalve shell. Inferiorly this sheath
preserves an unbroken transverse border posteriorly. The
enlarged posterior region is free beneath, but dorsally termi-
nates in a canal (6) connecting it with the intestine at d. The
outline given in Pl. 7 differs considerably from the organ in
60 W. OC. MCINTOSH.
Schmarda’s Euphrosyne polybranchia, in which the region
presents a series of frills posteriorly and only two papille or
bosses in front.
Behind and above the rounded muscular mass of the pro-
trusible pharynx is a chamber with thinner walls, and having
laterally a series of well-marked ruge (e). This chamber may
represent the stomach, as Ehlers states, and it is connected
with the intestine at d as above mentioned.
The mouth (Pl. 6, figs. 3 and 7, w) opens on the ventral
surface between the third and the fifth segments, as Ehlers
describes in E. racemosa, and the walls of the buccal chamber
are thrown into many complex folds (wf.) which externally
have a cuticular coat. These folds are continuous with the
walls of the proboscis. In transverse section the latter pre-
sents in front a closely interwoven series of muscular fibres
chiefly circular and oblique, though vertical also occur. More-
over, large blood-vessels are visible ventrally at the sides. The
eversible portion of the organ has both inner and outer sur-
faces coated with the cuticle (which is stained), and the finely
granular and streaked hypoderm beneath is well marked,
besides certain muscular fibres passing into the bases of the
papille. In all probability it is the latter processes which
Schmarda describes as horny teeth in Euphrosyne poly-
branchia.
The buccal chamber gradually enlarges into a flattened canal
above (i.e. dorsal of) the posterior portion of the great mus-
cular “stem” of the proboscis. The latter (stem) is com-
posed of a complex series of fibres, the ventral being chiefly
arranged in parallel and vertical bundles, and bounded by a
definite investment, a few longitudinal fibres being clasped in
the interstices. Moreover, this region is cut off by a thin cuti-
cular septum from the part above. Then the median region is
occupied by a dense mass of glandular tissue (Pl. 6, fig. 7, gi.),
the glands being large and granular and extending to the hypo-
dermic coat which, with the cuticle, bounds the chamber now
present in the organ at this part. These glands apparently
perform an important part in the functions of the region. The
A CONTRIBUTION TO THE ANNELIDA. 61
upper arch again has transverse muscular fibres close to the
hypodermic border. A powerful muscular mass exists above,
ending in the thin hypodermic and dense cuticular layers
bounding the floor of the upper canal (Pl. 7, fig. 3). The
opposite or upper face of the canal has a very thin cuticular
coat and a thick glandular (hypodermic) layer. Folds or
ridges next appear at the sides (Pl. 6, fig. 7, e) where the
dorsal arch joins the muscular stem of the proboscis, and they
pass dorsalwards on this arch in the form of an extended ridge
on each side in transverse section, the surface being hypo-
dermic, while beneath is areolated glandular tissue. Proceed-
ing backward these lateral ridges increase in size, and instead
of an even surface show prominent ridges, while by-and-by a
belt of this folded tissue passes entirely across the upper
chamber. Inferiorly the stem of the proboscis now presents a
somewhat regularly interwoven field of cross-fibres like mesh-
work, bounded ventrally by a rim of longitudinal fibres, the
upper edge being defined by a cuticular investment. Beyond
the longitudinal belt at the ventral border of the decussating
fibres is a broad belt of vertical fibres.
The chamber soon shows a regular series of transverse ridges
of hypodermic glandular tissue, and then terminates in the intes-
tine. Externally is a layer of longitudinal fibres, internally a
coat of circular fibres, and the glandular lining has cilia on its
surface.
In longitudinal section (Pl. 6, fig. 7) the mouth has various
prominent folds of the lining membrane in front, but the ante-
rior region of the proboscis in this species did not present the
regular series of papillz indicated in the figure of Professor
Ehlers. The main channel passes dorsally, and reaches the
thin-walled chamber marked ds. in fig. 7, Pl. 6. The dorsal
wall agrees with that shown in Pl. 7, fig. 3, though the cuti-
cular layer is less marked, and extends to the opening into the
intestine. In this chamber the ventral surface is formed by
a portion with bold ridges (r., fig. 7, Pl. 6, where a part only
is shown, as the section is not median). In the ordinary or
retracted condition of the parts the food would thus be acted
62 W. C. MCINTOSH.
on by the area with the ridges and the powerful muscles
beneath, so that—between muscles and glands—a considerable
alteration probably ensues. There are grounds, therefore, for
thinking that this chamber represents the stomach, since its
posterior end leads by a short canal directly into the intestine.
The intestine follows the short canal just mentioned, and
consists of a wide passage, usually thrown in the preparations
into deep folds—with a few shallow diverticula of its walls; but,
so far as spirit-preparations can be relied on, the diverticula
do not attain the dimensions shown by Ehlers in his HE. race-
mosa.! The canal has a thick wall of densely granular glan-
dular tissue, papillose internally, and of circular fibres ex-
ternally. It continues backward and terminates posteriorly
in the vent, which is situated between two prominent lips just
in front of the caudal papille. A differentiation of the gut
occurs at the rectum, so distinct that in transverse sections it
appears at first sight that an independent channel exists pos-
teriorly. The continuity ofthe mucous membrane is, however,
easily made out. In this region, which is somewhat triangular -
in transverse section, the thin membranous investment has
beneath it a layer of longitudinal fibres, and the adjacent
granular cells more resemble those of the hypoderm than the
homologous cells of the intestine. Portions of food and sand are
occasionally observed in the centre of this portion of the canal.
The food consists in some of a soft mass in which cells,
spicules of sponges, chitinous fragments, and other débris are
present ; while in others little else than sponge-spicules can
be seen. Posteriorly the cylindrical feecal mass usually occu-
pies the centre of the gut. In Schmarda’s species calcareous
fragments, bristles of Annelids, and sponge-spicules were
found. It may readily be concluded, therefore, that it is diffi-
cult to secure either perfection or continuity in transverse
sections of the body.
The digestive system of S pinther differs from the foregoing
in regard to the much more largely developed lateral ceca of
the gut, and the less massively muscular proboscis. The
Op. cit., Taf, ii, fig. 1.
A CONTRIBUTION TO THE ANNELIDA. 63
presence of an extensive dorsal blind-gut is another important
divergence. Both, however, feed for the most part on the
same substances.
Nervous System.—The central mass of the nervous
system lies immediately under the caruncle, and consists dor-
sally chiefly of nerve-cells, ventrally of fibres (Pl. 6, figs. 38 and
Sye92)s :
The position of the eyes (PI. 6, fig. 3, ocd. and ocv.) in
Kuphrosyne diverges from that in Spinther, for in the
latter the four eyes are confined to the dorsum, whereas in
the former two are dorsal and two anterior and ventral.
The subcesophageal ganglia form a single mass behind the
mouth, with a fibrous central region and two lateral cellular
masses as in EH. cirrata.
The ventral nerve-cords (which Ehlers states are non-
gangliated) lie on each side of the median line enveloped in
their sheath, and partly separated superiorly by a fascicle of
longitudinal muscular fibres (PI. 7, fig. 4, ne.). The oblique
muscles pass down by their external borders to decussate in-
feriorly. With the exception of the capsule of connective
tissue, the cords in section present a granular surface from the
cut extremities of the fibres. Beneath the fascicle of longi-
tudinal fibres is a curved strand of transverse fibres (Pl. 7,
fig. 4, em.), which forms a commissure between the cords. It
is concave from above downwards. In many sections a strand
of fibres closes in the space for the longitudinal muscular
fibres superiorly. External to the cords are flattened bands
of muscular fibres, the decussating fibres of the oblique mus-
cles, and the circular coat with the cutaneous tissues (Pl. 7,
fig. 4).
In comparing these cords with those of Euphrosyne
capensis, the same fascicle of longitudinal muscular fibres and
the firm investment superiorly are characteristic. In Euphro-
syne borealis the muscular parts are specially massive. In
E. cirrata, again, the nerve-cords, which lie close to the
hypoderm, are flattened in section and much more widely se-
parated. A transverse commissure, however, is present.
64 WwW. C. MCINTOSH.
In Spinther ganglionic enlargements and transverse com-
missures similarly placed are shown by von Graff. He also
describes optic, pharyngeal, esophageal, and other nerves
from the cephalic ganglia.
Circulatory System.—Schmarda was of opinion that in
Euphrosyne polybranchia the circulation agreed with the
general type of the Annelids. In the midventral line a large
vessel occurs, with two smaller lateral trunks which supply the
ovaries. The unpaired vessel gives many branches to the
alimentary canal. On the dorsum of the alimentary canal are
two vessels, from each of which large branches go to the
branchie. He finds that the vessels have two coats, an outer
longitudinal and an inner circular (his transverse) ; while the
blood is red, and shows corpuscles ;4,™™ in size. Ehlers
adds nothing to this description except to observe that the
blood in Euphrosyne racemosa is colourless.
In E. foliosa the ventral vessels are distended by the yel-
lowish and minutely granular blood (in the preparations), and
various branches are observed amongst the reproductive ele-
ments, but the sections are less satisfactory as regards the
dorsal vessels. All that can be said is that a dorsal trunk was
seen in certain sections over the alimentary canal, and most
frequently in a state of contraction. The inner coat of the
larger vessels appears to be a homogeneous elastic one, no cir-
cular striz having been observed, though such may be present
in other forms. The outer coat certainly has longitudinal
fibres.
At the anterior part of the intestine the sections’ of four or
five trunks are seen in the mid-dorsal line close to the gut or
amongst the reproductive elements. All these are empty, the
collapsed vessel having a dotted or cellular appearance, and
thus different from those distended with blood in the other
parts of the reproductive masses. Proceeding forward so as
to bring the duodenal channel with its dense walls and
inner surface covered with cilia into view, a grouping of these
1 Which, unfortunately, are less complete than 1 could have wished, and no
additional examples could be procured.
A CONTRIBUTION TO THE ANNELIDA. 65
vessels above the latter channel (which lies over the intestine)
takes place, so that a large mass, composed apparently of thick
striated walls round a central lumen, is soon formed. It is
difficult to say what this is, though it possibly may be a
specially contractile region of the vessel. It is shown in its
fully-developed condition in Pl. 7, fig.5,v.g. It then splits
into two trunks which gradually separate from each other as
we proceed forward, but they always preserve a dorsal position
to the duodenal gut, stomach, and proboscis. They diminish
in size as they go forward, but they retain the same structure
and contractility. Their mode of termination could not be
ascertained.
The structure and arrangement of these parts would point to
their connection with the vascular system, but as no vessel
contains blood, and as their structure differs from the distended
trunks, some doubt exists. Their highly contractile condition
may be associated with the functions of the proboscis, which,
as formerly shown, has large blood-vessels inferiorly.
Segmental Organs (Nephridia).—Schmarda does not
refer to these organs, but Ehlers describes them in his form as
reaching to the twentieth segment. They lie under the dorsal
wall of the body, and have their external openings in the
median line between the branchiz. Lach is in the form of an
elongated bifid tube, extending over two or three segments.
The two inner openings are trumpet-shaped, with orange pig-
ment and internal cilia. The only structures observed in the
sections of E. foliosa are the bifid vessels passing over the
proboscidian region from the single dorsal tube as described
above, and such would appear to pertain to another system.
Generative Elements.—Schmarda describes the ovaries
as funnels or tubes with a blood-vessel in the middle. These
tubes end in an oviduct which opens near the vent or near the
inner branchie. The specimens examined by Ehlers had no
generative elements. These elements in E. foliosa occur in
the perivisceral space around the alimentary canal, and at the
bases of the feet (PI. 6, fig. 1, ov., and Pl. 7, figs. 5 and 6,9. p.).
In the male the dense masses of sperm-cells are often arranged
voL. 36, PART 1.—NEW SER. E
66 W. O. MCINTOSH.
round the blood-vessels, and with the cells somewhat regularly
placed in rows, so as to form long loops, concentric or slightly
radiate areas. The linear arrangement of these cells, indeed,
is characteristic. The same regions are occupied by the large
eggs in the female. The sexual elements were best developed
in those obtained in July and August, both in Britain and in
Norway, and they probably escape dorsally by the nephridia.
Many ova occur amongst the branchie and bristles of the
dorsum. In regard to the escape of these elements, the most
likely channels appeared to be those that passed to the skin
near the bristle-bundles.
The ova in Spinther are proportionally as large as in
Euphrosyne, but no nephridia have been observed, though it
is possible they may yet be found.
On Certain YounG Staces oF MAGELONA.
The occurrence of several stages of the young of Magelona
in the bottom trawl-like tow-net in St. Andrew’s Bay gives an
opportunity for taking a brief survey of our knowledge of the
subject.
Claparéde, in 1863,! described some young stages of this
species which he had procured at St. Vaast la Hougue in the
summer of 1861. His youngest form was about 1 mm. in
length, and cylindrical. The anterior end formed a wide
funnel, and thus differed from any larve hitherto found at
St. Andrews. The border of the funnel was beset with cilia,
and cilia occurred behind it. Four reddish eyes occurred
transversely on the dorsum of the head. Between the broad
cephalic region and the first body-segment is an expanded
region bearing a tuft of long cilia. There are from fifteen to
twenty body-segments—which are widest anteriorly. From
the projecting process of the first body-segment proceeds the
tuft of long smooth bristles, which are almost as long as the
body. The following segments had touches of brownish pig-
ment, a pair of small hook-pads, and short bristles. Each
1 «Beobachtungen iib. Anat. u. Entwicklungs. Wirb. Th.,’ p. 74, Taf. x,
figs. 9—14, and Taf. xi, figs. 1, 2,
A CONTRIBUTION TO THE ANNELIDA. 67
segment had a ventral band of cilia, whereas none occurred on
the dorsum. The terminal segment had a ring of long cilia.
The alimentary canal was roseate. This larva is distinguished,
he says, from a Leucodore larva by the dilated anterior end,
by the brownish pigment-touches, the smooth bristles, by the
veutral cilia on all the segments, and by their absence on the
dorsal surface.
It exhibits by-and-by a modification of the head, which
becomes elongate heart-shaped, and by the development of
lateral papillee from which the tentacles grow behind the eyes.
The provisional swimming apparatus, viz. the bands on the
head, the lateral tufts, the ventral bands, and the anal ring are
all increased.
When it reaches the length of about 2 mm. it approaches
a Spio larva, and swims freely in the water by a rapid
wriggling of the body. The heart-shaped snout is narrowed in
front, concave below, and convex above. The four reddish
eyes are larger, and the middle pair placed in advance. The
first body-segment overlaps the head, and from its sides spring
the short tentacles, which are borne like a pair of horns, the
inner side of each being marked by brownish parallel striz,
and the slender papille are 0°017 mm. long. The tentacle has
a central blood-vessel which ends blindly, and its blood con-
tains corpuscles. The first segment bears the huge bristles,
while the last is elongated, cylindrical, and is devoid of bristles.
The rows of hooks occur from the ninth segment, some of
them in sacs. They are more numerous posteriorly (to the
number of fifteen). The posterior end is pointed and bears
small, colourless, pyriform (birnformige) papille. The diges-
tive canal begins with an oval mouth and a muscular proboscis,
which has a pair of opaque (brownish) glands at its posterior
end. The canal, which is constricted in front, dilates at this
point. The colour of the larva is delicate brownish, with
parallel bands of the same hue on the tentacles ; longitudinal
and transverse brownish touches on the snout. The trans-
lucent anterior region and the three following segments bear
similar touches. The middle fifteen segments of the body have
68 W. CO. MCINTOSH.
brownish pigment-rings. The following segments have lateral
touches of the same colour, which posteriorly almost assume the
form of rings. The tip of the tail is brownish, and the same
tint occurs anteriorly at the lips of the proboscis.
His next stage presents elongated tentacles with long
papilla and coiled like ram’s horns, yet the snout is even
shorter than in the foregoing stages, so that a certain amount
of variation exists. The length is 8 mm., and it swims like an
eel. Claparéde compares the elongated papille with the
peculiar “stabchen ”’ on the tentacles of Spio as described by
Strethill Wright, but the differences are considerable. The
first body-segment bears the long bristles. From the second
to the eighth segment the lateral bristles have disappeared.
From the ninth to the fifteenth long provisional bristles as well
as hooks with an elongated shaft are present. The brownish
glands behind the proboscis are larger. The blood has a bluish
appearance. The end-segment is hoof-like and somewhat
resembles that of Leucodore, and it has small coloured
warts.
He concludes by mentioning that further stages of Mage-
lona were not got in his nets, and that they probably took to
the bottom to burrow in the sand at the extreme margin of low
water. He was inclined to relegate the larva to the Spionidze.
On the present occasion only the more advanced S pio-like
forms, approaching Claparéde’s two last stages, figured in his
pl. x, figs. 10 and 12, will be dealt with. The first was
obtained on the 28th May, and had about twenty-five segments
exclusive of the head and tail. The animal (PI. 8, fig. 1) is
nearly translucent, faint touches of white pigment appearing
on the sides of the proboscis posteriorly, at the commencement
of the narrow part of the gut, and at the base of the tentacles.
An opaque white mass (the contents) also marks the centre of
the gut towards the tip of the tail. Lateral opaque white
specks in groups further indicate the segments in the anterior
(dilated) region of the body. ‘The latter consists of nine seg-
ments, the tenth being opposite the whitish opacity (a, Pl. 8,
fig. 1) marking the commencement of the constricted portion
A CONTRIBUTION TO THE ANNELIDA. 69
of the gut. Besides the anterior long bristles, there are
several pairs on the sides posteriorly, as shown by Claparéde,
who figures four pairs in Magelona (ibid., Taf. x, fig. 12). In
a coloured drawing of one at this stage by Mr. J. Pentland
Smith, M.A., B.Sc., about seven pairs are present. The
tentacles are longer than the body, and contain only a single
vessel, as in Claparéde’s stage of Magelona of 2 mm., and no
circulation is yet visible, though a few stationary corpuscles
are observed on the wall of the vessel. In this condition the
Annelid stretches itself freely, with the tentacles widely ex-
panded, and apparently draws in water by the mouth at
intervals, to judge from the movements of the gullet. When
irritated it coils its tentacles like springs, and wriggles rapidly
through the water, as noticed by Claparéde in Magelona.
Only slender spine-like papille occur on these elongated
organs, as in S pio, to which the young form has close affinities.
No forward growth of the snout has yet taken place, and
therefore this example would appear to belong to another form
or to be less developed than Claparéde’s fig. 12. Moreover,
the eyes are black, not red, and much smaller than he shows.
The pigment of the body further is white, not brownish, such,
perhaps, being due to variation. As in other species, such
forms do not follow the younger in regular succession as
regards date, for the spawning period is evidently prolonged.
Thus, for instance, the foregoing example was much larger
than some examples of Magelona procured at the same time,
or even than others found in the middle of June. Moreover,
the shedding of the long larval bristles takes place at different
periods in specimens of the same or nearly the same age.
They are absent, for instance, in the example figured in Pl. 8,
fig. 2, though in other respects it agrees in structure with the
foregoing.
On the 17th of October (similar forms, however, having been
seen earlier) a more advanced specimen of Magelona than
described by Claparéde was obtained in the bottom trawl-like
tow-net (PI. 8, fig. 3). The basal part of the tentacles is now
furnished with larger papille, while the slender processes
70 W. CO. MCINTOSH.
formerly mentioned still exist on the terminal region, as,
indeed, was observed in some examples in May. The eyes
corresponded with those already figured in the Spio-like
larvee (figs. 1 and 2), and are still much less than in Claparéde’s
representations. The snout now forms a flattened spathulate
process of considerable dimensions, though it is less in propor-
tion than in the adult.! Moreover, no blood yet enters this
region (snout), though the circulation in the tentacles is com-
plete. The latter organs can be fixed to the glass vessel by the
papille, which thus have adhesive properties. Each of the
nine anterior segments is furnished with slightly clavate feet
and bristles of considerable length. Then follow the constric-
tion of the alimentary canal (which Claparéde does not indicate)
and twenty segments, the first eight or nine of the series having
still longer bristles than the foregoing division. The dorsal
vessel (or vessels) sends powerful currents forward by regular
contractions, which form a fold of the trunk opposite the fifth
bristle-bundles.”, The opaque white glands of the posterior
part of the proboscis were distinct on capture, but they became
less visible after confinement in the laboratory.
The Spio-like forms figured in Pl. 8, figs. 1 and 2, thus
diverge from the unmistakable larve of Magelona figured
by Claparéde and in fig. 3 of the plate just mentioned in the
present paper. They show no transverse striz at the base of
the tentacles, and the proboscidian region of the gullet is much
shorter. Opaque white glands, however, occur at the sides of
the latter posteriorly, and the arrangement of the eyes, the
general contour and the number of segments in the anterior
region of the body, are similar. Again, some examples agree
with the two figured (Pl. 8, figs. 1 and 2), but have at the
base of the tentacles indications of the ruge which fore-
shadow the papille characteristic of the species in its adult
1 Claparéde (op. cit., fig. 10) figures the younger stage with a considerably
larger snout than the later stage, a feature perhaps due to individual variation.
In a dying specimen, somewhat smaller than the present form, a large
pinkish oil-globule occurred in the anterior region—the product of de-
composition of the blood or other fluid.
A CONTRIBUTION TO THE ANNELIDA. ral
condition ; but such have never presented at St. Andrews the
long slender larval papille simultaneously in the same region,
and which Claparéde shows in his fig. 12. Though the young
stages of allied members of the Spionide are not yet sufficiently
known, yet the weight of evidence inclines to the view that the
Spio-like larve pertain to Magelona, at a stage previous to
the appearance of the rugose ridges ushering in the thick
cylindrical papille. An interesting feature is the disparity in
size between the latter examples without forward growth of the
snout, and others, fully a third less, with a considerable snout.
On Cxuaparepre’s Unknown Larvat SPtio.
The first notice of a form apparently identical with the larval
Annelid which forms the subject of these remarks is given by
Maximilian Miller,! who alludes to a fragment of the posterior
end in connection with his observations on bacillary corpuscles.
When sojourning at St. Vaast la Hougue, between July and
September, Claparéde” found not unfrequently an unknown
larval Annelid (pertaining to a common form) which has also
occurred in considerable numbers in the bottom-nets at St.
Andrews from July to October. The same form has been met
with on the Norwegian coast at Christiansand.
The youngest stage measures about 0:045 mm., and has
about twelve bristled segments, besides several without bristles.
The head is short and is divided into symmetrical lobes. The
ridge is richly ciliated, and differs from the arrangement in the
larva of Leucodore. The larger bosses have short cilia and a
pair of longer ones. The eyes are arranged more or less on a
trapezoid (the posterior pair more widely separated), and have
reddish pigment. The mouth lies on the ventral surface be-
tween two ciliated lobes. Two bands of cilia occur on the
ventral surface behind the mouth. The first body-segment has
a larger lateral process than the rest, and bears a long tuft of
slender bristles, minutely spinose. The succeeding segments
1 *Qbservat. Anatom. de vermibus quibusdam maritimis,’ 1852, p. 29,
pl. ii, fig. 29.
"2 Op. cit., p. 77, Taf. vi, f. J—11.
72 WwW. C. MCINTOSH.
have shorter bristles. Each segment has a short tuft of cilia,
somewhat resembling the ventral tuft in the Magelona-larva.
The terminal segment is ring-like and bears long cilia. The
alimentary canal is pale and wide, but constricted at each
segment-junction.
An older stage with from eighteen to twenty-four segments
showed dorsal and ventral foot-papille, and a pair of rudi-
mentary tentacles on the head, the posterior part of which is
elevated (as in Pl. 7, fig. 7).
When the larva reaches 24 to 3 mm., and has from thirty-
five to fifty segments, the tentacles are longer and show an in-
ternal cavity, and the rows of cilia are longer, though shorter
than in the larve of Leucodore. The lips are richly ciliated,
and a tuft of cilia occurs on each side behind the head. The
anterior pair of eyes are blackish, as in the older larve, whereas
the posterior pair are reddish. The foot-processes are larger
at this stage and somewhat conical. Moreover, a brownish
speck occurs between them, whereas in the examples at St.
Andrews it was whitish or yellowish white. The bristles have
the same rough aspect (from minute spikes), and the first seg-
ment bears longer bristles than the following. In all the seg-
ments the ventral row of cilia is present, while posteriorly is
the anal ring of long cilia. The pale alimentary canal contains
sea-water. The larva wriggles through the water for some
time and then settles on the bottom. It is translucent like
Tomopteris, the only pigment being the coloured specks on
the sides (between the feet).
When supplied with sea-water, development proceeded, the
dorsal and ventral divisions of the feet from the seventh to the
eleventh becoming longer, thicker, and with a slender tip, all
the others remaining as before. Further, the dorsal branch
of the foot showed reddish pigment, whereas the ventral re-
mained pale. The opaque speck (black?) remains between
them. By-and-by the long cilia of the lateral regions of the
head and the anal ring disappeared. Moreover, the presence
of the longer feet from the seventh to the eleventh segments
was somewhat inconstant, for in one in which only thirty-five
A CONTRIBUTION TO THE ANNELIDA. 73
segments existed they were well developed ; whereas in another
with forty-five to fifty segments they were not longer than the
others.
The geographical distribution of this larva is extensive, and
Claparéde gives a figure of one from Christiansand which for the
most part corresponds. He does not know any adult Annelid
with longer feet from the seventh to the eleventh, nor with the
peculiar spinose bristles, which are probably provisional organs
The peculiarity is that they existed so long.
The most advanced larva procured by Claparéde was thus
only supplied with short tentacles, that is, little more elongated
than in Pl. 8, fig. 4, of the present paper; whereas several
procured at St. Andrews in October had these organs consider-
ably elongated—stretching backward to the fifth (Pl. 8, fig. 5)
and even to the eighth bristled segment. They are large and
comparatively massive organs resembling those of the S pio-
nid. The unpaired process in front is considerably shorter
than that figured by Busch,! and Claparéde and Mecznikow,?
in the larval Nerine cirratulus. Perhaps it only develops
in the later stages, for in Pl. 7, fig. 7, it is not visible in a
lateral view. Moreover, several with long tentacles had shorter
bristles than in the earlier stages. The tail terminates in a
somewhat ovoid tip with a dimple in the centre and a ring of
cilia towards the tip (Pi. 7, fig. 8). In lateral view, however
(ibid., fig. 9), the tip occasionally assumes a conical condition.
A dorsal and a ventral blood-vessel (Pl. 8, fig. 6) are evident
in the same view (lateral). The conspicuous pigment-speck
between the bases of the feet is either opaque white or yellowish
white, and is often finely ramose. Towards the end of October
(23rd) an advanced example had peculiar globules which re-
fracted the light like oil along the lateral region (Pl. 8, fig. 7)
from the tenth foot backwards. They were absent from both
dorsal and ventral aspects till near the tip of the tail. Their
Beobachtungen iib. Anat. u. Entwicklung. &c.,’ p. 65, Taf. viii,
figs. 1—4.
2 « Beitrage zur Kenntniss der Entwickelung. der Chetopoden,” ‘ Zeit.
wiss. Zool.,” Bd. xix, separ. Abdr., p. 11, Taf. xii, fig. 4.
74 W. 0. MCINTOSH.
nature is doubtful, but it is possible they were due to degene-
ration, though the animal appeared to be active and healthy,
the only feature of note being the great length of the bristles
flanking the sides and the comparative shortness of the ten-
tacles. Some of the advanced specimens were three-eighths of
an inch long and had thirty-two segments behind the head.
The bristles of the long larval tufts in front seem to be more
or less smooth in spirit-preparations. The minute spikes on
the stronger bristles of the feet are readily seen. The bristles
are generally in groups of three or four on each foot.
In section the cuticle is found of considerable thickness, and
beneath is a feebly developed circular coat, then a boldly
marked layer of longitudinal muscular fibres, arranged in two
dorsal and two ventral bands. ‘The long processes characteris-
ing the feet from the seventh to the eleventh segments have
large hypodermic cells internally, with their long axes parallel
to that of the process. The nerve-cords form two flattened
granular bands on each side of the middle line ventrally.
They have the hypoderm externally, and apparently a space
over each internaliy. The oblique muscles pass to their outer
edges, and probably go beneath them.
No further light has been thrown on the relationships of
this form except that the tentacles in the most advanced con-
firm the opinion of Claparéde that it pertains to the Spionide.
It is apparently the larva of a species not uncommon at St.
Andrews.
EXPLANATION OF PLATES 6—8,
Illustrating W. C. McIntosh’s paper, “ A Contribution to
our Knowledge of the Annelida.”
PLATE 6.
Fic. 1.—Longitudinal (horizontal) section of the caruncle of Euphrosyne
foliosa, Aud. and Ed., behind and on a level with the eyes. x 350.
Fic. 2.—Vertical section of the caruncle, showing a somewhat radiate
arrangement of the fibres in the organ, while inferiorly strong fibres pass
from the trunk into it. ™ 55.
A CONTRIBUTION TO THE ANNELIDA. 75
Fic. 3.—Vertical longitudinal section of the anterior region of the same
species on one side of the median line, showing the muscular fibres, m.,
passing to the caruncle, car. ocd. Dorsal eye. cv. Part of a ventral
eye. c.g. Cephalic ganglia. v. Section of blood-vessels. . Mouth.
wf. Anterior folds of the lining of the mouth. From the slightly oblique
nature of the section certain of the anterior bristles, &c., are seen in front.
x 55.
Fic. 4.—Transverse section of a branchial stem, towards the base. x 350.
Fig. 5.—Transverse section of the same organ, towards the tip. Zeiss,
obj. D; oc. 1.
Fic. 6.—Vertical (transverse) section of a branchial stem as it leaves the
surface of the body, showing fibres entering the region. X 350.
Fic. 7.—Vertical longitudinal section of the anterior end of E. foliosa.
bs. Gastric chamber. 47. Transversely ridged region, only a portion being
seen in this section. e¢. Muscular ridges. g/. Glandular area. w. Mouth.
wf. Folds of the buccal membrane. x 24.
Fic. 8.—Slightly oblique section (though more or less vertical and trans-
verse) of the cephalic ganglia of Euphrosyne foliosa in the region of the
dorsal eyes. xX 55.
PLATE 7.
Fic. 1.—Vertical (transverse) section of the region of the caruncle in
Euphrosyne cirrata. a. Vertical fibres from the ventral region, above
nerve-trunks. car. Caruncle. ¢. Tentacle. The knife has passed through
the caruncle after the central space has disappeared. The section is probably
more or less oblique. x 55.
Fic. 2.—Slightly enlarged view of a softened example of Huphrosyne
foliosa, with the alimentary canal exposed. a. Anterior glistening region
of the proboscis. 4. Dorsal continuation of the canal. ce. Striated or ridged
muscular region. d. Pyloric part of the passage 4. e. Intestine.
Fic. 3.—Vertical transverse section of the proboscidian stem near the
glandular region. A thin line cuts off a lateral area at each side. ds. Ali-
mentary canal. x 80.
Fic. 4.—Vertical transverse section of the anterior ventral region. da.
Wall of intestine. x. c. Nerve-cords. cc. m. Circular muscular coat; the
same letters mark the nerve-commissure. The oblique fibres decussate beneath
the nerve-trunks, and join the circular coat. Sections of vessels are seen at v.
Fic. 5.—Transverse vertical section of the anterior region, where the
canal, Js, from the gastric region joins the intestine, da. vd. One of the
dorsal trunks. vg. Central dorsal vessel. gp. Reproductive elements—in
this case male. x about 60.
76 W. OC. MCINTOSH.
Fic. 6.—Transverse vertical section, anterior to Fig. 5. The dorsal
vessel has now split into two trunks, vy. Zeiss, obj. A, oc. 1.
Fic. 7.—Head of Claparede’s larval “Spio” of 16th July. x 52.
Fic. 8.—Posterior end of the foregoing form, in a specimen procured in
October (?). x 52.
Fic. 9.—Lateral view of the tip of the same example. x 52.
PLATE 8.
Fic. 1.—Young form resembling Magelona papillicornis in which the
head has not yet extended beyond the tentacles, and in which the long larval
bristles are present anteriorly. «. Constricted region of the alimentary canal.
gl. Glands of the pharyngeal region. The anterior region of the alimentary
canal is distended with water. 28th May. x 55.
Fig. 2.—A similar specimen after the long larval bristles have been shed,
and viewed from the ventral surface with the tentacles fully extended. x 55.
Fig. 8.—A more advanced example, in which the basal part of the tentacles
has larger papillee, and the snout has considerably increased in size; yet the
long larval bristles are still present. 17th October. x 50.
Fic. 4.—Dorsal view of the anterior region of Claparéde’s larval ‘*Spio”
of 16th July. ph. Pharynx. x 52.
Fig. 5.—Anterior end of the most advanced form of the same species yet
found. The tentacles are considerably larger. 23rd October. x 52.
Fic. 6.—Lateral view of an example of the foregoing, showing the chro-
matophores and feet. 16th July. x 52.
Fic. 7.—A similar view of a specimen (23rd October), in which peculiar
refracting bodies, like oil-globules, occurred along the lateral aspects. xX 52.
SPOLIA NEMORIS. Ve
Spolia Nemoris,
By
A. A. W. Hubrecht, LL.D., C.M.Z.S.,
Professor of Zoology in the University of Utrecht.
With Plates 9—12.
Ir was in the summer of 1889 that an invitation reached me,
coming from the Royal Physical Society (Koninklijke Natuur-
kundige Vereeniging) in Batavia, to undertake a trip to the
Indian Archipelago for purposes of scientific research. This
invitation opened the prospect of realisation of a wish long
cherished and for a naturalist not exorbitant,—the wish to have
a direct glimpse and a personal impression of animal and
vegetable life in the tropics. And so it was accepted with
alacrity.
Now that I am going to give a summary account of my
investigations during this temporary sojourn in India, the
results of which are gradually taking a shape that will permit
of their successive publication, I cannot refrain from expressing
my grateful indebtedness to the above-named Society and to its
Council. Although the funds that were required for these
researches have been granted by the Government, to whom I
am for that reason equally indebted, still it was the Society
who transferred the responsibility for the way in which the
money was to be spent entirely to me, with what I would be
inclined to call a blind confidence.
In consequence of this I was quite free in the choice of any
research I might wish to undertake, and also in the method
78 A. A. W. HUBREOCHT.
according to which I should desire to conduct both the collect-
ing and the working out of the subject-matter of these investi-
gations. I resolved to extend certain researches with which I
had been occupied for the last few years, and which had refer-
ence to the earliest developmental stages and the formation of
the germinal layers of mammals, as well as to the numerous
and often unexpected points of difference which we observe in
the first origin and in the detailed anatomy of the placenta
(afterbirth) of different mammals.
Of late years the mammalian placenta has been more closely
studied by numerous anatomists, but nevertheless its highest
stage of differentiation as found in the human subject is yet so
imperfectly understood (genetically) that a comparative investi-
gation of the more primitive orders of mammals is an imperious
necessity. As in all other attempts at comparative analysis,
so in this case the selection of the material that is to furnish
the bases of comparison is most important.
Now the lowest mammals (Ornithodelphia, Didelphia) are as
yet deprived of a placenta; this organ has only become developed
in later, more highly differentiated orders. It is thus the very
youngest organ which we meet with in mammals, the latest
acquisition by the gradual perfecting of which they have
obtained a considerable advantage over the lower Vertebrates.
' The order of the Insectivora is regarded as being the most
archaic among the Mammalia Placentalia, both on account of
paleontological and of anatomical data. And so the objects of
comparison had to be chosen in the first place among these
more primitive forms.
Several years ago I commenced to study the process of
placentation in three European representatives of the order
Insectivora—the hedgehog, the mole, and the shrew,—and have
published part of the results of these investigations.
In the Indian Archipelago other genera of the same order
occur which are entirely absent in Europe. Towards these my
attention had in the first place to be directed during my stay
in the Archipelago. They are the genera Tupaja and
Gymnura, of which the latter very soon proved too rare to be
SPOLIA NEMORIS. 79
available for this investigation. Tupaja, on the contrary, is
much more common, and I might safely feel hopeful to collect
a rich harvest of Tupaja javanica.
Besides the additional genera of the order of the Insectivora,
the investigation had in the second place to be directed towards
another order which is said to occupy an intermediate place
somewhere between the Insectivora and the highest order, that
of the Primates, to which man and monkeys belong. This
intermediate order is that of the Lemuride or Prosimie. In
Europe it is no longer represented by living genera, although
in earlier geological periods it did occur in this part of the
world. A small number of genera compose this order, by far
the majority of them being found in Madagascar.
Two representatives of the Prosimiz occur in the Indian
Archipelago, viz. Nycticebus and Tarsius. A peculiar genus
of mammals, the so-called flying maki or Galeopithecus—dif-
ferent in organisation as well as in mode of life—was at one
time regarded by zoologists as being more closely allied to the
Lemurs, at another time to the Insectivora or to the Cheiro-
ptera, or even as an order by itself (Dermoptera). This genus
also occurring in the Indian Archipelago, it had similarly to
be included in the sphere of the projected investigation.
Finally, I was interested in the only representative of the
order of the Edentata that has as yet been brought to light
in the Indian Archipelago, viz. Manis javanica, and desirous
to obtain a complete series of the different stages of placenta-
tion of this animal; the Edentates presenting considerable
differences among themselves with respect to their placenta-
tion.
Coloured drawings of the above-named mammals were distri-
buted, a few months before my arrival in India, by the Royal
Physical Society amongst a number of persons with whom
readiness to co-operate appeared probable.
To this was added a circular, answers to which successively
arrived. My friend Dr. P. C. Sluiter, librarian to the Society,
to whose energetic assistance I am deeply indebted, entered
into a preliminary correspondence with the writers, and placed
80 A. A. W. HUBRECHT.
the outcome of this at my disposal when I arrived in Batavia
in November, 1890.
In this way matters were made easy for me, and I could
form a provisional opinion as to the question in which part of
the Archipelago I would probably find the most favorable
collecting spots.
Only certain general data were available as to the habitat of
the above-named genera of mammals, but detailed accounts
about their comparative rarity, by which certain regions might
at the outset be considered as less favorable than others, were
deficient, as were also reliable data about their time of repro-
duction, &c.
Moreover, different aspects of the question must be kept
sight of. Suppose the animals to be numerous in a region
without European inhabitants, I could not then expect a rich
harvest. Similarly it might be presumed that in parts where
the population is scarce, the inhabitants could hardly give any
important aid in the collecting of a great number of speci-
mens.
On the contrary, it was most probable that in strongly
populated districts where a large proportion of the soil is culti-
vated, the mammals in question would be very rare or extinct.
It was, in short, unavoidable to spend the months of my stay
in India in as considerable a number of different places as
possible. In this way I was able to enter into personal con-
nection with very many who might be willing to continue the
collecting business even after my departure. For this purpose
I left behind me, wherever I had succeeded in enlisting co-
operators, printed instructions, chemicals, glass tubes, &c., as
well as cash for the payment of premiums to the natives by
whom the collecting of the live material was to be done.
This method of going to work must appear to be a tedious
and slow one. At the same time it was in the commencement
most undoubtedly disheartening. Still I have conscientiously
applied it wherever I have stayed, after generally demonstrat-
ing by means of more common animals than those I was in
search of how the extirpation of the uterus had to be effected,
SPOLIA NEMORIS. 81
and how the preservation was to be done. Now that three
years have passed by, I may safely say that the results have far
surpassed my expectations.
Among the hundreds of persons with whom these roamings
through the woods and mountains of Java, Sumatra, Banka,
Billiton and Borneo have brought me into close connection,
and who have been interested in the object of my investigations
as explained to them by me, it is only natural that the great
majority has been unable by various circumstances to contri-
bute in any way towards the increase of my embryological
collection.
They, however, who have thus contributed can hardly have
imagined how their apparently small collections—but which are
being forwarded from numerous parts—can together constitute
a very considerable array of important material for research.
Such has undoubtedly been realised on this occasion, consider-
ing that at this present moment I already dispose of—
469 uteri of Tupaja,
137 » Nycticebus,
72 » Galeopithecus,
198 ee Lars,
150 >» Manis,
making the respectable total of 1026. This collection is yet
increasing continually by new arrivals.
The majority of these uteri are pregnant in one stage or the
other ; many have been preserved very shortly after parturition ;
only a very few are virginal.
The pregnant uteri contain the most divergent stages, from
the earliest phases of segmentation to the nearly ripe or newly
born foetus. Several newly born young have also come into
my possession, as also a few in the very act of birth, the nearly
born foetus being still connected by its umbilical cord with the
as yet adherent placenta.
The numerous microscopical preparations which have already
1 While correcting this proof, new arrivals have again increased this total
to 1072.
VOL. 36, PART 1.—NEW SER. K
82 A. A. W. HUBRECHT.
been made of the rich and varied material demonstrate the
perfect care which most of my correspondents have bestowed
on the preservation. Consequently the histological details
of the placentation process, of the formation of the ger-
minal layers, and of the ontogenesis can be studied from
these preparations quite as satisfactorily as if the preparations
had been freshly made in the laboratory.
Again in this respect Kleinenberg’s mixture (picro-sulphuric
acid) has proved to answer to a very high standard of excel-
lence; in the case of the preservation of uteri in toto it gives
the best chances for the finer details of early blastocysts therein
enclosed, or of the placentary structures in the course of forma-
tion, to be perfectly preserved ; always on this sole condition,
on which I have everywhere laid particular stress, that the
extirpation be made instantly after death. Preparations made
from animals that had been dead even for only a very short
time have already undergone so considerable an alteration that
they are of very inferior value for comparative and especially
for histological research.
A point which had more particularly puzzled me before I
commenced my peregrinations was the question at which
period of the year the animals I was going to search for repro-
duced their species. As was already noticed above, the litera-
ture on the subject leaves us entirely in the dark with respect
to this point. And though the alternation of seasons is much
less marked in the tropics than in the temperate regions, still
the regular succession of the “rainy” monsoon and of the
‘‘ dry’? monsoon—more marked, however, in certain parts of the
Archipelago than in others—might be expected to have a certain
influence on the birth-rate and on the association of the sexes
in these animals.
If in the commencement I have been inclined to believe
that it would be possible to detect any such parallelism, still,
as the collections have increased, it has become more and more
evident that reproduction of the species investigated occurs all
the year round.
In the same months the most divergent stages of pregnancy
SPOLIA NEMORIS. 83
have been observed to occur; in no month have they been
deficient. My different correspondents have come to the same
conclusion as soon as the material they brought together
became more extensive, and allowed them to compare the
results of different months.!
Another general conclusion, which has more especially been
verified for Tupaja and Tarsius, is that pregnancy is repeated
at rapid intervals, very early stages of development being often
found in the same uterus simultaneously with the yet indubi-
table remains of a preceding pregnancy, as judged from the un-
mistakable traces of a preceding placentation, from the nature
of the uterine wall and the uterine vessels, &c.
In the case of Galeopithecus it twice occurred that a young
animal was yet being suckled by the mother and was found
attached to her breast, whereas autopsy showed an already
fairly advanced younger foetus to be present in the uterus of
the same specimen.
The fact that all the species here mentioned bring forth only
one young ata time (Tupaja, which regularly carries two foetus
simultaneously, alone excepted) may perhaps account for the
prolific properties here referred to, being developed as a counter-
balancing agency to this restriction of the number contained in
one litter. In our European Insectivora, whose time of repro-
duction is limited to only a few months or even weeks in the
year, the litter normally amounts to eight (Sorex) or six
(Erinaceus, Talpa) young ones.
Of all the cases that have as yet come under my observation
I know of only one case of twins in Nycticebus. They were
enclosed each in a different horn of the uterus, whereas in the
normal cases one of the two horns is always barren.
In Tarsius, Galeopithecus, and Manis I have never noticed
more than one young at a time. In Tupaja never more and
1 It should be here noted that I have on more than one occasion heard it
reported by sportsmen and natives that for the Indian deer, periods of
heightened and lessened sexual activity do exist. I will by no means gene-
ralise any further than my acquaintance with the species here investigated will
allow me to do.
84. A. A. W. HUBRECHT.
never less than two are present, occupying the right and the
left half of the uterus.
There, as formerly in Sorex, I have, however, been able to
establish without doubt that the number of fecundated eggs
and even yet of early blastocysts is constantly found
to be more considerable than the number of ripe
foetus that attain maturity and form the normal contents of a
litter.
Thus in Tupaja four and sometimes more blastocysts are
found in early stages, apparently all of them in equal conditions
of vitality. A struggle between these blastocysts for the
definite attachment to the maternal uterine wall is thus
inevitable. How this struggle is brought about and what
points finally decide between those that shall thrive and those
that shall perish is at present obscure. Still the fact has no
doubt a definite significance, considering that it is not a casual
observation, but a most regular occurrence in at least two genera
of Insectivora.
My preparations are not yet numerous enough to allow me to
speak with the same emphasis for the other genera.
Attention will of course have to be directed to this point, in
order to make out whether it may be regarded as a general rule
in mammals that more blastocysts than can partake in the
normal course of intra-uterine development are present in the
earliest days after fecundation has taken place.
One question to which my preparations do not allow me to
reply is that concerning the duration of pregnancy in the five
species investigated. ‘The lapse of time that occurs between
the date of fecundation and that of parturition is in no way
indicated even by the most complete set of intermediate stages
between the cleaving egg and the ripe foetus. On the other
hand, it is in no way of any importance for the correct interpre-
tation of the different and successive ontogenetical processes to
be acquainted with the exact rate at which these stages succeed
one another, or with the age of any particular stage as expressed
in a fixed number of days.
With animals bred in domesticity this is of course easily
SPOLIA NEMORIS. - 85
accomplished. But then, on the other hand, the domestic
animals have in later years been often shown to furnish us with
data that are more liable to a certain amount of divergence
than those which have been obtained from animals living in
absolute freedom. There can hardly be a doubt that the
inevitable pammixia which accompanies domestication can
alter and render variable parts of the organism both internal
and external, which have a more fixed standard in their non-
domesticated congeners.!
For this reason the study of mammalian ontogeny, not from
the rabbit and the Cavia, but from specimens of other species
and genera captured in their natural haunts, deserves special
recommendation.
The object of this paper being to establish certain general
facts that come to light when the pregnant stages of the five
genera in question are compared macroscopically, and before
the microscope is as yet brought to bear on the numerous and
intricate questions of histological detail, it will recommend
itself to treat the five genera separately.
Tarsius spectrum. Figs. 1, 2, 18—21, 47—49.
There can be no question that all the specimens obtained
by me belong to Tarsius spectrum, Pall., and not to Tarsius
fuscomanus, Fisch. The differences between these two species
have lately been fully discussed by Weber in vol. iii, p. 260,
of his ‘ Zoologische Ergebnisse einer Reise in Niederlandisch
Ost-Indien,’ Leiden, 1893. None of the uteri in my collection
were obtained from the localities to which Tarsius fuscomanus
is restricted. The name by which Tarsius is known to the
natives in South-west Sumatra is singo puar; those of Banka
call it the berook puar, or mentiling ; those of West Borneo,
tempiling.
Renson, ‘Contribution a l’embryologie des organes d’excrétion des
oiseaux et des mammifeéres,’ Bruxelles, 1883, p. 37.
C. K. Hoffmann, “ Die Bildung des Mesoderms, &c.,” ‘ Verh. v. a Kon
Akad. v. Wetenschappen te Amsterdam,’ 1883, p. 2.
J.v. Erp, Taalman Kip, ‘ De ontwikkeling der Miillersche gang by de
zoogdieren,’ Dissert. inaug., Utrecht, 1893, p. 77.
86 A. A. W. HUBRECHT.
In a yet higher degree than the other Prosimiz, Tarsius was
recognised by the older anatomists to be intermediate between
Insectivora and Primates. Burmeister, in the preface to his
‘ Beitrage zur naheren Kenntniss der Gattung Tarsius,’ writes
as follows (p. 6) :—“ ‘Tarsius possesses, in addition to its con-
siderable external similarity to monkeys, the most complete
insectivorous dentition which Quadrumana can boast of, for
even the incisors have adopted the type of the canines, and
have thus become eminently like the true dentition of the
Insectivora. In this Tarsius differs from all other Prosimie.”
The non-pregnant uterus of Tarsius has been figured on
pl. 6, fig. 22, of the above-mentioned work.
The author thus describes the internal female organs :—
«‘ They consist of two small ovaries, the coiled oviducts, and the
two-horned uterus. . . . The ovaries are small spherical
bodies, half a line in diameter; their surface is quite smooth,
and their inner substance is of the ordinary condition of that
of the higher mammals. . . . The uterus is two-horned,
each horn being three inches long ; then follows the unpaired
portion, which attains to half an inch, and externally passes
into the vagina without any interruption. On the inner sur-
face I could, however, detect a faint boundary as an ostium
uteri. The uterus horns, as well as the unpaired portion,
have thick walls, and show numerous considerable folds.”
I have now before me several dozens of non-pregnant and
early pregnant Tarsius uteri, and I have little to add to Bur-
meister’s observations. There is, however, very often a strongly
marked difference in size between the two ovaries, one swelling
up to the size of a pill, the other remaining considerably
smaller. I was inclined to believe that this difference in size
might go parallel with fecundation, and thus indicate the
presence of an early developmental stage in a uterus with one
of the ovaries thus swollen. Series of sections in which the
uterus lumen and that of the oviduct have been most care-
fully scrutinised, oblige me to give a negative answer to this
conjecture. The cause of this swelling of one of the ovaries
was investigated, and will be treated of elsewhere, It was
SPOLIA NEMORIS. 87
more than once noted in the fresh animal before preservation
by one of my correspondents, to whom I am indebted for most
valuable material.
Concerning the aspect of the internal genital organs of
Tarsius when fresh and in situ, he tells me that the colour of
the ovaries is often very different. Sometimes pink, they are
at other times of a lighter and darker yellowish hue; and in
young specimens they have the appearance of a small row of
spherical or rod-like bodies of a light yellow colour. I have
not yet found time to study the sections of the young stages
of the ovaries thus characterised.
The body of the uterus with its double horns of the pre-
served specimens in my possession is extremely variable in
shape according to circumstances. It is difficult to detect the
very early stages of pregnancy at first sight.
Yet long before the embryo has proceeded so far that the
medullary groove has made its first appearance on the surface
of the blastoderm, there is a very marked swelling of the
uterine half in which the blastocyst has come to adhere.
This uterine swelling is in no way perfectly spherical, but
more saddle-shaped, in accordance with the fact that even in
these early stages the blastocyst adheres to the maternal tissue
in one particular region, and not along any more extensive sur-
face, as, for example, in the shrew, the mole, the hedgehog, &c.
The details of this process will be fully described elsewhere.
I may here add that this early point of attachment corresponds
in situation to what will, in a later stage, become the placenta,
and that no omphaloidean attachment precedes as a temporary
structure the definite placentary connection.
When pregnancy advances it can be noted that the placenta
does not occupy a varying but, on the contrary, a fixed position
with respect to the different regions of the uterus. Itis always
situated close to the apex of the horn on the mesometrical side,
and the swelling of the uterine walls is not most conspicuous
close to this point of attachment, but more towards the vaginal
portion of the horn (ef. fig. 1). It is in this more extended
part of the uterus that the head of the full-grown fetus is
88 A. A. W. HUBRECHT.
situated, which is thus normally the first to pass outwards at
birth (cf. fig. 18).
When a uterus containing a ripe or nearly ripe foetus is care-
fully opened by a longitudinal incision, there is seen to be no
adhesion whatever except in the placentary region (figs. 18—
21). The uterine walls are stretched to an extreme degree of
tenuity ; indeed, so thin have they become that even in the
specimens that were preserved in spirits, and have thereby
considerably increased in opacity, the limbs, the ears, the
fingers, and the tail of the foetus can be distinguished through
this thin layer of maternal tissue.
Immediately beneath the stretched uterine wall the feetal
envelopes form a very tight sac containing the foetus. This
sac is so transparent that in spirit specimens the individual
hairs on the head, body, and limbs, the nails, &c., can be
recognised through it (figs. 18 and 47).
Towards the tail end of the foetus the foetal envelopes pass
into a button-like projection, which constitutes the placenta.
Figs. 18, 19, and 47 show this both in the front view and in
profile ; in fig. 20 the longitudinal section indicates still more
clearly the way in which the thin foetal envelopes merge into
the placentary tissue. It is, moreover, visible, both in fig. 18
and in the longitudinal section (fig. 20), that the placentary
knob itself adheres with the maternal tissue only along a very
limited extent of its total surface, viz. a squarish area in which
numerous lumina are visible (figs. 18 and 47), when the
placenta is loosened from the maternal tissue by a slight
shaking. These lumina are formed by the tracts which convey
maternal blood to and from the placenta. Microscopical
examination of thin sections through this region reveal without
any doubt that, indeed, this limited area is the only point of
fusion, the remaining surface of the placenta being as little
fused with the maternal tissues opposite to it as are the foetal
membranes themselves. About the histology and the genesis
of the placenta of Tarsius I will treat in a later paper; suffice
it to say that, according to the nomenclature now in use, the
Tarsius placenta would be directly classed with the discoid
SPOLIA NEMORIS. 89
type. It has not the faintest trace of any relation whatever to
the diffuse type, which was hitherto considered as being the
type of placenta to which the Lemuride belong.!
The umbilical cord by which the embryo is connected with
the placenta is comparatively short; it is represented in figs.
20 and 21, containing very prominent vessels.
In fig. 21 the ramification of these vessels on the placenta is,
moreover, indicated as this is seen (in a spirit specimen)
after the removal of the foetus. Fig. 49 represents the foetal
membranes and the placenta with severed umbilical cord after
they have been removed out of the uterus, and the foetus has
passed out of its envelopes.
These envelopes having here been preserved after the foetus
had been expelled are less stretched and transparent than those
of figs. 18 and 47. The afterbirth of Tarsius (which is expelled
in the customary way and not resorbed in situ, as that of
Talpa) consists of these same parts; the envelopes are then
more folded together against the knob-like placenta than in
fig. 49. ;
Embryos of Tarsius are in my possession from the earliest
stages of segmentation up to the newly born young. Two of
them are represented in figs. 46 and 48. In the first the
comparatively large size of the head is worthy of note; in the
second the way in which limbs, fingers, and tail are tightly
folded together against the body in a small compass deserves
special attention.
The details of the ontogeny of Tarsius, which as yet has
never been investigated embryologically, I hope to be able to
work out soon with the aid of the very complete material now
at my disposal.
With respect to the details of the placentation process I will
also have to refer to a later publication, and can only state
that the trophoblast of the very early two-layered blastocysts
undergoes a most considerable amount of proliferation at the
spot where the uterine surface has in its turn undergone certain
1 Cf. a preliminary notice in the ‘ Proces Verbaal van de Koninkl. Akademie
van Wetenschappen te Amsterdam,’ Zitting van 2 April, 1892.
90 A. A. W. HUBRECHT.
differentiations intended for the future attachment of the
blastocyst. This proliferation, the products of which undergo
remarkable further developmental changes, eats its way very
deeply into the maternal tissue between the tubular uterine
glands.
Vascularisation of this proliferated region, which fuses
in a particular way with the surrounding maternal elements,
is then brought about, for maternal blood circulates in it
freely and copiously, and soon another system of vascular
channels connects the growing embryo with this rich source of
energy. 3
A very early and profuse growth of mesoblastic tissue plays
an important part in this secondary connection between the
growing foetus and its chorion, and accentuates in a suggestive
way the several features by which Tarsius approaches the
Primates.
However, I shall have to postpone a detailed description of
this point to a later publication.
Nycticebus. Figs. 3—5, 22, 23, 30—40, 50—56.
This second genus of Prosimiz, represented in the Archi-
pelago by the species Nycticebus tardigradus and N. javanicus,}
is known by a series of names which have much the same
sound, but in which the consonants vary according to the
different regions. These names are—kukang, tukang, pukang,
and huhang. In East Sumatra and Banka the name of berook
semoendi is in vogue among the natives. In East Java speci-
mens were especially difficult to procure because the skeleton
is said to be most efficacious in bringing about death and de-
struction among the unfortunate inhabitants of a house in
‘front of which it has been buried overnight. It is thus in high
1 Of this species I have obtained but very few specimens in East Java,
and no pregnant uteri. Although no specific determination was ever made
by those who so kindly collected and preserved the uteri at present available,
T have no doubt that they all belong to the only species which is known to
occur in the islands from which my collections have come (Sumatra, Banka,
and Borneo), viz. Nycticebus tardigradus,
SPOLIA NEMORIS. 91
demand among the wealthier natives who have family quarrels
to settle, and I have known exorbitant prices, with which a
collecting embryologist could not possibly compete, to be
stealthily paid for one specimen, for this unfriendly though
perhaps harmless purpose.
As will be seen, Nycticebus differs most considerably from
Tarsius in several important respects.
The stages of pregnancy, as studied from the unopened uterus,
are not characterised by any very marked peculiarity. In the
three uteri figured on PI. 9 the ovary is seen to be more or less
concealed by a mesenterial fold, which contains the Fallopian
tube, whereas the two horns of the uterus have a peculiar asym-
metrical shape, being rounded dorsally and pointed ventrally.
This latter detail, which can be easily recognised in the uteri
that are young or in early stages of pregnancy, is of course lost
as the swelling of the pregnant horn increases. Still, even
then it can yet for a very long time be detected in the non-
pregnant horn.
In the literature on the Mammalia I do not find any other
representation of the uterus of Nycticebus than those contained
in Kuhl’s “ Einiges tuber die Splanchnologie von Stenops
gracilis ” (Beitrage zur ‘ Vergl. Anatomie, zweite Abtheilung,’
p. 37, pl. 6, Frankfurt, 1820) ; and in Schroeder van der Kolk’s
papers, ‘ Bijdrage tot de Anatomie van den Stenops kukang’
(‘Tijdschrift voor Nat. Gesch. en Physiol.,’ vol. viii, pl. 5, figs. 8
and 9, Leiden, 1841). This latter figure is most insufficient,
and does not in any way indicate the peculiarity just mentioned.
Moreover in these figures other peculiarities—for example, an
abnormal extremity of the Fallopian tube (I. c., fig. 9)—are
represented, and a total absence of fimbria is noticed which
does not conform to the actual facts, and which differs markedly
from what figs. 3 and 7 teach us. V. d. Kolk’s specimens
must have been somewhat mutilated and perhaps imperfectly
preserved.
The first pregnant uterus of Nycticebus which I opened was
the object of particular expectancy. Knowing that for the
Madagascar Lemuroids (Propithecus, Indris, Avahis) both
92 A. A. W. HUBRECHT.
Milne Edwards! and Turner? had described and figured a
diffuse placenta, which was, however, first distinctly recog-
nised as such by the latter, and that Tarsius in this respect
reveals such a totally different arrangement, it was of course
of a double interest to know whether Nycticebus would conform
with either of these types, or would represent one by itself.
The first dissection which I ventured to make was for this
reason effected with special precautions. It is represented in
figs. 22, 30, and 36, and from the first of these three figures
it will be seen that, to begin with, the muscularis was carefully
peeled away. ‘The outer surface of the mucosa thus brought
to light revealed (by transparency) the presence of a network,
the meshes of which are visible to the naked eye. The cha-
racter of this network could be better recognised as soon as the
incisions had been made that are represented in fig. 30, M
being the same flaps of the muscularis that are indicated in
fig. 22.
The mucosa (m) was seen to present projecting ridges
arranged in reticular fashion, and between which polygonal
areas were thus enclosed, into which villiform protuberances of
the underlying foetal envelopes were seen to fit. So loosely
did they fit, however, that no traction whatever was required
to sever the connection between chorion and mucosa all along
this spherical surface. The foetus with its envelopes could be
floated out of the mucosa the moment the preparation repre-
sented in fig. 80 were to be turned upside down.
The reticular surface of the mucosa is seen in a much more
natural connection of the parts in figs. 23, 31, and 32, where
the uterus has been opened and the flaps cut out of the wall
have been folded back. The mucosal network and the muscu-
laris have here remained unseparated. Stillit was quite as easy
to remove them from the subjacent foetal envelopes as it was
in the foregoing case.
! A. Milne Edwards et A. Grandidier, ‘ Histoire Naturelle des Mammiféres
de Madagascar,’ Paris, 1875.
* Turner, “On the Placentation of the Lemurs,” ‘ Philosophical Trans-
actions of the Royal Society,’ 1876, p. 569, pls. 49—51,
SPOLIA NEMORIS. 93
And so all these preparations leave no doubt that with
respect to the connection between mother and foetus Nycticebus
resembles ever so much more closely the Madagascar Prosimiz
than it does Tarsius.
However, there are differences between the Madagascar
genera and Nycticebus that deserve special mention. Firstly,
the maternal network in the former (Milne Edwards, 1. c.,
pl. 114, fig. 1) is much less decidedly reticular, and, on the
contrary, more lamellar than what is here represented (figs. 38,
51, 52, 56) for Nycticebus. Turner’s figs. 6 and 8 (also taken
from Madagascar lemurs) agree very closely with those of
Milne Edwards. Secondly, the outer surface of the fetal
envelopes is very much the counterpart of the maternal ar-
rangement, as can more especially be seen from Turner’s figs. 3,
4, and 12; but also from Milne Edwards’ pl. 114, 117 (3), and
118 (1). And in this respect Nycticebus presents the same
phenomenon of concordance between the foetal excrescences
and the maternal crypts, so that, instead of the lamelliform
arrangement of the chorionic surface so conspicuous in the
Madagascar lemurs, we here find circumscribed short columnar
villi, each one of them fitting into a corresponding depression of
the maternal reticulum. These columnar villi are quite equally
distributed over the whole surface of the chorion, as is more
particularly indicated in figs. 28, 31, 50, and 53. As preg-
nancy draws to its close, these chorionic villi disappear on a
restricted chorionic area, which covers the head of the foetus
and is directed towards that side where the corpus uteri and
vagina are situated. The maternal surface opposite this part
of the chorion is similarly non-reticulated. A flattened pro-
jection of the chorion, similarly without villi, is sometimes
found attached to this anterior surface of the chorion. Both on
the latter and on the projection here alluded to we find that
the epithelial recesses, which will be mentioned lower down,
are, all the same, preseut. The greater part of the chorion
just before birth is, however, densely covered with the particu-
lar villi that indent into the maternal reticular crypts. The
transitional region between the areas is represented in fig. 55.
94. A. A. W. HUBREOCHT.
In fig. 35 the foetal envelope is seen in natural size, and between
the villi numerous openings (ap.) are detected. In the earlier
stages these openings are also already present, and can be easily
seen with a lens or even with the naked eye. If we open the
chorion enveloping the foetus (fig. 30) we find the inner surface
of what was the villiferous covering of the foetus to be flat, and
this inner surface to be only here and there interrupted by round
patches (R.), each of which corresponds to one of the openings
(ap.) just meutioned. Of these relations of the parts, figs. 30,
32, 34, and 386 give further elucidation, whereas the definite
proof of the correspondence of the flattened and faintly
prominent recesses (#.) with the apertures (ap.) can of course
be more especially obtained in sections, as that of fig. 39.
The distribution of vessels on the inner surface of the chori-
onic envelope is more particularly visible in fig. 84; the attach-
ment of the umbilical cord to the same in figs. 32, 33, and 36.
The villi themselves are at first (fig. 50) more cylindrical ;
when they increase in age they become folded and wrinkled to a
not inconsiderable extent, as is visible in figs. 37 and 37a. It
may be expected that these folds and wrinkles correspond to co-
ordinated arrangements of the reticular layer of the mucosa,
the two thus fitting together in a very simple way.
The maternal folds on the mucosa are in the Madagascar
lemurs interrupted at regular distances by small bald patches,
both according to Turner (I. c., figs. 6, 8, and 9) and to Milne
Edwards (woodcut on p. 280).
In these spots the tubular uterine glands open out between
the folds that have arisen on the inner surface of the uterus in
the course of pregnancy. In Nycticebus I find a more equal
distribution, the gland openings being found in the centre of
nearly every separate compartment of the reticular arrange-
ment. In fig. 38 this is indicated, the darker shading at the
bottom of these compartments representing gland tracts.
Viewed with a pocket lens the openings are often visible as
a whitish spot near the middle, where they appear to be more
concentrated.
Figs. 39 and 40, drawn with the camera with very low power,
SPOLIA NEMORIS. 95
give the exact relation of the maternal and the embryonic
parts in a section through chorion and uterine wall. Of the
latter, muscularis and mucosa are indicated in fig. 40, the
elevated ridges of the mucosa that form the peculiar reticulum
referred to being visible as so many inward projections. They
are all covered by an epithelium which even in this far advanced
stage can be readily distinguished. Immediately below this
epithelium numerous finely branched maternal blood-vessels
take their course, in every respect comparable to those which
both Turner and Milne Edwards have made out by injections
for the Madagascar Lemuroids.
The chorionic villi of Nycticebus are seen to fit very exactly
into these cryptiform spaces; it is worthy of remark that the
epithelium on the villi is in many places ever so much thicker
and more considerable than what is found on the opposite
maternal surface.
In the villi numerous embryonic capillaries take their course
immediately below the epithelial layer. The two vascular sur-
faces are thus separated only by the thickness of two cell-layers,
of which the maternal one is less high and less columnar.
The above-mentioned recesses (#.)in the chorion are clothed
by a direct continuation of the chorionic epithelium. Smaller
vascular villi with a much less massive core of connective tissue
stand out into the lumen of these recesses, as can be seen both
in fig. 89 and fig. 40.
The amnion which enshrouds the foetus has been partly
removed in fig. 30, and is partly folded back (after removal
of the foetus) in fig. 82. Also in figs. 36 and 54 it has been
dissected away, whereas in these two latter figures the connec-
tion between the foetus and the villiferous chorion by means of
the umbilical cord is still retained, the chorion being partly
inverted in the act of stripping off the embryo.
In the preparations here figured no indication is given of the
yolk-sac and the allantois. In Milne Edwards’ figures of
Madagascar lemurs a very conspicuous place is allotted to the
allantois, which he has inflated, and which thus showed digitate
processes and a multilobulate shape. It is thus described as
96 A. A. W. HUBRECHT.
being non-vascular. The exact terms of Milne Edwards are
the following (1. ¢., p. 283) :—** Les parois de l’allantoide sont
délicates et transparentes, aucun vaisseau ne s’y distribue. Si
Von injecte un liquide coloré dans le pédoncule de cette enve-
loppe membraneuse on peut le suivre dans l’ouraque, & travers
le cordon ombilical, jusque dans la vessie urinaire; preuve
manifeste que cette poche, malgré ses caractéres anormaux,
représente exactement l’allantoide des autres mammiféres.”
The above description would suggest that in the Mada-
gascar lemurs the allantois plays a part which is to a certain
extent comparable to what Selenka! has described for Di-
delphia (l. ¢., pl. 16, figs. 1—5, pls. 17,18). But then in
Didelphia it is the yolk circulation by which the chorion is
vascularised, whereas in Milne Edwards’ lemur fcetus he
finds the umbilical vesicle to be extremely reduced. ‘Traces of
it can only be made out in embryos of very early develop-
mental stages.
This latter fact shows that a comparison with the Didelphia
does not carry us very far. The vascularisation of the chorion
of the Madagascar lemurs must be a phenomenon sui ge-
neris if Milne Edwards’ observations are confirmed; and it
will be understood that for this reason an exact insight into
the state of affairs as they present themselves in Nycticebus
is all the more desirable, especially if all the genetic stages be
closely followed, as the now available material promises to
admit of.
This will at the same time explain why I wish to refrain
from further discussing the point on this occasion.
Still I may be allowed to refer to an earlier publication in
which I have insisted on the advisability of restricting the use
in mammalian embryology of the name chorion.” I have there
argued at some length why I proposed ‘‘ henceforth to restrict
the use of the term chorion to man, and—dependent upon
1 *Studien z. Entwickelungsgeschichte der Thiere,’ Heft 4, “ Das
Opossum,” p. 136.
2 «The Placentation of Hrinaceus europeus, &.,” ‘Quart. Journ.
Micr. Sci.,’ vol. xxx, 1889, p. 382.
SPOLIA NEMORIS. 97
future researches—perhaps to the Primates.” What Selenka
has since made known with respect to monkeys, indeed, shows
a close resemblance between man and certain monkeys with
respect to these placental phenomena.’ And I would now
venture to insert in the above citation, after the word “ Pri-
mates”: “and to the Prosimiz.”
In accordance with this it will be seen that in the present
paper I have used the term chorion a few times only in refer-
ence to Nycticebus and Tarsius, whereas with respect to the
other mammals I prefer to employ the term “ diplotrophoblast ””
(l. c., p. 885). It is thereby testified that a foetal envelope is
present which is only secondarily vascularised, either by the
vessels of the allantois or by those of the yolk-sac.
And thus, for the present, the new data here adduced for
Nycticebus are restricted to the fact that the embryo of Nycti-
cebus is enclosed in a complete sac which is entirely covered
with thick villi, and which is very loosely attached to the
vascular meshes of the mucosa into which the villi fit.
I hope to be able to furnish ample information concerning
the ontogenesis of the chorion, &c., in a later publication.
A short reference to the two figures 55 and 56 should yet be
made. Fig. 55 is an enlarged photograph of part of the same
preparation represented in fig. 35. The actual shape of the
villi, their flattening and partial disappearance towards the
right extremity, is here better visible than in the lithographic
figure.
Fig. 56 shows very graphically what becomes of the earlier
network of the mucosa that was represented in fig. 51. The
frilling of the border of the ridges, which is not yet present in
the latter preparation but which becomes conspicuous in the
later phases of pregnancy, is better brought out in this photo-
graph than in the still more enlarged fig. 38.
The big folds that are visible in fig. 56 have arisen in conse-
quence of an intentional folding backwards of the uterine walls.
All around the central depression the reticulation is less
1 ‘Studien z. Entwickelungsgeschichte der Thiere,’ Heft 5, pl. 35, fig. 11;
pl. 36, fig. 5.
VoL. 86, PART 1.—NEW SER. G
98 A. Aw. W. HUBRECHT.
marked. Where the mucosa faces the flattened surface at one
of the poles of the chorion above alluded to, the reticulation is
also deficient.
Galeopithecus. Figs. 6—11, 24—29, 57, 58.
Concerning the ontogeny and the placentation of Galeo-
pithecus, I could find no data in the mammalian literature
but a few lines in an article of Gervais! on the cerebral
conformation of the Mammalia (1. c., p. 425). He does no
more than mention the fact that he examined a foetus of
Galeopithecus which was shown to possess a discoid placenta.
Without entering into any further details, he figures (I. c.,
pl. 22) the said foetus with outspread patagium and severed
umbilical cord (fig. 1), and the same folded together in its
intra-uterine position and attached by a thick and short
umbilical cord to a disk-shaped placenta on which a number of
radiating blood-vessels are indicated.
Gervais’s figure corresponds in a general way with fig. 29 of
this paper, only it is much smaller, and was probably not
figured natural size. In our fig. 29 the circular placentary
area is seen to lie as nearly as possible in the level of the
uterine surface, and not to form such a marked button-shaped
prominence as, for example, the placenta of Tarsius figured
close to it (fig. 20) does to such a considerable extent.
Though both discoid, these two placentas are, no doubt,
also in other respects profoundly divergent from each other.
Although I have as yet only a provisional acquaintance with
the chief stages of the placentation of Galeopithecus, I can
more especially call the attention to the peculiar aspect of the
placenta in figs. 24, 25, and 27.
It is already a discoid formation, but in these younger stages
it is less compact and less intimately soldered with the uterine
walls; the placental vessels, on the contrary, being mutually
interwoven in an intricate manner, and being applied as a
1 “Mémoire sur les formes cérébrales propres a différents groupes de
mammiféres,” ‘ Journal de Zoologie,’ vol. i, 1872.
SPOLIA NEMORIS. 99
delicate but prominent web (figs. 24 and 25) against the inner
uterine surface.
The foetus is connected with it by means of ashort umbilical
cord. Besides, there is a vascular connection between the foetus
and the remains of the yolk-sac.
The latter is represented in figs. 26—28, whereas in fig. 24
it has been dissected away in order to show the embryo enclosed
by the amnion in its attachment to the placentary region.
In fig. 26 nothing has been removed but the uterine wall.
The blood-vessels radiating over the yolk-sac are as distinctly
visible as a spirit specimen will admit of. To the right as well
as to the left the cut lumina of blood-vessels (cf. figs. 10 and
11) are seen to take their course in the thickness of the uterine
wall. re
On the right the placenta is represented by one free border,
which is, moreover, loosened from any uterine attachment ; the
rest of the placenta is hidden from view by the embryo and its
envelopes.
In the next figure (fig. 27), which has reference to this same
specimen, the embryonic sheaths have been opened and the
embryo is removed. ‘The membranes to the right are the yolk-
sac and the amnion. At the bottom ef the uterine cavity the
placenta can be distinguished.
In fig. 28 the embryo with its envelopes and with the
placenta has been wholly scaled out of the uterus. Of the
placenta an indented border is seen to the left of the figure,
whereas to the right only the yolk-sac has-been dissected and
turned over; the amnion, however, is still in its place, and
hides the embryo from view.
We have now to say a few words concerning the outer aspect
of the uterus before and during pregnancy. More than in any
of the species hitherto noticed the uterus of Galeopithecus may
be said to be double, the vagina being spacious and thick-
walled, and the two halves of the uterus (cf. figs. 2, 6a, 7a, 8a,
and 9a@) opening out into the vagina by separate openings.
There is no unpaired median cavity in common between the
two uteri, communicating by means of a single “ os uteri ” with
100 A. A. W. HUBRECHT.
the vagina. Still when this proximal portion of the vagina is
more closely examined, we find projecting into it a median
prominence carrying a uterine crescentiform ostium on its left
and on its right surface.
This fleshy projection must be looked upon as the partial
soldering in the median plane of the distal parts of the two
uteri, the fusion not having gone so far that it affects the
uterine cavities.
Pregnancy soon reveals itself by swelling of one of the uteri
(figs. 7—11). I have never noticed more than one feetus at a
time in Galeopithecus. The earlier swellings do not offer any
peculiarity that could not be gathered from the figures 6—9;
the later swellings, which come to take a marked ovoid shape,
are externally characterised by an uncommon distension of
vascular tracts in the uterine wall, which even in the preserved
specimens stand out—in relief—against the flat outer uterine
surface. This is no individual peculiarity, but is noticed in
all the uteri of later stages. In fig. 11 the phenomenon is more
marked than in figs. 10 a and 10 0; in all of them the central
parts of this radiating vascular arrangement correspond with
the mesometrium. The situation of the placenta is not in any
special relation to this vascular arrangement. The way in
which the ovary of Galeopithecus is partly hidden in a me-
senterial fold (figs. 75 and 80) has a certain resemblance
to what was noticed above for Nycticebus and represented in
fig. 4.
The foetus of Galeopithecus that are figured on Pl. 12 (figs.
57, 58) show that the patagium is already indicated at an early
moment. Fig. 58 represents, however, a not yet ripe foetus;
this is figured (natural size) in fig. 29.
After the young Galeopithecus is born it seems to remain
attached to the mother’s nipples for a not inconsiderable time,
considering that on more than one occasion a pregnant uterus
of the size of figs. 9—11 was prepared by one of my corre-
spondents out of a female in which a young animal of the
preceding litter was found clinging to the mother’s breast.
Vernacular names for Galeopithecus in the Archipelago are
SPOLIA NEMORIS. 101
kubin or kubing, krendéh-kentjeng, and walang kékkes (some-
times also applied to the flying squirrel or walang kdpo).
Tupaja. Figs. 12—17, 41, 59, 60.
This small Insectivore, which, as the vernacular name of
Tupaj indicates, might easily be confounded with squirrels, was
common in the plantations of coffee and cinchona in the
Preanger districts. It often goes by the name of coffee rat,
kekkés being the name by which the inhabitants of the above-
mentioned districts generally designate it.
Tupaja has never more than two young at a time, as was
noticed above. A uterus in an advanced stage of pregnancy is
represented in fig. 17, the vaginal portion being here cut away.
Most marked in this figure is the prominence of two reniform
regions in the uterine wall. If the uterus were turned over, two
exactly similar patches would be noticed. As in each of the
two swellings only one embryo is contained, it follows that
the placenta of Tupaja javanica must necessarily be double.'
This is, in fact, the case, the two placentas lying right and left
of the foetus. They are connected with it (as fig. 41 distinctly
proves) by an umbilical cord. This commences as a single
strand of tissue, then bends upwards along the fetus’ side, and
only divides into a quadruple set of blood-vessels above the
foetus’ back. Two of these latter strands (each containing two
vessels) continue in the same course, and vascularize the
placenta which is situated on the side opposite to that where
the umbilical cord passes upward, whereas the two other strands
bend at an angle of 180° and vascularize the placenta that is on
the same side as the umbilical cord. Fig. 41 will make all this
clear; it was taken after one of the two swellings of the uterus
was longitudinally cut open, the foetus being also halved.
The perfect regularity in the situation of the two placentas
of each foetus is a phenomenon in which the maternal tissue
plays a prominent part. If we examine transverse sections of
very much earlier stages of pregnancy, such as are represented
1 « Proces Verbaal der Koninkl. Academie van Wetenschappen te Amster-
dam,’ 27 Mei, 1893.
102 A. A. W. HUBRECHT.
in figs. 12—16, the lateral attachment of these early and
much younger blastocysts is seen to have come about in the
same spots where afterwards the reniform placenta will develop.
The attachment of the early blastocyst, long before any allan-
tois or allantoidean circulation has made its appearance, comes
about by means of a very considerable proliferation of the
trophoblast. The proliferating patches of trophoblast are
double, and face the two spots in the uterine wall alluded to.
It is beyond all doubt that even before this proliferation of the
embryonic trophoblast has commenced the maternal tissue has
become visibly modified in those four regions of the mucosa
which correspond to the future placentary region, i.e. the right
and left inner surface of the uterus-horn when cut transversely.
The tubular uterine glands are then more particularly
limited to the mesometrical and antimesometrical regions of
the lumen; on the spots in question the interglandular con-
nective tissue has proliferated with partial displacement and
partial obliteration of the glands there situated.
The uterine epithelium does not take part in this growth.
It is, on the contrary, destroyed by the trophoblastic prolifera-
tion as soon as the blastocyst commences to adhere. This
trophoblastic neoformation is then gradually vascularized
(maternal blood penetrating into it), and undergoes a series of
interesting but complicated histological transformations. In
a future paper I propose to treat this placentation process of
Tupaja more fully; it may here suffice to remark that against
these early placentary cushions the yolk circulation is first
applied, and that in further stages of pregnancy the yolk-sac
is again removed from thence and replaced by the allantoidean
blood-vessels which then constitute the definite double placenta.
The placentas are shed at birth as are those of Sorex and
Erinaceus ; they are not resorbed in situ as is that of Talpa.!
The fully ripe placentas, shortly before birth, are in Tupaja
connected with the maternal tissue by an area which all along
its outer circumference is most easily detached and very loosely
connected. ‘Towards the centre, where the principal blood-
1 Cf. ‘Quart. Journ. Mier, Sci.,’ vol. xxx, p, 346,
SPOLIA NEMORIS. 103
vessels pass in and out of the placentary structure, the adhesion
is more firm.
Foetus of Tupaja at a comparatively late stage of preg-
nancy are represented in their normal situation in their
envelopes and in the uterine horn in figs. 59 and 60. These
two were obtained from one and the same uterus. They mea-
sure (exclusive of tail) about 27 mm. At birth ee foetus has
grown to a length of about 40—50 mm.
Manis. Figs. 42—45.
This genus, of which I have an extensive collection of uteri
that were obtained from Manis javanica (the trengiling or
tangiling of the natives), has of late years been investigated
with respect to its placentation by Max Weber.’ This fact
enables me to restrict myself to a very short notice, the more
so as the earliest stages, of which Weber makes no mention,
have not yet been studied by myself at all, although they are
also well represented in my collection. This, again, has to be
reserved for a future publication. In explanation of the
figures given on Pl. 12 I may say that fig. 42, which is enlarged
twice, was an embryo that had been contained in the uterus of
which a portion of the inner surface is represented in fig. 43.
This inner surface is characterised by irregular villiferous
bands, which become more numerous and more closely approxi-
mated as pregnancy advances.
The foetus and its envelopes is very loosely applied against
the maternal surface, outgrowths on the outer layers corre-
sponding to and interlocking with the maternal villiferous
bands just noticed.
Sections of these arrangements are figured by Weber.
In the uterus of fig. 43 the embryo of fig. 42 was enclosed
in the membranes that are represented in fig. 44. . These mem-
branes are peculiar in so far as only a part of the sac appears
expanded, a considerable part being more collapsed. Only the
expanded portion carries villous bands that interlock with those
1 «Zoologische Ergebnisse einer Reise nach Niederlandisch Ost-Indien,’
vol. ii, 1891, pp. 1—118, pls. i—ix.
104 A. A. W. HUBRECHT.
on the uterine surface. The pregnant horn of Manis carries
but one embryo at a time, as was noticed by Weber. For the
external shape of the uterus and the very considerable size of
the ovaries I may also refer to that publication.
That the aspect of the foetal envelopes is not always that of
fig. 44 is shown in fig. 45, which represents a stage of about the
same age but of a symmetrical development. The spacious
yolk-sac is situated, as was already noticed by Weber, against
the lower concavity of this sac. The foetus enclosed in these
envelopes will measure about 20 mm. in length from the vertex
to the root of the tail; towards parturition it will have increased
to the size of about 14 cm. from the centre of the skull to the
root of the tail.
Additional Considerations.
Having terminated the description of the Spolia nemoris
as at this moment they lie before me, I may be allowed to give
a summary review of a couple of biological questions in their
present stage towards the solution of which I hope in the first
place to utilize the material collected.
These questions were already mentioned in the introduction,
and have reference to—
I. The origin and morphological significance of the cell
layers constituting the two-layered blastocyst of mammals.
II. The origin, the minute anatomy, and the morphological
significance of the placenta.
I;
Since the very youngest mammalian blastocyst has been
studied by means of sections and with the aid of the improved
methods of the last decades, our knowledge of those incipient
stages has grown very rapidly. For these earliest contributions
we are indebted to Rauber,! van Beneden,? Lieberkiihn,®
1 “Die erste Entwickelung des Kaninchens,” ‘Sitzungsberichte der
Leipziger Naturforschenden Gesellschaft,’ 1875, p. 103.
2 ¢ Bulletin de l’Acad. de Belgique,’ t. 60, 1875, p. 686; five years later
followed by ‘‘La formation des feuillets chez le lapin,” ‘ Archives de Biologie,’
vol. i, 1880.
5 “ Ueber die Kleimblatter der Siiugethiere,” ‘ Gratulationsschrift Nasse,’
Marburg, 1879.
SPOLIA NEMORIS, 105
and Hensen.' The rabbit and the bat were more especially
employed in these researches; Heape has added the mole,
Selenka several other rodents as well as the opossum, whereas
the hedgehog and the shrew were studied by myself. Of late
Duval and also Robinson have again investigated the rat and
the mouse. Nevertheless, we are at present very far from a
consensus of opinion as to the significance and the genesis of
the parts in the early didermic mammalian blastocyst.
In his well-known paper on the early development of the
rabbit, Ed. van Beneden was one of the first to give a com-
plete set of valuable illustrations of the segmentation of the
mammalian ovum and of the consecutive stages between that
process and the didermic blastocyst in which the mesoblast
begins to make its first appearance.
Several of his figures have since passed into every text-book,
although his interpretation, both of the earliest and of the later
stages, is not adhered to in the form in which it was originally
given. Concerning the earlier stages, Lieberkithn (I. c.) and
afterwards Kolliker* have demonstrated that not only the lower
layers, but also the epiblast of the embryo arises out of the
inner cell-mass—v. Beneden’s “ masse endodermique.” Con-
cerning the later stages, they pointed out that van Beneden
mistook (l.c., pl. 5, fig. 7, and pl. 7, fig. 2) for the mesoblast
what is in reality the embryonic epiblast. The latter mistake
was due to the presence of a trophoblastic layer of flattened
cells outside that embryonic epiblast.
As to the development of the hypoblast, van Beneden made
out that both in the rabbit and the bat it gradually extends
centrifugally around the inner surface of the monodermic
blastocyst, the centre of this irradiation being the thicker knob
of tissue where the embryo is being shaped.
Similar formation of the hypoblast has been described by
1 “ Beobachtungen iiber die Befruchtung und Entwickelung der Kaninchen
und Meerschweinchen,” ‘ Archiv f. Anatomie und Entwickelungsgeschichte,’
Bd. i, 1876.
2 “ Die Entwickelung der Keimblatter des Kaninchens,” ‘ Zoolog. Anzeiger,’
iil, 1880, pp. 370 and 390.
106 A. A. W. HUBRECHT.
v. Beneden and Julin for the bat, by Heape for the mole, by
myself for the shrew, and by Selenka for the opossum.
The extension of the hypoblast against the outer wall of the
blastocyst is obtained in a different way in the hedgehog, as I
have elsewhere described.! Instead of having to spread out
against the inner surface of the wall of the blastocyst, the
hypoblast of the hedgehog is from the beginning a solid knob
which develops into a closed sac by distension. Further dis-
tension goes parallel to further growth of the didermic blasto-
cyst.
The cause of the difference in development of the hypoblast
is most probably the ever so much smaller size of the hedge-
hog’s blastocyst when compared in corresponding phases with
that of the rabbit, mole, &c. This is in its turn caused by the
fact that the hedgehog’s blastocyst, instead of being located in
the uterine lumen, becomes included at a very early stage in
the midst of maternal proliferating tissue. (‘ Anat. Anz.,’ 11,
p- 906.)
In 1892 Dr. Arthur Robinson? published a paper in
which, starting from what he finds in the mouse and the rat,
which he has studied for himself, he looks upon the process of
hypoblast formation in the rabbit and bat in quite a different
light than has been done by former investigators. The
process in the hedgehog is, according to his views, more
directly comparable to what he finds in the mouse. He has
based on his observations a series of far-reaching theoretical
speculations that partly correspond to views propounded by
Sedgwick Minot in 1885.3 Robinson concludes that in
mammals it is not the hypoblast that spreads against the inner
surface of the epiblastic wall of the blastocyst ; but that, on the
contrary, the epiblast, at a quicker or slower rate, spreads over
the outer surface of a hypoblastic vesicle, which, according to
1 « Anat. Anzeiger,’ Bd. iii, pp. 511, 906; and ‘ Quart. Journ. Mier. Sci.,’
vol. xxx, p. 291.
2 «Quart. Journ. Mier. Sci.,’ vol. xxxiii, p. 369.
3 *Buck’s Reference Handbook of Med. Sciences,’ i, 528, 1885; and
‘ American Naturalist,’ September, 1889; also ‘ Human Embryology,’ 1893,
p. 107.
SPOLIA NEMORIS. 107
his views, is there from the very first and forms the greater
part of the wall of the monodermic blastocyst.
In support of these views the author discusses the existing
figures and descriptions of early mammalian blastocysts with
considerable ingenuity. A couple of very difficult cases, which
I see no possibility of including in Robinson’s speculative
attempt, are, however, by him passed over in silence. As
such I would, for instance, point out Selenka’s fig. 2, pl. 18,
of the opossum (‘ Studien z. Entwickelungsgesch. der Thiere,’
Heft 4), as compared both to earlier and later stages.
On p.46 of Merkeland Bonnet’s ‘Ergebnisse der Anatomie und
Entwickelungsgeschichte’ (vol. 11, 1892), G. Born, in referring
to Robinson’s paper, recognises that if the views therein con-
tained were confirmed, this would mean a total revolution of
our present interpretation of the earlier stages of mammalian
ontogeny. Born adds, “an der nothwendigen Nachprifung
der Resultate wird es nicht fehlen.”
Such a “ Nachprifung” can be fully instituted with the aid
of the material now in my possession. Already have I examined
continuous section series through more than sixty segmentation
stages and mono- and didermic blastocysts of Tupaja that
had not yet adhered to the uterine wall, and through fourteen
preparations of the same early stages of Tarsius.
I will elsewhere fully report about these preparations, but
may be allowed now already to assert that they go dead against
Dr. Robinson’s speculations, and that I have no doubt that
certain peculiarities observed in Tupaja will convince even
Dr. Robinson of the fact that the outer layer of the mamma-
lian monodermic blastocyst (i.e. the trophoblast) is not in
direct continuity with the hypoblast cells inside of it.
On the other hand, we must recognize in Dr. Robinson’s
speculations, as also in the preceding attempt of Minot (I. c.)
and Keibel (‘ Anat. Anzeiger,’ vol. ii, p. 770),! laudable efforts
1 Tcannot admit with Keibel the possibility of a ‘ Wachsthumsenergie
derjenigen Zellen des Hies welche friiher den Dotter umwuchsen,” which
would be unchecked for millions of generations after the disappearance of the
yolk, and which is by him meant to explain certain formative processes in the
blastocyst.
108 A. A. W. HUBRECHT.
to the solution of a puzzle which the comparison of the holo-
blastic ova of Mammalia and of the lower Vertebrates and of
Amphioxus present to us as yet. For my own part I hold that
the principal reason why so many divergent and conflicting
views have been consecutively adhered to with respect to the
early mammalian blastocyst is this, that the appearance of a
cavity in a segmenting ovum has never left room for a doubt
whether this cavity could be anything else than a segmenta-
tion cavity which, as such, was proclaimed to be homologous
with that of Amphioxus. This homology, I hold, does
not exist. It should be remarked that both epiblast and
hypoblast, that will build up the embryo, are in the early
monodermic stages of the mammals contained inside
this cavity; and that we have to expect the real segmenta-
tion cavity to arise between epiblast and hypoblast, as also in
mammals it actually does later on.
If the space enclosed in the earlier monodermic stages is not
the segmentation cavity, then there will be no more a priori
difficulties to understand that a great portion of it is afterwards
converted into the archenteron. In fact, as little difficulty as
that a part of the cubic space inside the shell of a hen’s egg
becomes converted into the chicken’s cerebral ventricles.
A comparison with another point in the embryology of the |
higher Vertebrates will show that the above conclusion about
the cavity of the monodermic mammalian blastocyst is less
hazarded than might appear at first sight.
Suppose for a moment that the details of the development of
the amniotic Vertebrates were absolutely unknown to us, and
that we were only fully acquainted with that of the anamnia.
And suppose then some embryologist to teach that the differ-
ence between the development of these anamnia and the as yet
unknown higher Vertebrates would, for instance, prove to be
this, that the latter manage to become suspended for a time in
their own body-cavity, he would be in some danger of provoking
hilarity, if not worse.
Still we find no difficulty in interpreting the latter pheno-
menon, thanks to the gradual steps by which embryology has
SPOLIA NEMORIS. 109
advanced. The segmenting ovum of mammals may thus be
said to present the peculiarity, that among the products of the
holoblastic segmentation only one or very few cells represent
the real embryo; whereas a very considerable number that
rapidly expand into a vesicle (against the inner wall of which
the hypoblast becomes applied later on, either in one way or
another), are segregated at an uncommonly early period in
order that they may help to bring about a satisfactory attach-
ment between the blastocyst (which sensu strictiori isas yet
enclosed in this early vesicle) and the mother.
Only when the inner cell-mass shows the first traces of
differentiation between those elements that will become hypo-
blast and those that will become epiblast cells, is the stage
reached that corresponds to the blastula of Amphioxus; only
then there can be question to look for the homologue of the
segmentation cavity. As hypoblast and epiblast are at first
firmly pressed together, this segmentation cavity is even then
not yet present. The monodermic mammalian blastocyst is
thus a pseudo-blastula stage, its cavity is not the real segmen-
tation cavity, but a cavity which could not fail to arise ever
since, for purposes of attachment and nutrition, an extreme case
of precocious segregation of certain epiblast cells has come to
occur in the mammalian ontogeny. These cells arrange them-
selves into a vesicle even before the two primary germinal
layers of the embryo have differentiated.
There can be no doubt, however, that, phylogenetically, it is
the epiblast from which these cells have been segregated ; and
this explains the intimate fusion which, after a certain time,
obtains between this outer layer and the embryonic epiblast at
the periphery of the latter.
If I am right in upholding that the cavity inside the mono-
dermic mammalian blastocyst is not the segmentation cavity,
and that this blastocyst is only a pseudo-blastula, then we
must similarly conclude that it is not a real holoblastic
segmentation which the mammalian ovum undergoes. Even
the name of tertiary holoblastic, which Rabl proposes to
apply to the mammalian ovum (‘Theorie des Mesoderms,”
110 A. A. W. HUBRECHT.
‘Morph. Jahrb., vol. 15, p. 165), does not yet sufficiently
express the fundamental difference by which the mammalian
segmentation process is characterised.
There is no shred of evidence that with the disappearance of
the yolk, which took place at a comparatively late stage when
the mammalian character had already become predominant, the
process of segmentation immediately fell back into the lines of
-the so infinitely more distant alecithal ancestral forms.
A further reason for the distrust with which we may look at
this apparently holoblastic segmentation process is the fact that
it finally results in the appearance of a tridermic blastocyst
with elliptical blastoderm, primitive streak, &c., entirely
corresponding to the arrangement of the Sauropsida. So these
later stages have not returned to the earlier modes of develop-
ment, but have continued along the lines laid down by the
hereditary transmission of characters that were peculiar to-those
ancestral forms that possessed a considerable amount of food-
yolk.
If the mammalian (pseudo) morula and (pseudo) blastula
were indeed comparable with the same stages, 7. e. with the
true morula and true blastulain Amphioxus and the Amphibians,
about one half of the segmentation spheres would represent
potential epiblast, and the other half potential hypoblast. Now
this is evidently not the case. By far the greater portion of
these segmentation spheres gives rise to what will afterwards
be not any integral portion of the embryo, but a part of the
foetal envelopes and of the membranous expansion by which
the embryo is connected with the mother. Suppose we were
able to repeat Roux’s or Chabry’s most important experiments
on the partial destruction of one or more of the segmentation
spheres with the earliest mammalian stages, then we might
predict with great certainty from the data that Rauber, van
Beneden, Heape, Selenka, and others have brought to light that
only if the mother-cell of the inner cell-mass were attained, the
normal development would be interfered with ; and that in the
case of other segmentation cells being punctured, only a local
defect in the foetal membranes would ensue. This hypothetical
SPOLIA NEMORIS. Ti
experiment may still further drive home what is meant by the
non-homology of the two cases of holoblastic segmentation and
of the cavities which arise in these two cases inside the mono-
dermic vesicles.
New and valid reasons are thus accumulated for designating
the outer layer of precociously segregated epiblast-cells that
form the wall of this vesicle by a separate name, which at the
same time gives expression to the consideration that adaptation
to nutritive conditions of an entirely novel nature has initiated
this phenomenon of precocious segregation, simultaneously
with the diminution and the final disappearance of food-yolk,
a phenomenon that was consequent upon the passage from the
Hypotherian to the Eutherian stage.
Already in 1888! have I proposed—and several embryologists,
Dr. Robinson included, have since accepted—the name of
trophoblast for this outer layer of the mammalian blastocyst.
Only lately? I have given a fuller definition of the term,
which is, however, only in one respect an amplification of the
original definition of 1888. Then already (‘ Anat. Anz.,’ iii,
510), I remarked that to the trophoblast belonged all those
peculiar cellular structures of the mammalian blastocyst which
had been indicated by different authors as Reichert’s cells and
Rauber’s “ Deckschicht ” (Kolliker), as “ Trager ”’ (Selenka),
Ektodermawulst (Kolliker), horseshoe-shaped proliferation
(van Beneden). To this list may yet be added Duval’s
** formation ectoplacentaire.”
The amplification of 1893 just alluded to was this, that I
not only defined the trophoblast as “ the epiblast of the mam-
malian blastocyst that does not take part in the formation of
the embryo,” but that I added to this, “ or of the inner lining
of the amnion cavity.”
The difference which obtains between the trophoblast and
between the embryonic epiblast contributing to the formation
of the embryo and of that inner lining of the amnion cavity, is
1 « Anatomischer Anzeiger,’ July, 1888, p. 510.
_? © Proces-verb. van de Kon. Akad. van Wetenschappen te Amsterdam,’ 27
Mei, 1898,
113 A. A. W. HUBRECHT.
most distinctly brought out in such mammals as Pteropus,
Cavia, Tupaja, and others. Take, for instance, Selenka’s figure
of the guinea-pig’s blastocyst,! Gohring’s of that of Pteropus.?
In these latter figures, we see the epiblastic knob which is
enclosed between the trophoblast and the hypoblast of the
didermic blastocyst, hollowing out into a cell-mass, of which
the upper surface thins out and becomes the epiblastic clothing
of the amnion cavity, whereas the lower surface thickens and
becomes the epiblast of the blastodermic surface out of which
the embryo will be modelled.
I have no doubt that in the cases of Erinaceus and Sorex a
similar sharp line of demarcation may be drawn between the
epiblast that will develop into the lining of the amnion and
between the trophoblast, although here this distinction is not
so self-evident as in the preceding cases. And I suspect that
even such cases as that of the rabbit will some day admit of a
sharper delimitation of these two.
But even where such a sharp delimitation is not as yet
always possible in the later stages of the blastocyst, the earlier
stages are all the more evident.
The Ornithodelphia are not as yet affected by the causes
which determine the differentiation of a special trophoblast in
the higher placental Mammalia. In the Didelphia we may
hope to find certain transitory stages. Thus the early stages
of Phascolarctos, the ovum of which has been described by
Caldwell, may be expected to be especially instructive.
Already in the opossum Selenka has described the very
peculiar proliferation in the outer layer of the early blasto-
cyst (l.c., Heft 4, pl. 20, figs. 2,5, and ), which is no doubt
precursory to the ever so much more important proliferations
of the trophoblast which occur in most of the higher orders of
mammals.
In this paper it has already been noticed how both in
Tupaja and in Tarsius, portions of the trophoblast undergo
1 «Studien zur Entwickelungsgesch. der Thiere,’ Heft 3, pl. 12, figs. 13—
15, 73.
2 Tbid., Heft 5, pl. 41, figs. A—C, 1; 2, 4, and 6.
SPOLIA NEMORIS. 113
very active proliferating processes preparatory to the placentary
fixation of the blastocyst, whereas in my former papers I have
described the same activity for Erinaceus! and Sorex.*
Robinson’s speculations having tended to bring the part that
the hypoblast plays in the mammalian blastocyst more into
prominence, E. van Beneden has, on the contrary, upheld® that
the inner layer—his “so-called” hypoblast—of the mam-
malian blastocyst is not homologous with the hypoblast of
Amphioxus, but should be regarded as a yolk-envelope and be
no longer designated by the name of hypoblast, but by that of
lecithophore.
These views, tentatively accepted by Rabl,* have been com-
bated by Keibel,> by myself,® and by others. Also with
respect to this question I have no doubt that the material here
described will furnish very useful and perhaps decisive data.
Decisive, for example, with respect to the question whether
mesoblast takes its origin out of this hypoblast layer (v.
Beneden’s lecitophore), as Bonnet? and Hubrecht® have dis-
tinctly stated and figured it to do, although others (Keibel,® for
instance) deny this. It is clear that such participation in the
formation of the mesoblast is in itself sufficient to invalidate
v. Beneden’s considerations about the “lecitophore” and to
establish the homology of this layer with the hypoblast of
Amphioxus and the lower Vertebrates.
I have myself tried to explain the peculiarities that in mam-
mals also attend the formation of the hypoblast by the sugges-
' «Quart. Journ. Micr. Sci.,’ vol. xxx.
2 Thid., vol. xxxi.
3 * Anatomischer Anzeiger,’ ili, p. 713.
4 « Theorie des Mesoderms,” ‘ Morphol. Jahrb.,’ Bd. xv.
° “Zur Entwickelungsgesch. der Chorda bei den Saugern,” ‘ Archiv fur
Anat. und Physiol. Anat.,’ Abth., 1889.
6 «Development of the Germinal Layers of Sorex vulgaris,” ‘Quart.
Journ. Micr. Sci.,’ vol. xxxi, 1890.
7 © Beitrige z. Embryologie d. Wiederkauer,” 1 und 2, ‘ Archiv fiir Anat.
und Physiol.,’ Anat. Abth., 1884, p. 170, and 1889.
S “Ueber die Entwickelungsgesch. des Schweines,” ‘ Anat. Anz.,’ vi, 1891,
and Schwalbe’s ‘Morph. Arbeiten,’ Bd. ili, 1893, S. 69.
VoL. 36, PART 1.—NEW SER. H
114. A. A. W. HUBRECHT.
tion that precocious segregation of part of the hypoblast comes
into play and that we have to distinguish a coenogenetic and a
palingenetic hypoblast. This suggestion has been favorably
received ; the natural counterpart of it is the above sketched
precocious segregation of part of the epiblast. Both are adapta-
tions to similar external conditions.
if,
The origin, the minute anatomy, and the morphological
significance of the placenta have been of late inquired into by
a considerable number of independent investigators. It may
suffice to cite among the more recent ones Duval, Strahl,’
Frommel,’ Fleischmann,‘ van Beneden,® Masius,® Lisebrink,’
Heinricius,? Minot,? Hubrecht,!° and others. Questions that
were more particularly entered into are those that concern the
fate of the maternal epithelium at the spot where the blasto-
cyst comes to adhere against the uterine surface. In Erinaceus
it undergoes changes that are very different from those that
take place with it in the rabbit, and different again from what
happens with it in the Carnivora. In Sorex the fate of the
1 M. Duval, ‘‘ Le Placenta des Rongeurs,’ Paris, 1889-93.
2 H. Strahl, ‘‘ Untersuchungen iiber den Bau der Placenta,” I—ILV,
‘Arch. f. Anat. u. Physiol.,’ 1889, 1890. V. Anat. Hefte von Merkel u.
Bonnet, 1892.
3 R. Frommel, ‘ Ueber die Entwickelung der Placenta bei Myotus murinus,’
Wiesbaden, 1888.
4 A. Fleischmann, ‘ Embryologische Untersuchungen,’ Hefte 1—3, Wies-
baden, 1889-93.
5 i. v. Beneden, “De la formation et de la constitution du placenta chez
le Murin,” ‘ Bull. Acad. roy. Belg.,’ 3¢ ser., t. 15, 1888.
6 J. Masius, “De la genése du placenta chez le lapin,’ ‘Archives de
Biologie,’ vol. ix, 1889.
7 ¥. W. Liisebrink, ‘‘ Die erste Entwickelung der Zotten in der Hunde-
placenta,” Anat. Hefte von Merkel u. Bonnet, ii, 1892.
8 Heinricius, ‘‘ Ueber die Kutw. u. Struct. d. Placenta beim Hunde,” ibid.
“bei der Katze,” ‘ Arch. f. mikr. Anat.,’ Bde. 33 u. 37.
9 C.S. Minot, “ Uterus and Embryo,” ‘Journal of Morphology,’ ii, 1889.
10 Hubrecht, “ Erinaceus,” ‘Quart. Journ. Mier. Sci.,’ xxx, 1889 ; “Sorex,”
ibid., xxxv, 1894; and ‘ Verhandel. k. Akad. v. Wetensch. Amsterdam,’
2¢ Sec., vol. iii, 1893.
SPOLIA NEMORIS. 115
maternal epithelium is yet more peculiar, considering the fact
that an uncommonly marked proliferation of this epithelium
precedes its definite disappearance.
Secondly, the question has been much ventilated which part
the trophoblast plays in the attachment of the blastocyst.
Both in Insectivora (Erinaceus, Sorex, by myself) and in
Rodentia (rabbit, mouse, rat, Meriones, Cavia, by Duval) this
has been fully inquired into, and overwhelming evidence has
been forthcoming to show that this epiblastic layer and no other
but this layer contributes in an unexpected measure to the
genesis of the tissues that constitute the placenta.
It may even be said that since the penetration of maternal
blood into lacunar spaces of the hedgehog’s trophoblast that
are devoid of any vascular endothelium has been described as
occurring even at stages as early as the didermic blastocyst,!
and since Duval made his first communication about the rabbit
and other Rodents to the Paris Société de Biologie,2—communi-
cations that have soon after been worked out in his masterly
volume, ‘Le Placenta des Rongeurs’ ( Paris, 1889-92),—a conflict
of opinion has arisen about the real nature of the placenta, in
which on one side are a majority of the above-cited German
anatomists, and on the other the two authors just named and
also E. van Beneden with reference to the bat (*‘ Comptes
Rendus de la Société de Biol.,’ vol. v, Novembre, 1888), and
J. Masius, his pupil, with reference to the rabbit (I. c.).
The question centres in the way in which the osmotic
interchange between the maternal and the embryonic blood
comes about, influenced as it is by preparatory processes that
take place in those regions where the trophoblast of the
blastocyst comes in contact with the inner lining of the uterus.
Now this phenomenon is easy enough to understand in the
horse, the pig, and several other mammals on which the
1 Hubrecht, ‘Kleimblatterbildung und Placentation des Igels,” ‘ Ver-
handlungen der Anat. Gesellsch.; Versammlung zu Wirzburg,’ Mai, 1888 ;
‘Anat. Anz.,’ ili, p.512; and ‘‘ The Placentation of Hrinaceus europeus,”
‘Quart. Journ. Micr. Sci.,’ vol. xxx.
2 *Comptes-rendus de la Société de Biologie,’ Mars et Juillet, 1887;
Octobre et Novembre, 1888, vols. iv et y.
116 A. A. W. HUBRECHT.
researches of Turner, Ercolani, &c., have already years ago
thrown a flood of light.
We find there what we find repeated in two of the genera
which are treated of in this paper, viz. Manis and Nyc-
ticebus. The outer layer of the blastocyst acquires numerous
villiferous processes that are vascularized and fit into vascular
crypts of the maternal wall, out of which they are retracted at
birth with the greatest facility. In Nycticebus the two epithelia,
both the embryonic and the maternal, remain intact, and the
osmotic interchange takes place through two cell-layers of
different origin and of different physiological significance (phy-
logenetically).
As soon as the complications in this arrangement commence
to make themselves felt, which are so varied and so charac-
teristic in the different and so-called “deciduate” orders of
mammals, a clear insight is much less easily obtained.
Partly because as yet only a restricted number of genera has
been examined sufficiently in detail; partly because when such
investigation has taken place the different observers do not
always concur in the interpretation of the phenomena which
present themselves on examining the microscopical prepara-
tions of the same species.
A significant cell-layer is by the one declared to be maternal,
by the other to be of embryonic origin, Maternal blood is by
the one said to be enclosed in vascular spaces, that never lose
their real character of further extensions of capillary vessels,
whereas the other pretends that the maternal blood penetrates
sometimes at a very early, sometimes at a later stage of the
ontogenesis into lacunar spaces that are wholly surrounded by
tissue that is exclusively of embryonic origin,
Duval very tersely expresses the latter view, of which he is
himself one of the staunchest advocates, as follows :—** Le
placenta représente a son origine, une hémorragie maternelle,
circonscrite ou enkystée par des éléments fcetaux ecto-
dermiques.” The fact that certain interpretations based on
older researches, that could not yet profit by the modern
technical improvements, have been adopted in the text-books,
SPOLIA NEMORIS. 117
gives a long vitality to views which would most probably be
soon abandoned if the problem were now-a-days brought for-
ward for the very first time. Similarly, generalisations that
were based on incomplete data, although fully justified at the
time when they were made, are now found to obstruct the way
to a certain extent.
One of the mammals that will facilitate the real under-
standing of the method according to which the very simple
manner of foetal interchanges above alluded to has been
converted into the more complicated placentary structures,
is the mole. Some years ago I called attention to the fact
(Quart. Journ. Microsc. Science,’ vol. xxx, pp. 346 and 388)
that here, too, embryonic villi that cover the foetal envelopes
are easily drawn out of their sheaths at birth, and that no
afterbirth is shed, although the animal has a discoid placenta,
which up to lately was held to mean that it was also deciduate.
I then expressed the opinion that not only the mole is not
deciduate, but that even embryonic tissue is left behind against
the uterine surface, and is gradually resorbed in situ.
According to the patient investigations made by Mr. Vern-
hout, a pupil of the Utrecht Zoological Laboratory, which are
at present in the press, this is actually the case. Mr. Vern-
hout has cleared up the early details of the mole’s placentation,
and comes to very different conclusions from those of Strahl.
We may say that in the mole the epithelial connection, as
it was described above for Nycticebus and others, is a phase
that is very rapidly passed over, and that it is followed by the
application of a trophoblastic cell-layer against the maternal
epithelial layer. According to Mr. Vernhout’s investigations,
based upon preparations which I have myself repeatedly had
occasion to compare with the drawings which he is about to
publish, the maternal epithelium is very rapidly destroyed, the
trophoblast now becoming a pseudo-epithelium by which the
denudated mucosa and its deepening crypts are covered. Into
these crypts, which are in fact of embryonic origin, the allantoic
villi penetrate and are withdrawn out of them at birth, the
trophoblastic pseudo-epithelium, and the further derivates it
§
118 A. A. W. HUBRECHT,
has given origin to, remaining in connection with the maternal
tissues.
I hold this to be not a secondary modification which has
arisen among mammals that were already frankly deciduate,
but, on the contrary, a more primitive developmental phase.
In very many cases it may have preceded that more complete
arrangement in which the uterus, after having expelled the
foetus, also rids itself (be it even at the cost of some of its own
elements—rapidly renovated after parturition) of the growths
(afterbirth) by which the embryo has succeeded to obtain so
firm a hold on the maternal sanguiniferous tissues.
If we look at the Carnivora, at the bats, the rodents, the
Primates, and the Insectivora, we find their more complicated
placentary structures to belong to very divergent types. Inthe
latter order there is no common type, but a different one for
nearly every genus. The shrew, the mole, the hedgehog, and
the Tupaja are all most incredibly divergent with respect to
their placentary arrangements. Only when the comparative
investigations shall have covered a more considerable number
of different genera, the time for new theoretical generalisations
will have arrived.
Towards the accumulation of material that would be thus
available I hope the Spolia Nemoris here described may
contribute.
SPOLIA NEMORIS. 119
EXPLANATION OF PLATES 9, 10, 11, & 12,
Illustrating Professor A. A. W. Hubrecht’s paper, “Spolia
Nemoris.”
ov. Ovary. Jig. Uterine ligament. J. Muscularis of the uterine wall.
m. Mucosa of the uterine wall. 2. Hollow recesses clothed with epithelium
in the chorion of Nycticebus. ap. Apertures by which these open to the
exterior. amz. Amnion. w. Umbilical cord. V. Chorionic villi of Nycti-
cebus. cr..Crypts clothed with epithelium, in which these villi fit. /.
Uterine glands.
PLATE 9.
All figures natural size.
Fig. 1.—Tarsius spectrum. A pregnant uterus in the latest stages.
In Fig. 1 the barren horn of the uterus, with coiled oviduct aud ovary, are yet
visible on the top of the swelling that contains the foetus. The other ovary
protrudes in the left lower border of the figure. At the right lower border
the uterine wall shows a rupture ; here the vagina formed its continuation.
Utr. Mus. Cat. n® Tarsius 10.
Fic. 2.—Tarsius spectrum. Harlier phase of pregnancy. One ovary
(ov.) visibly more considerably swollen than the other.
Utr. Mus. Cat. n® Tarsius 11.
Fies. 3—5.—Three uteri of Nycticebus tardigradus. Figs. 3 and 5
front views. Fig. 4 viewed from above to show the peculiar shape of the
uterine horns. In the latter figure the vagina and the two ligamenta ro-
tunda are bent forwards from under the horns. Ovaries partly hidden
from view by fold, including oviduct. Fig. 5 is the stage furthest advanced ;
the fully ripe foetus reaches up to four times this size.
Fig. 3.—Utr. Mus. Cat. n° Nycticebus 6.
Fig. 4.— 55 » be Ue
Meno uss p abet
Fies. 6a and 64.—Galeopithecus variegatus. The double uterus in
a very early stage of pregnancy. 6 seen from behind, 64 seen from above.
The two halves of the uterus open out into the vagina by separate canals and
openings. ‘There is no median portion in common.
Utr. Mus. Cat. n° Galeopithecus 3.
Fies. 7a and 7 4.—The same, in a somewhat later stage of pregnancy.
Utr. Mus, Cat, n° Galeopithecus 13.
120 A. A. W. HUBRECHT.
Fics. 8a and 84.—The same, with one of the uteri already very markedly
swollen.
Utr. Mus. Cat, n° Galeopithecus 27.
Fics. 9a and 94.—The same, in a later stage.
Utr. Mus. Cat. n* Galeopithecus 18.
Fies. 10a and 104.—The same, with the indication of very much widened
blood-vessels in the uterine wall. 10a seen sideways, 104 seen from below.
Utr. Mus. Cat. n® Galeopithecus 16.
Fic. 11.—Side view of a pregnant uterus of Galeopithecus, at nearly full
term; the blood-vessels in the uterine wall yet more prominent.
Utr. Mus. Cat. n® Galeopithecus 14.
Fics. 12—16.—Five uteri in early stages of pregnancy of Tupaja
javanica.
Utr. Mus. Cat. n° Tupaja 251, 62, 254, 17, 39.
Fic. 17.—Pregnant uterus of Tupaja at full term, both halves containing
a foetus, the right placenta of the left foetus and the left placenta of the right
foetus being visible as a reniform thickening in the uterine wall. The other
placentas are situated quite symmetrically on the opposite side, invisible here.
Utr. Mus. Cat, n° Tupaja 170.
PLATE 10.
All the figures (with the exception of Figs. 20, 24, and 25) natural size.
Colour as shown by spirit specimens.
Fic. 18.—Tarsius spectrum. Fully developed fcetus folded together in
foetal membranes with discoid placenta, viewed from above, on the left side of
the drawing. The placenta is actually attached to the maternal tissue only in
the central angular spot.
Utr. Mus. Cat. n° Tarsius 10.
Fie. 19.—The same, seen in profile to show the relative height of the
placenta.
Fic. 20.—Tarsius spectrum. Part of the uterine wall after removal
of the foetus. Umbilical cord and placenta in situ. The latter cut longi-
tudinally. Enlarged 2.
Utr. Mus. Cat. n°: Tarsius 15.
Fic. 21.—The same, as seen from below before the placenta was cut
in two.
Utr. Mus. Cat. n° Tarsius 15,
Fic. 22.—Highly pregnant uterus of Nycticebus, with only the muscularis
peeled off. Cf. Figs. 30-32, 52.
Utr. Mus. Cat. n°* Nycticebus 24,
Fic, 23.—Another pregnant uterus of Nycticebus, with three incisions in
SPOLIA NEMORIS, 121
the uterine wall. Two triangular flaps of muscularis and mucosa are folded
backwards and reveal the foetus enclosed in its villiferous envelope. Cf. Figs,
31, 32, 50, 51.
Utr. Mus. Cat. n° Nycticebus 23.
Fie, 24.—Galeopithecus variegatus. Pregnant uterus, opened opposite
to the placenta. Embryo in amnion. The yolk-sac has been removed, to-
gether with the portion of the uterine wall. Enlarged twice.
Utr. Mus. Cat. n® Galeopithecus 18.
Fie, 25.—Placentary area of the same, enlarged three times, after removal
of the embryo.
Utr. Mus. Cat. n° Galeopithecus 18. .
Fic. 26.—Another uterus of Galeopithecus, in which the wall opposite
the placenta has also been removed, but in which the foetal membranes, &c., are
as yet all of them in situ. The blood-vessels on the yolk-sac are clearly
visible. To the right of the figure the section of the uterine wall has passed
through a portion of the placentary region.
Utr. Mus. Cat. n° Galeopithecus 19.
Fic. 27.—The same stage as that of Fig. 26, after the foetal envelopes have
been opened and turned over (yolk-sac and amnion) to the right. The embryo
is removed ; the placenta is visible.
Utr. Mus. Cat. n* Galeopithecus 19.
Fic, 28.—A similar stage, peeled out of the uterus. The placenta is partly
visible on the left. The yolk-sac has been cut and turned over to the right,
the embryo is yet enclosed by the amnion.
Utr. Mus, Cat. n° Galeopithecus 1.
Fie. 29.—Uterus of Galeopithecus at full term, opened. The ripe foetus
is attached by the umbilical cord to the discoid placenta, which presents a
smooth surface, continuous with that of the uterine wall in which it is im-
planted.
Utr. Mus. Cat, n® Galeopithecus 17.
PLATE. 11.
Figs, 30—33, 35, 36, and 41, natural size. Figs. 34 x 3,37 and 88 x 27,
39 and 40 x 16.
Fie. 30.—Nycticebus tardigradus. The same uterus as that of Fig.
22. The flaps of the muscularis in the same position; mucosa opened;
villiferous chorion inside this opened likewise; amnion partially removed.
Utr. Mus. Cat. n° Nycticebus 24.
Fic. 31.—Uterus of Nycticebus in somewhat earlier stage of pregnancy,
opened by a circular incision. Muscularis and reticular mucosa have here been
left in their natural connection, and the portion of the uterine wall that is
132 A. A. W. HUBRECHT.
here bent to the left has been removed from the subjacent villiferous chorion
without any effort of traction.
Utr. Mus. Cat. n° Nycticebus 84.
Fic. 32.—A similar stage of the same specimen, but in which not only the
uterine wall but also the foetal envelopes have been opened and have also
been folded back. Embryo removed.
Utr. Mus. Cat. n° Nycticebus 45.
Fie. 33.—A ring-shaped section of a Nycticebus uterus, nearly at full term.
The embryo alone has been removed. The umbilical cord is seen to divide
into a number of vasiferous strands, attached to the inner surface of the
chorion. The fcetal envelopes (with-the exception of the amnion, which has
been removed with the fcetus) have been left in their natural position.
Utr. Mus. Cat. n° Nycticebus 41.
Fie, 34.—The inner surface of the chorion enlarged three times, showing
finely ramifying blood-vessels, both afferent and efferent, the two of different
colour in the preserved specimens. Chorionic recesses (cf. Figs. 39 and 40)
form conspicuous round projections inwards. The radiate spots correspond
to the chorionic villi present on the opposite side.
Utr. Mus. Cat. n° Nycticebus 41.
Fic. 35.—Nycticebus foetus, wholly enveloped by the villiferous chorion,
very shortly before birth. Between the villi the apertures (ap.) of the
chorionic recesses (cf. Fig. 39) are visible to the naked eye. ‘To the right the
villi are larger, but also more flattened and wider apart.
Utr. Mus. Cat. n° Nycticebus 34.
Fie. 36.—The same Nycticebus embryo of Fig. 30, to show its attachment
by means of the umbilical cord (w.) to the chorionic envelope, which is
partially turned inside out.
Utr. Mus. Cat. n* Nycticebus 24.
Fies. 37 and 37 a.—Three chorionic villi of Nycticebusas seen from
above, enlarged x 27. They were taken from the specimen of Fig. 30, and are
seen to be multilobulate.
Utr. Mus. Cat. n° Nycticebus 24.
Fic. 38.—The prominent network of the mucosa in which the chorionic
villi fit. Also taken from the same specimen and enlarged x 27,
Utr. Mus. Cat. n® Nycticebus 24.
Fie. 39.—Transverse section of a portion of the chorion of Nycticebus.
Blood-vessels are red. Epithelial covering of chorionic villi here and there
thickened, more especially on the tops of the villi. Chorionic epithelium con-
tinuous in the round and flattened recesses (#.) that open out in the extra-
chorionic space by the apertures (ap.).
Utr. Mus. Cat. n° Nycticebus 24.
Fy. 40.—Another section of the chorion of Nycticebus, but with the
SPOLIA NEMORIS. 123
portion of the uterine wall against which the chorion is applied in situ.
The numerous indentations and reticularly arranged spaces into which the
chorionic villi fit are also covered by an epithelium which is generally some-
what flatter than that of the chorion. The maternal as well as the foetal
blood-vessels are indicated by a red colouring. It can here be seen that the
separation which in Figs. 22 and 30 was brought about between muscularis
and mucosa must have been facilitated by the intervening glandular region
here indicated. The chorionic recess in this figure protrudes further inwards
than those of Fig. 39.
Utr. Mus. Cat. n° Nycticebus 45.
Fie. 41.—One of the two compartments of the pregnant uterus at full
- term of Tupaja javanica (cf. Fig. 17), opened by a longitudinal incision.
The foetus was cut in two by this operation, the one half that is figured in
-outline fitting in the uterine segment to which it remains attached. The
vessels of the umbilical cord (which passes towards the dorsal side of the
fcetus) are there seen to divide into four principal tracts, two for each
placenta. The placenta which was situated to the right of the foetus is
figured in the lower, that which was situated to the left of it in the upper
segment. The latter has thus to be placed in situ by revolving downwards
around its base line by 180°. The cut vessels at the top of the figure will
then be seen to become continuous with those at the bottom of it.
Utr. Mus. Cat. n® Tupaja 258.
PLATE 12.
All the figures natural size with the exception of Figs. 42, 46,55 and 56,
which are enlarged twice.
Fic. 42.—EHarly embryo of Manis javanica prepared out of the fetal
envelopes that are represented in Fig. 44. Enlarged x 2.
Utr. Mus. Cat. n* Manis 29.
Fic. 43.—View of the inner surface of a pregnant uterus of Manis
javanica that contained the foetus and fcetal envelopes of Figs. 42 and 44.
Villosities on the inner uterine surface united into irregular bands.
Utr. Mus, Cat. n° Manis 29.
Fic. 44.—Fetal envelopes of Manis javanica that contained the foetus
which is represented (enlarged twice) in Fig. 42. These foetal envelopes were
obtained intact (after the uterus had been opened) by simply floating them
out. The foetus was contained in the left part. The twisted projection
stretching to the right was devoid of villosities, and measures about twice the
length of the villiferous portion in which the foetus and yolk-sac were found.
It is an example of an asymmetrical arrangement of the fcetal envelopes in
contrast to those of Fig. 45.
Utr. Mus. Cat. n° Manis 29,
124 A. A. W. HUBREOHT.
Fic, 45.—Manis javanica. Embryo of about the same age in its foetal
envelopes, the latter more symmetrically developed than in Fig. 44. The
streaks and bands on the surface corresponding to villous bands on the uterine
wall are clearly visible. The yolk-sac is internally applied against the lower
concave surface.
Utr. Mus. Cat. n°* Manis 71.
Fic. 46.—Tarsius spectrum. Young embryo removed out of its enve-
opes, seen in profile. Enlarged 2.
Utr. Mus. Cat. n® Tarsius 11.
Fic. 47.—Nearly ripe foetus of Tarsius spectrum enclosed in all its
membranes. The discoid placenta is here visible at the top. The only point
of adhesion with the uterine wall is found in the midst of this placentary
disc (cf. Figs. 18 and 19).
Utr. Mus. Cat. n® Tarsius 10.
Fic. 48.—Feetus of Tarsius of about the same age removed out of its feetal
membranes.
Utr. Mus. Cat. n° Tarsius 15.
Fie. 49.—The foetal membranes of a Tarsius at full term after the removal
of the foetus. Discoid placenta and umbilical cord distinct.
Utr. Mus. Cat. n® Tarsius 101.
Fig. 50.—Feetus of Nycticebus tardigradus enclosed in its villiferous
chorion. Obtained by very gently turning upside down the opened uterus
(Fig. 51) in which it was enclosed.
Utr. Mus. Cat. n° Nycticebus 84.
Fie. 51.—One half the uterus of Nycticebus in which the foetus of Fig. 50
has been enclosed. View of the inner surface.
Utr. Mus. Cat. n° Nycticebus 84.
Fie. 52.—The mucosa of Nycticebus of the stage represented in Figs. 22
and 80, peeled off from the muscularis and seen from the inside.
Utr. Mus. Cat. n°: Nycticebus 24.
Fic. 53.—Nycticebus fcetus in all its envelopes, the latter being more folded
than in Fig. 50.
Utr. Mus. Cat. n° Nycticebus 28.
Fic. 54,—A later foetus of Nycticebus prepared out of its envelopes, part
of which are still in connection with the umbilical cord and visible above the
head of the foetus.
Utr. Mus. Cat. n° Nycticebus 54.
Fic. 55.—The villiferous chorion in a very late stage of pregnancy. En-
larged x 2. To the right the villi are more flattened (ef. Fig. 35).
Utr. Mus. Cat, n°* Nycticebus 34.
SPOLIA NEMORIS. 125
Fie. 56.—The reticulated mucosa of a similar late stage of pregnancy.
Enlarged x 2.
Utr. Mus. Cat. n° Nycticebus 34.
Fic. 57.—Embryo of Galeopithecus removed from its envelopes, front
view. The severed umbilical cord is seen protruding between the claws.
Utr. Mus. Cat. n° Galeopithecus 54.
Fic. 58.—Much younger embryo of the same, viewed in profile.
Utr. Mus. Cat. n* Galeopithecus 19.
Fig. 59.—Feetus of Tupaja javanica in its half of the uterus. This latter
was slit open longitudinally, and the left placenta visible. The right placenta
is hidden from view by the embryo.
Utr. Mus. Cat. n* Tupaja 302.
Fig. 60.—The same, the foetus from the other half of the same uterus.
The head of the foetus is seen to be directed distally towards the vagina.
Utr. Mus. Cat. n* Tupaja 302.
VoL. 36, part 1.—NEW SER. !
Ry
AY el cy Ade pile i
Cm tall
‘
7 ing v
At
,
; P lia -
~ 2on
<i wie leeh we ae 1B sg
5 = eo
r ~~»
nN i¢ 91 us 19 LE Seeks
val’ dard) QRS taal
a
eh) daereag tb va wend © ce
7 _
ao a. 4 7
van f "es a. =m i *
:
‘ 7 i Uw gk oe !
’ wv <: t
GUS werh) 2) ae we
oe
‘Se a)
ge
fete re.) tad Ce *
STUDIES ON THE COMPARATIVE ANATOMY OF SPONGES. 127
Studies on the Comparative Anatomy of
Sponges,
VI. On the Anatomy and Relationships of Lelapia australis,
a Living Representative of the Fossil Pharetrones.
By
Arthur Dendy, D.Sc.
With Plate 13.
I. IntRopuctory REMARKs.
In my memoir on the ‘Structure and Classification of the
Calcarea Heteroccela,’ recently published in this Journal
(1), I had occasion to refer to that very remarkable calcareous
sponge Lelapia australis. Unfortunately this species is
extremely rare, only two specimens being as yet known.
Both of these were dredged by Mr. J. Bracebridge Wilson,
M.A., off the Victorian coast, and both were sent to Mr. Carter,
by whom their external characters and spiculation were de-
scribed (without illustration) in the ‘Annals and Magazine of
Natural History’ (2). At the time when I wrote I had never
had the opportunity of personally investigating this sponge,
but since then Mr. Carter, with his usual generosity, has most
kindly sent me a portion of the better of the two specimens,
preserved in spirit.1_ He has also sent me an unpublished sketch
of the entire sponge, and has permitted me to make use of it
in the present memoir (fig. 1). For this and many other kind-
1 The specimen itself is in the British Museum.
VOL. 386, PART 2,—NEW SER. K
128 ARTHUR DENDY.
nesses of a similar nature I desire to express my most sincere
thanks to Mr. Carter. The material was fortunately in suffi-
ciently good condition to enable me to make out the structure
of the canal system, and to establish the correctness of my sup-
position that it belonged to the Leuconoid type. By far the
most interesting feature of the species, however, proved to
be the very remarkable reticulated fibrous character of the
skeleton, which appears to have hitherto escaped observation.
This character is unknown in any other living calcareous
sponge, while it forms the most prominent feature in the great
fossil group “‘ Pharetrones”’ of Zittel (3), hitherto regarded as
entirely extinct. Lelapia australis may therefore be looked
upon as the only known living representative of this important
group,! and a minute study of its anatomy thus acquires an
exceptional interest.
I have much pleasure in again expressing my most sincere
thanks to Professor G. B. Howes for kindly undertaking the
correction of the proof sheets in my absence from England.
II. Anatomy or LELAPIA AUSTRALIS.
A. External Form.
The larger of the two specimens (fig. 1) measured 3} inches
in length by 1 inch in greatest diameter, and is thus described
by Mr. Carter (2, p. 148) :—‘ Cylindrical, clavate, the largest
part upwards, somewhat curved or bent upon itself, rugose
longitudinally. Consistence firm. Colour dark grey. Surface
even, smooth, interrupted by the projection of crooked ridges
extending from the free to the fixed end, subspirally and longi-
tudinally, in broken lengths, sometimes reduced to mere scat-
tered tubercular points, most pronounced on the concave side
towards the mouth, least so on the opposite side: largest and
most continuous ridge + inch long, ;4 inch broad, and 3;
3
inch high. Pores plentifully scattered over the surface, not
1 Whether the group, as it stands, is a natural one, appears to me
doubtful.
STUDIES ON THE COMPARATIVE ANATOMY OF SPONGES. 129
remarkably large. Vent single, terminal, represented by a
narrow elliptical opening about + inch in its longest diameter,
so constricted in the centre as to be closely approximated by
an infolding of the lip on each side; provided with a peri-
stome, whose spicules here are broken off short; leading intoa
cloaca corresponding in shape with the specimen, that is, wide
above, narrowed to a point below (after which the stem becomes
solid).”
The specimen thus described is obviously a single Leuconoid
individual. The smaller specimen, however, as described by
Mr. Carter, showed some indication of a tendency to branch,
and possibly the species may sometimes form branching
colonies.
B. The Skeleton. 1. The Spicules.
The following account of the spicules is taken from prepara-
tions boiled out with caustic potash, as it is extremely difficult
to obtain a satisfactory view of entire spicules in their natural
position. I have not had the opportunity of examining the
spicules of the peristome, but we learn from Mr. Carter’s
writings that there is no important modification amongst
these which is not also met with elsewhere. All three prin-
cipal types of calcareous spicules are met with, but the
quadriradiates are rare, and the apical ray is very feebly
developed.
Oxeote Spicules.—Three varieties may be clearly dis-
tinguished :
(1) Large, stout, fusiform; usually slightly curved and
slightly irregular in diameter; tapering gradually to a sharp
point at each end (fig. 2, a). Size variable, when fully grown
about 1:9 by 0°11 mm.
(2) Long, straight, and very slender, gradually and sharply
pointed at both ends, and sometimes slightly hastate or
bayonet-shaped (fig. 2, 4). Size variable, say about 0:9 by
0:008 mm., but often less.
(3) The so-called “ mortar-spicule.’ Minute; straight or
slightly crooked ; gradually and sharply pointed at both ends,
130 ARTHUR DENDY.
lanciform or hastate at one (fig. 2, c). Size variable, say
about 0:08 by 0:004 mm.
Triradiate Spicules.—Here again three principal varie-
ties may be clearly recognised :
(1) Normal sagittal triradiates ; with wide oral angle; with
long straight shaft (or basal ray) and much shorter lateral (or
oral) rays which may be straight or slightly curved away from
one another (fig. 2, e). All rays rather slender, and gradually
and sharply pointed ; orals measuring about 0°25 by 0°016
mm. ; basal about 0°46 by0°016 mm. (The subgastral sagittal
triradiates usually have somewhat longer and stouter rays.)
(2) Laterally extended sagittal triradiates ; with oral rays so
widely divergent as to be almost in the same straight line, and
basal ray very much shorter, reduced almost to insignificance
(fig. 2, f). Oral rays almost straight or curving slightly away
from one another, gradually and sharply pointed, measuring
about 0°25 by 0°012 mm.; basal ray short, straight, conical,
about 0°05 by 0:0082 mm, but of course variable.
(3) The “ tuning-fork”’ spicules; with all three rays long,
straight, and slender, gradually and sharply pointed; the
basal ray longest and stoutest, and the two oral rays running
straight forwards, parallel to and almost touching one another,
so that the entire spicule is much elongated in the oro-basal
direction (fig. 2, d). The two oral rays are commonly slightly
unequal in length. Total length of an average example, 0°74
mm.; basal ray, 0°42 by 0°01 mm.; longest oral, 0°32 by
0:007 mm.
Quadriradiate Spicules.—These are exactly like the
laterally extended sagittal triradiates, with the addition of a
very short, straight, sharply pointed apical ray (fig. 2, g).
In addition to the principal forms of spicules thus described,
various intermediate as well as more or less abnormal forms
occur, but these are neither numerous nor important.
2. The Arrangement of the Skeleton (figs. 3—5).
As in all the more highly organised Calcarea Heterocela,
we can divide the entire skeleton into three principal parts,
Gy
STUDIES ON THE COMPARATIVE ANATOMY OF SPONGES. 151
viz. that of the dermal cortex, that of the gastral cortex, and
that of the chamber-containing layer of the sponge-wall
between the two. To these may be added the skeleton of the
peristome, but this is hardly of sufficient importance to deserve
special consideration. The principal part of the skeleton is, of
course, that of the chamber layer, which occupies nearly the
entire thickness of the sponge-wall, and it is here that we meet
with the most surprising peculiarities in structure.
Skeleton of the Dermal Cortex.—The dermal cortex
is very thin, and its proper skeleton consists of a thin, confused
layer of rather small, slender-rayed, normal sagittal triradiates,
lying parallel with the dermal surface, together with an im-
mense number of the minute oxeote “ mortar-spicules.” Mr.
Carter notes the occurrence of small quadriradiates on the
dermal surface, but these I have not detected in the small
piece which I have examined. They certainly do not play any
important part in the formation of the skeleton, and, unlike
the large subdermal quadriradiates of the Amphoriscidz, cannot
be regarded as of any systematic importance. The huge
oxeote spicules take no part in the formation of the dermal
cortex proper, although many of them lie just beneathit. Here
and there the dermal cortex is pierced by very stout and
densely packed bundles of the long slender oxeotes (inter-
mingled with “tuning-fork” spicules. These bundles (fig. 3,
d. t.) constitute the expanded (but densely packed) ends of
some of the spicular fibres of the chamber layer, which just
pierce the dermal cortex and give rise to the characteristic
ridges on the outer surface of the sponge. The outermost
ends of the spicules are commonly broken off short.
Skeleton of the Gastral Cortex.—The gastral cortex
is very much more strongly developed than the dermal, having
a thickness of about 0°2 mm. It is composed almost entirely
of very densely packed, laterally extended sagittal triradiates,
with very long oral rays and very short basals. These spicules
are arranged with their longest axes parallel to the gastral
surface, but otherwise in the greatest confusion, with the
short basal rays pointing in all directions instead of constantly
132 ARTHUR DENDY.
towards the base of the sponge. This is the only situation in
which I have detected quadriradiate spicules, and even here
they are few in number, and the apical rays, which project
towards the gastral cavity, are very feebly developed. The
widely extended oral rays of the subgastral sagittal triradiates
may also be regarded as taking part in the formation of the
gastral cortex, but these spicules are best considered in con-
nection with the next portion of the skeleton.
Skeleton of the Chamber-bearing Layer.—The part
of the sponge lying between the gastral aud dermal cortex and
containing the flagellated chambers is, in the piece examined
by me, a little over 4 mm. in thickness, the total thickness of
the sponge-wall being about one sixth of an inch. Its skeleton
is very strongly developed, and (excluding the triradiates which
line the large exhalant canals, and which resemble more or less
those of the gastral cortex) it consists of the following parts:
(a) The subgastral sagittal triradiates. Very well
developed and abundant; occupying the normal position, with
widely extended oral rays lying beneath the gastral cortex, and
long straight basal rays penetrating the chamber layer more
or less vertically or obliquely (fig. 4).
(b) The spicular fibres (figs. 3—5, fi.). These consist
of long bundles of the characteristic slender, elongated,
“tuning-fork” spicules. The component spicules are closely
packed together side by side, parallel with one another (fig. 4).
So closely are they packed together that in a stout fibre it is
very difficult to make out the outlines of the individual
spicules. There does not seem to be any special connecting
substance analogous to the spongin of siliceous sponges, but
the spicules appear to be held together simply by the gela-
tinous ground-substance of the mesoderm. The arrangement
of the fibres is very similar to that of the spicular fibres in
many siliceous sponges. They do not simply run through the
wall of the sponge from gastral to dermal surface, but they run
in every direction, and, by frequently coming in contact
and crossing one another at all sorts of angles, give rise to a
loose, irregular network (fig. 3). The thickness of the fibres
STUDIES ON THE COMPARATIVE ANATOMY OF SPONGES. 1833
s
is very variable, according to the number of spicules entering
into their composition at any given point. They are seldom
more than about 0:07 mm. in diameter, except towards the
dermal surface, where their thickness may be greatly aug-
mented by the addition of numerous slender oxeote spicules as
already mentioned. Towards the gastral surface the fibres
often appear, as it were, to spring from the long basal rays of
the subgastral sagittal triradiates (figs. 4, 5), a relation which
is of considerable importance in considering the derivation of
this peculiar type of skeleton from the primitive articulate
type. Except just close to the two surfaces of the sponge-wall
the fibres appear to consist solely of the remarkable tuning-
fork-shaped triradiates. All the spicules of any one fibre, so
far as I have been able to make out, have their basal rays
pointing in the same direction. Usually the fibres have a
distinct, though more or less oblique trend from gastral to
dermal surface, and it is extremely interesting to note that the
basal rays of the component spicules iu such cases almost
always point towards the outside of the sponge, a fact which
has already been noted by Mr. Carter.
(c) The huge Oxeote Spicules. These occur in im-
mense numbers, disposed in the utmost confusion between the
spicular fibres (fig. 3). The thickness of each one is greater
than that of an average fibre, and, indeed, the fibrous portion
of the skeleton can only play a part of secondary importance
in strengthening the sponge-wall as compared with these giant
spicules.
Skeleton of the Peristome.—This appears, from Mr.
Carter’s description, to present no very characteristic or i1m-
portant features. It consists of long, straight, slender oxeote
spicules arranged perpendicularly (parallel to the long axis of
the sponge). The lower ends of these spicules are crossed at
right angles and supported by the outspread lateral rays of the
gastral triradiates.
134 ARTHUR DENDY.
c. The Canal System.
The canal system of Lelapia australis conforms iu all
respects to the typical Leuconoid arrangement, the entire
sponge being, as already pointed out, a single Leuconoid in-
dividual. The flagellated chambers are spherical or ovoid,
only about 0:06 mm. in diameter and frequently less. They
are thickly scattered in the transparent, gelatinous, meso-
dermal ground-substance which separates the branches of the
inhalant and exhalant canals. Their exhalant openings are, as
usual, circular and well-defined, each with a delicate chamber-
diaphragm. The prosopyles, which are not very easy to make
out in small Leuconoid chambers, I have not succeeded in de-
tecting.
The inhalant canal system is irregular and more or less
lacunar. Owing to the feeble development of the dermal
cortex, there is no separately recognisable cortical canal
system. The inhalant pores are small and scattered over the
dermal surface. They open into short canals which unite to
form larger trunks before penetrating the deeper parts of the
sponge-wall, but there appears to be nothing definite about
the arrangement. The smaller exhalant canals collect into
large trunks, which run to open on the gastral surface, pierc-
ing the gastral cortex more or less at right angles. The wider
parts of these trunks are lined by a layer of laterally extended
sagittal triradiates, similar to those of the gastral cortex, and
amongst them may be seen sagittal triradiates like the sub-
gastrals, with long basal rays projecting into the surrounding
tissue at right angles to the course of the canal. These facts
argue in favour of the supposition that the larger exhalant
canals in Lelapia may be formed by pitting in or folding of
the gastral surface. The openings of the exhalant canals (fig.
6, ex. ap.) into the wide gastral cavity are abundantly scattered
over the inner surface of the gastral cortex, and are provided
with membranous diaphragms, as Mr. Carter has already
pointed out. From .the gastral cavity, of course, the water
STUDIES ON THE COMPARATIVE ANATOMY OF SPONGES. 135
finds its way out of the sponge through the wide, terminal os-
culum (fig. 1, osc.).
The histology of the sponge, so far as the condition of the
specimen will permit of investigation, offers no features of
special interest, and appears to agree with that of other He-
terocela. Beyond the transparent gelatinous ground-substance
of the mesoderm, the contracted collared cells, and the nuclei
of the pavement epithelium lining the canals, I have not been
able to make out any details.
III. Revationsuirs oF LELAPIA.
A. Relationships to other recent Heterocela.
The canal system of Lelapia australis, as already pointed
out, offers no features of peculiar interest and, as regards its
probable derivation from the more primitive Syconoid type,
stands on exactly the same footing as the canal system of any
other Leuconoid Heterocele. As the probable mode of deriva-
tion of the Leuconoid from the Syconoid type has already been
discussed in my memoir on the structure and classification of
the Heteroccela (1), it is unnecessary to enter further into the
question in this place.
The skeleton, however, is very peculiar, and, at first sight,
may seem to place great difficulties in the way of believing
in the Syconoid ancestry of Lelapia. These difficulties, how-
ever, disappear upon closer examination.
The peculiar form of the “tuning-fork” spicule is not, in
itself, of much significance, and, as already pointed out by
other writers, it is paralleled more or less closely in Haeckel’s
Leucandra (Leucortis) pulvinar and L. (Leucetta)
pandora (4), both of which are recent species ; while it is also
met with in the fossil Sestrostomella rugosa and S.
clavata described by Dr. Hinde (5). In none of these, how-
ever, does it appear to attain to anything like the degree of
development met with in Lelapia australis. The tendency
of the triradiate to vary is well known, and we meet with a
136 ARTHUR DENDY.
modification perhaps even more remarkable in my Gran-
tiopsis cylindrica (1).
As I have previously pointed out, the arrangement of the
skeleton in the Calcarea Heterocela is of more importance
for purposes of classification than the mere form of the com-
ponent spicules, and the difficulty lies in explaining how the
very peculiarly arranged skeleton of Lelapia can have
been derived from a Syconoid ancestor with its characteristic
articulate tubar skeleton.
We may, however, at once confine our attention to the
skeleton of the chamber-bearing layer, for that of the dermal
and gastral cortex differs in no essential points from the corre-
sponding parts in other corticate Heterocela. Taking first the
subgastral sagittal triradiates, we find in these a strong argu-
ment in favour of our view, for they exactly correspond to
the similar spicules of the typical articulate skeleton. These
spicules, indeed, seem to be wonderfully persistent, being also
met with, as I have already pointed out, in the genus Leu-
candra, after all other traces of the articulate tubar skeleton
have disappeared.
The huge oxeote spicules are probably, like the very similar
spicules of some species of Leucandra (e. g. L. cataphracta,
Haeckel), to be regarded as incursions from the dermal cortex,
the dermal surface being the characteristic position for oxeote
spicules.
We have left the spicular fibres, whose presence distinguishes
Lelapia australis from all other known recent sponges.
These I believe to be derived from the articulate tubar skeleton
of a Syconoid ancestor. The arguments in favour of this view
are as follows:—(1) The position of these spicules with regard
to one another is the same as in the typical articulate tubar
skeleton, i. e. with their basal rays parallel, overlapping, and
all pointing in the same direction. (2) The position of the
spicules with regard to the gastral and dermal surfaces is,
when the position of the fibre as a whole allows of it, almost
invariably the same as in the articulate tubar skeleton, i.e.
with the basal rays pointing to the dermal surface. (3) The
STUDIES ON THE COMPARATIVE ANATOMY OF SPONGES. 1387
relations of the subgastral sagittal triradiates to the fibres is,
in many cases at any rate, identical with that of the corre-
sponding spicules to the articulate tubar skeleton of a
Syconoid.
Thus, while in Leucandra the spicules of the primitive
articulate skeleton become scattered and disjointed, in Lelapia
they retain their mutual relationships, and indeed become
much more intimately associated with one another to form
spicular fibres,—this formation of fibres being greatly facilitated
by their very peculiar shape. The fibres as a whole, however,
become irregularly arranged, as do the individual spicules of
Leucandra.
Thus, then, I see no reason for altering the systematic posi-
tion of the genus Lelapia as given in the genealogical tree at
the end of my previous memoir (1). In other words, I regard
Lelapia as an offshoot from the great family Grantide,
coming off from the same branch which gave rise to the genus
Leucandra. At the same time this way of thinking would
not prevent us, if necessary, from accepting the Pharetrones
as a distinct family and including Lelapia therein.
B. Relationships to the Fossil Pharetrones.
Professor Zittel, in his classical “ Studies on Fossil Sponges”
(3), accepted Haeckel’s division of the Calcarea into Ascones,
Leucones, and Sycones, but added thereto a new family,
Pharetrones, which he regarded as of co-ordinate systematic
value with Haeckel’s three groups. The following diagnosis
was given of the new family :—“ Wand dick, mit ungeraden
Astcanalen oder ohne alle Canale. Skeletelemente zu anasto-
mosirenden Fasern angeordnet. Dermalschicht haufig yor-
handen.”
In this family Zittel placed, with one exception (Proto-
sycon), all the then known fossil Calcarea, amounting to no
less than fifty genera, and ranging from the Devonian to the
Upper Chalk.
These numerous fossil genera are naturally very imper-
138 ARTHUR DENDY.
fectly known, and even their calcareous nature has been dis-
puted. This was largely owing to the fibrous character of the
skeleton, which was not then known to occur in any living
calcisponge, and certainly appeared to indicate a siliceous
nature for the Pharetrones.
In 1882, however, Dr. Hinde published a valuable paper (5)
in which this question was fully discussed, and described more
or less in detail the spiculation of five species of so-called
Pharetrones from the Cretaceous and Upper Greensand forma-
tions. These observations sufficiently proved the calcareous
nature of at any rate these five species, for characteristic tri-
radiate and quadriradiate spicules were detected.
The first species described by Dr. Hinde is Verticillites
d’Orbignyi. The description and figures clearly prove this
to be a calcareous sponge, but I do not think that there is the
slightest evidence in favour of regarding it as a Pharetronid.
It is a very thin-walled sponge, and the portions of the skeleton
described are evidently simply the dermal and gastral cortex
respectively. The latter is very strongly developed and differs
in no essential respect from that of many living Heteroccela,
being composed of a dense feltwork of triradiate spicules with
aborted basal rays. This layer is pierced by the circular ex-
halant apertures leading into the gastral cavity, which, being
placed near together, give the gastral cortex a reticulate cha-
racter. To apply the term fibrous to this skeleton appears to
me to be a mistake; it is no more fibrous than that of many
living Heteroceela, and the so-called fibres are certainly not
homologous with the spicular fibres of Lelapia. Verti-
cillites d’Orbignyi ought, then, to be removed from the
Pharetrones altogether. The thinness of the sponge-wall
and the regular disposition of the exhalant openings on the
gastral surface suggest that it may have possessed a Syconoid
canal system, while the presence of quadriradiates in the
dermal cortex, with inwardly projecting apical rays, suggests
a position amongst the Amphoriscide of my classifica-
tion.
In the next two species, again, Corynella rugosa and C,
STUDIES ON THE COMPARATIVE ANATOMY OF SPONGES. 139
socialis, the evidence placed before us does not seem to be
conclusive as to the truly fibrous character of the skeleton,
apart from the so-called fibres of the cortical one.
In the two species of Sestrostomella, viz. 8S. rugosa and
S. clavata, described by the same author, there appears, on
the other hand, to be little doubt as to the existence of a
truly fibrous skeleton distinct from the cortical one. The fibre
itself, to quote the words of Dr. Hinde, “ exhibits an altogether
different character in the form and arrangement of the com-
ponent spicules from that which prevails in the examples of
Verticillites and Corynella, already described.’ The
ensuing description, however, also shows that the fibre is very
different from that of Lelapia. ‘The central portion of the
fibre generally appears to be occupied by a large tri- or quadri-
radiate spicule, one ray of which extends along the central axis
of the fibre. . . . Beyond this centrally-placed large spicule
the remaining portion of the fibre appears to be composed of
several different forms of triradiate spicules, whose rays are so
compactly and intricately interlaced together that it is ex-
tremely difficult to ascertain their complete forms with any
degree of precision.” Amongst these smaller spicules occur
fork-shaped triradiates somewhat similar to those of Lelapia,
but they appear to be irregularly arranged. In view of the
occurrence of these spicules in other recent Heteroccela, as
already mentioned, their systematic value must be considered
as very questionable.
Lelapia, then, does not appear to be very closely related
to any of the fossil Pharetrones described by Dr. Hinde.
Whether it is more closely related to any of the other nume-
rous fossil sponges which have been included in that group, it
is impossible, in the absence of detailed information as to the
form and arrangement of the spicules in the latter, to decide.
It has, however, a truly fibrous skelelon ; and in this respect it
agrees with the main character of the family Pharetrones as
laid down by Zittel. We may therefore regard it as a living
representative of the group, but whether the group itself, as it
stands, is a natural one is another question altogether, and one
140 ARTHUR DENDY.
which, owing to the state of preservation of the fossils, will
perhaps never be decided.
Zittel (3) regarded his Pharetrones as the ancestral forms from
which the living calcareous sponges (Ascones, Leucones, and
Sycones of Haeckel) originated. I need hardly say that I do not
agree with this view, and I have endeavoured to show in this
paper how a “spiculo-fibrous” skeleton may have been derived
from the more primitive Syconoid type. My arguments, how-
ever, obviously apply only to the case of Lelapia, for the
minute structure of other calcareous sponges with a fibrous
skeleton is not sufficiently well known to justify speculation as
to their origin. Thus it is possible that the fibres of those
Pharetrones which are described by Zittel as consisting of
bundles of parallel oxeote spicules are really composed of
elongated, fork-shaped triradiates. Even in Lelapia it is
hard to distinguish these, in situ, from oxeote spicules, owing
to the closeness with which they are packed together; and in a
section of a fossil sponge it would be impossible to do so unless
one happened to get a section passing just through the fork of
the spicule and in the plane of all three rays.
All things considered, then, it seems unadvisable, in the
present state of our knowledge, to introduce the family Phare-
trones into our system of recent Calcarea, and Lelapia may
be regarded simply asa very specialised type of the Grantidz.
MeExLBourNE, November, 1893.
REFERENCE List or LITERATURE.
1. Denpy.—“ Studies on the Comparative Anatomy of Sponges: V. Observa-
tions on the Structure and Classification of the Calearea Hetero-
cela,” ‘Quart. Journ. Micr. Sci.,’ vol. xxxv, N.8., p. 159.
2. Carter.—“ Descriptions of Sponges from the Neighbourhood of Port
Phillip Heads,” ‘ Annals and Magazine of Natural History,’ vol. xviii,
ser. 5, p. 126.
3. ZitTEL,— Studien iiber fossile Spongien, Dritte Abtheilung,” ‘ Abhand-
lungen der k. bayer. Akademie der W.,’ II Cl., Bd. xiii, Abth. ii, 1878.
STUDIES ON THE COMPARATIVE ANATOMY OF SPONGES. 141]
4. Harcxet.—‘ Die Kalkschwamme.’
5. Hinpe.—‘ Notes on Fossil Calcispongiw, with Description of New
Species,” ‘Annals and Magazine of Natural History,’ vol. x, ser. 5,
p. 185.
DESCRIPTION OF PLATE 13,
Illustrating Dr. Dendy’s paper ‘ On the Anatomy of
Lelapia australis.”
Reference Letters.
a. Large oxeote spicules. &. Long slender oxeote spicule. c. Small oxeote
spicules (‘‘ Mortar-spicules”). d. Tuning-fork-shaped triradiates. e. Normal
sagittal triradiates. f Laterally extended sagittal triradiates. g. Quadri-
radiate. d.c. Dermal cortex. d. ¢. Dermal tuft of triradiates and slender
oxea. ex.ap. Openings of exhalant canals into central gastral cavity.
ex. c. Exhalant canal. i. Spicular fibres formed of tuning-fork-shaped
triradiates. g.c. Gastral cortex. osc. Osculum. s. g. s. Subgastral sagittal
triradiates.
(The spicules are delineated in blue throughout.)
Fie. 1.—Lelapia australis. (From a sketch by Mr. H. J. Carter,
F.R.S.)
Fic. 2.—lIsolated spicules, from a preparation boiled out with caustic
potash. Only the two ends of the large oxeote (a) are drawn. All are drawn
to the same scale, under Zeiss C, ocular 2, camera.
Fic. 3.—Skeleton arrangement, as seen in a thick, unstained, longitudinal
section. x 25.
Fig. 4.—Portion of the gastral cortex and adjacent skeleton of the
chamber-bearing layer, as seen in a thick, unstained, longitudinal section.
Drawn under Zeiss C, ocular 2.
Fic. 5.—Portion of a subgastral sagittal triradiate, showing its relation to
a very slender fibre formed of tuning-fork-shaped triradiates. Drawn under
Zeiss C, ocular 2.
Fic. 6.—Portion of a transverse section, stained with borax carmine and
cut by the paraffin method. x 25. Each flagellated chamber is represented
by a red spot.
142
ARTHUR DENDY.
APPENDIX.
I take the present opportunity of correcting a few slight
misprints which occur in the last paper of this series, “ On
the Structure and Classification of the Calearea Hetero-
Page 163, line 2, for “caused” read “lined.”
cola.”
pee UGG;
GOs
a LOd,
3 168,
op lets)
cy he
Agaiel b/ (i F-
Sie Los
> #9,
ye ds
29 206,
» 208,
» 216,
gx! LY,
» 229,
», 232,
33
16, erase the comma after the word “ to.”
34, for “strobulus” read “ strobilus.”
17, for “ Bauerbank ” read “ Bowerbank.”
15, for “tuber” read “ tubar.”
24, for “fig. 8” read “fig. 5.”
34, for “cram” read “ crown.”
11, for “non-” read “ inter-.”
2, for “situated” read ‘‘ inflated.”
3, for “ Syllectoid ” read << Sylleibid.”
11, for “strangely” read “strongly.”
28, for “ liycetta” read ‘ Sycetta.”
4, 5, 6, for ““ Heterocela Calcarea” read “Cal-
carea Heterocela.”
3, for “into” read “on to.”
12, for “permanent ” read “ pavement.”
9, for “van” read “von.”
16, for *Thabden” read “ Rhabden,”
BILL AND HAIRS OF ORNITHORHYNCHUS PARADOXUS.
143
The Structure of the Bill and Hairs of Ornitho-
rhynchus paradoxus; with a Discussion
of the Homologies and Origin of Mammalian
Hair.
By
Edward B. Poulton, M.A., F.R.S., «&c.,
Hope Professor of Zoology in the University of Oxford.
(With Plates 14, 15, and 15a.)
ContTENTS.
I. IntrRopuctTion
II. Tur PusH-Rops. :
III. Tue GLanpD-DUCTS OF THE nek AND THE ren eennne ASSOCIATED
WITH THEM . 3
IV. Toe Harrs oF OLD aNnD Yours open ee
1. Historical, by W. Blaxland Benham, D.Sc. (Lond.), Hon.
M.A. (Oxon.), Aldrichian Demonstrator in ane
Anatomy in the University of Oxford
. General Structure and Arrangement of ine
. Comparison between the Hairs of Old and Young
Animal .
4. Minute Structure Al meamation Gh air baa
its Sheaths
5. Mode of Succession ae sie, ais
. Recapitulation of Essential Peculiarities of ‘Hair
of Ornithorhynchus as compared with that of
Higher Mammalia
V. THe Homotocizs an» ORIGIN OF Teena) Hen
1. Historical, by W. Blaxland Benham
2. Conclusions derived from the Study of Onno:
rhynchus
vou. 36, PART 2.—NEW SER, L
oo %
7)
144 EDWARD B. POULTON.
I. IntTRoDUCTION.
In the summer of 1884 I made a communication “On the
Tactile Terminal Organs and other Structures in the Bill of
Ornithorhynchus” to the meeting of the Physiological Society
at Oxford. A short account, containing however the chief
results, was published in the Proceedings of the Society (‘ The
Journal of Physiology,’ 1884, pp. 15, 16). In searching for
the terminations of the abundant medullated nerves I found
two remarkable structures in the epidermis of the bill :—
(1) Epithelial rods which appear to convey the effect of surface
pressure to a group of nerve end organs resembling Pacinian
bodies. These rods, from their obvious analogy to the push
of an electric or, better still, pneumatic bell, may be con-
veniently called “ push-rods.”’ (2) Other epithelial rods
along the axis of which passes the duct of a gland. These
rods resemble in certain respects a shortened and truncated
hair.
The interpretation of these latter rods as modified hairs was
disputed at the meeting by Dr. Klein, but supported by Pro-
fessor Schafer. This controversy and the obvious interest of
the inquiry led me to undertake further work, which included
an investigation of the hairs covering the body of this animal.
From time to time this work has been continued from 1884 up
to the present date (December, 1893), and many drawings
have been made and discussed with friends, although until now
T have published nothing more upon the subject. During the
summer of last year (1892) I worked for some weeks in the
laboratory of my friend Professor Lankester, at which time the
drawings were arranged and the description of Pl. 14 written.
Within the last few days the recent description of the push-
rods by Professor J.T. Wilson and Mr. C. J. Martin (‘ Macleay
Memorial Volume,’ pp. 190—200) has reached me. I was
prepared to find that their investigation would have rendered
any further publication on this point unnecessary, inasmuch as
fresh material had been available, the want of which had
confronted me throughout, and especially in the attempt to
BILL AND HAIRS OF ORNITHORHYNCHUS PARADOXUS. 145
trace the endings of the nerves. The authors have, however,
relied almost entirely upon photomicrography for their illus-
trations, with the result that the figures upon two (xxiv and
xxv) out of the three plates are utterly worthless, and convey
a most inadequate conception of the appearances presented by
the sections. It is evident from the descriptions that the
sections relied upon were adequate, and perhaps the negatives
and the first impressions were also satisfactory: but in the
form in which the paper reaches the public there is something
inconsistent in the confident references made in the text to the
representation of detail in a figure which turns out on inspection
to be the merest smudge (see especially fig. 5, pl. xxiv, which,
among all figures ever published in scientific papers, must surely
take the palm for obscurity). My sympathies are entirely with
those who have been misled into trusting a process which for
this purpose appears to be entirely useless. The authors have,
however, made some drawings, the photographic representa-
tions of which are infinitely better than those of the sections.
Fortunately, too, these figures, which are all to be found on
the last plate (xxvi), deal with points of structure which needed
the fresh tissues for their adequate treatment, and are there-
fore unrepresented among my illustrations.
I therefore propose to publish my figures together with their
explanation (in the “Description of Plates”) exactly as I wrote
them in 1892, except for mere verbal corrections, &c. Thus
my description of the push-rods is entirely independent of that
recently published in Australia, although it will be found that
the two accounts are in substantial agreement.
The gland-ducts and their associated hair-like and nervous
structures are not alluded to by the above-mentioned authors.
This is to be regretted, inasmuch as many of the most important
points can only be decided by the use of the fresh tissues. A
detailed description of these, so far as it is possible with the
material at my disposal, is also given below, and the same is
true of the hairs covering the body of the young and mature
animal,
The material upon which all the work described below has
146 EDWARD B,. POULTON.
been done consisted of (1) parts of the bill hardened in chromic
acid by the late Professor Moseley in 1874, at the same time
that he prepared the tongue which I have described in this
journal (see ‘ Quart. Journ. Micr. Sci.,’ July, 1883, p. 453,
for further details concerning this material), and the horny
teeth (see ‘ Quart. Journ. Micr. Sci.,’ July, 1888, pp. 31, 32) ;
(2) a specimen kept many years in spirit in the Oxford
University Museum, the ovary of which was described in this
journal (January, 1884, p. 118); other similarly preserved
specimens in the Museum were also employed; (3) the young
Ornithorhynchus of which I was enabled to describe the true
teeth through the kindness of the late Professor W. K. Parker
(‘ Quart. Journ. Micr. Sci.,’ July, 1888, pp. 10, 11). This
latter specimen was 8°3 cm. long in the curled-up attitude in
which I received it, and the larger hairs had alone appeared
above the skin.
Since the above was written I am enabled, through the
kindness of Professor W.N. Parker, to supply figures of the
natural size of the young specimen of Ornithorhynchus referred
to above. The appearance from the left side is seen in Plate
15a, fig. 1, from the ventral aspect in fig. 2. ‘The latter
has been already published, reduced one half, by the late
Professor W. K. Parker. The drawings from which Pl. 15a
was prepared were made by Mr. M. P. Parker. The considerable
amount of work which has been already published upon material
from the single specimen here figured is indicated in the
following list of memoirs :
G. B. Howrs.—‘ Journ. Anat. and Physiol.,’ 1886, xxi, p. 190. Describes and
figures the shoulder girdle.
G. B. Howes.—‘ Journ. Anat. and Physiol.,’ 1892, xxvii, p. 543. Describes
_the pelvis, and figures a section of it.
The late W. K. Panker.—‘‘ Mammalian Descent,’ London, 1885. Describes
some of the external characters (p. 48, et seq.), and figures the ventral
view of the specimen one half the natural size (p. 25). This same
view is represented the full size in Pl, 15a, fig. 2, accompanying this
paper,
BILL AND HAIRS OF ORNITHORHYNCHUS PARADOXUS. 147
W. ON. Parker.— Brit. Association Reports,’ 1891, p. 693. Compares the
snout with that of the young Echidna.
W. N. Parxer.—‘On some Points in the Structure of the Young of
Echidna aculeata,” ‘Proc. Zool. Soc. Lond.,’ read January 16th,
1894, and to be published shortly. Compares with young Nchidna,
especially as regards ‘“ Jacobson’s organ,” a section of which is
figured.
BE. B. Poutron.—‘ Proc. Roy. Soc.,’ 1888, xliii, p. 353. Describes the true
teeth, and figures a section of one.
E. B. Poutron.—‘ Quart. Journ. Mier. Sci.,’ 1888, vol. xxix, p.9. Describes
the true teeth, and figures their appearance, in sections, and in
minute structure.
R. WirepersHEm.—‘ Zeit. f. wiss. Zool.,’ liii, 1892, suppl., p. 43. Describes
and figures the pelvis.
This paper having been prepared for the press at a time of
great pressure, my friend Dr. Benham has very kindly relieved
me of the labour of writing the historical part of the two latter
sections. I also wish to thank my friend Professor Howes for
kindly suggesting certain lines of recent research which bore
upon the subject. So far as the earlier sections are concerned,
with the exception of Wilson’s and Martin’s recent paper there
is no other record worthy of mention.
II. Tue PusH-rops.
It will not be necessary to describe the histological details
of these structures at any great length. Reference to Pl. 14
and to the full descriptions will be sufficient. I shall therefore
chiefly dwell upon the points which are not touched by Wilson
and Martin, or in which my interpretation differs from theirs.
The proportion of these structures to the gland-ducts varies
in different parts. Thus on the presumably highly sensitive
ridges within the lower jaw the former are the more numer-
ous, as may be seen by glancing at fig. 5, where the push-
rods are seen in section as circles. On the external surface
of the bill, however, these proportions are usually reversed.
In relation to the sensitive condition of the bill we learn
with great interest that it is, in the living state, covered
with “smooth, soft, and humid skin” (“Anatomy of the
148 EDWARD B. POULTON.
Muzzle of the Ornithorhynchus,” by Professors J. T.
Wilson and C.J. Martin; ‘Macleay Memorial Volume,’ p. 180).
In this country, where we see only dry or spirit-preserved
specimens, it has generally been looked upon as a tough, horny
structure. The glands to be described below are doubless
responsible for the humidity, although they may also, as I
have suggested (‘The Journal of Physiology,’ 1884, p. 16),
secrete a substance which protects the skin from the prolonged
action of water.
Wilson and Martin describe structures in the horny teeth
which they believe to represent push-rods which have under-
gone cornification (Il. c., p. 192).
In offering this interpretation I believe that they have mis-
taken for push-rods the columns of cells which extend from the
summits of the numerous papillz towards, and often as far as,
the surface. These cells are much less cornified than those
around, and stain readily; furthermore it may be said that
the columns “ show a series of imbricated superposed cells,”
although the arrangement differs in various details from that
of the push-rods. The fact that a column invariably extends
from the summit of a papilla, while a push-rod is invariably
surrounded by two or more papillz, is sufficient proof that
the two structures are not homologous. The structure of the
cell-columns and their relation to the papille are clearly shown
in my paper already quoted (‘ Quart. Journ. Mie. Sci.,’ July,
1388, PL-TYV, figs. 4, 5,.8, 10,11);
Speaking of the dermal sheath which surrounds the lower
part of each push-rod, the above-quoted authors regard the
rod as an epithelial downgrowth which causes “a depression
in the summit of a dermal papilla” (1. ¢., p. 193). Inasmuch
as two or more distinct papille arise from the upper edge of
the dermal sheath (to be seen in section in PI. 14, fig. 3, accom-
panying this paper ; and, by the eye of faith, in pl. xxiv, fig. 6,
of the ‘ Macleay Memorial Volume ’), I should prefer to regard
the rod as a modified interpapillary process, with the surround-
ing papillz united into a continuous sheath below, while they
remain free above.
BILL AND HAIRS OF ORNITHORHYNCHUS PARADOXUS. 149
The shaft of a push-rod consists of four concentric layers
of cells, which are rendered quite distinct by their unequal
staining (figs. 3 and 4). Wilson and Martin regard the
outermost of these layers as belonging to the general epidermis,
and forming “a kind of follicle’ round the rod proper. I
took the same view in 1884, speaking of the third layer as
“followed by concentrically arranged fusiform cells belonging
to the general epidermis” (lI. c., p. 16). The innermost im-
bricated cells, constituting the first layer, are looked upon by
Wilson and Martin as hollow truncated cones made up of three
or more cells. This interpretation is certainly supported by
the teased preparation represented in their pl. xxvi, fig. 17.
I could never make out any cell outlines in this layer when
studied in transverse sections of the rods, although three or
more nuclei were often seen (fig. 3). Longitudinal sections,
on the other hand, show the outlines of successive cells very
distinctly (fig. 4). My observations to this extent support
Wilson’s and Martin’s conclusions as to the nature of the
innermost cell-structures. As to the second layer of imbri-
cated cells, and the larger far less sloping cells constituting
the third layer, we are in entire agreement, and my fig. 4
would, in this respect, serve for the illustration of Wilson’s
and Martin’s description. It should be noted, however, that
the nuclei of the cells of all the layers are surrounded by
pigment masses and granules.
The central group of fine filaments occupying the axis of the
rod, together with the ring of similar structures placed between
the first and second layers of imbricated cells, are very con-
spicuous, and early attracted my attention (1884, 1. c., p. 16).
They are clearly seen in figs. 3 and 4, Longitudinal sections
show that they pursue a parallel course along the rod except
at the convex lower extremity, where I could just make out,
in the most favorable preparations, that the filaments of the
central group diverge in a brush-like manner (fig. 1). The
most careful examination, under very high powers, of the
best material I could command, displayed a structure which
is represented in fig. 4. Each filament appeared to be
150 EDWARD B. POULTON.
made up of short sections corresponding in length to the thick-
ness of the adjacent imbricated cells. From this structure I
formed the conclusion that the filaments are probably built up
of units contributed by the cells which lie outside them. Wilson
and Martin describe them as made up of bead-like varicosities,
each of which is placed between a pair of cells (simple or com-
pound) in vertical succession (I. c., pl. xxvi, fig. 15, for a diagram-
matic representation). At the meeting of the Physiological
Society (1884) I suggested that the filaments might be nerve
terminal organs, and that the imbricated arrangement of cells
around them might result in pressure whenever the free sur-
face of the rod was in contact with external objects, thus
effecting the stimulation of the nerve. Professor Schafer pointed
out, in the discussion, that the highly refringent character of
the filaments is opposed to the view that they are terminal
' nerve-fibrils. I have often tried to find a connection with the
abundant nerves below the base of the push-rod, but such an
investigation needs the fresh tissues. Wilson and Martin now
describe the filaments as naked axis-cylinders prolonged from
the nerve-fibres below the rod, which suddenly lose their me-
dullary sheaths soon after entering the epithelium. Although
their drawings from gold-stained preparations (pl. xxvi, fig. 21)
appear to leave no doubt that the filaments are connected with
nerves and form some kind of terminal organ, it is obviously
erroneous to speak of them as “naked axis-cylinders” (p. 196) or
“ nerve-fibrils” (p. 197). They are highly refringent and have
none of the characters of these nerve structures. It is note-
worthy, too, that although the filaments are represented in
pl. xxvi, fig. 22, as black varicose threads, the authors accurately
state in the description (p. 200) that the fibrils “ are uot black,
but only highly refracting.” Furthermore, if the photomicro-
graphs establish nothing else, they certainly prove that this
conclusion as to the nature of the filaments is erroneous ;
neither naked axis-cylinders nor nerve-fibrils could have caused
the appearances seen in transverse section in pl. xxv, fig. 13,
and in longitudinal section in fig. 8.
These authors state (p. 196) that I failed to recognise the
BILL AND HAIRS OF ORNITHORHYNCHUS PARADOXUS. 151
nervous character and connections of the filaments. I am in
entire agreement with this statement: ‘‘ the nervous charac-
ter” does not exist, and the nervous connections could not by
any possibility have been seen in the tissues with which I was
supplied. Although I did not recognise connections which
could not have been seen, I certainly inferred them, as the
title of my communication indicates (‘‘ On the Tactile Terminal
Organs,” &c., ‘Journ. of Physiology,’ 1884).
The filaments must be regarded as remarkable terminal
organs, entirely distinct in histological nature from the axis-
cylinders which terminate in them; and Wilson and Martin’s
fig. 25 seems to suggest that the change in nature takes place
at or close to the point at which the medullary sheath dis-
appears. The appearance presented by the filaments is very
clearly shown in figs. 3 and 4 accompanying this paper. They
must be looked upon as a new and interesting form of nerve
terminal organ, probably epithelial in origin.
As regards the Pacinian-like bodies, Wilson and Martin
support my previous account, but they also describe certain
larger forms of these structures, rather more deeply placed
than those immediately below the push-rods. Bodies similar
to those beneath the rods occur in the mouth, and were
described and figured in some detail in my paper on ‘“ The
Tongue of Ornithorhynchus” in this Journal (‘ Quart. Journ.
Mier. Sci.,’ July, 1883, Pl. XXXII, fig. 5).
The constant occurrence of a group of Pacinian-like bodies
at the base of each push-rod (figs. 5 and 6) is of great phy-
siological interest, as it strongly supports the view—widely
held but as yet unproved—that the function of this form of
nerve end-organ is to aid the nervous system in the apprecia-
tion of pressure. The obvious use of the push-rods is, as I
stated in 1884, “to supply specially moveable areas yielding
to surface pressure, which is thus communicated to the terminal
organs below.”
Another interesting end-organ described by these authors is
placed among the epidermic cells of the base of the rods. In
these “lenticular bodies” the axis-cylinder is described as
152 EDWARD B. POULTON.
ending in a disc which separates “two clear vesicular cell-like
structures.” This form of ending evidently needs the fresh
tissues for its demonstration.
At first sight the push-rods appear to be modified hairs, but
the examination of their minute histological details does not
support this comparison. It is possible, however, that they
may be found to have some bearing upon the recent suggestion
that the Mammalian hair corresponds to an epidermic tactile
organ of the lower Vertebrata. However this may be, the re-
semblance to the Mammalian hair as it now is, in my opinion, is
far less close than that of the epidermic structures associated
with the ducts of glands which open on the surface of the bill.
Souza Fontes, in 1879 (‘ Beit. z. Anat. Kenntniss der Haut-
decke des Ornithorhynchus,’ Inaug. Dissert., Bonn), men-
tioned and figured these structures and the gland-ducts
described below, but the paper is quite unworthy of mention.
Indeed, the principal feeling evoked by a glance at the Plate is
one of surprise at the system which can confer a University
Degree for such a production.
III. Tue Guanp-pucts or THE BILL AND STRUCTURES
ASSOCIATED WITH THEM.
The gland-tubes of the bill, and, indeed, of the general body
surface, closely resemble the Mammalian sweat-glands, the
secretory part of the tubule being wider than the duct, and
lined with short columnar cells surrounded by a longitudinal
layer of smooth muscle-cells (fig. 8, g/d.). The wall of the
duct is composed of nucleated, probably polyhedral, cells,
indistinctly marked off from one another in my sections.
These cells are separated from the lumen by a cuticle, repre-
sented as a row of thin, deeply staining, plate-like structures
resembling nuclei (fig. 8, d’, transverse section; below d longi-
tudinal section). Externally the tubes are surrounded by a
membrana propria, in which nuclei are especially distinct in
the transverse section of the duct (d’).
The existence of such typical structures in the most primi-
tive mammal indicates that sweat-glands are among the most
BILL AND HAIRS OF ORNITHORHYNCHUS PARADOXUS. 153
ancestral features of the Class. Although the hairs of Ornitho-
rhynchus will be shown below to present many peculiarities
which are, as I believe, ancestral, the sweat-glands are essen-
tially similar to those of Mammalia generally.
In the bill, the deeply placed coiled gland-tube is succeeded
by a coiled duct which, as in many other mammals, enters the
base of an epidermic downgrowth—the interpapillary process
(fig. 8). The epidermic process itself, however, is by no means
typical, but presents many special peculiarities, some of which
support the conclusion that it is a modified hair,—sharply cut
off above at the level of the uppermost epidermic layer,
shortened below by retraction of the hair-bulb, so that the
latter descends but a short distance beneath the lowest layer of
the epidermis (figs. 8, 13—16). Nevertheless, in the young
Ornithorhynchus the bulb-like part of the structure extends
to a somewhat deeper level (compare fig. 16). These hair-like
structures were briefly described in 1884 (1. c., p. 16, where,
however, in line 11 from bottom, the words “hair papilla”
are obviously intended for “ hair-bulb’’).
The epidermic processes are, like the push-rods, surrounded
below by a continuous dermal sheath, the upper edge of which
gives rise to several papillary upgrowths (figs. 8, 13—15
for longitudinal, figs. 9—12 for transverse, section). The
process is continued upwards through the stratum corneum
as a cylinder—either straight or with §-like curves,—which
remains perfectly distinct from the epidermis around, being
separated by a downgrowth of cells so marked that their
direction becomes vertical. In this respect the structures
in question resemble the push-rods (compare fig. 1 with
fig. 8). At its upper free end this cylinder is sharply trun-
cated so as to be flush with the surface of the bill. But in
favorable examples it is surrounded by a distinct circular
depression, and it may even project a little above the general
surface (figs. 8, 18, and 15). At the posterior part of the
upper bill the upper ends of the cylinders are remarkably ex-
panded, so that their outline becomes funnel-shaped (fig. 14).
The §-like curves into which many of the cylinders are thrown
154 EDWARD B. POULTON.
in their passage to the surface are exceedingly characteristic
in appearance (figs. 13 and 15).
The cylinder itself is hair-like in structure, being composed
of elongated fusiform cornified cells in which traces of a nu-
cleus, surrounded by pigment, can be detected (fig. 8). The
lumen of the duct traverses the axis of the corneous cylinder,
and is star-like in transverse section (figs. 9 and 10). The
cells near the cylinder are disposed in concentric circles round
it (fig. 9). Tracing the cylinder downwards into the inter-
papillary process, we find many points of resemblance to a hair.
The general epidermis is continued over it as a sheath, which
strongly suggests the outer root-sheath of a hair, and between
it and the cylinder itself a line of separation tends to appear
(fig. 8, 0.7..s., and also at level c; see also fig. 10). Below,
this sheath forms the outer part of the bulb, and is separated
from the inner part by a space containing small branched cells,
the nature of which could not be determined in my material
(fig. 8, sp., and fig. 13). The cells of this outer sheath are
richly pigmented, like, or even more than, those of the lower
layer of the general epidermis with which they are continuous
(figs. 8 and 10). Within this sheath the cylinder is surrounded
by a distinct layer composed of flattened cells, shown by trans-
verse sections to be two or more deep, but varying in thickness
in different parts (fig. 10, fig. 11, c.; compare also fig. 8, c.).
This layer may represent the inner root-sheath, or the cuticle
of the hair, or both of these together. It is continued over the
inner part of the bulb, and separates the latter from the space
described above (fig. 8, sp.). Within this layer the cylinder
consists, as described above, of fusiform corneous cells arranged
longitudinally ; below, these elements pass into polyhedral cells,
staining in carmine, &c. The inner part of the bulb is made
up of these latter, and presents the strongest possible
resemblance to the bulb of a hair (figs. 8 and 11). In
certain slender cylinders which occur intermixed with the
others, this part of the bulb appears to be wanting (fig. 15).
In the upper part of the bulb, and for a variable distance
above it, the duct is lined by a single layer of distinct cubical
BILL AND HAIRS OF ORNITHORHYNCHUS PARADOXUS. 155
cells (fig. 8), and it coils in a very characteristic manner in the
neck and upper part of the bulb (fig. 13; in fig. 8 the coiling is
only slightly marked ; figs. 14 and 15 represent peculiar forms
in which this character is not present). The coiled duct passes
downwards into a straight section resembling the gland duct in
the dermis with which it is continuous below the bulb (fig. 8).
It was, however, impossible to establish the existence of a
cuticle to the duct in the bulb, although the inner part of the
cells immediately round the lumen stained far more deeply
than the outer part (fig. 11, d’.).
Immediately below the whole bulb the duct is invariably
surrounded by a large ganglion containing abundant medul-
lated nerve-fibres showing Ranvier’s nodes, and large ganglion-
cells (figs. 8 and 12). This ganglion is surrounded by a fibrous
sheath which appears to be prolonged from the dermal sheath
(corresponding to the hair-sac) of the hair-like cylinder. Large
nerves are seen entering the ganglion, the sheath of which is
continuous with their epineurium (fig. 8). The duct which
pierces the ganglion is separated from it by a distinct fibrous
sheath, clearly shown in transverse section (fig. 12) but very
thin in the longitudinal (fig. 8). This sheath appears also to
be derived from the layer corresponding to the hair-sac, and to
be continuous with its inner part, while the ganglionic sheath
is continued from its outer (fig. 8). This, at least, appears to
be the probable interpretation of the appearances represented
in fig. 8; but the whole structure of these hair-like epidermic
cylinders, and the nervous tissues evidently associated with
them, is so remarkable and complex that the fresh tissues are
required for their satisfactory elucidation and for the discovery
of the nerve terminations which we must believe to exist in
connection with the apparatus.
I believe that this account represents all that can be ascer-
tained by the careful examination of the available material,
and that it supports the conclusion I suggested in 1884—that
the gland ducts of the bill reach the surface by entering the
bulbs and by advancing along the medulla of shortened and
degenerate hairs. Such an opinion is further confirmed by the
156 EDWARD B. POULTON.
fact which will be established below that the corresponding
glands of the general body surface of this animal bear a con-
stant relationship to the larger hairs.
TV. Tue Hairs oF Oty anv YounG ORNITHORHYNCHUS.
1. Historical, by W. Blaxland Benham, D.Sc. (Lond.),
Hon. M.A. (Oxon.), Aldrichian Demonstrator in Com-
parative Anatomy in the University of Oxford.
The fact that Ornithorhynchus possesses two kinds of
hair, larger and smaller, was known to Blumenbach (1) and to
Home (2). The latter also recognised the peculiar character
of the larger hairs. He writes (p. 69), “‘The hair is made up
of two kinds: a very fine thick fur half of an inch long, and a
very uncommon kind of hair three-quarters of an inch long;
the portion next to the root has the common appearance, but
for a quarter of an inch towards the point it becomes flat,
giving it some faint resemblance to very fine feathers.” Later,
Glockner (3), in a very brief note describes the larger hair in
somewhat similar terms. In 1823 van de Heeven (4) repro-
duced Péron’s (5) figures of these hairs; the figures are very
small, but are the first published, and show the characteristic
flattening, the narrow stalk, and pointed free end.
In his monograph, Meckel (6) makes no addition to our
knowledge of these structures.
In 1859 Leydig (7), in his classic paper on the Mammalian
coat, is the first to record the fact that the smaller hairs are
in bundles—several in each follicular neck—and points out
that each hair has nevertheless its own special follicle opening
into the bottom of the common pit. He also states that the
large spiny hair is surrounded by the bundles of small ones,
and he gives a figure (pl. xx, fig. 7) of the arrangement, and
shows the sebaceous glands to each hair and the sweat gland
accompanying the large one.
These facts are confirmed by Welcker (8), a few years later,
who gives measurements of the hairs, the smaller being ‘007
mm. in diameter, the larger ones ‘(045 mm. across. He too
BILL AND HAIRS OF ORNITHORHYNCHUS PARADOXUS. 157
gives a figure (pl. 1, fig. 5) showing the grouping of the hairs,
and points out the fact that the roots of the small hairs diverge,
that the hairs then converge in the common follicular neck,
whence they issue in a bundle (p. 69). Whilst Leydig describes
only four or five hairs in a bundle, Welcker found fifteen to
thirty of them.
No author gives any histological details of the structure of
the hair till Waldeyer (9) in 1884, who only describes and
illustrates by photographs the difference in the character of
the medulla and cortex in the various parts of the larger
hair (pl. vii, figs. 1OO—104). According to him (p. 100), the
“bristle hairs”? commence basally in a rounded sliaft of mode-
rate diameter with a thick medulla; then follows a more flat-
tened but very slender “isthmus,” which further on widens
out considerably to form the main part of the hair; this
narrows again at the tip to a point. The isthmus is non-
medullated, and the medulla of the broad part differs from that
in the basal shaft. The paper dealing with the skin of Ornitho-
rhynchus, by Souza Fontes (10), consists of little but quotations
from Leydig. It contains nothing worth recording.
These authors give no information as to the bulb, papilla, and
root-sheaths, nor any other histological details of the hair itself.
As to the arrangement of the hair, Meijere (11), who has
recently examined the arrangement of the hair throughout the
Mammalia, gives a figure on p. 49 showing the hairs on the
back of Ornithorhynchus in more or less irregular rows—
the large hairs, represented as round in section, being separated
from each other by three or four bundles of small hairs: on
the ventral surface, however, the number of the latter is
greater. His measurements closely agree with those given by
Welcker, being ‘048 mm. diameter for the large ones, ‘(008 mm.
for the small hairs.
1. BLumrensbacu.—“ Anatomical Observations on the Structure of Orni-
thorhynchus paradoxus,” ‘ Philosoph. Mag.,’ xi, 1802, p. 366.
2. Home.— A Description of the Anatomy of Ornithorhynchus para-
doxus,” ‘ Phil. Trans.,’ 1802, p. 67.
8. GLtockneR.—‘ Ueber die Haare des Ornithorhynchus,” ‘Isis,’ 1819,
p- 651.
158 EDWARD B. POULTON.
4 Van dE Haven.—‘ Mémoire sur le genre Ornithorhynque,” ‘Nova
Acta’ (Bonn), 1823.
5. Péron and Lesurur.—‘ Voyage de découvertes aux Terres australes
pendant les années 1800—1804.’
6. Mrecxet.—‘ Ornithorhynchi paradoxi descriptio anatomica,’ 1826.
7. Lryp1e.—‘“‘ Ueber die ausseren Bedeckungen der Saugethiere,” ‘ Arch. f.
Mikr. Anat.,’ 1859.
8. WetcKer.— Ueber die Entwickelung und der Bau der Haut und der
Haare bei Bradypus,’’ ‘ Abhand. Naturf. Gesell. zu Halle,’ ix, 1866.
9. Wa.pEYER.—‘ Atlas den Menschlichen und Tierischen Haare,’ 1884.
10. Souza Fontes.—‘ Beit. z. Anat. Kenntniss der Hautdecke des Orni-
thorhynchus,’ Dissert. Inaug., Bonn, 1879.
11. MeErEe.—‘Over de Haren der Zoogdiere,’ Inaug. Dissert., Amsterdam,
1893.
2. General Structure and Arrangement of the Hair.
The hair-like appearance of the epidermic cylinders in the
bill led me to investigate the hairs of the body generally, and
although the results seem to have no special bearing on the
significance of the former structures they are of much value
and interest on their own account.
This research was undertaken in the winter of 1887-8,
when all the figures (except fig. 24) represented on P1]. 15 were
drawn. Since then it has been renewed from time to time,
special attention having been devoted to the subject in Pro-
fessor Lankester’s laboratory dnring the summer of 1892.
Over the general surface of the body and head of Ornitho-
rhynchus we meet with large hairs, each of which is attended
by bundles of small hairs. Furthermore, on the dorsal area at
least, each large hair is attended by a constant number of four
bundles of small ones. The latter commonly vary in number
in the dorsal region from seven to twelve in a bundle, and all
emerge in a sheaf from a common follicular mouth. The
four bundles are arranged on each side of and behind the
larger hair, and, as all slope obliquely backwards, the shafts
of the small hairs lie beneath that of the large one belonging
to the same group.
The protective large hairs are evidently subject to much
wear and tear, and succeed each other very rapidly; the new
BILL AND UWAIRS OF ORNITHORHYNCHUS PARADOXUS. 159
successional hair, which is to be met with in nearly every sec-
tion, emerging from the same follicular mouth in front of, and
therefore overlapping, the base of the old one. The succession
of smaller hairs is less rapid, but one or more younger, growing
hairs are to be seen in every bundle. The exact relationship
in two instances can be made out by ascertaining the propor-
tion of dark circles (sections of young hairs) to the smaller
unpigmented circles (sections of bases of old hairs) in each of
the four bundles in figs. 17 and 18 of Pl. 15, which contains
all the figtires illustrating the structure of the hair of Ornitho-
rhynchus. The proportions in the four bundles of three other
groups were as follows:
First GRovp.
Bundle I. ; . 1 young hair to 9 or 10 old.
ea Waa : On sg, Shales’ £0 Se,
ro TE ; ea. nal te 1 5
arte LNs 3 5, hairsito By aay
SEcOND GROUP.
Bundle I. : : . 1 young hair to 8 old.
Het ae : Oe oe a DBIS:LOlS. 3s,
hie : : Sle ae Tal cOn Ory 55
as ing ge HAIESL6O 7) 55
THIRD GROUP.
Bunde I. : ; . 2 young hairs to 7 old.
vat pili tien F ; tbs sees Lair etO, gaya
” TE . . ° = ” oF) 8 ”
cpallNat . P aoe. hairsstonl 0)...
The proportion of young hairs was ascertained in each case
by counting the numbers of dark and light circles in a section
taken sufficiently near the surface of the skin to bring all the
hairs of a bundle into a common follicular neck, as in fig. 17.
These figures, together with the number shown in fig. 17,
prove that, at any rate in the dorsal region, Leydig’s estimate
of the number of small hairs in a bundie is too small, while
that of Welcker is far too large.
Each group of hairs is attended by a single gland exactly
VOL. 36, PART 2,—NEW SER. M
160 EDWARD B. POULTON.
resembling those of the bill of the same animal or the sweat-
glands of mammals generally. On the head and back, the duct
invariably opens just in front of the large hair, so that it is the
most anterior member of the whole group.
Such a group is distinctly represented in transverse section
immediately beneath the skin in fig. 17, where the young
successional hairs, both large and small, are at once distin-
guished by the presence of pigment from the colourless bases
of the shafts of older hairs. At this level the gland-duct (d.)
is seen in section in front of the large hair. At a lower
level, as shown in fig. 18, the duct is replaced by a secretory
tube of typical structure (g.), which is often placed, as in this
figure, between the large hair and one of the bundles of smaller
hairs. The latter, at this level, have become separate, but their
outer root-sheaths are continuous peripherally, forming a single
epithelial mass. Furthermore, the two masses on each side
have approached and tend to coalesce (compare figs. 17 and 18),
soon doing so completely. At this level, and just below it,
the small unpigmented bases of the older hairs come to an
end (one is thus ending in each of the right-hand bundles in
fig. 18). The younger growing shafts, with their outer root-
sheaths no longer continuous, descend much deeper, gradually
converging to form a single bundle, which lies in the middle
line under the follicle of the large hair. The growing small
hairs then end in bulbs at various levels, but the most deeply
placed do not, as a rule, descend so far as the bulb of the large
hair.
The large hairs terminate superiorly in large, although
narrow, flattened leaf-shaped expansions borne by a shaft long
enough to carry them just beyond the ends of the small hairs.
As the neck of this shaft is comparatively thin, it is probable
that in the living state the terminal shields tend to fall over,
and lie flat on the finer hair beneath, their tips pointing back-
wards and overlapping the basal part of the shields behind
them.
The tip of the shield is beautifully formed and free from
pigment. Looked at from the side or in section, the pigment
BILL AND HAIRS OF ORNITHORHYNCHUS PARADOXUS. 161
is seen to be more developed on the under side, and a broad un-
pigmented margin is always found along the upper surface.
This is in part due to the cuticle of the upper surface, which
is much thicker than that of the under, as can be well seen in
sections (fig. 18). The degree to which the pigment extends
towards the upper surface differs, however, in individual hairs.
Passing along the centre of the shield is a very distinct medulla
containing abundant gas-bubbles, entering as a result of the
contraction of soft protoplasmic cells, which are present and
stain in logwood, &c., in the shields which have not yet
appeared above the surface (fig. 18). Traces of this medulla
can be followed nearly to the tip of the structure. The pigment
is sometimes so far restricted to the under side of the shield as
to be entirely beneath the level of the medulla in the middle
line, but it extends further at the sides (fig. 18).
Below the shield there is a constricted neck in which the
medulla is apparently wanting and the pigment scanty, but
immediately below, at the beginning of the long shaft, the me-
dulla is strongly developed suddenly and the pigment assumes
the characteristic ladder-like appearance, being arranged in
bands alternating with spaces containing colourless shrunken
cells and gas bubbles. The arrangement is not, however, so
regular as that of the corresponding part of the smaller hairs.
Below, the shaft passes into the somewhat more slender unpig-
mented base in which a medulla is wanting.
Hence in one of the larger hairs we can distinguish a shield,
neck, shaft, and base.
The small hairs can be divided into similar regions. The
free tip 1s unpigmented, and is followed by a section corre-
sponding to the shield, with the same longitudinal arrange-
ment of small granules and large fusiform masses of pigment.
Traces of a medulla marked by gas bubbles are to be seen in
places. Then follows the part which represents the neck,—not,
indeed, more slender, but with the same diminution in the
amount of pigment and cessation of the medulla. To this
succeeds the shaft with an extremely uniform alternation of
medullary dark and light bands, the latter containing distinct
162 EDWARD B. POULTON.
gas bubbles. As in the larger hairs, the shaft passes into a
pigmentless and, in this case, far more slender base (fig. 18,
in which the sections of the colourless bases of small hairs
are seen to be much smaller than the dark circles which
represent the sections of shafts and upper parts).
There is no trace of either pigment or medulla in the bases
of both larger and smaller hairs (see figs. 17 and 18 for
sections of bases).
Such is the appearance and arrangement of the groups of large
and small hairs which apparently cover the whole surface of
the body and head.
On the upper surface of the tail the large hairs become stiff
and bristly, but still flattened. They probably correspond to
the shields only of the large hairs already described. Passing
from the back on to the tail the small hairs become short and
scanty, and towards the tip disappear altogether. The succes-
sional hair emerges as on the body, and the overlap is the
same, but the gland duct opens beneath and behind instead of
in front of the hair. The under surface of the tail in a male
individual not quite fully grown was covered with short
flattened large hairs set very obliquely, in fact almost hori-
zontally. They appear to want the medulla and the hair pig-
ment. In several full-grown individuals this part of the tail
was more or less bare, but the bases of large hairs could be
detected, together with patches of hairs having frayed and
worn ends.
The manus and pes, except for the bare palmar and plantar
surfaces, are covered with hairs very similar to those last
described, but even shorter.
Everywhere the groups of large and small hairs, or the large
hairs alone, appear to be set in irregular rows, transverse to
the long axis of the body which they clothe.
3. Comparison between the Hairs of Old and Young
Animal.
Very interesting results follow from the comparison of the
hairs of the mature animal] with those of a young one in which
BILL AND HAIRS OF ORNITHORHYNCHUS PARADOXUS. 168
only the tips of the larger hairs had appeared above the surface,
while the smaller ones were still at some distance below it (see
Pl. 15a).
The parts of the large hairs which had been formed evidently
corresponded to the terminal shields of the adult, from which
they differed in their greater thickness, although the differen-
tiation of an upper surface with a greatly thickened cuticle
from a lower surface in which the pigment was concentrated,
‘was equally marked (figs. 19, 20, and 23).
The gland, which was quite short in the young animal, opened
in front of the large hair, as in the adult.
The small hairs differ from those of the adult and in such a
manner as to indicate the probable origin of the four bundles
emerging from common mouths.
When longitudinal sections or successive transverse sections
of the young skin are examined, each large hair is found to be
accompanied by four tubes, exactly like the ducts of glands, open-
ing on the surface. These tubes correspond in position to the
four bundles of small hairs in the adult. Tracing them down
wards, each tube gives rise to a bundle containing a much smaller
number of hairs than in the adult. Each bundle nearly always
contains a single hair which is specially prominent, and it is
this latter which occupies the lumen of the tube itself, the
apex being however, in animals of this stage of growth, a con-
siderable distance below the surface of the skin. In some
cases, however, the four bundles of a group appear to contain
only three chief hairs between them, in others as many as five
or six. Four tubes appear to be always present, evidently
representing the four common follicular necks and mouths of
the adult. There is great disparity in size between the chief
hair in a bundle and the smaller ones which are grouped
around it. The hairs of the four sheaves terminate in bulbs,
placed, as in the adult, at very different depths below the
surface.
From this structure it may be inferred that each of the four
bundles is, in the course of development, at first represented
by a single hair (the chief one), formed in a follicular
164 EDWARD B. POULTON.
tube which is open to the surface, and that the smaller
hairs are developed in follicles which are outgrowths from
some part of this open tube, and which do not tend to coalesce
as in the mature animal. The distinction between chief and
smaller hairs is subsequently lost, probably in the hairs which
succeed those described above. In an animal of this age, how-
ever, the first-formed hairs have not approached their full size,
and no trace of any successional hair—either large or small—is
to be seen.
The existence of a distinct lumen opening on the surface at
some distance above the apex of the hair is of such great
morphological importance and interest in relation to the origin
of hairs and their homology with feathers and scales, that the
observation was confirmed again and again, by examining both
longitudinal and transverse sections, until there could be no
doubt about the matter. I was then anxious to ascertain
whether the same fact held true with the larger hairs. The
general surface of the body was valueless for this inquiry, as
the animal was rather too old, and the tips of the large hairs
were visible. On the under side of the tail, however, many, at
any rate, of the large hairs (here unaccompanied by small ones)
had not reached the surface, and there was an open tube above
them, as in the case of the small hairs elsewhere. The probable
bearing of this and many other peculiarities will be discussed
at the end of the paper.
The hairs upon the tail bear the same relation to those on the
body as that already described in the adult.
4. Minute Structure and Formation of Hair and its
Sheaths.
The hair is developed from a bulb composed, as in other
mammals, of polyhedral cells of the rete mucosum connected
with the superficial epidermis by an outer root-sheath,
Protrusions of the latter give rise to apparently typical sebaceous
glands, situated, in the smaller hairs, at the level at which the
bundle unites to enter a common follicular neck, The upper
BILL AND HAIRS OF ORNITHORHYNCHUS PARADOXUS. 165
part of the outer root-sheath often contains cystic growths like
those described in other mammals.
The bulb is entered by a dermal papilla which is, at any
rate in the large hairs of the young animal, of enormous
size and length, extending beyond the bulb far into the base
of the hair proper, where the cells are fusiform, pigment
abundant, and the cuticle well defined (see fig. 23, in which,
however, the two halves have fallen apart, so that the hair
appears to be thicker than the bulb). The bulbs and papille
are especially large in the tail.
A very long, well-marked papilla also penetrates the long
narrow bulbs of the smaller hairs. The relations of the adult
bulb and papilla appeared to be very similar, but were made out
with greater difficulty than in the young animal, which was
chiefly employed for the histological side of this inquiry.
From the tip of the papilla, at any rate in the larger hairs,
an axial rod of soft protoplasmic cells, deeply staining in
reagents, is continued (fig. 18, mature; figs. 19 and 24,
young). This, when dried and shrivelled, admits the air and
forms the characteristic medulla.
Around the papilla the inner zone of cells of the bulb forms
the hair proper with its cuticle, the structure and mode of
formation being typical except for the bilateral symmetry of
the larger hairs indicated by their shape, thickened upper
cuticle, and predominant pigment on the under surface (fig.
18, mature; figs. 19, 20, and 23, young).
External to this zone the cells form the inner root-sheath,
while the outer root-sheath appears to be continuous with
the lower part of the bulb. The latter sheath is probably
always typical, but it was extremely hard to make out in cer-
tain sections of the young animal (figs. 19, 20, 23), although
in others it was perfectly distinct and of normal appearance
(figs. 21, 22, 24). This discrepancy is due to the facts that the
thickness of the sheath varies greatly at different levels, and
that the animal was not prepared for histological investigation.
The inner root-sheath is always present in the developing
hair, and is a structure of great importance, throwing much
166 EDWARD B. POULTON.
light upon the corresponding sheath as it is described in other
mammals. As in the latter, the inner root-sheath surrounds
that part of the hair which is enclosed in the follicle, but grow-
ing less rapidly it does not extend to the neck through which
the hair protrudes; hence we do not find it at all in sections
of the upper part of the follicle (figs. 17 and 18). It is far
thicker and more important in the hairs of the young animal,
and especially in those of the tail (figs. 21 and 22,7.7.s.). In
structure it consists of a network of corneous fibres enclosing
fusiform meshes with a longitudinal direction, through which
the outer root-sheath can be seen (fig. 24). Hence in certain
cases Henle’s original description of his inner root-sheath as a
fenestrated membrane is certainly true of Ornithorhynchus.
Henle’s account is also followed by Mertsching (‘ Archiv f.
Mikr. Anat.,’ 1888, p. 32), who shows in pl. v, fig. 8, that this
sheath in the human hair possesses a fenestrated structure. In
transverse section the corneous fibres of Ornithorhynchus are
seen to be polyhedral and irregular in outline (figs. 19, 21, 22,
i. 7. 8.), and if the section be taken at some little distance above
the bulb, distinct spaces appear between them (figs. 19 and 21).
Round the small hairs, on the other hand, only a uniform layer
of the proportions usually found in Mammalia could be detected
(fig. 19). In some of the longitudinal sections of larger hairs I
could make out a thin internal layer exhibiting a serrated
edge, with teeth the reverse of those on the hair cuticle. This
evidently represents the so-called cuticle of the inner root-
sheath, and it is probably shown in section in fig. 21. Well above
the bulb (fig. 21) this thin internal lamina was the only trace of a
separation of the sheath into layers, and even this could be
detected only occasionally ; and I gained the impression that it
is not a distinct and definite layer, but merely the condensation,
as it were, of the innermost part of the inner root-sheath upon
the exterior of the hair and the moulding of its surface by con-
tact with the cuticle of the latter. But at a lower level, just
above the bulb, there is seen what I believe to be the homologue
of Huxley’sand Henle’s layers. Thus in fig. 22 the fibres of the
inner zone have not become corneous and take the stain readily,
BILL AND HAIRS OF ORNITHORHYNCHUS PARADOXUS. 167
while still nearer the bulb in fig. 20 the same inner layer shows
distinct traces of cells with their nuclei. In both cases this
inner zone probably represents Huxley’s layer, but it does not,
at least in Ornithorhynchus, imply any differentiation of the
sheath into layers, and when we consider the immense develop-
ment of the structure in this animal, it seems probable that
the distinction can hardly be sustained throughout mammals
generally. The appearance which has led to the separation of -
the sheath into two layers is, I believe, due to the fact that the
inner and outer zones do not arise from the cells of the bulb
at one and the same horizon, but that the inner zone rises at a
higher level than the outer. It therefore follows that, in sec-
tions at a certain horizon, the cells of the bulb will have
undergone complete transformation into the corneous substance
of the sheath in the outer zone, while they will still remain
unchanged, and at a rather higher level only partially changed,
in the inner zone. At any rate, I am convinced that this is
the explanation of the two layers in figs. 20 and 22.
In certain sections the inner part of the sheath appears to be
made up of fibres and the outer part homogeneous (fig. 18), but
in the thickest and best developed sheaths (fig. 21) the fibrous
structure is distinct in the outer as well as the inner part (with
the exception of the thin ‘‘cuticle’’).
5. Mode of Succession of the Hairs.
The succession of the hairs already alluded to presents
some features of great interest. The appearance of two hairs
in one follicle is spoken of as an occasional appearance in other
mammals. In the large hairs of Ornithorhynchus it is the in-
variable rule, and the succession follows the bilateral symmetry
of the hair itself, the new hair always overlapping the anterior
surface of the old. Hence a section of the neck of the: follicle
commonly shows the terminal shield of the emerging young hair
anteriorly, and the circular or oval base of the old hair opposite
the middle of its posterior surface (fig. 17). We do not find
two shafts in one neck, because the old hair is doubtless shed
by the time that the new shield has risen above the surface,
168 EDWARD B. POULTON.
In deeper sections it is found that the base of the old hair,
becoming irregular in outline, first enters (fig. 18), and then
soon comes to an end, in the posterior wall of the outer
root-sheath, terminating as a knob, a rounded end, or in
diverging lines of corneous cells, which spread among the cells
of the sheath; while the new hair passes far deeper and ends
in a bulb. When, similarly, successively deeper sections of a
- bundle of small hairs are examined, it is found that the pig-
mentless bases of old hairs soon come to an end, together with
their follicles, while the much larger coloured shafts of the still
growing hairs pass down far deeper to end in bulbs. These
appearances are confirmed by the examination of longitudinal
sections, which show that the extreme ends of the bases thin
away, become sinuous, and sometimes branch among the cells
of the outer root-sheath.
The probable interpretation is as follows:—AlIl the parts of
a hair which possess the typical structure are developed froma
bulb, which is nearly exhausted when the hair has almost
attained its full length. The hair is, however, further pro-
truded by the development and growth of a solid and pig-
mentless base possessing none of the typical features of a hair,
from the outer root-sheath, which is thus used up, the whole
follicle being shortened and retracted towards the surface.
When the base of the old hair has reached a certain height the
lower blind end of the outer root-sheath begins to descend
again, forms a new bulb and a new hair, which therefore
ascends and passes the base of the old one. It is perhaps
significant that each fresh hair which develops during the life
of the individual should originate in an epidermic down-
growth, thus repeating the development of the first-formed
embryonic hair.
It is, I think, improbable that the outer root-sheath which
forms the base of the old hair, and not the typical structure of
the shaft, should itself possess the power of originating a new
bulb. It is more probable that the new bulb is developed
from cells which are the genetic descendants of the old one,
and which retain similar potentialities.
BILL AND HAIRS OF ORNITHORHYNCHUS PARADOXUS. 169
In the case of the small hairs, however, we do not meet with
two in any of the single follicles; and here the old hair is
shed before the tip of the new one reaches its base, or, as I
think more probable, a new follicle is formed, perhaps from the
point which represents that at which the earliest follicles were
budded out from the first single tube.
6. Recapitulation of Essential Peculiarities of Hair
of Ornithorhynchus as compared with that of
Higher Mammalia.
In order to draw the conclusions which, as I believe, follow
from the investigation of these structures in the lowest mammal,
it will be convenient to recapitulate, making a short statement
of the points in which the hairs differ from those of other
Mammalia.
In the GRouPs OF HAIRS we notice a marked bilateral sym-
metry (fig. 17), and a definite relation to the bilateral sym-
metry of the body itself.
In the tance Harrs there is (1) a most distinct bilateral
symmetry in shape, in structure (differentiation of upper and
lower surface), and in succession, and a definite relationship
to the bilateral symmetry of the body. The terminal shields
are scale-like, especially in the tail, where they become
sessile.
(2) The inner root-sheath is of immense size (especially in
the tail), and possesses a definite and peculiar structure, being
formed of longitudinal corneous fibres united into a network.
(3) The bulb and papilla are extremely large, the latter
penetrating the base of the hair.
(4) The hair is developed in a tube, which is open to the
surface,—-in a tubular and not a solid downgrowth from the
epidermis.
The sMALL HAIRS are (1) arranged in bundles, which are bi-
laterally disposed in relation to the large hair, and have a
definite relationship to the bilateral symmetry of the body,
although the shape and structure of the individual hairs is not
170 EDWARD B. POULTON.
bilaterally but radially symmetrical, while the succession is
probably asymmetrical. In these characters and in the general
appearance the individual smaller hairs are not scale-like, but
resemble those of Mammalia generally.
(2) The inner root-sheath has about the same relative size
aud the same appearance as that of Mammalia generally.
(3) The bulb and papilla are both very long and narrow,
but apparently not so remarkable as those of the large
hairs.
(4) The hair is developed in a tube open to the surface.
(5) The fact that the bundles of small hairs reach the surface
by a common follicular mouth is probably shown by the mode
of development to be a comparatively recent feature of no
ancestral significance. ,
Comparing these features of the large and small hairs, we
may omit the last-mentioned characteristic of the latter, which
is probably unimportant. There is similarity in the most im-
portant point of all—development in open tubes ; but in every
other respect there is divergence, and always of such a kind as
to bring the small hairs nearer to those of mammals generally.
Furthermore, the characters which separate the large hairs from
those of other mammals are, as I hope to show in all cases,
characters which suggest homology with one or other of the
epidermal structures of the higher Vertebrate Classes.
V. Tue Homotociets AND OrniGiIn oF Mammatian Harr.
1. Historical; by W. Blaxland Benham, D.Sc., &e.
There are two main suggestions as to the relation of hairs to
other Vertebrate epidermal structures :
(1) The first is that hairs are homologous with the scales of
reptiles and the feathers of birds.
(2) The second regards them as not homologous with these
structures.
In the latter case two other epidermal structures have been
suggested as the ancestors of hairs: —(@) Maurer would derive
hairs from epidermal sense-organs of fishes and Amphibia; (6)
BILL AND HAIRS OF ORNITHORHYNCHUS PARADOXUS. 171
Emery regards hairs as “substitution-derivatives” of certain
elements of the placoid scales of fishes; whilst (c), finally,
there are authors who have contented themselves with con-
testing the hair-feather-scale theory, but have not attempted
to homologise hairs with anything else, appearing to regard
them as structures sui generis.
All authorities appear to be in agreement that scales and
feathers are homologous structures. The mode of origin is
closely similar in the two cases—the first forecast being an up-
growth of the corium giving rise to a papilliform projection
carrying the epidermis outwards ; the cells of the epidermis
then proliferate to a slight extent, and later become horny.
In the scale the papilla is more or less flattened, with its free
edge (apex) directed backwards, and the cornification becomes
more marked on the upper (outer) surface than on the lower
(inner) surface of the papilla.!_ In the case of the feather, the
papilla becomes more or less cylindrical and more upright;
further, its base sinks down into the corium at an early age, so
that the root of the feather comes to lie in a follicle. The axis
of the papilla (corium) gives rise to the pith of the feather
axis ; the enveloping cornified epidermis to the barbs. But the
most superficial coat of the papilla (the original outermost
layer of epidermis) forms a sheath which closely surrounds the
feather forecast. As the latter grows it breaks through the
apex of the sheath, which then dwindles and is ultimately cast
off. This feather-sheath consists of two layers of cells—a more
superficial cornified layer, and a deeper layer of granular cells.
The feather, then, may be derived from a scale by supposing
that the original papilla, after growing outwards for a time, has
sunk downwards into the corium so as to give rise to a follicle
and a root-sheath, for the purpose of better support and nutri-
tion.
Those authors who take the view that the hair is homologous
with the feather, believe that this process of sinking has gone
1 Tt is noteworthy that this statement is in every detail an exact description
of the formation of the large hairs of Ornithorhynchus, save that they are
developed as an upgrowth at the bottom of an open pit.—H. B. P.
172 EDWARD B. POULTON.
on further, and that it takes place at an earlier stage in the
ontogeny of the hair; while the proliferation of the epidermis
also commences much earlier than in the feather-forecast, and
is more extensive and rapid.
On the other hand, it is argued that if the hair is homologous
with the feather, there should be a more or less close agree-
ment in (a) the final structure of the two, (d) their mode of
development, and (c) their arrangement on the body. It is
unnecessary here to compare the structure of hair and feather
in detail. In both there is a (1) root (embedded in the skin)
embracing a vascular papilla, which is of much greater extent
in a feather than in a hair; and (2) a projecting shaft of horny
cells (which in the feather is flattened like a scale, but bears
secondary processes upon it, while in the hair it is cylindrical
or flattened). With regard to the arrangement of these struc-
tures, nothing need, I think, be said here. But in dealing
with the development we are at once met with the fact that
two divergent accounts have been given with regard to the very
first forecast of the hair. Some authors, amongst them Gotte
(2), Reissner (1), Studer (4), Kerbert (8), describe the first
trace of a hair as showing itself in the form of a projection of
the corium into the epidermis, just as in the feather and scale.
This is the view which is taken by all those who regard feathers
and hairs as homologous (Claus, Wiederheim, Hertwig, and
others).
It may be well to refer to the observations of Davies (5) on
this point, who finds the earliest forecast of a hair—or rather
a spine, for he studied the matter in the hedgehog—marked
out by a proliferation of “ intermediate cells,” and not by the
deeper columnar cells. He carefully describes the mode of
origin of feather and spine, and regards both of them as de-
scendants from a common scale-like structure. He holds that
the “scales”? covering the feet of birds are not homologous
with Reptilian scales ; the former frequently carry feathers, just
as the “scales” of Dasypus carry hairs, and he argues that if
the reptilian scale is equivalent to a feather and to a hair, then
two of these homologues cannot be superposed. These scales
BILL AND HAIRS OF ORNITHORHYNCHUS PARADOXUS. 173
in birds are merely secondary thickenings round the base of
the feather.
He compares the hair rather with the definitive feather, which
arises from the root of the down papilla, and he emphasises the
fact that in this case, as in the hair, the feather forecast has
to make its way up through the neck of a solid follicle of de-
generating cells (cf. his figs. 30—39), as on the appearance of
the definitive feathers there has been an active downgrowth of
the primary feather follicle.
The spine of the hedgehog is, of course, derived from a more
simple hair, and is not to be compared with the quill of a
feather.
On p. 629 he summarises his ideas on the mode of origin of
the feather :—1. A simple thickening of the skin. 2. A radially
symmetrical knob. 3. A backwardly directed papilla whose
horny layers become thickened round the apex. 4. A back-
wardly directed papilla whose point ended in a short, thick,
hair-like process. 5. A longer hair-like structure, which con-
sisted of a firmer cortical layer and a looser axial tissue, and
whose base became sunken with the cutis papilla into the skin.
6. By the bursting of the wall of the freely projecting part of
this structure the enclosed tissue became free, and, separating
into distinct filaments, gave rise to the primitive “ down.”
But even admitting the general similarity in the mode of
their development, one of the earliest to draw attention to the
striking difference between hairs and feathers was Gegenbaur,
who contrasted the early outgrowth of the feather forecast, in
which the greatest share is taken by the corium papilla, with
the early downgrowth of the hair forecast, due to active proli-
feration of the epidermic constituents.
This doubt as to the strict homology of feather and hair is
accentuated by the statements made by other observers—
Kolliker, Romer (12), Maurer (9), &c.—that the corium papilla
is not the first to become defined, but that the epidermis is for
some time the only representative in the hair forecast. This
statement, which appears from Maurer’s work (1892) to hold true
for a variety of forms, may be explained in either of two ways:
174 EDWARD B. POULTON.
the upholders of the hair-feather homology interpret it merely
as a precocious development of the epidermic constituent, owing
to the greater importance of the root in the hair, while the
corium papilla, which only becomes of importance for nutritive
purposes at a later stage of development, is correspondingly
delayed. Such blearing of the record is not unknown in onto-
geny; indeed, the fact that the follicle of the hair commences
as a solid structure with no morphological communication with
the exterior, and only later shows a differentiation into outer
and inner root-sheaths as the hair pushes its way through the
superficial layers of epidermis may be, and has been, explained
by invoking the same causes—precocious and retarded develop-
ment.
But another meaning has recently been given to the early
activity of the epidermis in the formation of the hair. This
and the various other divergences from the process occurring
in feather-formation have been regarded as pointing to an en-
tirely different ancestor for the hair, viz. to the epidermal sense-
organs of fishes and Amphibia. Maurer (1892) is the parent
of this ingenious theory, and discusses the question in a very
complete and thorough manner. In his earliest paper he con-
trasts the development of hair, on the one side, with that of the
feather and scale on the other. He has examined the matter
for himself, and finds what he regards as very essential differ-
ences between them.
He describes the early stages in the development of the hair
in a variety of mammals,—cat, mouse, mole, hedgehog, Dasy-
urus, and Perameles; and though the details may differ, he
finds that invariably the first trace of a hair is expressed by the
elongation of the cells of the deepest layer of the epidermis ;
so that the forecast is sharply marked off from the surrounding
epidermis. These elongated cells may, as in Talpa, reach
the surface, which is here slightly pitted (giving an appearance
exceedingly similar to that of an early stage of an Ichthyopsid
sense-organ). Usually, however, the superficial flattened cells
of the epidermis are continued over the tops of the elongated
cells. The modification of cells indicating a corium papilla
BILL AND HAIRS OF ORNITHORHYNCHUS PARADOXUS. 175
appears distinctly later—in Dasyurus very much later.
Papillee of the corium are scattered about, in many of the cases
studied, quite independently of hair forecasts; and Maurer
believes that these papille, even if they happen to underlie such
forecasts, are not the true root papille of hairs.
In the development of feather (or scale) he finds very marked
differences ; the papille—projections of corium—are distinctly
the first to appear, and the epidermis covering a papilla is for a
considerable time not different from the surrounding epidermis.
There is no elongation of cells of the Malpighian layer; the
deepest cells are cubical, as elsewhere. The papilla, it is well
known, as it grows outwards becomes oblique, so that upper
(outer) and lower (inner) surfaces are distinguishable ; the epi-
dermis is scarcely changed, but is thicker on the outer side,
and this thickening is due to proliferation of cells in the middle
layers—not of the deepest layers. With regard to the root-
sheath of a hair,—suppose, he says (1893, 4), that an epidermic
scale or feather, with its papilla, were to sink downwards into
the corium (as a hair forecast does) then the resulting sheath
will be simple. A deep-lying position does not necessitate a
complex sheath enclosing the horny structures ; so that an epi-
dermal structure, borne by a deep-lying papilla and possessing
a complex sheath (hair), is not directly explained as a papilla
which has sunk downwards (feather). Further, although the
feather-sheath is fairly complex, it is very early cast off after
the feather forecast has ruptured the apex of it; whereas the
root-sheath of a hair persists throughout life and grows so long
as the hair continues to grow. He gives (1893, c) an explana-
tion of this feather-sheath—he compares it with the moulted
cuticle of reptiles. In the case of these animals, more or less
extensive portions of the stratum corneum are from time to
time cast off; but before this occurs the underlying epidermis
has formed a new “cuticle” and new stratum corneum, so that
in section the old stratum corneum does not rest directly upon
the new one, but is separated from it by a layer of granule-
containing cells; below this is the new “cuticle”’ and then the
new stratum corneum. The feather-sheath consists of an
VOL. 36, PART 2.—NEW SER. N
176 EDWARD B. POULTON.
outer layer of cornified cells and a deeper layer of granule-
containing cells, immediately surrounding the feather forecast.
This sheath, then, that is shed later on, he believes to be the
same thing as the old “cuticle ;’—instead of a new “cuticle”
below it, there has been formed the feather. He thus explains
this feather-sheath in much the same way as the believers in
the feather-hair theory might explain the inner root-sheath.
But, he points out, there is no such periodical shedding of the
stratum corneum in mammals, preceded by the formation of
definite layers below; there is a gradual transition from the
deepest layer to the most superficial cells which drop off from
time to time.
Having in this way contested the supposed homology be-
tween hair and feather, he proceeds to elaborate his theory
as to the homology between a hair and an epidermal sense-organ
of fish and Amphibia. He believes that there is no essential
difference between an ‘‘ Endknospe” and a “ Nervenhiigel ;”
each is essentially a collection of nerve end-cells (which may
or may not traverse the whole depth of the epidermis) sur-
rounded by elongate supporting cells, which separate the organ
from the general epidermis. He describes (1892, @) these
in various fish from different parts of the body. In the Peren-
nibranchiate Amphibia and in several Caducibranchiate (e. g.
Triton) they exist throughout life; while in others (Sala-
mandra) and in Anura they are present only in the larva.
In the newt, after the metamorphosis, these sense-organs
have a definite relationship to the wart-like papille which
have made their appearance. At the top of the wart, in a cup-
like depression, lies the sense-organ, which is protected by the
‘overhanging lips of the cup, formed of horny epidermal cells.
This removal of the sense-organ from the surface and its loss of
function are evidently related to terrestrial life, and differ from
the sinking of this same organ in fishes, where it still retains
its function.
In Menopoma, Menobranchus, and Cryptobranchus,
though aquatic, the sense-organ is similarly removed from
the surface; this he explains by the suggestion that these
BILL AND HAIRS OF ORNITHORHYNCHUS PARADOXUS. AG
forms were originally terrestrial and have returned to an
aquatic life. Now in Cryptobranchus the condition of the
sense-organ and its relations are very interesting. The epi-
dermis contains a considerable number of layers, the three or
four outer ones being horny. The sense-organ at the apex of
the wart-like papille is oblique to the surface; there is a
corium papilla below it, on the apex of which the sense-cells
are grouped. The supporting cells are greatly increased in
number, and leave only a very narrow channel between them
leading to the centrally-placed sense-cells.
After describing these organs, he proceeds to compare the
various parts with those of a hair. The axially-placed sense-
cells, having no longer any purpose to serve, have disappeared,
and with them the nerve; nevertheless, the cells are repre-
sented by the medulla of the hair. The supporting cells have
given rise to the cortex; the outermost, or covering cells, as
he terms them, become the cuticle, while the protecting cells,
overhanging the sense-organ in Triton, &c., are represented
by the inner root-sheath in the hair. This gives what in
Maurer’s opinion is the only intelligible explanation of this
inner root-sheath.
In reply to this theory as to the origin of the Mammalian hair,
Leydig (8) admits that there is no doubt a remarkable similarity
in the mode of development of the two structures, but points
out other structures which might with greater probability be
taken into consideration in dealing with its phylogeny, viz. the
“ perle-organe ” of certain Cyprinoid fishes and the “ femoral
glands” of Lizards, which he had previously described (7). He
rejects a view which he was tempted at one time to hold, that
hairs may be derived from horny projecting papille on the skin
of Amphibia which are mere local thickenings of the epidermis.
Maurer again (c) takes up the cudgels in favour of his theory,
and describes the results of his own examination of these two
organs, which he has made with an open mind, putting
himself, as it were, in the position of a sceptic of his own
theory. But he finds no reason to alter his previous opinion,
indeed is strengthened in it: he adduces confirmatory evi-
178 EDWARD B. POULTON.
dence and launches other arguments against the feather-hair
theory.
Still more recently, Emery (15) has given in a preliminary
note another view as to the origin of hair, promising further
details later, Hairs are not homologous with scales and
feathers, but all three are descended from the placoid scales of
fishes by the substitution of horny material for bony substance.
In other words, the horny part of these structures in the
higher animals represents the enamel, and the underlying
papilla (whether ossified or not) represents the basal (cement)
plate of the placoid scale. Starting with this assumption, for
which he at present brings forward no justification, he seeks to
show that he is supported by the relation of the hairs to the
scales in the scale-bearing mammals, for he finds, as did Romer
for Dasypus, that hairs arise on the papillz or forecasts of
scales, and not between them, as Weber believes. In cases
where hairs are, in the adult, situated between scales (as on
parts of the body of Dasypus, Chlamydophorus, and other
mammals) Emery supposes that the scales which originally
bore the hairs have been crowded out of existence by the
greater development of neighbouring scales, which in their
turn may lose the hairs; these, however, may commence to
develop, as Romer has shown in the case of Dasy pus.
With reference to the “scales” of certain mammals (Manis,
Dasy pus) it has been stated, though without sufficient founda-
tion, that they are “ fusions of hairs.” ‘There is, however, not
the slightest foundation for this view, as Max Weber (13),
Romer (12), and others have pointed out. They are true scales,
similar to those of reptiles! and having a similar developmental
history, the chief difference being the shedding of the scale in
reptiles and its permanence in mammals. Both the above-
mentioned writers agree in regarding the scales and hairs as
different structures. Thus Romer says, “ Hairs have nothing
whatever to do with scales ;” whilst Weber, in a later paper
(14), hesitates to adopt either Maurer’s view or the earlier theory.
Weber, however, regards the scales of Manis, those on the tail
1 Davies, however, denies this.
BILL AND HAIRS OF ORNITHORHYNCHUS PARADOXUS. 179
of the rat and other mammals, as directly descended from those
of reptiles, and as remnants of an original scaly covering of
mammals; the arrangement of hair on the body of these ani-
mals being dependent upon the scales. He also holds that
even in scaleless mammals the arrangement of hair retains the
impress of the original scaly covering. Romer, on the other
hand, regards the scales as secondarily acquired. These scaly
mammals are, according to him, descended from hairy mam-
mals; the epidermis merely retains the power, inherited from
reptiles, of forming scales, which are thus not directly descended
from those of reptiles and not strictly homologous with them
(rather they are homoplastic).
As to hairs, Romer does not attempt to refer to any other
epidermal structure as their ancestor, but follows Haacke in
his suggestion that an explanation of the epidermic down-
growth of the hair forecast may be found in the fact that the
scales of reptiles touch or overlap, and between two scales
there is therefore a depression of the epidermis. Of this he
says, ‘‘In this epidermal depression I see the spot for the
origin of the hair, whence the hairs will develop by a cornifica-
tion of the epithelial cells.”
LITERATURE.
1. RetssnER.—‘ Beit. z. Kenntniss d. Haare d. Menschen und d. Saugethiere,’
1854.
2. Gottr.— Zur Morphologie der Haare,” ‘ Arch. f. Mikr. Anat.,’ iv, 1868.
8. Krerpert.—‘‘ Ueber die Haut d. Reptilien und anderer Wirbelthiere,”
* Arch. f. Mikr. Anat.,’ xiii, 1877.
4. StupEr.—“ Beit. z. Entwickelungsgeschichte d. Feder,’’ ‘ Zeit. f. wiss.
Zool.,’ xxx, 1878.
5. Davins.—‘‘ Die Entwick. der Feder und ihre Beziehung zu anderen
Integumentbilden,” ‘Morph. Jahrb.,’ xv, 1889.
6. Haacxr.— Ueber die Entstehung d. Saugethiere,”’ ‘ Biol. Centralbl.,’
vili, 1889.
7. Lzypic.—(a) “ Integument briinstiger Fische und Amphibien,” ‘ Biol.
Centralbl.,’ xii, 1892, p. 205.
8. Leypic.—(d) “Besteht ein Beziehungen zwischen Hautsinnesorgane
und Haaren ?” ‘Biol. Centralbl.,’ xiii, 1893, p. 359,
180 EDWARD B. POULTON.
9. Mavrer.—(a) “ Hautsinnesorgane, Feder- und Haare-anlage,” ‘Morph.
Jahrb.,’ xviii, 1892.
10. Mavrer.—(d) “ Zur Phylogenie der Saugethierehaare,” ibid., xx, 1893,
p. 260.
11. Mavrer.—(c) “ Zur Frage von d. Beziehungen d. Haare d. Saugethiere,”
ibid., p. 429.
12. Romer.—“ Ueber d. Bau und Entwick. d. Panzers d. Giirtelthiere,”’
‘Jen. Zeit.,’ xxvii, 1892.
13. WrBer.—(a) “Beit. z. Anat. und Entwick. d. Genus Manis,” ‘ Zool.
Ergebnisse einer Reisen in Niederl. Ostind., 1892.
14. Wesrr.—(4) “ Bemerkungen tiber d. Ursprung d. Haare und iiber
Schuppen bei Saugetieren,” ‘Anat. Anzeig.,’ viii, 1893, p. 413; and
translated in ‘ Annals and Mag. Nat. Hist.,’ July, 1893.
15. Emery.— Ueber die Verhaltnisse d.Saugethierehaare zu Schuppenartigen
Hautgebilden,” ‘ Anat. Anz.,’ vill, 18938, p. 731.
2. Conclusions derived from the Study of
Ornithorhynchus.
Referring to the interesting historical account written by Dr.
Benham, I will first endeavour to give reasons which seem to me
to oppose Gegenbaur’s and Kolliker’s distinction between hairs
on the one hand and feathers and scales on the other, founded
on the fact that the former are developed from the base of a solid
epithelial downgrowth, the latter from an epithelial upgrowth.
I have never regarded this distinction as a very important
one, and I believe that it may be successfully opposed even if
we had not the arguments which follow from the condition
here shown to obtain in Ornithorhynchus. The papilla and
epidermic bulb forming the hair are clearly parallel to the
papilla and epidermic cells over it forming the feather and
‘scale, the one projecting upwards at the bottom of a solid
downgrowth, the other projecting upwards from the free
surface. The only morphological distinction worthy of con-
sideration is the fact that the downgrowth is solid. But even
this difference is, in all probability, to be explained as a simple
result of the formation of follicle first and hair afterwards
as contrasted with feather first and follicle afterwards, while the
reversal of order is not in itself at all difficult to understand,
BILL AND HAIRS OF ORNITHORHYNCHUS PARADOXUS. 18]
If the cells of the bulb were to form the merest trace of a
hair at the surface, and were then carried down in the sinking
follicle, the hair rudiment descending in the axis would
necessarily leave a lumen behind it—an open tube connected
with the surface. If such were the case there would be no
morphological distinction between hair and feather, for the
follicular downgrowth would still be secondary in the hair as
it is in the feather, although preceding the development of
everything except the hair tip. And such a hair, developed,
except for its tip, from an epithelial upgrowth rising from the
bottom of an open tube connected with the surface, would still
be, from a morphological standpoint, at the surface, and would
be entirely and strictly homologous with a structure developed
from an epithelial upgrowth rising from the surface.
This will be admitted, and yet the transition from such a
condition to that now met with in Mammalia generally is so
easy that no important morphological distinction can be
founded on it. It is merely that the same tendency which has
- gradually increased the relative proportion of development at
the bottom of the tube carries this process one small step
further, so that the whole hair is formed beneath the
surface. There would then no longer be such a slight up-
growth from the bottom of the sinking tube as would ensure a
free lumen, and sooner or later a shorter simpler method,
leading to a solid downgrowth, would replace the open follicle,
But all this does not imply any great morphological distinction.
The case is, in fact, exactly parallel to the development of a
gland—which at an earlier stage was produced from a tubular
depression of the surface—from a solid downgrowth which sub-
sequently becomes tubular; and the chief principle at work
would probably be the same in both cases, viz. the tendency
towards abbreviation and simplification of development.
As a confirmation and a test of this argument it is important
to inquire whether the solid cylinder above the primitive bulb
forms any part of the hair itself, or whether it gives rise only
to the epidermic walls of the follicular tube (outer root-sheath)
which is afterwards formed. If the solid cylinder represents
182 EDWARD B. POULTON.
an abbreviated development of the epidermic tube, we should
expect its constituents to be histologically equivalent to those
of the tube, and of nothing else.
Different writers have taken different sides upon this question,
some accepting such an account as that summarised by F.
Balfour,! following Kolliker,’ that the hair develops as ‘‘a cor-
nification of the cells of the axis of one of the.... [down-
growing, solid, epidermic] processes, and is invested by a
sheath similarly formed from the more superficial epidermic
cells ;” and that given by Gegenbaur’—“ the shaft is differ-
entiated from the invaginated epidermis by the cornification of
its cells, while the other cellular parts of the follicle form the
root-sheaths.” The other point of view is concisely expressed
by Klein,t who describes the hair and inner root-sheath as
formed by the bulb alone, and pushed up the axis of the solid
cylinder of cells which connects the bulb with the superficial
epidermis, and who expressly states that “the cells of the
primary solid cylinder represent the rudiment of the cells of
the outer root-sheath only.”
The latter account is supported by the fact that, as soon as
the tubular follicle has been formed, it possesses no power of
development into a hair. This power is only present in the
epidermic upgrowth from the bottom of the tube, which repre-
sents, on the view expressed above, the epidermic upgrowth at
the surface from which the primitive hair was developed. The
walls of the tube, the outer root-sheath, have no such power,
but, as has been previously pointed out, can only form the far
less complex structure of the hair-base.
In the case of successional hairs some authorities consider
that the new bulb is formed from the outer root-sheath of the
old, while others regard it as formed from the old bulb. Even
on the former hypothesis there is no difficulty in looking upon
1 «Comparative Embryology,’ London, 1881, vol. ii, p. 328.
2 ¢« Rntwickelungsgeschichte d. Mensch, u. der hoheren Thiere,’ Leipzig,
1879.
3 *Comparative Anatomy,’ translation by Bell, London, 1878, p. 420.
4 © Atlas of Histology,’ London, 1880, p. 325.
BILL AND HAIRS OF ORNITHORHYNCHUS PARADOXUS. 188
it as formed from some cells of the old bulb which have not
been used up in forming the old hair, and which have followed
the shortening tube.
I therefore believe that, accepting Gegenbaur’s account of the
development of a hair, the most probable interpretation is still
to regard it as morphologically equivalent to a scale or feather.
When, however, we study this development in the very
direction from which most help is to be expected, and ascertain
the process which takes place in the most ancestral mammal,
there can surely be no doubt about the matter. For the solid
downgrowth of the higher mammals is replaced by an open
tube in Ornithorhynchus. To an evolutionist this fact cannot
fail to be the clear demonstration that the solid cylinder is the
abbreviated representative of the open tube. With this fact,
the significance of Gegenbaur’s distinction between feather and
hair falls to the ground. -
It is, indeed, by no means improbable that the first and
earliest trace of the hair is formed at the surface in Ornitho-
rhynchus, and subsequently sinks with the deepening tube.
Material by which this suggestion could be tested is unfortu-
nately wanting.
But the open tube is not the only, although it is the most,
important point about the development of the hair of Ornitho-
rhynchus. The great length of the papilla projecting through
‘the bulb into the lower part of the hair is also very significant,
suggesting a previous development like that of a scale or feather
from the surface of the epidermic covering of a papillary core
traversing the structure from base to apex.
Further confirmation is afforded by the axial rod of soft pro-
toplasmic cells forming the medulla of hair; for a shortening
papillary core, surrounded by cells of the rete mucosum super-
ficially undergoing cornification, would tend to leave just such
an indication of its former presence.
This interpretation is strongly confirmed by the beautiful
and detailed figures of A. Mertsching (‘ Beitrage zur Histologie
des Haares und Haarbalges,’ ‘ Archiv fur mikrosk. Anatomie,’
1888, p. 32). Thus in pl. iv, fig. 1, he shows in human hair,
184 EDWARD B. POULTON.
and in pl. v, fig. 4, still better in that of the guinea-pig, that
the papilla terminates superiorly in a long slender process,
which becomes actually continuous with the interspaces be-
tween the medullary cells, while the latter cells cover the lower
part of the process as a single layer and are represented as
continuous with the lowest layer of columnar cells covering
the papilla in the region of the bulb. Until we reach the
region where the medullary cells have shrunk and degene-
rated, the apex of the papilla is represented as forming an axial
tube in the base of the hair, and the medulla as its epithelial
covering in the form of a single layer corresponding to the
lowest layer of the epidermis.
The gradual increase in relative length and diminution in
diameter brought about by the necessity for warmth, have im-
plied solidity of structure, while firmness of attachment has
demanded the follicular invagination. Development at the
bottom of the follicle, together with the changes in shape,
have led to the present condition in which the structure is
formed from the lower end instead of from the surface of a
cylinder, the axial traces of rete remaining as a vestige of the
past.
Turning to more recent work, we find that Maurer seeks to
establish a distinction between hairs and scales or feathers, in
the elongation of the lowest epidermic cells which precedes the
development of the former but not the latter. But he has
never made this observation upon a Monotreme, and until this
has been accomplished no very great weight can be attached to
his argument. Indeed, the results which I have obtained
afford grounds for the belief that the earliest stages of develop-
ment will approach those of scales and feathers far more closely
than is the case with the higher mammals.
Some of the features especially pointed to by Maurer as
characteristic of feathers as opposed to hairs, are equally cha-
racteristic of the hairs of Ornithorhynchus. Thus, both in
obliquity of direction and in distinction between an upper and
a lower surface, the large hairs of Ornithorhynchus resemble
feathers,
BILL AND HAIRS OF ORNITHORHYNCHUS PARADOXUS. 185
T entirely agree with Maurer that any explanation of the
hair which fails to account for the inner root-sheath is a failure ;
but I hope to show that a reasonable explanation is quite
possible on the hypothesis that hairs and feathers are homo-
logous. Indeed, when Maurer comes to criticise the parallel
between a hair and a feather sunken in a follicle, I wonder that
the homology between inner root-sheath and the appendicular
parts of a feather did not suggest itself to him. Had he seen
the strongly developed thick fibrous sheath of Ornithorhynchus,
I venture to think that he might have come to a different con-
clusion.
The true significance of any peculiar structure in the higher
members of a Class is to be best understood by the study of
those lower forms in which its undoubted homologue is far
more strongly developed. If we only knew of the hairs of the
higher mammals, there would be a great deal to be said for
Maurer’s theory. But it fails because it is framed to account
for the Mammalian hair as it is, and not to account for it as the
condition in Ornithorhynchus shows that it has been. Thus,
his explanation of the inner root-sheath may fairly account for
the structure as it is usually described, but it fails to account
for it in the lowest mammal. Again, he explains hair as a
radially symmetrical structure, but Ornithorhynchus shows that
it is primitively bilateral. The same objection may be raised
to his explanation of papilla and solid epidermic downgrowth;
the explanations do not apply to Ornithorhynchus.
As to his objection that mammals have nothing comparable
to the moulting of scales in reptiles, it may be replied that the
succession of hairs affords as close a parallel as the mechanical
conditions of the case admit. Replace the flattened scale
growing from its under surface by the solid slender cylindrical
hair growing from its lower end alone, and it becomes clear
that the flaking off of the corneous surface of the one can only
be paralleled by the loss of the whole corneous cylinder of the
other and its replacement by a fresh one. Furthermore, the
new large hair is, in Ornithorhynchus, far advanced in develop-
ment before the old one is shed. Hence there is exactly what
186 EDWARD B. POULTON.
Maurer denies to mammals—a periodical shedding of the
stratum corneum (hair) preceded by the formation of a new
stratum corneum (new hair) below. The succession of hairs is,
in fact, the one exception to the gradual wearing off of the
superficial corneal cells in mammals which is so important a
difference between them and reptiles.
Although quite accepting Max Weber’s explanation of the
scales of Manis, I should agree with Romer in thinking it
more probable that they are homoplastic rather than homogenic
with reptilian scales. We must remember that the tongue of
mammals possesses in its papillz structures quite as comparable,
except as to situation, with the scales of reptiles as are those of
Manis; while protection, attrition, or any other advantage which
can be secured by a papilliform upgrowth from the surface, is
obtained so simply that it seems unnecessary to explain it by
genetic relationship.
My objection chiefly turns, however, on the conclusions forced
upon me by the study of Ornithorhynchus; for if hairs are
modified scales and homologues of feathers, it is very unlikely
that the scales of Manis can be the same. Romer and Emery
both show hairs developing on the scales of Dasypus, while
they also prove that these scales are homologous with those of
Manis.
Both feather and hair point to some ancestral scale which is
very different from that now found in reptiles and in Manis.
I believe that some of the features of this primitive structure
can be reconstructed with a fair degree of probability.
Inasmuch as I have argued that the hair of Ornithorhynchus,
developed in an open pit, was in some ancestral phase formed
upon the surface, it is legitimate to regard it as a superficially
placed structure, and then to inquire whether new homologies
or discrepancies between it and feathers or scales, are thus
revealed.
Woodcut fig. 1 is a diagrammatic rendering of a longitudinal
section of the developing shield of a large hair in Ornithorhyn-
chus, while woodcut fig. 2 represents its appearance if it were
formed at the surface instead of ina pit. The wall of the pit,
BILL AND HAIRS OF ORNITHORHYNCHUS PARADOXUS. 187
the outer root-sheath, has of course disappeared, while the
inner root-sheath forms a thick fibrous network round the base
of the hair. No longer confined in the narrow space between
hair and outer sheath, its fibres would be looser and their free
ends would curl up.
Giovanni has recently (“ De la regeneration des poils aprés
Vépilation,” ‘Archiv f. mikr. Anat.,’ 1890, p. 528) published
most beautiful illustrations of the earliest phases and growth of
the successional hair in man. He shows in pl. xxv, figs. 138,
19, 25, that the tip of the inner root-sheath is the first product
of the bulb, and that the hair itself is formed later and pierces
it (pl. xxvii, figs. 1,6,11). ‘The appearance suggests a possible
homology of inner root-sheath with the sheath of a developing
feather, but the specialised character and great size of the
former in Ornithorhynchus favour rather the homology with
the appendicular parts. The growth of the inner root-sheath
round the base of the hair is to be explained by the shortening
of the cone round which both the primitive hair and sheath are
here supposed to have developed. The final result of such
shortening leaves the hair to develop in the axis and the sheath
188 EDWARD B. POULTON.
round its periphery. Giovauni’s account is important in this
relation, proving that the inner root-sheath is at first formed
over the apex of the future hair, and only assumes its normal
relationship when the hair has grown through it.
We have in woodcut fig. 2 the representation not only of
the expanded tips of the larger hairs of Ornithorhynchus as
they would be if they grew at the surface, but of flattened
Ihe, Ye
¢ py) ,
= ZA { Epidermis
——— ——— ——t
imbricated protective scales with a rich fibrous growth round
their bases and sides, a growth which in the form of a thick
continuous felted under-coat would be of the greatest assistance
in maintaining a constant temperature.
By far the most bird-like structures in Archeopteryx, as we
know it, are its feathers. While its skeleton is profoundly
modified from that of the typical bird, its feathers remain
entirely typical.
Among the most Mammalian structures in Ornithorhynchus
are its hairs, and although they differ in some important
respects from those of other mammals, their divergence is
small compared with that of development, ovary, skeleton, &c.
This extraordinary persistence of the epidermic characters of
the Class when other Class characters are failing, suggests most
strongly a persistence altogether beyond the limits of the Class,
BILL AND HAIRS OF ORNITHORHYNCHUS PARADOXUS. 189
We are thus led to believe that both feathers and hairs were,
so far as their essential structure is concerned, existent in the
Reptilian ancestors of birds and mammals.
Now in woodcut fig. 2 we have the representation of a scale
which contains everything essential to the structure of both
hair and feather—equally capable of attaining, and attaining
by no very important changes, the simplicity of the one and
the complexity of the other.
On this hypothesis the hair represents the axial, its inner
root-sheath the appendicular part of a feather; and thus an in-
telligible morphological significance is given to the mysterious
inner root-sheath—a true part of the hair itself, and with it
arising from the bulb—but which, owing to the mode of
development, is buried deeply beneath the surface.
It is, indeed, possible that the existence of such a thick
under-coat may have conferred upon the ancestors of homo-
thermic mammals and birds the power of becoming themselves
homothermic. ‘
In birds the non-conducting coat is still supplied by the
appendicular parts of feathers, while in mammals the homolo-
gous structure is not available, being invaginated into the fol-
licle as the inner root-sheath. Under these circumstances a
non-conducting coat has been supplied by a fresh formation of
fine hairs. The transition was probably very gradual, and the
latter are to be regarded as much later products than the large
hairs. It is in harmony with these supposed changes that the
small hairs of Ornithorhynchus depart in so many respects
from the ancestral characters of the large hairs, and that they
should be entirely wanting from those parts where the large
hairs are, except as regards the medulla, the most scale-like, and
possess the thickest inner root-sheath, and where the necessity
for warmth is less imperative.
The invaginated inner root-sheath has, I believe, an important
function in retaining the hair in its follicle. The hair itself,
growing from the soft cells of the bulb, cannot be fixed very
firmly, but the inner root-sheath pressed tightly between the
hair and the outer sheath, and with its innermost cells imbri-
190 EDWARD B. POULTON.
cated downwards and interlocking with the cuticular cells of
the hair, which are imbricated upwards, gives to the hair a
swollen base which prevents it from being drawn with ease
through the narrow neck of the follicle. If it be forcibly torn
out, the inner root-sheath accompanies it. When the hair is
ultimately shed, its attachment has risen to a higher level in
the wall of the outer root-sheath, so that it is no longer fixed
in the manner described above.
The suggestions and hypotheses here put forward are not the
result of a hasty consideration of the subject, but were delibe-
rately adopted years ago as the best explanation of the new
facts brought out by the investigation, and their relation to
old facts which were imperfectly understood.
It may be fairly claimed for the views here expressed that
they suggest a simple and probable morphological explanation
of every structure in the hair or associated with it. The dermal
papilla, epidermic bulb, medulla, hair-shaft, and inner root-
sheath, all follow naturally from the invagination of a proto-
Mammalian scale like that diagrammatically represented in
woodcut fig. 2, while the outer root-sheath is clearly the walls
of the pit into which the invagination took place. The further
fact that fig. 2 is, except for the want of invagination, essen-
tially similar to the hair of the young of the lowest mammal,
lends additional support to the hypothesis here suggested.
if, however, the interpretations offered in this and{other parts
of the present paper be dismissed by the results of further in-
vestigation upon better material, the main facts upon which
the interpretations rest will always remain, and, yielding as
they do a considerable store of fresh knowledge about this
most interesting of all mammals, must always have a value.
BILL AND HAIRS OF ORNITHORHYNCHUS PARADOXUS. 191
EXPLANATION OF PLATES 14, 15, & 15a,
Illustrating Professor E. B. Poulton’s paper on ‘‘ The Struc-
ture of the Bill and Hairs of Ornithorhynchus para-
doxus, with a Discussion of the Homologies and Origin
of Mammalian Hair.”
PLATE 14.
The figures in this Plate illustrate the structure of the bill. The structure
of the epidermis is indicated in Figs. 1 and 8. Figs. 1—7 deal with the
tactile organs of the bill, Figs. 8—16 with the glands and their associated
structures.
Figs. 4, 8, 9, 10, 11, and 12 were drawn from sections stained with car-
mine; while Figs. 3, 5, 6, 7, 18, 14, 15, and 16 were from sections stained
with logwood. Fig. 1 was from a section stained with logwood and picric
acid, Fig. 2 from one stained with carmine and picric acid. Cornified epi-
dermic cells are indicated by a yellow colour.
Fie. 1.— x nearly 200 diameters. Vertical section through the surface of
the bill (probably near the edge of the superior surface of the upper bill),
showing a complete push-rod and parts of two others. The convex upper
end of each rod, surrounded by a slight ridge, is clearly seen. Each rod is
seen to consist, for the upper fourth of its length, of extremely thin curved
epidermal plates, retaining traces of nuclei surrounded by pigment (compare
Fig. 2). The corneous layer is not nearly so thick in the push-rod as in the
general epidermis of the bill. Below the upper fourth of their length the
rods are formed of imbricated cells with distinct nuclei, the layers being better
shown in more highly magnified figures (Figs. 3 and 4). The same applies
to the axial group of longitudinal filaments and the circle of similar struc-
tures shown in longitudinal section on each side of the group, and separated
from it by a row of imbricated cells. The filaments are seen to diverge
slightly at the lower end of the rod. Pigment is chiefly present at the
sides of the rods, and especially among the lowest cell-layers forming the
extreme base; but it also exists in the inner layers, as is shown in Fig. 4.
The lower part of the rod is free from the general epidermis of the bill,
being separated from it by a tubular dermal upgrowth, the upper part of
which sends up papillary processes. Hence the dermis seen on each side of
VOL, 36, PART 2,.—NEW SER. 0)
192 EDWARD B. POULTON.
the rods in Fig. 1 is partly a longitudinal section of the tube and partly that
of the papillary processes (as seen in transverse section in Fig. 3). The
lower end of the complete rod is seen to rest on three touch-bodies, of which
two are in longitudinal and one in transverse section. Medullated nerve-
fibres approach the touch-bodies, and a fibre is shown entering the base of one
of them.
Fic. 2.— x rather over 400 diameters. Vertical section through the
upper part of a push-rod and the adjacent corneous epidermis of the bill.
From the same locality as Fig. 1. The thin flattened cell-plates form nearly
hemispherical shells which, placed one over the other, build up this part of
the structure. The form and arrangement suggest a structure which would
easily yield to pressure, but would be instantly restored to the former level on
its withdrawal.
Fic. 3.— x rather over 400 diameters. Horizontal section of the epidermis
of the bill (locality unnoted, but appearance typical) taken rather below the
corneal layer. The push-rod in transverse section is clearly seen to consist
of the following layers, the epithelial cells being represented by their nuclei.
(1) Centrally, the axial group of filaments: these are very highly refringent,
and stain with great difficulty; they are probably partially or completely
corneous. (2) A single layer of deeply staining epithelial cells. (3) The
circle of filaments, similar to those of the axial group; they are arranged with
great regularity. (4) A single layer of rather deeply staining epithelial cells.
(5) A single layer of very faintly staining epithelial cells. (6) The outermost
epithelial circle, seen by comparison with Fig. 4, to consist of several layers
of cells which stain deeply. Outside this are the deeply staining fusiform cells
of the general epidermis. Three papillary outgrowths are seen in section.
I'ie. 4.— x over 600 diameters. Longitudinal section of a push-rod, with
the transverse section diagrammatically rendered. From nearly the same
locality as Fig. 1, but, in this case, from the very margin of the bill. The
layers described in the last figure are here indicated by the corresponding
numbers. The appearance represents the details as they can be made out
with high powers and the most favorable sections. The filaments are seen
to have a ladder-like appearance, apparently due to their constitution out of
short sections, the length of which corresponds to the thickness of the cells
of layer 2. The appearance suggests that each section may be contributed
by the inner end of the adjacent cell. The same is true of the filaments of
the circle (3) as regards the cells of layer 4, although in some cases the
sections appeared rather to correspond with the outer ends of the cells of
layer 2. The correspondence between the sections and adjacent cells is not
well represented in the figure. The flattened imbricated cells of layers 2 and
4 are seen to be swollen at the spot where the nucleus is situated, where
also there is a special accumulation of pigment. In transverse section of the
rod, the cells of all the layers have the form of flattened curved plates, and
BILL AND HAIRS OF ORNITHORHYNOCHUS PARADOXUS. 1938
not the radial form indicated in the diagrammatic transverse section. They
are, furthermore, far less numerous than in the latter, each plate curving
round the rod for a considerable distance, as is indicated by the arrangement
of the nuclei in Fig. 3.
Fie. 5.— xX about 50 diameters. Horizontal section of one of the highly
sensitive transverse ridges on the outer side of the upper surface of the lower
bill (inside the mouth). The push-rods seen in transverse section are ex-
tremely numerous. Nearly all the layers shown in Figs. 3 and 4 can be made
out, even with this low power. Fifteen gland-ducts are seen in transverse
section, scattered between the twenty push-rods. In other parts these pro-
portions are generally reversed. Many of the gland-ducts are surrounded by
pigment. The right-hand upper push-rod has a distinct papillary upgrowth
on each side of it. “Higher powers show that two or more such papillz are
usually arranged round the other rods.
Fic. 6.— x about 50 diameters. A deeper horizontal section, nearer the
middle of the length of the same ridge as that which supplied the material
for Fig. 5. The section passes just below the lower ends of the push-rods,
and shows the numerous groups of Pacinian bodies (from two to five ina
group), each of which is placed under a rod (compare Fig. 1). Twenty-four
groups are represented, together with nineteen gland-ducts, the proportion
being about the same as in the last figure. When only a single Pacinian body
is seen, it is probable that the rest of the group was not included in the section.
Many of the gland-ducts are surrounded by pigment. Between the struc-
tures described is the dermis of the bill, which is concentrated to form a
fibrous investment to both ducts and groups of nerve end-organs.
Fic. 7.— xX rather over 400 diameters. A single group of Pacinian bodies
from one of the ridges of the lower bill. The figure shows the essential
similarity of these nerve end-organs to those previously described in the
tongue of the same animal (‘ Quart. Journ. Micr. Sci.,’ July, 1883).
Fic. 8.— X nearly 200 diameters. Vertical section through the surface of
the upper bill, probably from the same locality as that described in Fig, 1,
including a longitudinal section of a gland-duct on its way to the surface,
together with the remarkable structures associated with it. The duct passes
along the axis of an epithelial cylinder, which has many points of resemblance
to a hair, shortened at both ends, so that it is flush with the general epidermic
surface above, while its bulb is but little below the lowest layer of the epi-
dermis. In favorable examples, however, as in that selected for figuring,
the upper corneous end of the cylinder projects a little, and is surrounded by
a distinct trench. Below the corneous stratum of the general epidermis the
cylinder is surrounded by an epithelial layer (0. 7. s.), which presents many
points of resemblance to an outer root-sheath, from which the cylinder is
seen to be in part free, a distinct space between the two being visible on both
sides a little below the level A and on one side at the deeper level C. Below,
194 EDWARD B. POULTON.
this layer is continuous with the outer and lower part of a large mass of
epithelium bearing strong resemblance toa hair-bulb, but separated from the
rest of the mass by a space (sp.) containing small branched cells. For the
proper understanding of this part of the structure better material is much to
be desired, for it is probable that considerable post-mortem changes had
occurred. Within the supposed outer root-sheath the cylinder (the supposed
hair) is surrounded by flattened cells (c.), which suggest the cuticle or possibly
the inner root-sheath or both of these. This layer also arises from the sup-
posed bulb. The central part of the cylinder consists below of polyhedral
cells like those of the bulb, passing above into corneous fusiform scales with
a longitudinal arrangement. The gland-duct (d.) passes into the lower end
of the bulb, remaining distinct from the epithelium of the latter. In this
lower part of the bulb the cells stain more deeply than those of the bulb,
and their outlines are less distinct (compare Fig. 11). The duct is at first
straight, but subsequently pursues a twisted course in the bulb, generally to
a more marked extent than in the example figured. In the upper part of
the bulb the duct becomes surrounded by a single layer of cubical cells, still
staining rather more deeply than those external to them, but with distinct
outlines. Above the point where the cylinder becomes corneous the duct
appears to traverse its axis without any proper investment, the walls being
extremely irregular and the lumen of a radiate star-like form in transverse
section (compare Figs. 9 and 10). The whole arrangement suggests that the
duct may have penetrated the bulb of a shortened and rudimentary hair by
way of the papilla, and found its way along the medulla. Immediately below
the bulb the duct is surrounded by a ganglion (gz.) containing large nerve-
cells and medullated nerve-fibres, in which nodes of Ranvier are commonly
noticed. Below this a large medullated nerve (z.) and the gland-duct (d’.) are
seen in transverse section, while a curved gland-tube (g/d.) is shown to the
left. The structure is evidently that of a Mammalian sweat-gland, a single
layer of cubical gland-cells being surrounded by delicate fusiform smooth
muscle-cells longitudinally arranged.
The following four sections represent horizontal transverse sections at
successively deeper levels of a gland-duct like that shown in Fig. 8, together
with the associated structures. All probably come from a very similar locality
on the upper bill. All are magnified rather over 400 diameters.
Fic. 9.—Transverse section of the duct at about the level marked A in
Fig. 8. Six papillary upgrowths mark the line of separation between the
epithelium round the duct and the general epidermis of the bill. The corni-
fication is more extensive than in Fig. 8.
Fic. 10.—Transverse section of the duct at about the level marked B in
Fig. 8. The lumen appears stellate in this and the previous figure. Of the
six papille seen at a higher level, three have coalesced, while three remain
distinct. Just below this level all six pass into the continuous dermal sheath
BILL AND HAIRS OF ORNITHORHYNCHUS PARADOXUS. 195
which surrounds the lower part of the epithelial cylinder traversed by the
duct. The cells immediately round the duct tend to split away from those
placed more peripherally, suggesting the separation of a hair from its outer
root-sheath.
Fic. 11.—Transverse section of the duct at about the level marked C in
Fig. 8. This is the region of the bulb of the supposed hair. The difference
between the epithelium of the duct (d.) and that of the supposed bulb is
clearly shown, although the duct resembles that which is seen at a rather
lower level than C in Fig. 8. The deeper staining of the duct (d.), and
especially of its innermost layer, is very characteristic in these sections. The
supposed cuticle (c.) and outer root-sheath (0. 7. s.) are clearly seen.
Fic. 12.—Transverse section of the duct at about the level marked D in
Fig. 8. The duct is surrounded by the ganglion containing many ganglion
cells. One of these is seen to give off a long process; a medullated nerve-
fibre with a node of Ranvier is also present. Both the duct and the ganglion
are surrounded by a fibrous sheath.
The following three figures represent various forms of the duct; all are
magnified about 50 diameters.
¥1G. 13.—The coiling of the duct in the lower end of the epithelial cylinder
(the supposed hair) is well seen. The resemblance of the upper end of the
cylinder to a truncated hair is especially distinct in this and Fig. 15. An S-like
curve of cylinder and duct in the upper layers of the epidermis is shown in
this and Fig. 15. It is a very common appearance.
Fic. 14.—A very common form at the posterior part of the upper surface
of the upper bill.
Fic. 15.—A form in which the epithelial cylinder is very slightly developed.
It occurs intermixed with those of the commoner type (Figs. 8 and 13).
Fic. 16.— x nearly 200 diameters. This figure shows the form of two
ducts, &c., in the bill of a young Ornithorhynchus, 8°3 centimetres long in
the curled-up attitude. The ‘ bulbs” at the lower end of the cylinders are
not only extremely well developed, but are more deeply placed than in the
mature animal. Just below the level of the bulbs, large medullated nerves
(diagrammatically rendered in the figure) extend in a horizonta layer, uniting
to form ganglia round the bases of the bulbs themselves. It is very un-
fortunate that histological details could not be made out.
196 EDWARD B. POULTON.
PLATE 15.
The figures in this plate illustrate the structure of the hairs and the struc-
tures associated with them. Figs. 17, 18, and 25 represent structures in the
mature animal; the remaining figures represent those of the young animal
83 centimetres long, in which the large hairs alone had appeared above
the surface of the skin.
Fie. 17.— x nearly 200 diameters. Transverse section of a group of hairs
from the middle of the back. The section is taken just below the epidermis.
Groups like this apparently make up the fur which clothes the body of the
animal, although modifications occur on the tail and limbs. The constituents
of the group pierce the subepithelial tissue obliquely, and overlap each other
on emergence. ‘Those in the lower part of the figure are overlapped by those
in its upper part, which are situated anteriorly to them. Most anteriorly of
all is usually found the duct (d.) of a gland resembling the Mammalian sweat-
gland. It and the more deeply-placed gland-tube are found along the anterior
margin of the outer root-sheath (0. 7. s.) of the large hair, or between it and
one of the two anterior bundles of small hairs (Fig. 18, g.). Behind the duct
is the most important member of the group—the large hair contained in its
outer root-sheath (0. 7. s.). These hairs succeed each other very rapidly and
always in an antero-posterior direction, so that the great majority of sections
show two of them associated with one outer root-sheath. The shield-like end
of the developing large hair (. 4.) is seen anteriorly in the lumen of the follicle,
while behind it is the attenuated unpigmented basal part of a fully formed
hair (/. 2’.) Behind, on either side, are two bundles of small hairs, each
bundle being enclosed in a single epithelial ring. The majority of sections in
each bundle are those of the attenuated pigmentless bases of fully formed
hairs, while the minority are those of various parts of the shaft of hairs which
are still growing from a bulb, and in these pigment is seen to be present.
Fie, 18.— x nearly 200 diameters. A transverse section through a similar
group from the same locality taken at a somewhat lower level. The duct has
now passed into a gland-tube (g.) in which peripheral longitudinal smooth
muscle-cells and central gland-cells are seen. The expanded end of the young
large hair (¢. 4.) is here cut nearer to the bulb, and is only pigmented in its
lower part. The medulla is distinct as a group of cells which still remain
protoplasmic and stain deeply. After emerging at the surface these soft cells
dry up, and give rise to the ordinary hair medulla with its enclosed air-bubbles.
The cuticle 4o which the reference letters (/. 2.) point is seen to be far thicker
on the exposed anterior surface than elsewhere. Comparison with Fig. 17
shows that nearer the tip of the hair it possesses a uniform thickness, The
base of the mature large hair (/’. #’.) is, at this deeper level, enclosed in the
posterior wall of the outer root-sheath, and is irregular in outline. A little
below this point it terminates in an irregular rounded extremity, sometimes
co
BILL AND HAIRS OF ORNITHORHYNCHUS PARADOXUS. 197
sending out processes which pass between the adjacent epithelial cells. The
younger hair, on the other hand, passes down to a great depth and ends in
a bulb. The small hairs in the four bundles are seen to have separated, each .
being enclosed in an epithelial tube corresponding to an outer root-sheath, but
the walls of the tubes are at this level continuous peripherally in each bundle.
Below, the pigmented younger hairs are continued to a far deeper level, their
outer root-sheaths expanding into bulbs, while the unpigmented bases of fully-
formed hairs terminate without bulbs very slightly bélow this horizon. Indeed
two of them (one in each of the right-hand bundles) are already terminating,
and have ceased to be corneous.
Fic. 19.—x rather over 400 diameters. Transverse section of one large
and many smaller hairs from the middle of the back of the young Ornithorhyn-
chus. ‘The sections are of the first-formed hairs, and as these are by no means
fully formed, only the tips of the large ones having appeared above the skin,
no traces of the successional hairs are to be seen. The upper part of the large
hair in the figure represents its anterior and upper side on emergence, and it
is seen that the cuticle (c.) is here thicker, while the pigment is contained in
the hair-cells which are undermost in the natural position. The medulla is repre-
sented in section by four protoplasmic cells, and the nuclei of the hair-cells
between this and the thickest cuticle stained with logwood. Outside the
cuticle is a layer (¢. 7. s.), which is very characteristic of Ornithorhynchus,
surrounding the hair above the neck of the bulb for a considerable ‘part of its
passage through the dermal tissues. It consists of anastomosing, apparently
corneous fibres, representing the inner root-sheath to which the homogeneous
deeply staining layer outside it probably also belongs. The outer root-sheath
(o. vr. s.) is at this horizon extremely delicate and thin, and is succeeded by
alymph space. The entire sections of four small hairs are seen, as well as
parts of four others. Four small hairs are at this age usually found beneath
a large one, and each of them appears to represent one of the four groups of
small hairs found beneath each large one in the mature animal. The four
entire sections probably represent those which attend the-large hair in the
section. The two inner ones are cut through at tie level of their bulbs, and
centrally the papille are seen in section. The two outer indicate the struc-
ture at a higher level as seen in section, the shaft being composed of six
radially arranged pigmented hair-cells surrounded by a cuticle followed by a
thick inner and irregular outer root-sheath and lymph space. It must be
borne in mind that the delicate tissues of the young Ornithorhynchus had not
been prepared for histological investigation, and that it was therefore extremely
difficult to form-an opinion upon some of the layers. This caution applies
especially to the layers interpreted as outer root-sheath and the outer part of
inner root-sheath in Figs. 19 and 20 and the former in Fig. 23.
Fic. 20.— xX rather over 400 diameters. A section of one of the large hairs
from the same part of the same animal as Fig. 19, taken at a rather deeper level,
198 EDWARD B. POULTON.
The reference letters correspond to those of Fig. 19. In the centre of the
hair-cells is a cavity in which the papilla was doubtless contained, for in
Ornithorhynchus the papillee extend above the bulb into the ueck (compare
Fig. 23). The inner root-sheath (7. 7. s.) is apparently divided into two layers ;
which, however, are less sharply marked off from each other than in Fig. 23.
The cellular nature of the inner part is apparent. The upper surface of the
hair is rendered very distinct by its immensely thickened cuticle and by the
absence of pigment.
Fic. 21.— x rather over 400 diameters. Transverse section of large hair in
tail of young Ornithorhynchus. The pigmented hair-cells constitute the
thickness of the shaft, and no medulla is visible. A superficial pigmentless
layer is probably the-cuticle. The inner root-sheath (7.7. s.) is exceedingly
thick and made up of fibres which are polyhedral and irregular in section, and
its innermost part appears to be condensed into a continuous layer imme-
diately surrounding the shaft of the hair. The outer root-sheath (o. 7. s.) is
very distinct. A lymph space separates the last layer from the hair-sac.
Fic. 22.—x rather over 400 diameters. A transverse section of a rather
small hair from the tail of the same animal, taken at a somewhat deeper level
immediately above the bulb. The inner root-sheath is here divided into two
well-marked layers, the inner of which is not corneous and stains very deeply
in logwood. Ata still deeper level this would be true of the whole sheath.
In other respects the sections are essentially similar.
Fic. 23.—x nearly 200 diameters. A longitudinal vertical section through
the bulb and neck of a large hair from the middle of the back of the same
animal. The immense size and upward extension of the papilla is the most
remarkable feature in the section. The tip, to which the reference letter p.
points, is well above the bulb of the hair, and has reached a horizon at which
the hair-cells have become distinctly fusiform. The papilla itself contains
capillaries and branched cells. The upper surface of the hair is on the left
side, and is at once distinguished by its thick cuticle (c.) and pigmentless
cells. The root-sheaths are thin, and their structure indistinct. The section
being exactly along the middle line of the hair there was nothing to keep the
two halves together, consequently they have fallen outwards so that the struc-
ture appears to be wider above the papilla. If this had not happened it would
have been much narrower than the bulb, and the chink in which the letter p.
is placed would have been obliterated, although in estimating the diameter we
must allow for the medullary cells which have been accidentally removed in
the manipulation.
Fic. 24.— x nearly 200 diameters. A longitudinal section of a large hair
on the head of the same animal. The section was somewhat oblique, passing
through the middle line of the hair at the point where the latter is represented
in diagrammatic transverse section. At this point and just below it the
deeply staining medullary cells are seen surrounded by the pigmented hair-
cells, and these by the cuticle (¢c.). The inner root-sheath (¢. 7. s.) is seen
BILL AND HAIRS OF ORNITHORHYNCHUS PARADOXUS. 199
from within as well as in longitudinal section, and is clearly shown to consist
of a network of corneous fibres enclosing longitudinal fusiform spaces through
which the outer root-sheath (0. 7. s.) is seen. At the upper end the section
becomes more tangential, and hence the more obliquely cut edge of the outer
root-sheath gives an erroneous impression of greater thickness at this point.
Fie. 25.— x nearly 200 diameters. A gland tube and duct opening into the
bay between the flange-like posterior extension of the upper bill and the hairy
surface of the head. The secretory part of the gland is seen to possess the
structure typical of the Mammalian sweat-gland. The duct is, however,
unusually wide. In this position the glands were peculiar in opening on the
surface independently of hairs or hair-like epidermic processes.
PLATE 15a.
The young Ornithorhynchus paradoxus of which the hairs are de-
scribed in this paper, and are represented in Plate 15, figs. 19—24. The
animal is represented of the natural size, as seen from the left side in Fig. 1,
from the ventral aspect in Fig. 2. The skin appeared to be quite smooth and
bare, but closer examination showed that the tips of the large hairs had
emerged from the surface, the small ones being still concealed beneath it.
Hab TAG el eel eles
*
eV eCae Y {% . v2 :
OTe iar eo: A hse ms ©
of (Worle ae eae ieee 1 alt
he Be oy plik, aS Tyo
_ : als ry
ft deep hte tag ae
‘) Shel ea vay ini) 125 ue
Uy bial oda Aes
pure 9 OLS 1) iy Vales (yeh ink
ai fni MA iy’ LPosrea a i
a ol ee
shitty he:
OLIGOCHATA OF TROPICAL EASTERN AFRICA. 201
A Contribution to our Knowledge of the Oligo-
cheta of Tropical Eastern Africa.
By
Frank E. Beddard, M.A., F.R.S.,
Prosector to the Zoological Society of London.
With Plates 16 and 17.
TABLE oF CONTENTS.
(1) Introductory, p. 201. (4) Note on the Substitution of Or-
gans as illustrated in the Eu-
(2) Descriptions of new Species,
drilide, p. 263.
Das __ __ | (6) Classification of the Eudrilide,
(3) The Calciferous Glands in Eudri- p. 264.
lidee, p. 255. (6) Descriptions of the Plates, p. 268.
1, Introductory.
Tuere seems to be no doubt that at present tropical Africa
furnishes the most remarkable and interesting representatives
of the terrestrial Oligocheta. The family Eudrilide have
their headquarters there; indeed, with the exception of the
almost cosmopolitan genus Eudrilus, the family is confined
to the Ethiopian region, not even extending, so far as we know
at present, into the more northern part of the continent.
Eudrilide are already known from both the west and from
the east side of Africa; they appear to abound principally in
the equatorial region, though by no means unknown from
202 FRANK E. BEDDARD.
more southerly districts. The forms from the east have been
mainly described by Dr. Michaelsen!; those from the western
side of the continent by myself.? It has been shown by these
investigations that, as a rule, the east and west of tropical
Africa are inhabited by different genera, and always by differ-
ent species. It also appears that, on the whole, the number of
worms belonging to this family is greater on the east coast
than on the west.
The peculiar interest attaching to this group of Oligocheta,
independently of their distribution, concerns the structure of
the reproductive organs; in most of the members of the group
there are no spermatothece homologous with those of other
Oligocheta ; the place of these organs is taken by cclomic
sacs which acquire an opening to the exterior; rarely are there
true spermatothecz in addition to these coelomic sacs ; when
such spermatothece are present they are partially or entirely
enclosed by the sacs in question. In the more specialised
members of the family the ovaries are also enclosed in sacs
which communicate with the egg sacs and sometimes also with
the spermatothecal sac. The Eudrilide, in fact, are an alto-
gether remarkable family of Oligocheta, and it is desirable
that our knowledge of them should be perfected. I am there-
fore glad to be able to contribute towards that end by the
following account of a number of new forms collected in
Zanzibar and Mombasa. I was enabled to acquire this mate-
rial through the liberality of the Government Grant Committee
of the Royal Society, who awarded me £100, with which I
paid the costs of a collector; I was so fortunate as to secure
the assistance of Mr. Frank Finn, to whom I am indebted for
the careful way in which he carried out my instructions in the
matter of preservation, &c.
1 « Beschreibung der von Herrn Dr. F. Stuhlmann im Miirdungs gebict
des Sambesi gesammelten Terricolen,” ‘J. B. Hamb. wiss. Anat.,’ vii.
‘Beschreibung der von Herrn Dr. F. Stuhlmann auf Sansibar und dem
gegeniiberliegenden Festlande gesammelten Terricolen,” ibid., ix.
7 “On the Structure of Two New Genera of Harthworms, &c.,” ‘ Quart.
Journ. Mier. Sci.,’ vol. xxxii. ‘‘On the Structure of an Karthworm allied to
Nemertodrilus, &c.,’’ ibid.
OLIGOCHATA OF TROPICAL EASTERN AFRICA. 203
I received from him a very large number of Oligocheta
belonging to thirteen species at least; of these I describe eight
in the present paper as new. The eight new species are the
following :
Eudriloides Cotterilli, n. sp.
Eudriloides brunneus, n. sp.
Polytoreutus violaceus, n. sp.
Polytoreutus Finni, n. sp.
Polytoreutus kilindinensis, n. sp.
Pareudrilus stagnalis, n. gen., n. sp.
Gordiodrilus zanzibaricus, n. sp.
Alluroides Pordagei, n. sp.
In addition to these there were numerous examples in the
collection of species apparently identical with Michaelsen’s
Stuhlmannia variabilis. This species is indeed to all
appearance the most abundant form of Eudrilid in the regions
visited by Mr. Finn. Besides the specimens preserved by Mr.
Finn, he brought me a large number of living Oligocheta ;
the bulk of these were of this species.
Another species abundant in the gatherings was a Ben-
hamia of small size, which I have not minutely studied as it
showed no noteworthy differences from the smaller species
described by Michaelsen.
In the mud from swamps brought home there was an abun-
dant supply of a Dero with two long processes in addition to
the four “ gills;” this species seems to be identical with that
named Dero Miilleri by Bousfield ; in the same mud I found
a Nais and an Enchytreid, neither of which have I identified ;
these two species, however, were immature, and they did not
exhibit any characters of particular interest.
The worms were all of them, with the exception of the species
of Kudriloides, the Benhamia, and the Polytoreutus,
found in, or at the margin of swamps. The aquatic character
of these Eudrilids is perhaps to be noted in connection with
the total absence of dorsal pores, a character already known to
distinguish the Eudrilide from the majority of “ earthworms.”
The other forms were collected in soil outside a bungalow,
204 FRANK -E. BEDDARD.
which was kept continually moist with the “slops” of the
household. In no other situations were any Oligocheta to
be found even after or during rain. It seems probable that in
the dry season they retire deep within the ground or take
refuge in swampy ground. As to the former suggestion, it is
only supported by the presence of the worms in moist ground,
for Mr. Finn informs me that he examined a deep trench in
the course of its digging, and did not come upon any traces
of these animals at all. It is possible that the comparative
rarity of terrestrial Oligochzta in other tropical countries is
to be accounted for by the fact that they lurk in swamps, and
only come forth when the ground is thoroughly soaked, and
fit for them to traverse. On the west coast of Africa the
earthworms seem to be more purely terrestrial in habit. I
hear from Mr. Finn that I am indirectly indebted to the
kindness of General Mathews, principal minister of His High-
ness the Sultan of Zanzibar, to Mr. Pordage, to Mr. Macalister,
aud to Mr. Cotterill, for some of the material. To these
gentlemen I beg to tender my thanks.
2. Description of New Species.
Eudriloides Cotterilli, n. sp. (figs. 1, 15, 16, 18—20).
A considerable number of examples of this species were
collected by Mr. Cotterill outside a house at Kilindini. I
have studied the species partly by the section method, and partly
by dissection, and subsequent examination in glycerine.
The species is a small one—an inch or so when preserved.
The specimens were preserved with Perenyi’s fluid and had,
after preservation, hardly any definable colour, but the paired
nephridia were very conspicuous as white masses shining
through the transparent skin.
The prostomium of the worm is not large, nor does it
appear to be continued by grooves on to the buccal segment.
In longitudinal sections the prostomium seemed to be divided
by a transverse fissure, a character which is known to distin-
guish Phreoryctes but has not been met with elsewhere in
OLIGOCHATA OF TROPICAL EASTERN AFRICA. 205
this group. The setz begin on the 2nd segment; they are
strictly paired, and rather small in size, of the usual form,
without any ornamentation. The two ventral pairs are rather
further apart from each other than is either of them from the
lateral pair of its side, but all the sete are distinctly upon the
ventral surface of the body. They appear to be absent from
a few segments of the body. I did not find the ventral pair
on the 13th segment on either side of the body, nor were there
any traces to be discovered on any of the clitellar segments of
either ventral or lateral pairs. They re-commence just on a
level with the posterior boundary of the male papilla.
The clitellum is saddle-shaped. It occupies the Seg-
ments x1v—xvil. The median ventral area which is not part
of the clitellum is chiefly occupied by the prominent pores of
the reproductive organs. The whole area is hour-glass-shaped,
being narrower in the middle than at either of the two ends.
Independently of the very prominent papille which bear the
reproductive orifices, there is a single papilla upon the 11th
segment of an oval form. It lies on the posterior half of the
segment, just causing a slight convexity of the boundary line
between this segment and the 12th. The papilla les entirely
between the ventral sete. It ends on each side just on a level
with the inner of the two set of that pair. The transverse
diameter of the papilla is greater than its antero-posterior
diameter. The median spermatothecal pore lies on the
line between Segments x111/x1v. The actual orifice itself is
circular and not large, not so large as the male pore. It is
borne upon a very prominent swelling of the body-wall, which
may extend as far forwards as the posterior boundary of Seg-
ment x11. In some specimens the terminal section of the
spermatothecal sac is protruded from the aperture.
The oviducal pores are not conspicuous until the worm is
carefully examined with a tolerably high power of the micro-
scope; they are invisible under a low power; each aperture is
on the 14th segment; it lies in a position exactly correspond-
ing to the lateral setz.
The single median male pore is a Danspersely elongated
206 FRANK E. BEDDARD.
aperture on the boundary line between Segments xvi1/xVIII;
it is curved, the concavity being forwards; the two ends of the
orifice are wider than the middle part—like the mouth of an
Actinozoon ; out of this orifice protrude the penial sete; the
aperture is upon a conspicuous papilla, which is larger than
that which bears the spermatothecal pore; there are two small
supplementary papille at the posterior angles of the cushion.
The external characters distinguish this species from the
three others, viz. Eudriloides parvus, E. gypsatus, and
EK. titanotus.
In the first of these species there is a median papilla upon
the 19th segment, but none is mentioned by Michaelsen upon
the 11th. No papille are mentioned in the remaining two
species; Eudriloides Cotterilli has not, as have the other
species, a complete clitellum.
Vascular System.—tThere is at present no information as
to the vascular system of this genus. Michaelsen’s two papers,!
which deal with the only species of Eudriloides known
previously to the publication of this memoir, contain not a
syllable about the blood-vessels. My own notes upon the sub-
ject, though few, will therefore be of some use. The dorsal
vessel has a thick coating of peritoneal cells, and its walls are
of some thickness; in this it contrasts with the supra-intes-
tinal vessel, which, although of greater calibre, has thin
walls; the ventral vessel has also a covering of quite conspi-
cuous cells. In the 10th and 11th segments there are a pair
of perivisceral vessels which arise from the supra-intestinal
vessel alone. I have ascertained that they have no connection
at all with the dorsal vessel. These ‘“ hearts” have thick
muscular walls, which commence abruptly at their origin from
the supra-intestinal trunk; they are furnished along their
course with frequent valves, and the opening into the ventral
vessel is guarded with valves, as is also that into the supra-
intestinal; in front of these are a series of perivisceral trunks
of less calibre which arise from the dorsal vessel only, and
have no connection with the supra-intestinal. The dorsal
' Loe, cit. (on p. 201).
OLIGOCHHTA OF TROPICAL EASTERN AFRICA. 207
vessel has valves at the points where it perforates each
septum.
The alimentary canal has a gizzard in Segment v1; the
cesophagus has a number of pairs of peculiar glands, which
are separately described in connection with those of other
Eudrilids below. The intestine begins in Segment xv. The
cesophagus is very vascular.
The first septum separates Segments tv/v; it is very delicate;
behind this are five very thick septa, and then follow two others
which are somewhat thicker than those which follow them.
The nephridia commence in Segment 1v. They appear to
open in front of the ventral sete; in any case their duct was
traced into the body-wall on a level with the ventral sete; but
whether it expands into a plexus like that of Libyodrilus, or
opens at once on to the exterior, I am unable to say; the
nephridial duct is not a conspicuous sac, as in some other
Eudrilide, but a tube having an intra-cellular duct.
Reproductive organs.—This species has, as have the
other species of the genus, only a single pair of testes; as is
customary when there is but a single pair of these gonads, they
lie in the 11th segment. Lach testis is attached to the front
wall of Segment x1; its shape is a little unusual. The testis is
a curved rod of narrow dimensions, only a little broader than
elsewhere near to its origin from the septum; if it were to lie
in a perfectly straight line in its segment, as the testes usually
do, it would reach as far as the funnels of the sperm ducts ; it
therefore has to be coiled in order to get it out of the way.
The cells of which the testis are composed are not very distinct;
the nuclei, however, are; these nuclei are larger and clearer at
the base of the gonad.
The sperm-sacs lie in Segment x11; they depend from the
anterior septum of this segment; they are fused together at
their origin, and the question therefore arises as to whether
we are to consider that there are two sacs or only a single
bifid sac. The latter view would bring the sperm-sacs into
correspondence with other parts of the generative apparatus ;
in any case they are not to be distinguished at their actual
VOL. 36, PART 2.—NEW SER. P
208 FRANK E. BEDDARD.
origin. The sperm-sacs have a slightly racemose appearance ;
they are attached to the septum by a stalk; this stalk is hol-
low, and seems to open into the cavity of the 11th segment ; it
is of course lined by peritoneum, which is especially thick on
that part of septum near to the origin of the sacs. The walls
of the sacs are muscular and thin; there are also trabecule of
muscular fibres which divide the interior of the sacs into com-
partments ; these spaces are full of developing sperm. Michael-
sen has described in Eudriloides gypsatus, in addition to
these sacs, a sac in the 11th segment. I have looked carefully
for this, but can only find a mass of sperm in that segment.
This mass of sperm is not enclosed by any walls, and there are no
blood-vessels for the supply of any sac, such as are abundantly
obvious in the case of the sperm-sacs of the 12th segment.
The sperm duct runs along the body-wall just at the line
of implantation of the ventral sets; it is accompanied in its
course by a blood-vessel. Near to the external orifice of the
atria the sperm-duct perforates the outer coat of the atrium of
its own side, and comes to lie where it is represented in fig. 1.
Anteriorly the sperm-duct of each side opens into the cavity of
the 11th segment by a large and much folded funnel; the two
funnels together occupy a great deal of the space in the 11th
segment; above the nerve cord the two funnels become fused
together. The atria extend through three or four segments ;
they lie side by side, and are never coiled; this coiling, which
is often seen in the long atria of other species, is rendered
impossible in the present case by the thick muscular coat
which forms the outer layer of the atria. This muscular layer
consists of longitudinal and circular fibres; the circular fibres
are internal. The epithelium lining the atrium appears to
consist of two layers, but these are not by any means so thick
as in other Eudrilide ; in transverse sections of the atrium the
lumen is seen to be crescentic; the convexity of the crescent
is dorsal; below this lies the sperm-duct, which, as already
stated, lies within the atrium: the sperm-duct passes just be-
tween the epithelium and the muscular layers; it is accom-
panied by a blood-vessel. Towards the external pore of the
OLIGOCHATA OF TROPICAL EASTERN AFRICA. 209
two atria the lumen of each gets to be more and more oval in
outline. As in other species of the genus Eudriloides, the
aperture of the atria is furnished with penial sete; there are
two of these, one on each side; the penial seta is curved into
an §-shape; the curvature is much more pronounced than that
of the ordinary set; the free tip of the seta appears to be
bifid ; it is ornamented with a series of minute denticulations ;
these occupy a restricted area just in front of the actual tip of
the seta; their shape can be understood by an inspection of
the accompanying figure (fig. 20).
There is in connection with the male efferent apparatus a
very complicated arrangement of muscles, recalling in many
particulars the corresponding structures in Eudriloides
brunneus. These structures are best elucidated in glycerine
preparations of the worm from which the alimentary tract has
been removed. In such a preparation the two penial setz will
be seen on either side of the male pore lying somewhat
obliquely ; near to the distal end of each seta a strong bundle
of muscles is attached ; the actual attachment is not extensive,
but the muscles rapidly fan out, and where they are inserted
on to the body-wall form a bundle of considerable dimensions.
The muscular fibres forming these two bundles are bound up
into separate muscles, as is shown in the figure illustrating
this copulatory apparatus (fig. 18). On the opposite side
of each seta there is another strip of muscles which runs
obliquely almost in the same straight line as the first mentioned
set; this bundle of muscles is, however, of equal diameter
throughout, and is not nearly so wide as the first. It seems
clear that these two sets of muscles have an opposite effect,
moving the penial seta in different directions. We now come
to the muscles of the bursa copulatrix. The two atria join to
form a tube, which is much narrower than either of themselves ;
this unpaired tube is covered by a layer of muscles which runs
transversely across the body, being attached to the body-wall
in the immediate neighbourhood of the penial seta on each
side. From the same point of origin arise a few fibres which
are inserted upon the tube itself, and a still more slender
210 FRANK E. BEDDARD.
bundle which is inserted upon each of the two atria just in
front of the point where they become fused. On the opposite
side of the male pore is a bundle of muscular fibres running
longitudinally. The total effect of the contraction of all these
muscles, as it appears to me, would be the protrusion of the
terminal apparatus of the male organs. I think that any of
these muscles could play the part of retractors when the
terminal sac is protruded ; they could, I should imagine, serve -
both as retractors and protractors, according to the position
of the organ.
The ovaries are in the 13th segment ; they are unenclosed
by any sac; the oviducts opposite to them open partly into the
cavity of the 13th segment and partly into the egg-sac. In the
independence of the ovaries, egg-sacs, and spermatothecal sac,
the present species agrees with other Eudriloides.
The spermatothecal sac is much like that of Eudri-
loides brunneus; it opens on to the 13th segment and ex-
tends a little way in front of its external pore and reaches for
some way behind it—as far as the 17th segment. I have
studied the minute structure in a nearly mature specimen (quite
mature except as regards the clitellum), and in a much younger
specimen. The part of the spermatothecal sac lying in front
of the aperture is not in any way different in structure from
the rest. The walls of the sac are much thinner posteriorly
than anteriorly ; they are lined by a layer of large cells which
are covered externally by a muscular coat; at the pore the
structure is a little difficult to understand; it is shown in
fig. 16. I could find no actual orifice, perhaps to be accounted
for by the worm not being fully mature. The epidermis is
thin just below the pouch and the muscular layers of the
body-wall have disappeared. ‘The spermatotheca is lined by
the thick layer of cells referred to. ‘These get so close together
where the lumen narrows towards the pore that the lumen is
entirely obliterated. It may be that this arrangement of the
cells means that the protrusible termination of the male
efferent apparatus can be thrust into the spermatotheca, but
that the sperm cannot escape from the sac. The ventral half
OLIGOCHAHTA OF TROPICAL EASTERN AFRICA. 211
of the pouch is encircled by a cup-shaped layer of epithelium
whose cells are of an altogether different character. These cells
are regular columnar cells with the nucleus at about the
middle. They have every appearance of being epidermic cells
invaginated, but I could not find that they were at any point
continuous with the epidermis; there was a slight break on
either side. This, however, is not fatal to the view that they
are invaginated epidermic cells. These cells, so easily dis-
tinguishable by their characters from the thick mass of cells
occupying the inner surface of the spermatothecal sac, do not
immediately abut upon the latter; the two layers are separated
by a thick, non-staining membrane which has every appear-
ance of being of a chitinous nature; it is homogeneous and of
a faint horn colour; it completely divides the two layers of
cells spoken of. This species has, like Eudriloides brun-
neus, glands at the sides of the spermatothecal sac (see
fig. 19). These have, apparently, precisely the same structure
as in that species, and need not therefore be particularly de-
scribed ; there is, however, but a single pair of them. The
ducts of these glands open partly on the exterior direct, and
partly through the layer of cells already spoken of and pre-
sumed to be formed by an invagination of the epidermis.
This is additional evidence of the justice of this interpretation
of the cells in question.
In a younger specimen of the species the spermatothecal sac
was less fully developed ; the posterior part of the sac is very
thin-walled, and lies to one side of the nerve-cord ; it looks like
a piece of a septum detached, and has a roughly circular con-
tour. Following the course of the sac in a series of sections, it
is seen to get beneath the nerve-cord and to lose its lumen; at
the same time it decreases in diameter ; at the external aperture
there is a protuberance of the body-wall in which there is a
slit-like crescentic lumen bordered by tall columnar cells; a
lumen also suddenly appears in the dorsal part of the sac which
is quite independent of the crescentic lumen referred to; this,
indeed, is merely the border line between the tall columnar
cells, which I believe to be an invagination of the epidermis,
212 FRANK E. BEDDARD.
and the cells lining the spermatothecal sac. This specimen
was remarkable for a plug of cells filling the lumen of the sac
just where it narrows towards the pore. The plug of cells pro-
jects a little way into the lumen, as shown in the figure (fig. 16),
and is furthermore rendered obvious by its less staining and by
a fibrous appearance with scattered nuclei. In the more
mature worm I could distinguish no such plug. Underneath
the epidermis the mass of cells forming this plug spreads out
into a more extensive layer.
One of the three individuals which I examined by longitudinal
sections showed a difference from the typical structure of the
genus in the presence of two pairs of sperm-duct funnels. To
these corresponded two pairs of testes. The additional pair
was in Segment x. This segment, like the following, was filled
with a mass of developing sperm not contained in a sperm-sac
or sperm reservoir. In other particulars I could detect no
differences from other specimens. It should be stated, how-
ever, that the penial setz could not be studied. These setz
are, of course, frequently most useful in distinguishing species.
It must, therefore, be left undecided as to whether this worm
is a distinct species of Eudriloides or is only a variety of
Eudriloides Cotterilli.
Eudriloides brunneus, n. sp. (figs. 2, 10, 21—28).
Of this new species of Eudriloides I have examined four
examples; two were dissected and two examined by means of
transverse sections.
As compared with the first species of Eudriloides described
in the present communication, this is large ; itis not, however,
quite so large as Michaelsen’s Eudriloides gypsatus.
The colour of the species was of a uniform greyish brown
(after preservation in alcohol) ; the sete are exceedingly minute,
so as to be quite invisible when the worm is examined by a
hand lens alone. Michaelsen has commented upon the minute
sete of Eudriloides gypsatus.
The clitellum was not very distinct ; it appears to occupy
the greater part of Segments xtv—xvu1; the clitellum is not
OLIGOCHATA OF TROPICAL EASTERN AFRICA. 215
developed upon the ventral surface of the body; the segments
of which it is composed are a little difficult to map on account
of the fact that the segments are subdivided by numerous
transverse furrows, which do not correspond to the actual
segments themselves. The arrangement is shown in the
accompanying figure (fig. 10). Behind the median spermato-
thecal pore, which lies upon the 13th segment, there is a
furrow which is the boundary line of Segments x111/xIv.
This latter segment bears a median genital papilla of an oval
form, the long axis of the papilla being disposed transversely ;
in front of and behind this papilla is a short furrow ; then follow
five other transverse furrows, the last of which marks the line
between Segments xv1/xvit ; the clitellum is developed laterally
in the region which is occupied by the various furrows which
have been described; on the 16th segment is a pair of genital
papillze, one on either side of the median line; behind the male
pore is a single and median papilla upon the boundary line of
Segments x1x/xx.
The spermatothecal pore is upon the 13th segment ; it
is very conspicuous, being borne upon a prominent projection,
which is no doubt retractile. The male pore is even more
conspicuous, and is of about the same area.
The nephridia of this species appear on a dissection to be
paired structures like those of other Eudrilide. This is,
indeed, the case, but the duct leading to the exterior forms a
network within the integument, as in the genus Libyodrilus,
where I have lately described it.1 I have mainly studied
the nephridial system of this worm in the anterior segments of
the body ; in transverse sections through one of the anterior
segments—the ninth, I think—two longitudinal ducts are
observable running just to the inside of the longitudinal mus-
cular layer in the peritoneum on each side of the body. These
ducts correspond in position with the pairs of setz ; the inner
of the two on each side is the larger. Near to the septum
dividing the segment from the one in front the inner duct
1 On the Structure of an Earthworm allied to Nemertodrilus, &c.,”
‘Quart. Journ. Mier. Sci.,’ vol. xxxii, p. 553.
214 FRANK E. BEDDARD.
became continuous with the nephridium, and at that point
gave off a duct which penetrated the body-wall and apparently
opened on to the exterior, though I did not succeed in dis-
covering the actual pore. Following the duct back, the
longitudinal duct was found to vary in calibre from place to
place; it was sometimes so reduced as to be very nearly
invisible; at other times it became of much greater calibre
than the average. In these places it seemed as if the duct
formed a kind of rete. In the one segment I counted three
or four of these retia; in the same segment three or four
branches arose from the longitudinal duct to apparently reach
the exterior. At the point where the septum was attached to
the body-wall the nephridial duct penetrated still deeper within
the longitudinal muscular coat. When the septum lifted away
from the body-wall, the inner of the two longitudinal ducts was
found to have resumed its original position. Immediately after
the septum a branch was given off from the vessel which,
penetrating into the body-wall, passed round the circumference
of the body, ultimately joining the outer of the two longitudinal
ducts. I confess that the actual junction was not observed,
but the tube was traced up to a very minute distance away
from the second of the two longitudinal vessels; in front of
this there is a connection by way of the peritoneum—a duct,
that is to say, runs from one longitudinal duct to the other in
the peritoneum. Here, again, I was not able to find the actual
opening at both ends; but I have no doubt, from the appear-
ances presented, that this takes place. The second longitudinal
duct—that running on a level with the outer pair of setee—is
not dilated, and shows no such retia as the inner duct; nor
does it appear to give off any branches penetrating the integu-
ment and reaching the exterior; in other segments, and more
particularly in the region of the body occupied by the clitellum,
I could not find the connection of the nephridia from segment
to segment, such as undoubtedly occurs in some, at any rate,
of the anterior segments. The nephridial system of this
species, therefore, is constructed on the plan of that which
characterises the genus Libyodrilus, but is in a less differ-
OLIGOCHATA OF TROPICAL EASTERN AFRICA. 215
entiated condition. The integumental plexus appears to be
much less developed in Eudriloides than in Libyodrilus.
After finding that this species of Eudriloides showed the
above mentioned resemblances to Libyodrilus, I examined
other species, with a view to discovering how far they agreed
with the present. In Pareudrilus something of the same
kind seems to occur (see below). In Stuhlmannia varia-
bilis I traced the nephridial duct into the body-wall at the
inner pair of sete ; instead of opening on to the exterior opposite
to the point where it entered the body-wall, the tube passed
along the body-wall below the circular muscular layer, and
eventually opened on to the exterior on a level with the outer
pair of sete. In one instance, at any rate, I feel sure that the
tube did not extend further than this point, but whether there
was more than one opening on to the exterior for each
nephridium I am not able to say with certainty. I think
that in Eudriloides Cotterilli the same state of affairs
is found. In the two latter cases the small size of the
worms was a matter of difficulty ; the corresponding fineness
of the tubes rendered their discovery a matter of greater diffi-
culty than in Eudriloides brunneus. Iam almost inclined
to think that all those species of Eudrilids in which Michaelsen
has described ventral nephridiopores will be found to have a
nephridial system like that of the species described here. One
is tempted, in the instances brought forward here, to associate
the complexity of the integumental plexus with the size of the
worm. In Libyodrilus and in Eudriloides brunneus,
which are the two largest species in which the nephridial
system has been investigated, it is apparently the most com-
plex; in Stuhlmannia, at any rate, it seems to be less
complex. The only time that I was able to follow out the
tube from the point where it entered the body-wall to its
external pore, it seemed to me to pass straight from opposite
the inner pair of setz to the aperture on a level with the outer
pair of sete. In this connection Professor Hubrecht’s! inter-
1 «The Nephridiopores in the Earthworm,” ‘ Tijdschr. Ned. Dierk. Ver.,’
ser. 2, vol. ii, p. 226.
216 FRANK E. BEDDARD.
esting discovery of the course of the nephridial tube in certain
species of Lumbricus (or Allolobophora) will occur to the
reader. The facts discovered by Hubrecht, which I can
confirm by my own experience, seem to be analogous to those
described here. In those worms Hubrecht found that the
nephridial duct, after entering the body-wall, passed along it
between the two muscular layers to the opening, which is
situated in some cases beyond the outer pair of set; there is,
however, no question, in Lumbricus, of an integumental
plexus. Still it is possible that there may be in this peculiar
disposition of the nephridial duct in some Lumbricide a re-
semblance to the Eudrilide. We may have here a clue to the
affinities of the Lumbricide, which has been hitherto wanting ;
but I do not propose to follow up this matter at present.
Reproductive Organs.—The sperm mass of the 11th
segment is, like that of the next segment, apparently not
enclosed in a sac. When the worm was dissected the mass
of sperm could be easily disturbed by the dissecting needle.
There was nothing to offer any resistance to the needle.
There is, however, seemingly a functional equivalent of the
missing sac. The septum which bound this segment pos-
teriorly is comparatively thin, though thicker than the septa
in the posterior region of the body. It is inserted on to
the body-wall in the usual way. Just before its insertion a
sheet of muscular tissue, of precisely the same thickness and
general appearance as the septum, arises from the said sep-
tum and passes obliquely forwards, ultimately joining the
anterior septum of the segment which encloses the mass of
sperm. There is thus formed a chamber which encloses
the mass of sperm, but which does not seem to be the exact
equivalent. of the sperm-sacs of other earthworms. It is
well known that in many earthworms the successive septa are
bound together by muscular bands running in various direc-
tions. This state of affairs is more especially characteristic of
the anterior and often thickened septa. It is probable, there-
fore, that the materials for the formation of the septum above
described already exist, and that the sheet of tissue which
OLIGOCHATA OF TROPICAL EASTERN AFRICA. 217
bound the sperm mass of the 11th segment anteriorly is to
be regarded as a development of these muscular threads. In
another specimen, however, this peculiar disposition of the
septa was absent.
The spermatothecal panel of this species is single and
median, extending from its opening on the border line of the
12th and 13th segments for three or four segments posteriorly.
Anteriorly, also, it extends for a short distance beyond the
external pore. This portion of the sac is conical in form, as
is shown in the figure (fig. 22). The general appearance
of the spermatotheca is closely similar to that figured by
Michaelsen for Notykus emini (loc. cit. on p. 1, Taf. ii,
fig. 8). The spermatothecal sac is also, as in that worm,
without lateral branches, such as occur in Stuhlmannia
and form a ring round the gut. There is, moreover, a further
point of resemblance to Notykus illustrated in my figure,
which may be compared with that of Michaelsen. Round the
base of the sacin Notykus are represented in Michaelsen’s
drawing a pair of small glands lettered “nt.” It is described
by Michaelsen as “ein kleines muskuloses Polster ....
wahrscheinlich mit einem Hohlraum versehen, welcher durch
die oben erwahnten spaltformigen Offnungen neben der Samen-
taschen-Offnung ausmiindet.”’ I shall revert to the probable
nature of these organs later. The anterior part of the spermato-
thecal pouch which lies in the 12th segment is histologically
different from the region which it precedes. There is in this
particular a remarkable analogy between these sacs which
function as spermatothecz and the true spermatothece of, for
example, the Perichztide. In the latter the diverticula of the
spermatothece have invariably a different structure from the
pouches of which they are diverticula. This, it will be noticed,
is also the case with the spermatothecal sac of the present
worm. When the sac is examined by a series of transverse
sections, from in front backwards, the first pair to appear in
the series is of course the anterior diverticulum already
referred to. This (fig. 23) is seen to have a circular form in
section, and it presents a most curious resemblance to a
218 FRANK E. BEDDARD.
section through the csophagus of a worm. Its walls are
muscular and stout. They are lined within by a layer of
darkly staining cells which have much the appearance of a
low columnar epithelium ; the lumen, however, is not simple.
Numerous folds of the lining membrane are visible, which
project far into the lumen, and nearly meet in the centre.
These folds have a fairly regular arrangement. They are not
only folds of the lining epithelium, but also of the strong
muscular layer. This anterior diverticulum gradually passes
into the bursa. The latter is a somewhat flattened sac, with
strong muscular walls. It is lined by a regular columnar
epithelium. This epithelium is markedly different in appear-
ance from the epithelium lining the diverticulum of the sper-
matothecal sac; the two layers do not in any way pass into
each other. In a-series of sections the epithelium of the pouch
terminates more or less abruptly; near to its end it gets to
be covered by the cells of the diverticulum, which ultimately
replace it. In a series of sections quite complete through the
anterior of the two genital pores it was impossible to find any
external aperture. This was also the case with a second
series. Both worms, it should be said, were sexually mature.
The epithelial lining of the pouch already referred to dips
down towards the epidermis. Some little way above the
epidermis it ends abruptly, and the lining of the sac is made
up of cells of a quite different appearance. There is, however,
no communication that I could discover between the inside of
the bursa and the exterior of the body. A moderately thick
layer, chiefly cellular, blocks the spot where the aperture
should be. The principal part of the spermatothecal pouch
lies behind the aperture. At first the pouch is lined with
epidermic cells entirely similar to those which line the bursa.
These cells form numerous folds in the interior of the sac.
The folds, however, are not, as in the anterior diverticulum,
supported by upgrowths of the muscular layer. Further back
still the folds die away. At first, as already said, the pouch
is lined by cells which resemble those of the bursa, and must
be, I should imagine, formed by an ingrowth of the epidermis.
OLIGOCHATA OF TROPICAL EASTERN AFRICA. 219
These cells are in places partly covered by isolated groups of
cells much smaller than themselves. These cells lie to a great
extent loose in the cavity of the spermatotheca; only here
and there are they congregated into little heaps covering the
epidermic lining of the pouch. Presently the covering of
peritoneal cells gets to be closer, and coincidently with this
the cells which they cover diminish in height, though they are
still clearly separated from each other; indeed, the separation
is usually marked by a darkly staining but structureless layer,
which seems to be thrown off by the epidermic cells. The
peritoneal cells which cover the subjacent epidermis are long
and filamentous cells, which proliferate freely at their free
extremity. The whole structure of the spermatothecal sac of
this species seems, asin the case of Eudriloides Cotterilli,
which it greatly resembles, to indicate that it is formed out of a
peritoneal sac into which an epidermic invagination has grown.
The fact that there seems to be no external pore is very
remarkable and unintelligible. It is a further point of agree-
ment with the other species of this genus described in the
present paper, and also with Stuhlmannia. All the worms
that I examined were sexually mature, but in no one of them
were there any bodies of any sort within the spermatothecal
sac, except the detached cells evidently derived from the lining
peritoneal epithelium. In no one of them was there any pore
leading to the exterior. It is, of course, possible that at stated
seasons there is a pore which is at other times blocked; but
this is only supposition, and I have no facts which suggest
any interpretation of the use of these sacs. In front of and
behind the point where the external aperture of the sper-
matothece ought to be, were it visible, is a glandular mass
lying upon the bursa. This mass reaches down on either side,
thus forming two rings, one anterior and one posterior, nearly
enclosing the spermatothecal sac. These two ring-shaped
glands are composed of small nucleated cells, which above the
spermatotheca are arranged in a continuous mass. As the
gland comes to lie at the sides of the sac its cells become
arranged in a series of columns, which anastomose here and
220 FRANK E. BEDDARD.
there. In transverse section these columns are seen to be
formed of about eight or ten cells whose nuclei lie peripherally.
I could not detect any lumen for the most part. In places a
lumen appeared to exist. Hach disc of cells (as seen in trans-
verse sections) is enclosed in a delicate sheath not closely
adherent to the cells. These rows of glandular cells appeared to
open on to the exterior, but their external apertures were not
very evident. They were plainer in Eudriloides Cotterilli.
Glands in the neighbourhood of the spermatothecal orifices
are common in the Oligocheta; but the form of these glands
in the present species is unlike that which I have seen in any
other, excepting, of course, in the other species of the same
genus described in the present paper. It may be that
Michaelsen’s Notykus is furnished with similar glands. The
figure given by him seems to indicate a similarity.
The atrium (see fig. 2) differs in structure from that of
Eudriloides Cotterilli in the fact that the muscular layers
surrounding the glandular lining are thinner; but, though thin,
both layers are there. The cellular lining of the atria re-
sembles that of other Eudrilids and of nearly all other earth-
worms in being composed of two layers of cells. The layer
immediately abutting upon the lumen is composed of columnar
cells. In certain tracts of the atrium near to the external
orifice this layer of cells is very plain, and resembles such
columnar cells as line the sperm-duct and their funnel, &e.
Elsewhere the inner lining of the atria consists of cells which
are loaded with granules, and of which the nucleus has got
obscured. The sharply marked layer of epithelium referred
to is furthermore remarkable by the presence of cilia.
Cilia are also visible in other parts of the atrium, but more
obvious along this tract of unmodified epithelium referred to.
I believe that this is the first record of ciliation of
the atrium in any earthworm. Thesperm-duct comes to
lie within the muscular sheath of the atrium as in other Eudri-
lids, but I have not ascertained the exact spot at which it opens
into the lumen.
The terminal pouch of the male efferent system is furnished
OLIGOCHATA OF TROPICAL EASTERN AFRICA. 221
with a rather complicated arrangement of muscular bands.
These are illustrated in the accompanying figure (fig. 21).
Just behind the muscular bulbus arise by a common stalk the
two atria. At the angle on either side of their entrance
into the terminal bulbus is a bundle of muscular fibres
running transversely to the longitudinal axis of the body.
These muscular fibres appear to be mainly concerned with
the penial seta, which is enclosed by them. From each
of these bundles of fibres arises a flat strap-shaped band of
muscles, which passes forwards obliquely and is inserted on to
the body-wall just in front of the edge of the muscular bulbus.
Besides these muscles, a number of delicate fibres pass from the
parietes to be inserted round the periphery of the bulbus.
The latter muscles are, I take it, retractors, while the single
strap-shaped muscle on each side, perhaps, by its action when
contracted, protrudes the bulbus by drawing back the body-
wall just in front of it.
The penial seta of each side of the body lies, as already
stated, in the mass of muscles lettered in my figure. These
sete are very strong and rather short, in fact very short if
compared, for example, to those of Eudrilus Cotterilli.
They are strongly curved into an S-shape, and are a deep
yellow colour. The free extremity is not at all ornamented,
but it thins off like the blade of a knife, and on this thin
edge a few transverse strie are to be noticed. The genital
sete of this species are decidedly peculiar in form and very
characteristic of the species.
Pareudrilus stagnalis, n. g., n. sp. (fig. 9).
I have had for examination a considerable number of indi-
viduals of this worm, which is referred to a new genus. It
presents upon a casual inspection all the characters of the
genus Eudrilus, to which genus I was at first disposed to
assign it. Not only are the colour and general appearance
quite similar to those of that genus, which also appears to exist
on the east coast of Africa, but the position of the reproductive
pores are identical, and they are paired—a character which is
222 FRANK E. BEDDARD.
only known in the genus Eudrilus, and in the not very nearly
allied Nemertodrilus. Nevertheless, a more detailed study
of the worm has convinced me that it is an entirely distinct
genus, showing, except for the paired apertures of the genera-
tive ducts, no particular likeness to any Eudrilus with which
I am acquainted. It is probably true that the genus Eudri-
lus requires revision. It is possible that the individuals from
Africa, America, New Zealand, &c., differ specifically among
themselves. I confess that I have not been able to detect any
such differences, and I am at a loss to understand by what
characters Dr. Horst distinguishes his Eudrilus jullieni,
though I admit that he only creates the new species in a very
diffident manner. On the other hand, Michaelsen’s Eudrilus
pallidus is, in my opinion, unquestionably a distinct form.
These two species of Eudrilus and Nemertodrilus griseus
are at present the only Eudrilids with paired genital pores.
These Michaelsen places in a distinct sub-family from the
members of the family, the great majority of which have
median unpaired pores. I shall attempt to show in the sequel
that this character is not alone sufficient to distinguish two
such groups; that the resemblances between the genus Par-
eudrilus and certain genera of the sub-family Teleudrilini is
closer than that which obtains between Pareudrilus and
Eudrilus.
The colour of this worm is a dark purplish brown upon the
dorsal surface, becoming light below. The colour was not well
preserved in specimens that had been treated with Perenyi’s
solution. In a single individual, which was brought home
alive and preserved by myself in gradually increasing strengths
of alcohol, the colour was very dark, as dark as during life ; it
was much bluer than the rest. All the specimens were found
either in the mud at the edge of a pond on Mombasa Island or
from a bog up country about four miles from the coast line
opposite to Mombasa. The length of a worm selected for
measurement was 63 mm., the breadth 4 mm.; this represents
about the average size. The prostomium is continued for a short
distance on to the peristomial segment by two grooves enclos-
OLIGOCHATA OF TROPICAL HASTERN AFRICA. 228
ing a narrow space. The sete are strictly paired, and ventral
in position ; the last segment of the body had four pairs just
like those of any other segment. The only modification of the
setze that I could see was that the ventral sete of Segment
XVII were wanting, being here replaced by the large penial
sete.
The male pores are paired. They lie on the border line of
Segments xvii/xvi11. In one or two specimens a round mass
was protruded from the pore which I take to be the partially
protruded bulbus atrii. The male pores are on a line with the
ventral sete.
The spermatothecal pore is on the border line of Seg-
ments xiv/xv. The clitellum I am unable to place. The only
specially noteworthy matter to be recorded about the epi-
dermis is that there is no trace whatever of the peculiar sense
bodies, so like the Pacinian corpuscles of Vertebrates, and so
commonly met with among the Eudrilide. The circular
muscular layer can be seen in longitudinal sections to show
a decidedly bipinnate arrangement of its fibres; these are
grouped into narrowish tracts which are two to four fibres
wide. ‘The fibres of the circular muscular layer are laxer in
their packing than the longitudinal layer. The interspaces are
filled up by delicate strands of an appearance like that of con-
nective tissue. The pigment is chiefly lodged in this outer
layer of muscles. It is disposed in tracts that follow the
direction of the fibres. Here and there threads of pigment pass
down through the longitudinal layer following the course of
the blood-vessels. These are so regular in their arrangement
that the longitudinal muscular layer is divided up by them into
a series of squarish blocks. The peritoneum lining the walls
of the body is also pigmented, and pigment also occurs in the
peritoneum covering the septa.
The body-cavity contains in parts numerous corpuscles.
The septa are, as in so many, if not in all, earthworms, not
attached to the parietes entirely along the grooves which cor-
respond to the external metamerism. The first septum divides
Segments v/v1. It is very delicate. The next six are thick-
VOL. 386, PART 2.—NEW SER. Q
224. FRANK K. BEDDARD.
ened, but not so greatly as is so often the case. The next
septum is also thicker than those which follow, but not so
thick as those which precede it. All these thickened septa lie
behind the gizzard. The anterior septa are cup-shaped.
The alimentary canal is differentiated into a gizzard
which lies in the 6th segment. It is long and rather narrow.
The cesophagus which follows extends as far as the 17th seg-
ment. It is very vascular throughout its whole length, and
the lining membrane is so folded that in cross sections the
edges look like a mass of tubes containing blood and cut
across. The cesophagus is entirely unprovided with glands of
any kind appended to it. The intestine has a very small
typhlosole.
The nephridia are paired structures. Those of the pos-
terior segments are, as is the general rule, much more obvious
in dissections than the few most anterior. This is due to the
development of the peritoneal cells which clothe them. These
cells are often filled with quite large lumps of an amorphous
secretion which stains darkly in borax carmine. It is these
secretions which give the white colour and therefore the con-
spicuous appearance to the nephridia in the posterior segments.
When the worm was examined with a hand lens, or even
mounted in glycerine on a slide and studied with compara-
tively high powers of the microscope, there was no indication
of nephridiopores. These are usually conspicuous in the
Eudrilide, but they were absolutely invisible in Pareu-
drilus.
The reason for the apparent absence of nephridiopores was
revealed by an examination of transverse sections. In such
preparations the duct of the nephridium could be easily traced
into the body-wall in the region of the ventral pair of sete ; the
duct of the nephridium was quite obvious on account of its
considerable width, and the fact that it has an intercellular
lumen. Directly it penetrates the body-wall it becomes of less
calibre, and instead of opening on to the exterior forms a duct
which runs right round the body-wall on the boundary line
between the two muscular layers, and in immediate proximity
OLIGOCHAMTA OF TROPICAL EASTERN AFRICA. 225
to the principal nerve. I have in many cases followed out
these ducts to the dorsal mid-line. I am inclined to believe
that there is a connection between those belonging to the
nephridia of opposite sides of the body. In any case the con-
nection below the nerve cord seems to exist, the duct as it
enters the circular muscular layer dividing into a right and
left half. The arrangement of the nephridia of this worm is
very similar to that which characterises the genus Libyo-
drilus, but in Libyodrilus there are not only numerous
branches of the circular ducts which lead to the exterior, but
numerous branches which lead in other directions, thus form-
ing a plexus of tubes within the body-wall. In the present
species I did not observe anything of the kind. In one case lL
found a branch apparently leading to the exterior. Such
branches cannot, I am convinced, be very numerous or they
would have been more obvious; on the other hand, fine tubes
would be very difficult to detect without a special mode of
demonstration, and it seems likely that the present species
has more resemblance to Libyodrilus than I have at present
been able to discover.
The vascular system has not been studied in great detail.
In the intestinal region the dorsal vessel gives off three pairs
of branches in each segment. The vessel itself lies upon the
chloragogen cells, and is covered superiorly by a layer of thin
cells, which are not continuous with those of the intestine.
Here and there are delicate muscular strands, two or three
fibres wide, attaching the dorsal vessel to the intestine ; they
pass from the muscular layer of one to the muscular layer
of the other. There are six or seven of them on each side
in each segment. The openings into the dorsal vessel of the
branches are guarded by valves. ‘Two pairs of vessels supply
the intestinal walls, the third pair spread out over the septum.
Reproductive organs.—Unfortunately there was only a
single specimen out of the ten or a dozen examples collected
for me that was sexually mature. In this individual the gene-
rative organs were dissected, parts of them being investigated
later by the section method.
226 FRANK E. BEDDARD.
The ovaries were not seen in this specimen, but I have
found that they occupy the normal position in the 13th seg-
ment. Apparently they are enclosed in a large sac which ex-
tends from the septum bounding the 11thand12thsegments. This
sac (see fig. 9) is somewhat pear-shaped; it gradually narrows
to a fine tube which opens into the large spermatothecal sac.
The latter opens on to the exterior between the Segments
xtv/xv. The pear-shaped sac, which I believe involves the
ovary, is the ‘ovarialblase”? of Michaelsen. A narrow and
short tube leads from it to the egg-sac, which occupies the
usual position attached to the anterior wall of the 14th seg-
ment. From the egg-sac leads another tube, which appears to
open into the ovarian sac. This tube, as 1 have ascertained
from a continual series of sections through the entire appa-
ratus, only lies within the ovarian sac; its lumen does not
communicate with that of the ovarian sac; the tube soon issues
again from the sac, and opens on to the exterior by the ovi-
ducal pore. The first tube mentioned which connects the
egg-sac with the ovarian sac is really a portion of the funnel
of the oviduct—that portion which in other Oligocheta does
not open into the egg-sac, but freely into the cavity of the 13th
segment. As in the present species, there is a sac developed
which involves the ovary; it must happen that the funnel
partly communicates with the sac in question. As the latter
grows a stretch is put upon the funnel which grows out in the
way that we see. The spermatothecal sac is a large pouch
which is somewhat bilobed at its free extremity. This region
lies posteriorly to the end which opens on to the exterior. In
sections it is seen to be lined by an epithelium which is
columnar in form, and shows no signs of that proliferation so
frequently seen in the spermatothecal sacs of these Annelids.
Its interior is folded.
The male organs of generation are as in many other
Eudrilids. There are two pairs of testes situated in Segments
x and x1; the sperm sacs are in Segments x1 and x11, and, like
the testes, are attached to the front walls of their segments.
The sperm-ducts retain their distinctness until their point of
OLIGOCHATA OF TROPICAL EASTERN AFRICA. 227
Opening into the atrium. The atria lie perfectly straight on
each side of the body ; they are comparatively short, and are
not twisted as is so frequently the case with the longer atria of
other species. Each atrium consists of two parts; the external
pore leads into a nearly spherical and very muscular bulbus ;
from this arises the atrium proper. This tube is quite as wide
as the terminal bulbus, and has a nacreous appearance. This
appearance is of course due to the stout muscular walls of the
organ. In longitudinal sections the atrium is seen to be com-
posed of four layers exclusive of the peritoneum ; beneath the
peritoneum is a layer of longitudinally disposed fibres ; beneath
this, again, a layer of circular fibres. These two layers are of
about the same diameter. The iining membrane of the atrium
is built up of two strata of epithelium ; the innermost layer is
composed of not very tall columnar cells, outside which are
several layers of slightly staining pear-shaped glandular cells.
The two layers of epithelium together are twice as thick as the
muscular coats.
Kach atrium is accompanied by a long, thin, muscular sac,
which is nearly, if not quite, as long as the atrium itself. This
sac is placed to the outside of the atrium, and is slightly
curbed. Each of the two sacs contains a single very long
penial seta, It is very thin, and curved more sharply at the
free extremity. The free end of the seta expands at the actual
extremity into a thin, flattened plate; just before this the seta
is beset with a few short spinelets.
I have been able to study the female reproductive system in
immature worms, and to ascertain that the spermatothecal
apparatus is formed from at least two sources. The sper-
matotheca appears to be formed by an epidermic invagina-
tion. Its lining epithelium is continuous with the epidermis.
This sac is large and has very thick walls. It appears at
first sight to be independent of any other part of the system
and I believe that originally it is so. However, in all the
specimens examined by me there is a strand of tissue, princi-
pally constructed of muscular fibres, which has in the imme-
diate neighbourhood of the spermatotheca no lumen, This
228 FRANK E. BEDDARD.
solid strand joins the spermatotheca. Traced in the opposite
direction it is seen to gradually develop a lumen, which gets
wider and wider until it expands into a trumpet-shaped orifice
which opens into the 13th segment. It seems in fact to be
clear that this tube, which ultimately forms the communica-
tion between the ovary egg-sac and spermatotheca, is merely
a backward growth of the septum separating Segments
xiui/xtv. The oviducal funnel lies just above the mouth of
this diverticulum of the septum, and is placed within the
mouth of the egg-sac. The oviduct itself, shortly after it
expands to form the funnel, projects into the interior of the
septal sac; it does not open into it, but is enclosed by the
walls of the said sac. At this stage the ovary is quite free, and
is attached in the usual position to the anterior septum of the
13th segment.
The development of the corresponding regions of the female
generative apparatus has been studied by myself in Libyo-
drilus violaceus; I was able to show that nearly the whole
of the large spermatothecal sac originated from the ccelom,
the septa being modified to form its walls; at most the
merest trace of an invaginated part was to be found at the
external orifice. I suggested that the large spermatothecal
sacs of Eudrilus and Teleudrilus were also probably to be
regarded as the homologues of the unpaired sac of Libyo-
drilus. The facts that I make known in the present paper
do not lead me to adhere to that opinion. For in Pareu-
drilus it seems to be more than probable that the whole of
the spermatothecal sac is an invagination, and that the egg-
conducting apparatus only is of mesoblastic origin, and has a
cavity which is an enclosed section of the celom. The nature
of the spermatothecal sacs of the Eudrilide was first proved
by myself to be different from that of other earthworms in the
paper upon Libyodrilus referred to; at least it was ren-
dered exceedingly probable that the conditions obtaining in
Libyodrilus were not confined to that genus, but were cha-
racteristic of the whole gronp. Rosa, however, independently
in point of observation, but subsequently to myself in date of
OLIGOCHATA OF TROPICAL BASTERN AFRICA. 229
publication, arrived at the same results. His results, how-
ever, were not the outcome of a study of the young stages,
but of a comparison of the structure of the several regions in
the adult. He found that the epithelial lining of the sac
stopped abruptly at a point not far removed from the external
aperture, and gave place to a layer of cells of an entirely differ-
ent character, and like the peritoneal epithelium. It may, I
think, be regarded as certain that a great part of the compli-
cated system of spaces surrounding the ovary, and communi-
cating with the exterior, in the Eudrilide are derivatives of the
celom, but it is also clear that a variable tract of what has
been termed the spermatotheca is really an invagination of
the epidermis, and is therefore comparable to the spermato-
thecee of other worms. I shall recur to this subject in describ-
ing the anatomy of some species of Eudriloides. The deve-
lopment of the sac a shows how the oviduct comes to be partly
enveloped by it; as it is simply a part of the celom, there is
nothing to be surprised at in the fact that the oviduct lies
within it.
§ Affinities of Pareudrilus.
This genus evidently is referable to the second of the two
sub-families into which I divide the Eudrilide (see below).
It has no integumental sense-organs of the characteristic form
found in Eudrilus and other allied genera. The sperm-
ducts are not dilated at their junction with the funnels.
There are no calciferous glands. It is the only genus in this
sub-family besides Nemertodrilus which has paired repro-
ductive apertures. A very marked peculiarity of the present
genus is arrangement of the nephridia. This is only paralleled,
so far as we know at the present time, in the West African
genus Libyodrilus. In that genus there is a complex
system of integumental nephridial tubes. Nevertheless it is
not perhaps the only other genus in which this peculiarity
occurs. In the description given by Michaelsen of Mega-
cheta tenuis, it is stated that the nephridiopores could
not be discovered; so, too, with Notykus and Metadrilus.
-
230 FRANK E, BEDDARD.
In the latter genus, however, Michaelsen was unable to see
the nephridiopores, but found (by transverse sections) that
the nephridiopores were placed in the neighbourhood of the
ventral sete. It seems to me to be possible that in these
species there are nephridia of the type which characterise the
genera Libyodrilus and Pareudrilus. When the nephridia
are of the usual form, the orifices are so conspicuous that it is
almost impossible to overlook them. In these three genera,
moreover, as well as in Libyodrilus, the atria are compara-
tively short and have very muscular walls. These are addi-
tional points of resemblance to Pareudrilus. Pareudrilus
differs from Libyodrilus mainly in the form of the female
reproductive apparatus, and in the absence of the three poste-
rior gizzards. With Metadrilus, the genus Pareudrilus
agrees in the position of the spermatothecal orifice. But in
Metadrilus the spermatothece are much reduced, and there
are besides only the rudiments of penial sete. The affinities
of Pareudrilus are with the genera mentioned, but no one
of them can be said to be much nearer than the others.
Polytoreutus violaceus, n. sp. (figs. 3, 7).
Two species of the genus Polytoreutus have been at pre-
sent described. The type species of the genus, P. ceruleus,
was described briefly by Michaelsen in a preliminary account
of the earthworms of Zanzibar and the opposite mainland,!
and subsequently re-described in a more thorough fashion."
A second form, P. magilensis, was afterwards described by
myself from Magila, Hast Africa. I have now to add a third
species, which I name P. violaceus, on account of the colour
of the worms.
There were four examples of this species in the collection ;
all of them were fortunately sexually mature, and all of them
were about the same size. The length of one individual which
was selected for measurement was 83 mm., its diameter 4 mm.
1 Loc. cit. (on p. 201).
2 “Some New Species and Two New Genera of Karthworms,”’ ‘ Quart,
Journ, Mier. Sci.,’ vol. xxxiv,
OLIGOCHETA OF TROPICAL EASTERN AFRICA. 231
The species is therefore fairly stout in build. This individual
had 190 segments. The size of the species is therefore about
the same as that of Polytoreutus ceruleus, and consider-
ably less than that of P. magilensis.
The dorsal surface of the preserved specimens was of a red-
dish purple, bluer perhaps in some than in others; but there
was nothing that could be fairly described as ‘ leuchtend
himmelblau,” a term which Michaelsen uses in writing of
Polytoreutus ceruleus. The under surface was yellowish.
The clitellum also was readily distinguishable from the rest of
the body by its yellowish coloration. The coloration of this
species was indeed almost, if not exactly, that of the genus
Eudrilus.
The prostomium is broad and does not indent in the least
the buccal segment. This is a generic character, and not
peculiar to the present species.
The clitellum occupies Segments xI1v—xviII or in some
specimens xvit only. It is less developed on the ventral than
on the dorsal or lateral surface, and it is here only that the boun-
daries of thesegments, which compose it, are clearly visible. The
clitellum may in fact be spoken of as ‘ saddle-shaped,” though
there is really no distinction to be drawn between a clitellum
of this kind and a “cingulum.” In both the remaining
species of the genus the clitell 1m lasbeen described as com-
plete, and consists of an additional segment, the 13th.
The nephridiopores are lateral in position.
The sete are paired, and are nowhere deficient except upon
Segment x11 (see below). The two setz of each of the ventral
pairs are, however, further apart from each other than are
those of the lateral pairs. This arrangement was apparent on
the posterior as well as on the anterior segments of the body ;
it is not peculiar to the present species, but also characterises
P. magilensis and P. ceruleus. Michaelsen found no
setee upon the clitellum of the last-mentioned species; they
were certainly not absent from this region of the body in
Polytoreutus violaceus.
The median genital pores lie, as in the other two species,
232 FRANK E. BEDDARD.
on Segments xvii—x1x. The middle region of the 18th
segment is occupied by a protuberant swelling of the body-
wall, which also extends for a short distance on to the segment
in front; this lies entirely between the ventral sete. The
posterior border of this projection appears to be the boundary
line of Segments xvir1/x1x. Anteriorly, however, the border
line of Segments xvit/xvilr is seen to end on each side behind
the anterior part of this protuberant pad. The anterior ori-
fice therefore lies on the 17th segment, and not on the border line
between this segment and the one which follows. On the other
hand, the posterior aperture, which is that of the spermato-
thecal pouch, is distinctly on the border line of Segments
xvilt/xtx. The latter orifice appears to be very much smaller
than the male pore, which has slightly crenated lips. The ventral-
most seta of Segment xvii is absent on both sides of the body.
The present species is also remarkable for a median unpaired
papilla situated upon Segments xxir and xxi. This is oval
in form, and is flattened with a raised margin. It commences
at the level of the sete of Segment xxt1, and extends back as
far as the end of the next segment, the border line of which
convex backwards. The median region of this segment is, in
fact, very much wider than that of the neighbouring segments.
The ventralmost seta of Segment xx1ir lies on the papilla on
both sides of the body ; on the 22nd segment that seta of the
left side is upon the papilla.
Reproductive Organs.—As in other species of the genus,
Polytoreutus violaceus has only a single pair of testes.
These lie in the 11th segment ; they are not, however, attached
to the front wall of their segment, nor are they, as is some-
times—though rarely—the case with earthworms, attached to
the posterior wall. They are attached to the wall of the
sperm-sac, which commences in this segment. It can hardly
be doubted that originally the testes were attached to the
front septum of the 11th segment; but the growth ‘of the
sperm-sac appears to have cut them off from this position, and
in the adult worm the heart lies between them and the septum
in question, The testes lie in the proximal end of the sperm-
OLIGOCHATA OF TROPICAL EASTERN AFRICA. 233
sac. The sperm-sacs are paired, and extend, as in Poly-
toreutus magilensis, through a large number of segments.
Their commencement is in the 11th segment, and here they are
somewhat dilated. The upper wall of the sperm-sac curves
round, and is attached to the dilated part of the sperm-duct,
the testes lying just in the angle formed by it and the wall of
the sperm-duct dilatation. The lower wall of the sperm-sac is
formed by the septum separating Segments x1/x11. The sperm-
sac then perforates this septum, and becomes a very narrow
tube not more than a quarter or less of its dimensions at first.
These two fine tubes pass along the dorsal surface of the gut
side by side, and in contact with each other. They are
partially concealed from view by the bulky atria, which also
lie—at least for the greater part of their course—upon the
dorsal lateral aspect of the intestine. At about the 30th seg-
ment of the body the two elongated and narrow sacs appear to
fuse together and form a much wider sac, which extends back
for about ten segments, and is deeply constricted at the points
where it passes through the intersegmental septa. The atria
are circular in section. The lining epithelium is, as usual,
composed of two kinds of cells; it is much folded. The
muscular layers are excessively thin. On the lower surface of
the atrium, but within the muscular coats, run two blood-
vessels, between which is a distinct thickening of the longi-
tudinal muscular coat. The two atria fuse together before
opening on to the exterior. The exact point where the sperm-
duct enters the atrium I have not discovered, but in any case
it is not at a very great distance from the external pore.
The illustration (fig. 7) will give some idea of the very re-
markable character of the spermatothecal sac. It extends
through five segments—from its beginning in the 14th to its
external aperture on the border line of Segments xvrt1/x1x,
and for two segments beyond this point—to the 21st segment,
in fact. The spermatothecal sac is single, but shows unmis-
takable indications of being the result of a fusion between
two originally separate sacs. Anteriorly it is in contact with
the wall between Segments x111/x1v. There are two diverging
234 FRANK E. BEDDARD.,
horn-like processes which meet in the middle line of this
segment. They end blindly in front. Just at the blind end
the egg-sac is attached to each whence the oviduct passes out-
wards, as shown in the figure referred to. The median sac
formed by the fusion of the two anterior sacs passes straight
down the body-wall below the ventral blood-vessel and the
nerve cord. It is a narrow tube, but is rendered conspicuous
by the fact that it gives off on either side a series of diver-
ticula. These diverticula are accurately symmetrical. They
are narrower at their origin from the median sac, and become
dilated at their free end, which is of course closed. Their
shape reminds one of the Polian vesicles of the Holothurians,
In another specimen, also mature, which I dissected, these
lateral vesicles showed an asymmetry. Those of the left side
were, with the exception of the fifth, much smaller than those
of the right side. I regard this as an abnormal or perhaps, in
spite of appearances, not a fully mature specimen. These
lateral czeca of the spermatotheca bore a distinct relation to
the metamerism of the body; there were, in fact, a pair of
them to each segment. The atria pass between the fifth and
sixth pairs. The two last pairs are beyond the aperture of the
spermatotheca. The most anterior pair of diverticula—those
which bear the egg-sacs—differ from all the rest in being a
little longer and narrower. I have investigated the minute
structure of the spermatothecal apparatus in this species by
longitudinal sections. The first point of importance to be
noted is that they contain spermatophores exactly like those
of Polytoreutus magilensis, a species recently described
by myself in this Journal.
These spermatophores do not for this reason need any
particular description. Their distribution, however, is remark-
able. They do not, as perhaps they might have been expected
to do, occur in the entire spermatothecal sac. They are
restricted to the end of the sac, which, as already mentioned,
‘lies in the 13th segment. These two sacs at the end of the
spermatothecal sacs are crowded with spermatophores. I have
also seen spermatophores at the external pore of the spermato-
OLIGOCHATA OF TROPICAL EASTERN AFRICA. 235
thecal sac, but never in between. I find it difficult to under-
stand this. The spermatothecal apparatus is so large that it
must, one would think, perform some function not performed
by spermatothecz which are smaller in size ; for instance, those
of Eudrilus. The presence of spermatophores needs more
room, supposing that there is about the same amount of sperm.
But the restriction of the spermatophores to the anterior end
of the spermatothecal apparatus is still unexplained. In Poly-
toreutus coeruleus and in Polytoreutus magilensis the
oviduct is dilated near to the opening into the egg-sac. In this
dilatation are lodged bundles of sperm. Michaelsen has sug-
gested that this looks as if the ova were fertilised in the egg-
conducting apparatus. The facts described in the present paper
support this contention, but unfortunately throw no further light
on the exact place where fertilisation takes place. The spermato-
thecal sac is lined throughout with a layer of very tall columnar
cells, the nuclei of which are near to the attached base of the
cells. In the anterior part of the spermatothecal sacs where
the spermatophores are lodged these latter are seen to lie
partly among the cells, having, as it were, thrust their way in
between them. There were, however, no indications that the
cells lining the sacs are in any way responsible for the forma-
tion of the spermatophores.
The ovaries of Polytoreutus ceruleus have been de-
scribed by Michaelsen as lying in a spherical chamber which
itself lies at the extreme end of each of the two branches in which
the spermatothecal sac ends anteriorly. The epithelium lining
this chamber is believed by Michaelsen to be the ovarian epi-
thelium. This spherical pouch is attached to the Septum
x11/x111 by a strand of connective tissue. In Polytoreutus
violaceus the position of the ovarian chamber—if I am right
in so calling it—is rather different. At the point where the
large terminal chamber of the spermatothecal pouch comes
nearest to the Septum x11/x111 there is a minute sac attached
to its wall, and formed of a muscular coat with a lining of epi-
thelium. Where this sac is in contact with the wall of the
spermatothecal sac there is no development of muscles, so
936 FRANK E. BEDDARD.
that the epithelium of both pouches is in actual contact. This
minute sac is attached to the Septum x11/xtrr by a strand of
fibrous tissue (muscular or connective tissue). It is situated on
the opposite side of the spermatothecal sac to that on which is
placed the orifice of the duct leading from the oviduct. The
sac was filled with a few small rounded cells which might be
immature germinal cells or might, indeed, be any kind of cel-
lular tissue in an immature state. I have no evidence to bring
forward that this tissue is really what is left of the ovary, ex-
cept its position and the fact that it is enclosed in a special sac.
The egg-sac and the oviduct appear to have the same struc-
ture and relations as in Polytoreutus ceruleus, The only
difference that I noted was the absence of any diverticula of the
oviduct lodging sperm masses such as Michaelsen has figured
and described in that species. The egg-sacs contained, besides
ripe ova, germinal cells in various stages of growth.
In describing the structure of Polytoreutus kilindi-
nensis I shall have some observations to offer about the de-
velopment of the different parts of the female reproductive
organs.
Polytoreutus kilindinensis, n. sp. (fig. 8).
The present species was represented by two individuals, both
of which were collected at Kilindini, on Mombasa Island ;
they were found, together with another species to be presently
described, in damp ground where the slops of a household were
deposited.
This species cannot be confounded with the next, nor with
either of the two remaining species of the genus. It is larger
than the last; the larger of the two individuals measures
120 mm. in length by a diameter of 5 mm. This specimen
consisted of 220 segments. The colour, too, is different ; it is
characterised by the same general coloration, but the violet is
less deep and less extensive.
The prostomium is broad, and does not invade in the least
the buccal segment ; the latter segment has numerous wavy
lines anteriorly.
OLIGOCHZTA OF TROPICAL EASTERN AFRICA. 237
The setz seem to be asin other species of the genus. There
seems to be a tendency for the ventral setz to be defective
upon the segments which bear the genital pores; I refrain,
however, from giving any details, since 1 am not certain how
far this deficiency may be normal.
The nephridiopores are lateral in position.
There is no median genital papilla such as occurs in P.
violaceus.
The clitellum extends over Segments x1v—xviul, being
only developed upon the anterior half of the last segment.
Behind the clitellum is a median area bounded by the ventral
setz, which looks like an extension of the clitellum ; it has the
same tumid appearance, and the boundary lines between the
segments are there uot apparent. This area reaches from the
18th to the 21st segments ; when the worm is seen in profile it
is seen to project somewhat. It is very possible that this area
serves the purpose of the genital papilla in Polytoreutus
violaceus.
The anterior of the two median and unpaired genital open-
ings is situated on the border line between Segments xvui/xvi11;
it is a widish aperture with crenated margins; behind it is a
transverse groove which runs for a considerable distance right
and left; this groove is not, as might be supposed, the boundary
line of Segments xv11/xvi11; it can be easily seen that the
furrow between these segments is anterior to it and bisects the
genital pore already described. Immediately behind the groove
referred to is the posterior genital aperture; this is much less
conspicuous than the anterior pore, and has not crenated
margins.
The internal structure of this species is very similar to that
of the last, but there are certain recoguisable differences.
There are six specially thickened septa; the first of these lies
behind the 5th segment, and the last behind the 11th segment.
As appears to be usual in the family Eudrilide, these septa are
not much connected together by threads.
The gizzard lies in the 5th segment, and is small; the
calciferous glands, of a white colour, are in Segment x11; the
238 FRANK E. BEDDARD.
unpaired calciferous pouches are in Segments 1x—x1, and are
reddish in colour.
Reproductive Organs.—The sperm-sacs are, as in the
other species of the genus, long, but they are by no means so
elongated as in Polytoreutus magilensis; they commence
in the same way in the 11th segment, and are at first thin tubes ;
in the next segment, however, they attain their ultimate size, and
extend back to about the 27th segment. The two sacs run close
together on the dorsal surface of the intestine, but they do not
become fused as is the case with Polytoreutus violaceus;
the sacs are constricted where they pass through the segments ;
their whitish colour contrasts with the orange colour of the
atria, which extend through the same segments that they do.
There is a single pair of sperm-ducts which open into the
llth segment; they show the usual dilatation before their
opening; the atria present no noteworthy particulars; they
extend as far back as do the sperm-sacs. It is in the disposition
of the spermatothecal pouches that the present species is chiefly
to be distinguished from its congeners.
Fig. 8 illustrates the arrangement of the sacs. From the ©
point of opening on to the 17th segment a narrow median
tubular sac passes forwards beneath the nerve-cord up to the
14th segment; here it divides into two sacs, each of which
immediately becomes dilated into a wide pear-shaped pouch
lying transversely to the longitudinal axis of the body; just
where this pouch narrows into the stalk which connects it with
the median spermatothecal sac a short tube arises, which very
soon dilates into the funnel of the oviduct; the latter is a
globular sac, as in other species of the genus, and is connected
on the one hand with the oviduct, and on the other with the
egg-sac, as is shown in the figure. I could find no sper-
matophores in the sac—not the least trace of the bundles
of spermatozoa figured by Michaelsen (Taf. iv, fig. 30, sk.),
and observed by myself in Polytoreutus magilensis, were
to be seen in the present species.
The spermatothecal sac of the present species is the simplest
that has yet been met with in the genus. The appendices of
OLIGOCHATA OF TROPICAL EASTERN AFRICA. 239
the median sac, that occur in the remaining species, are here
quite absent; the storage of the sperm is effected by a de-
velopment, not met with in the other species, of the anterior
end of the sac on each side.
I investigated the structure of the female reproductive
organs in an immature example of this species. The sper-
matothecal sac was almost filled by small rounded cells,
quite unlike the tall columnar cells which line the mature
spermatothecal sac of Polytoreutus violaceus. So nume-
rous were these cells that the lumen of the entire tube was
almost completely obliterated. I had hoped to find some
indication of the position of the ovary in this specimen; but,
unless the epithelium lining the two end pouches into which
the spermatothecal sac divides anteriorly is the germinal
epithelium, I could find nothing at all. There is not, as there
is in P. violaceus, a small sac attached to the main sperma-
tothecal sac, set apart for the lodgment of the ovary. It is
probable that the ovary is only free in very young specimens,
and it is also possible that it has a very transitory existence.
There is a precedent for this in Libyodrilus. In that genus
the ovary appears to exist only for a short time, its contents
being early transferred to the interior of the egg-sacs. In this
young specimen of Polytoreutus kilindinensis the egg-
sacs were quite fully developed as regards size, but they
contained only quite immature cells ; the germinal cells filling
the egg-sacs were exactly like the cells in the immature ovary
of other worms; it is evident, therefore, that the germinal
tissue is transferred en masse to the egg-sac, and that the
entire development of the ova goes on in those sacs. The
spermatothecal sac seems not to be formed by an invagination
from the epidermis; the epithelium lining it bears no resem-
blance at all to the epidermis; the structure of the sac is
exactly like that of the developing sperm-sac which lies in the
preceding segment; in the section of the worm the two could
be very well compared, as they almost came into contact.
Judging from structure only, no one would hesitate to regard
the two structures as of the same nature.
VoL. 36, PART 2.—NEW SER. R
240 FRANK E. BEDDARD.
§ Note upon an Immature Example of this Species.
From the same locality I have an example of a Poly-
toreutus which I regard as an immature specimen of P.
kilindinensis. It measured 98 mm., and consisted of 158
segments. The external characters, except those afforded by
the apertures of the reproductive organs, were as in the species
P. kilindinensis. The reproductive openings were only
represented by a pore upon the boundary line between Seg-
ments xvui/xvi1l. This pore was, however, extremely conspicu-
ous, and showed no indications of being in an immature con-
dition. Behind the pore was a short transverse groove such as
exists in the mature worm, but there were no signs of the
second pore. The internal anatomy, apart from its immature
state, showed one or two small differences from that of the
species of which I presume the present example to be an im-
mature one. There was a thick septum in front of the gizzard,
which, therefore, separated Segments 1v/v. The dorsal vessel
also was double, a rare condition in this family. The dorsal
vessel was formed of two tubes, at any rate in Segments yi11—
x11. The two tubes became fused at the septa. There are
other examples beside the present which appear to show that a
double or a single dorsal vessel is not necessarily a diagnostic
character of a species. Thus I described the dorsal vessel of
Megascolex ceruleus as double, while Bourne saw no signs
of any such doubling. Another point in which the present
specimen differs from the type of the species is in the fact that
the two sperm-sacs join together at their distal extremity.
The median calciferous pouch of Segment x1 was distinctly
smaller than the two which precede it.
The sperm-sacs, it should be said, are of precisely the same
form as in the mature examples; that is, they arise from the
septum bounding posteriorly the 11th segment. At first each
sac is thin, and this region extends through one segment only.
It may, therefore, be pointed out that the probability of this
shape of the sperm-sacs being characteristic of the species and
distinguishing it from, for example, Polytoreutus magi-
OLIGOCHATA OF TROPICAL EASTERN AFRICA, 241
lensis is increased by the fact that the immature worm
exhibits the same condition. The funnels of the sperm-ducts
were by no means so prominent as in the fully mature worm;
they, of course, occupied the same position. The only other
part of the male apparatus that was visible were two little
sacs lying one on either side of the extremity of the spermato-
thecal sac. Ido not think that these were the immature sacs
of the penial setz of a species of Polytoreutus provided with
these structures. It seems, therefore, that the first rudiment
of the terminal apparatus of the male ducts is double, which is
so far a demonstration that originally this apparatus was
double. Beyond these two minute sacs nothing was visible of
the male efferent ducts. On the other hand, the spermatothe-
cal sac was as well developed as in the mature worm. It
showed no differences that I could detect from the structure
already described. There was, however, no sperm in the sac.
We may, therefore, note that the female apparatus is de-
veloped before the male, and that the sperm-sacs are the first
part of the male apparatus to reach maturity.
Polytoreutus Finni, n. sp. (figs. 6, 17).
I have unfortunately only a single specimen of this worm
for examination. It is extraordinarily long and thin—perhaps
I may say even for an Eudrilid. The preserved specimen
measured 183 mm. by 3 mm. in breadth at the clitellum, which
is distinctly broader than any other region of the body. The
worm consisted of rather more than 500 segments, a most
unusual number. The clitellum is exceedingly conspicuous,
being raised above the level of the surrounding segments; it
occupies Segments x111—xvill. It is quite complete except
for the area which lies between the genital pores. The genital
pores (fig. 17) are, as is usual with this genus, situated on
the 17th, and on the boundary line between the 18th and the
19th segments respectively. These apertures are very large
and prominent, and are surrounded by thick tumid lips. The
integument at the actual orifice is marked by numerous slight
furrows which have a radiate arrangement. These pores,
242 FRANK E, BEDDARD.
together with the modified integument immediately surround-
ing them, occupy nearly the whole of the ventral surface of
the worm. Between the anterior and the posterior orifice is a
tract of integument of the same character as that which imme-
diately encircles the pores, and differing from the clitellar
tissue.
The sete are as in the other species of the genus.
The internal structure, no less than the external charac-
ters, distinguishes Polytoreutus elongatus from the re-
maining species of the genus Polytoreutus. These dif-
ferences mainly concern the spermatothecal sacs, which are
different in all the species of the genus. In other particulars
there is less difference. The last specially thickened septum
divides Segments xt and xu. The last pair of hearts are in
Segment x1. There are the usual three median calciferous
pouches in Segments 1x, x, and x1. The calciferous glands
are present, but have a very unusual form; they appear to
lie in the 15th segment, but I am not able to be quite certain,
as the segments just about this region were hard to fix. Not
only are the calciferous glands unusual by reason of their
position, they are also peculiar in shape. Lach gland is
situated at the sides of the cesophagus, and is curved up like a
ram’s horn.
As in the other species of the genus there is but a single pair
of funnels; andI presume, though I have not actually verified
the fact for the present species, they have only a single pair
of testes. The funnels of the sperm-duct lie in the llth
segment, and the funnel itself is preceded by a dilated section
of the sperm-duct, which has an opaque white appearance, and
is of large size. It is directed obliquely backwards. I have
not followed the course of the sperm-duct.
The atria are long; they open together into a terminal
bulbus which is median in position, and again opens on to
the exterior by the anterior of the two genital orifices already
described. The atrium belonging to the right side of the body
was extended at full length, while that of the opposite side
was looped once or twice. The fully extended atrium reached
OLIGOCHATA OF TROPICAL EASTERN AFRICA. 243
back to about the 25th segment behind the clitellum; a
peculiarity about it was the fact that the last half or rather
_ less of the gland was double, the two portions, however,
running in close contact. Whether, as seems likely, this is a
mere abnormality I am unable to say; but I may point out
that in Eudrilus each atrium is normally divided into two
separate tubes by a continuous longitudinal septum.
The sperm-sacs are very remarkable. On the dorso-
lateral surface of the intestine I observed a pair of fine tubes
running a fairly straight course, which I put down at first as
being the sperm-ducts, thinking that they terminated in the
atrium. They do as a matter of fact terminate close to the
atria, but quite independently of them. These slender tubes
are the sperm-sacs. One of them, that of the left side, was
distinctly varicose, being dilated here and there into oval
chambers. Traced forwards, they appeared to end in the
immediate neighbourhood of the funnels. Hach sperm-sac
was accompanied by a blood-vessel. It is a pecularity of this
genus to possess long sperm-sacs, which in Polytoreutus
magilensis are of enormous extent, but in no other species
are they of the extreme tenuity exhibited by Polytoreutus
Finni. This state of affairs may be simply due to the fact
that the sperm did not happen to be present in any great
amount; but this is unlikely, as the worm was in all other
respects fully mature. Besides, this is not the only case of a
worm possessing such extraordinarily long and thin sperm-
sacs. I have described elsewhere the sperm-sac of the Geos-
colicid genus Trichocheta, which are of precisely the same
character as those of the present species of Polytoreutus.
On the other hand, it will be recollected that there are a
number of different degrees in the development of the sperm-
sacs in this genus which may perhaps be interpreted as dif-
ferent grades of development of the sacs. In Polytoreutus
magilensis the sperm-sacs are at first extremely narrow,
and later become much wider. In Polytoreutus kilindi-
nensis, described on a preceding page of the present paper,
the narrow region of the sperm-sacs is reduced greatly, nearly
244 FRANK E. BEDDARD.
the whole of the sacs being wide. Finally in the present
species we have the other extreme. The entire sperm-sacs
are formed by the slender tubes referred to.
The female reproductive organs present a fourth variety, all
the species at present known being different in the form of
these organs. They are most like those of Polytoreutus
kilindinensis. P. Finni agrees most closely with P. kilin-
dinensis in the general form of the spermatothecal sacs. As
in the last-named species, there are only a single pair of diver-
ticula of the median unpaired sac. The latter runs beneath
the nerve cord until it reaches the anterior boundary of the
bulbus of the male efferent apparatus. Arrived at this point
it diverges to the left, and, forming a semicircle, again bends
to the middle line, and opens by means of a dilated terminal
sac behind the orifice of the atria. Anteriorly this median
sac extends as far as the 18th segment. Just below the septum
which divides this segment from the one in front, it divided
into two. Each branch swells out as in P. kilindinensis
and forms a largish oval sac. The two sacs are coiled to some
extent round the intestine. From the base of each, not far
from the point where it joims the median sac, a short tube is
given off, which passes into the receptaculum ovorum and
thence becomes continuous with the oviduct. The arrange-
ment of these parts is, in fact, precisely as in Polytoreutus
kilindinensis. They are illustrated in fig. 6.
Alluroides Pordagei, n. gen. (figs. 4, 5).
I shall describe this new form under the name of Allu-
roides Pordagei. It was collected along with a number of
examples of Stuhlmannia variabilis in a swamp four miles
up country, opposite to Mombasa Island. The species is re-
presented by only two individuals, measuring in the preserved
state about an inch in length. They had a delicate appear-
ance owing to their small size and the thinnish body walls,
and resemble somewhat, except in colour, an aquatic member
of the family Phreoryctidee which I have lately described from
New Zealand under the name of Pelodrilus violaceus. In
OLIGOCHATA OF TROPICAL EASTERN AFRICA. 245
fact, any one acquainted with this group of worms would pro-
bably assign the species from its general appearance to the
Lumbriculide or perhaps to the Tubificidee.
Reproductive organs.—The testes are a single pair
only, which are placed in the 10th segment attached to the
front wall of that segment. There appear to be no actual
sperm-sacs, but the 10th and the Lith segments are filled with
a mass of developing sperm. ‘This is so compacted together
that the appearance of a definite sac is produced, and the
sperm is so abundant and occupies so much of the interior of
the two segments in question that the septum dividing them,
which is thinnish, is hardly visible without a very careful
inspection.
The funnels of the sperm-ducts correspond in number to
the testes, that is to say there is only a single pair, which les
opposite to the testes in the same segment. They are much
folded.
The terminal apparatus of the male efferent duct is formed
by an atrium.
The atria (fig. 5) extend through more than one segment,
and are long enough to be coiled. They open on each side
on the 13th segment, the aperture being lateral in position,
showing therefore, which is remarkable, no relation to the
pores of the spermatothece. The tubular atria have, however,
not a close resemblance in structure to the tubular atria of
such generaas Acanthodrilus. Their structure is as follows:
—The internal lining of the tubes is formed by a single layer
of cells, which have a clear appearance, as they were not
stained by a long immersion in borax carmine. The cells
were certainly in some places ciliated. Towards the external
pore these lining cells got to be more and more like the epi-
dermic cells, and were also ciliated, until at the actual orifice
they became continuous with the epidermis. Outside the epi-
thelium is a layer of muscular fibres of some thickness. These
fibres are entirely circular in disposition. They do not form
an absolutely continuous covering of the epithelium; here
and there slight gaps are to be seen. These gaps correspond
246 FRANK E. BEDDARD.
to the exits of the ducts of a mass of glandular cells which
form the outer covering of the organ. As in Moniligaster,
the atrium is invested externally by a mass of pear-shaped
cells, which are loosely compacted into separate masses.
The structure of the atrium, therefore, is like that of Moni-
ligaster alone among “earthworms.” In fact, it only differs
from the atrium in that genus in its greater length. The
external aperture of the atrium is placed upon a fan-shaped
outgrowth of the body wall, which in all probability serves as
a penis. Whether or not these penes are in- and evaginable I
am unable to say. They were extruded in both the specimens
at my disposal. I should imagine that they are protrusible.
The ovaries are in the 13th segment.! They are attached
as usual to the front wall of this segment. From the 13th to
about the 20th, there are ova and masses of ovarian cells appa-
rently lying loose within the body cavity. In the most ante-
rior of the segments in question, there are only egg masses
consisting of immature ova surrounded by groups of small
cells, but-in the segments situated further back, there were
only ripe ova visible. These ova are of special interest on
account of their large size; they are also, like the ova of the
aquatic Oligocheta, generally crowded with yolk. The ova
are fully as large as those of such a genus as Rhynchelmis.
In longitudinal sections of the body the ova reached across
nearly from one side of the body to the other.
I could find no egg-sacs. There is a single pair of oviducts
which open into the 13th segment. The tube remains very
wide after it has entered the 14th segment, and has a much
folded lumen; it narrows rapidly before the external pore.
The spermatothece are present to the number of a single
pair, which are in the 8th segment. They are oval pouches
without any diverticula, and with perhaps unusually thick
1 The septum dividing Segments x111/xrv was largely deficient, and masses
of young egg-cells and non-differentiated germinal cells passed into the 14th
from the 13th segment. I should not like to be certain that these were not
developed in situ; younger specimens are required to clear up the matter.
In the meantime there is only one pair of oviducts.
OLIGOCHATA OF TROPICAL EASTERN AFRICA. 24,7
muscular walls. They open quite dorsally close to the median
dorsal line of each side of the body. This is a remarkable
but not unknown position for the spermatothecal pores.
Another instance of a similar position, which occurs to me,
is in the species Allolobophora fetida.
Other facts in the anatomy of the worm which are of some
little importance are the following.
The prostomium is, as is indeed usual, covered with a
thick columnar epithelium. This thickened pad is prolonged
for a very short distance into the mouth-cavity. This epithe-
lium is very possibly of a sensory nature. The ccelom is, of
course, divided up by transverse septa into a series of chambers.
Some of the septa which divide these chambers are thicker
than others. The first of these thickened septa divides Seg-
ments 1v/v. The following seven septa are, with the exception
of that which divides Segments x/x1, also thickened. The next
septum to the last of the specially thickened septa is rather
thicker than the excessively fine septa which separate the
following segments. As in so many of the lower Oligocheta,
there are septal glands present; these glands commence in
the present species in the 5th segment, and the last pair were
observed in the 9th. The brain hes in the 3rd segment.
From the brain one among several nerves which pass forward
ends in a medianly situated ganglion in close juxtaposition to
the epithelium of the prostomium, which consists of but few
cells. A median ganglion in this position has not, I believe,
been described as existing in any earthworm, but it has been
met with in certain aquatic Oligochzeta belonging to the family
Tubificidee. In this family Stolc! has figured such a ganglion
in Bothrioneuron and in Lophocheta.
The alimentary tract has no traces of a gizzard. The
cesophagus does not appear to be at all vascular; it termi-
nates in the 18th segment, in which segment begins the
intestine. There are no glands of any description appended
to the alimentary tract unless the septal glands can be referred
to this category.
1 “ Monogr. Ceskych Tubificidu,” ‘Abh. bohm. ges.,’ 1888.
248 FRANK E. BEDDARD.
The nephridia commence in the 16th segment. They open
on to the exterior by the second seta. They are clothed with
a thick layer of peritoneal cells.
The question now to be considered is the family into which
this new type should be placed. That it is generically distinct
there is in my opinion no doubt whatever. This genus Allu-
roides is one of those forms which render the distinction
between the old groups of the “ Limicole” and the “ Terri-
cole”? untenable. In some respects it is even more perfectly
intermediate than Moniligaster.
Aside from Moniligaster, the “waterworms,” all of them,
differ from any earthworm in the following characters :
1. Clitellum one cell thick.
2. Ova very large and full of yolk, few in number.
3. Genital aperture situated far forwards.
4, Egg-sacs occupy more than one segment.
These are positively all the distinguishing marks if we leave
aside the genus Moniligaster. Moniligaster itself, as I have
shown in several papers! dealing with the structure of this re-
markable worm, breaks down the first, third, and fourth of the
above distinctions. Moreover, it has eggs which, although they
are not greatly above the average size of the eggs in earthworms,
differ from those eggs in containing a great quantity of yolk in
the form of large spherules. Moniligaster, in fact, is only
an earthworm in having a gizzard or rather gizzards, and in
the comparative thickness of the body-wall. This latter
character, however, is seen in Phreoryctes, which is one of
the genera assigned by Claparéde to his division Limicole.
Besides the points enumerated in the above tabular statement,
Moniligaster has various other resemblances to several
Limicolous Oligocheta, which are not of first-rate importance
from the present point of view, inasmuch as they also occur in
other earthworms, though not to so marked a degree. The
atrium, for example, is almost exactly like that of the Lumbri-
culide. ‘'The protrusible penis is constructed more on the lines
1 For a list of literature see ‘‘ Description of New or Little Known Earth-
worms from various localities,” ‘ Proc. Zool. Soc.,’ 1892, p. 690.
OLIGOCHATA OF TROPICAL EASTERN AFRICA. 249
of the corresponding organ in the Tubificide than is the penis
in the Eudrilide and other terrestrial Oligocheta which
possess an organ of this kind. The atrium is lined by a single
layer of cells, a feature which is also found in Ocnerodrilus
and in some other genera. The sperm-duct has a very short
course, opening on to the exterior in the next segment to that
which contains the internal funnel. It is quite possible, how-
ever, that Tetragonurus shares this peculiarity with M onili-
gaster. So, too, with the oviducal pores. Moniligaster is
not alone among earthworms in the fact that they are in front
of the sperm-duct pores. In both Allurus and Tetragonurus
the oviducal pores are nearly certainly (in Allurus quite cer-
tainly) in front of the sperm-duct pores. Dr. Rosa has
rebuked me for laying much stress upon the fact that the male
pores of Moniligaster are so far forward as the boundary
line of Segments x/x1, a position which of course recalls the
very anterior position of the corresponding apertures in the
generality of the “ Limicolz.” It is true that the difference
between an opening upon the 10th and 11th segments and one
upon the 12th segment is not a very great one, but the differ-
ence, such as it is, is in the direction of the lower Oligocheta,
and not in the reverse direction.
So much, then, for Moniligaster. The only point in
which it differs in an important way from the aquatic forms is
in the relatively small size of the ova. The Annelid which
forms the subject of the present communication is the only
known example of an Annelid with marked affinities to the
terrestrial Oligochzeta which has that hitherto distinctive
character of the lower Oligocheta—large ova filled with yolk.
It resembles the aquatic Oligocheta in the following points:
1. Clitellum consisting of a single layer of cells.
2. Ova very large and full of yolk, few in number, and
occupying several segments.
3. Atrium lined by a single layer of epithelium, and covered
by masses of pear-shaped cells ; sperm-ducts open into it.
4. Longitudinal muscular layer of body-wall consists of a
single row of plate-shaped fibres.
250 FRANK IE. BEDDARD.
The above resemblances are in structures which are, with
the sole exception of Moniligaster, confined to the aquatic
Oligocheta. Besides these, the genus Alluroides departs
from the usual structure of the terrestrial Oligocheta in a few
other points, viz. :
1. There is no gizzard, no calciferous glands, and no typhlo-
sole.
2. The nephridia are deficient in the anterior segments.
3. There is no subnervian vessel. ;
These points do not absolutely distinguish the terrestrial
from the aquatic Oligocheta, but they occur in a few of the
former while characteristic of the latter. For example, there
is no gizzard in certain species of Microscolex; Ponto-
drilus has no calciferous glands, &c.
The points in which the present genus resembles the terres-
trial Oligocheta are by no means numerous. They are as
follows :
1. The segments occupied by the clitellum.
2. The position of the male pores, and the fact that the sperm-
duct traverses several segments on its way to the external pore.
3. The situation of the ovaries in Segment x1it.
In addition to these, there are some points in which Allu-
roides agrees with earthworms to differ from the majority of
the lower Oligocheta. The sperm masses in Alluroides are con-
- fined to the 10th and to the 11th segments; it is the rule among
the lower forms for the sperm-sacs to extend much further
back. The testes being limited to the 10th segment is rather
unusual among earthworms. When there are but a single pair
of these gonads they are, as a rule, in the following segment.
Among the Lumbriculide the testes are in the 9th segment,
or, as in Rhynchelmis—and possibly in other genera,—in the
9th and 10th. However, in Phreoryctes the testes are in
Segments x and x1, but here the sperm-ducts open on to the
exterior in the following segments. The same is the case with
the nearly allied Pelodrilus.!
1 « Anatomical Description of Two New Genera of Aquatic Oligocheta,”
‘Trans. Roy. Soc. Hd.,’ 1890.
OLIGOCHATA OF TROPICAL EASTERN AFRICA. 251
While we may, as it appears to me, term Moniligaster an
earthworm with numerous points of affinity to the “ water-
worms,” it is better to speak of Alluroides as a “water-
worm”? with affinities to the terrestrial worms. If an exchange
could be effected between these two genera of various characters,
we should get as a result either an obviously terrestrial genus
or an equally obviously “ Limicoline” genus. Thus Allu-
roides would be undoubtedly referable to the terrestrial sec-
tion of the Oligocheta if it possessed the body-wall and the
ova of Moniligaster; on the other hand, Moniligaster
would be an undoubted “waterworm” if we could transfer to
it the body-wall and the ova of Alluroides.
It is therefore, in my opinion, useless to attempt any
comparison with any particular family of terrestrial Oligo-
cheta; it is rather with some family of the aquatic Oligocheta
that Alluroides should be compared; be it noticed, however,
that, judged by external characters only, Alluroides would
probably be referred to the immediate neighbourhood of
Allurus.
The family of “ waterworms” with which Alluroides has
the closest affinities is that of the Lumbriculide. It agrees
with that family in the following characters :
(1) Setze paired and S-shaped.
(2) Atrium with thick peritoneal investment.!
(3) The great depth of the single layer of longitudinal
muscular fibres.
These two characters are found together in the Lumbriculide
alone among the aquatic Oligochzeta; in other respects, however,
there are not any striking resemblances between the genus
Alluroides and the Lumbriculide.
Two of the most characteristic features of this family are
wanting in Alluroides; these are (1) the absence (?) of the
vascular contractile ceca, and (2) the absence of a second pair
1 T have shown that in Moniligaster the cells enveloping the atrium are
prolonged through the muscular layer and epithelium to open into its lumen;
Vejdovsky’s figure (‘ Zeitsch. wiss. Zool.,’ Bd. xxvii, pl. xxiv, fig. 3) seems to
show that this is also the case with Rhynchelmis.
252 FRANK E. BEDDARD.
of sperm-ducts. As to the latter point, I have discovered that
in Sutroa the second pair of sperm-ducts are much thinner
than the first pair, and that coincidently with commencing
disappearance (?) of one of the two pairs of sperm-ducts the
testes belonging to the vanishing pair are absent. In my genus
Phreodrilus? there is a cecum of the sperm-duct, which is
possibly a still further reduced condition of a second pair of
sperm-ducts. Among the higher Oligocheta the absence of
one pair of testes and of the corresponding sperm-duct is not a
matter upon which great weight is usually laid. In any case it
appears to me that Alluroides shows no marked affinities to
any other family of worms.
Alluroides, gen. nov.
Setz simple, S-shaped, arranged in pairs; clitellum occupy-
ing Segments x11I—xv1, consisting of a single layer only of
cells; alimentary canal without a gizzard or any appended
glands ; some of anterior septa thickened ; testes, one pair in
X ; sperm-ducts open on to exterior on x111 through a moderately
long atrium, which has much the same structure as in the
genus Moniligaster; above the apertures of the atria is a
process of the body-wall (a penis?) ; ovaries in x1; ripe ova of
large size, and filled with yolk, occupy five or six segments of
the body; oviducts open on to Segment xiv; spermatothece,
one pair, without diverticula, in viii.
The genus contains one species, Alluroides Pordagei, of
which I shall not attempt a definition.
1. Gordiodrilus zanzibaricus, n. sp.
A large number of specimens of this species were collected
from damp mud at the edge of a pool. They are, when
preserved, an inch or so in length. Their colour during life
was red.
1“ A Contribution to the Anatomy of Sutroa,”’ ‘Trans. Roy. Soc. Ed.,’
vol. xxxvil, p. 195.
2 « Anatomical Description of Two New Genera of Aquatic Oligocheta,”
‘Trans. Roy. Soc. Hd.,’ vol. xxxvi.
OLIGOCHATA OF TROPICAL EASTERN AFRICA. 253
The setz are strictly paired, and are not in any way orna-
mented. The pairs are equidistant, and are all of them
decidedly ventral in position. The only modification of the
sete occurred on the 17th and 18th segments. On both of
these segments only one ventral seta was present on each
side; the remaining seta appeared to be the outermost of each
pair.
The atrial pores are two pairs, which open on both 17th and
18th segments. Lach pore is situated in a groove with raised
and somewhat folded margins, which connects the two pores of
each side.
The oviducal pores lie alittle to the outside of the outermost
seta of each ventral pair.
The spermatothecal pores occupy a corresponding position
between Segments vi1/ViE, vitt/Ix.
The clitellum occupies Segments x1v—xvitt, and is complete
except over the area lying between the atrial pores.
The nephridiopores lie in front of the inner seta of the outer
pair.
The alimentary canal shows the structure which is charac-
teristic of this genus.
In Segment 1x is the median ventral calciferous gland, which
to describe would be merely to recapitulate my description of
other species. From the 9th segment the walls of the cesopha-
gus are highly vascular; the intestine begins in the 13th
segment; the intestine is not at first so regularly constricted
in successive segments as it is posteriorly, and appears to be
of a rather wider calibre. The ciliation of the cesophagus
commences just in front of the opening of the calciferous
pouch.
The first distinct septum separates rv/v; this septum is thin,
but the four following are thickened ; the next three se pta,
though thinner than those which precede them, are thicker than
those which follow.
As in other species, there are masses of unicellular glands in
the neighbourhood of the pharynx, which have been termed by
myself and others “septal glands,” on the view that they
254 FRANK E. BEDDARD.
correspond to the septal glands in certain genera of aquatic
Oligocheta. These glands extend back as far as the 7th
segment.
The nephridia are paired structures. The first pair lie in
the 5th segment; they are not absent in any of the genital
segments; but in the llth, 12th, and 14th segments the
nephridia are more or less rudimentary. That they are present
can be made out without any difficulty, for the large vesicular
cells which clothe the nephridia from the 9th segment onwards
can be readily seen.
The degeneration of the nephridia in these segments must,
as it appears to me, be correlated with the development of the
genital ducts, or rather their funnels; so complete is this
degeneration in the case of the nephridia of Segment x1v that
nothing is left but a mass of vesicular cells to tell of the former
existence of a pair of nephridia in this segment.
As to the reproductive organs, the testes lie in Segments x
and xt, in which are also to be found the funnels of the sperm-
ducts;. there is nothing unusual in either their structure or
their position. The same segments, with the addition of the
12th, contain the sperm-sacs. The sperm-ducts and atria are
precisely like those of the West African Gordiodrilus
elegans, and call, therefore, for no particular remark. The
ovaries are in Segment x111, and there is nothing remarkable
about them or their ducts. There are however, and I have
not yet observed this in the genus, egg-sacs in Segment xtv.
The spermatotheca are in Segments viii and 1x.
§ Calciferous Glands in the Eudrilide.
I have studied with care the calciferous glands in two of the
species described in the present paper, viz. Eudriloides
Finni and Stuhlmannia variabilis. These two species
and another which I have lately described in a paper com-
municated to the Zoological Society, and named Eudriloides
durbanensis, show a peculiar form of these glands which
present various points of interest. I find also that Eudri-
OLIGOCHATA OF TROPICAL EASTERN AFRICA. 259
loides brunneus has glands of a similar character, but I
have not studied them in detail. Michaelsen has not men-
tioned the existence of these glands at all in the two genera
referred to. In fact, he distinctly states them to be absent in
the following genera:—Eudriloides, Notykus, Stuhl-
mannia, Megacheta, and Metadrilus. It is not the case
that calciferous glands are absent from at any rate two of the
genera mentioned in the above list. But the glands are so
little like the usual form of these glands in the Eudrilide that
it is not at all surprising that their existence has been over-
looked.
In Eudrilus and in other genera there are a pair of calci-
ferous glands in the 13th segment or thereabouts, which recall
in every particular the calciferous glands of other earth-
worms. In addition to these there are in Eudrilus, Polyto-
reutus, Heliodrilus, and Hyperiodrilus unpaired median
pouches which agree with the calciferous glands in structure,
and are clearly to be referred to the same category. The
only ways in which these glands differ from calciferous glands
are—(1) their unpaired character—which I am not able to
regard as of importance, and (2) the excessive complication of
the folded interior of the organ, which is so developed in some
forms that the lumen becomes partially intra-cellular. Of
these “‘ Chylus-Taschen,” as Michaelsen terms them, there are
never more than three. In longitudinal and transverse sec-
tions of Stuhlmannia, Segments vi—x1r are largely occupied
by whitish masses on either side of the intestine. These have
a paired arrangement, there being a pair to each segment.
The shape of these masses is more or less irregular. They
are roughly oval with indented margin, as shown in fig. 12;
they have in certain regions the form of a coiled tube, the
individual coils being closely pressed together. ‘The diameter
of each gland varies at different points. The white colour of
the glands appears to be due to the presence of innumerable
rounded granules which make up the tissue. ‘hese granules
suggest the yolk spherules of ova. Lying among them are a,
comparatively speaking, limited number of small darkly stain-
VOL. 36, PART 2.—-NEW SER. S
256 FRANK E. BEDDARD.
ing bodies which I take to be nuclei. There is, however, no
trace of cell limitations corresponding to the nnclei. The
whole gland is covered with a darkly staining but thin sheath.
Through the middle of the gland runs a stout blood-vessel,
which I found to be nearly everywhere filled with coagulated
blood. Its track through the gland was therefore not difficult
to follow. The blood-vessel runs through the gland from end
to end, but gives off very few branches; it is, however, of
such great width as compared with the gland that any branch-
ing seems to be unnecessary for the adequate blood-supply of
the surrounding tissue. The vessel belonging to each gland
could be traced in three directions. Anteriorly it leaves the
gland, and traversing the septum becomes the blood-vessel of
the corresponding gland of the segment in front. Posteriorly
the same thing happens. The successive glands of one side of
the body are therefore, as it were, strung upon a continuous
lateral vessel. In the middle of each gland a large branch
arises from this which communicates with the plexus round
the cesophagus. In the case of the last pair of glands, those
belonging to the 12th segment, I observed two such branches
communicating with the peri-cesophageal blood-plexus of which
one was distinctly smaller than the other. Nearly the whole
of the gland is made up of the peculiar tissue described and
illustrated in figs. 12—15. In places where the gland was
very slender in dimensions, this tissue has taken on a decided
resemblance to columnar epithelium, faint lines of demarcation
between the cells being apparent. Here the appearance pre-
sented is that of a tubular gland, but the lumen of the gland
is filled with blood. Although these structures are solid,
excepting for the blood-vessel which occupies so large a por-
tion of their interior, they are not without communication
with the gut. Transverse sections show the nature of their
connection with the cesophagus better than longitudinal sec-
tions. A transverse section is illustrated in fig. 12. On the
ventral surface of the cesophagus a pair of moderately long
cellular tubes arise close together from the lining epithelium
of the alimentary tract. These diverge, and each follows the
OLIGOCHAHTA OF TROPICAL EASTERN AFRICA. 257
course of the blood-vessel destined for the supply of the gland.
It terminates abruptly in the calciferous gland. Iam not at
all certain that these tubes really possess a lumen; it is at
least very inconspicuous, and also the duct is not always so
long as that represented in the figure quoted. The aperture
into the gut of a second gland is shown in fig. 11. Here it
will be seen the duct of the gland is excessively short, and it
appears to become solid a very short distance from the point
whence it arises from the gut.
Concerning the nature of the peculiar tissue which makes up
the greater part of the calciferous gland it is very difficult to
speak positively. In several preparations from specimens
which had been preserved with Perenyi’s solution, the layer of
peritoneum surrounding the esophagus appeared to pass with-
out a break into the tissue of the gland. The appearance of
this tissue is, indeed, more suggestive of peritoneal cells than
_ of epithelial cells derived from the intestine. On the other
hand, sections of a worm that had been killed and preserved in
gradually increasing strengths of alcohol did not show any
such gradual passage as has been indicated, for in these sections
the peritoneum clothing the intestine was coloured of a
greenish tint, and there was a sharp demarcation between this
tissue and that forming the bulk of the calciferous glands. Not-
withstanding this fact, the tissue in question has more likeness
to peritoneal than to any other tissue in the worm’s body.
The only possible alternative, as it appears to me, is to assume
that the cells have retained their embryonic state. In the
embryo (not of this species, which is unknown, but of others)
the cells of the mesenteron are charged with yolk spherules
exactly like those in the gland tissue of the calciferous glands
of the adult Stuhlmannia. The structure of the glands in
this species is not very widely different from what I have
described in Notykus (?) durbanensis in a paper recently
published in the ‘ Zoological Society’s Proceedings’ (1892).
It appeared to me, however, that the lumen of the glands in
that worm were rather more developed than in Stuhlmannia,
The lumen was quite obvious, though of little extent. The
258 FRANK E. BEDDARD.
glands, too, were of a more regular form, and showed no
modification of the peculiar cells of the gland such as occurs in
places in the gland of Stuhlmannia. I have now to record the
structure of the corresponding glands of Eudriloides Finni.
In this worm the glands are rather different in the details of
their histology. There are pairs, but there is not always an
absolute separation between the glands of adjacent segments.
In a few cases I have found that there is a communication
from segment to segment. The tissue composing the glands is
for the most part exactly asin Stuhlmannia. The glands,
however, are more irregular in form and the blood-vessel is
much more coiled ; where it (the blood-vessel) leaves the gland
the tissue surrounding it is reduced to a comparatively thin
layer. As the vessel with the surrounding tissue is much coiled,
the appearances of a transverse section through a portion of
the gland are much as is shown in fig. 15. This section pre-
sents a most curious resemblance to a section through the
thyroid gland. It has every appearance of tubes of columnar
epithelium surrounding a lumen which is filled with a homo-
geneous secretion ; this “secretion” is nothing but blood. I
have, of course, traced the supposed blood-vessels into connec-
tion with the vascular system. The modification of the tissue
of the gland is not gradual; here and there it suddenly passes
into the tissue illustrated at c in fig. 15. The tissue in ques-
tion stains much more darkly than the rest of the gland, the
granules in the cells which compose it are disposed in a radiate
fashion, and the cells have acquired a columnar appearance, of
which indications are observable in Stuhlmannia, as I have
already pointed out. The specialisation of the cells is much
more marked in the present species. It will be clear, at least
from the figures which illustrate the foregoing description,
that the glands which I call “calciferous” are not only different
in the three genera referred to from those of other Eudrilids,
but are also different—very different—from the corresponding
glands of nearly all other earthworms. The only genus which
at all approaches these Eudrilide in the structure of its calci-
ferous glands is my genus Gordiodrilus. In all the species
<OLIGOCHATA OF TROPICAL EASTERN AFRICA. 259
of this genus there isa single calciferous gland or rarely a pair of
these glandsin the 9th segment ventral in position. The genus
Gordiodrilus is mainly an African genus; it has been found
in West Africa, and I describe in the present paper a species
from Zanzibar. Gordiodrilus has no marked affinities to the
Eudrilidz, and for the present I place it in that unsatisfactory
family the Cryptodrilide. The only point of resemblance to
the Eudrilide is in the median and unpaired calciferous gland.
This gland is a diverticulum of the csophagus, which is sur-
rounded by a mass of tissue exactly like that which makes up
the greater part of the glands in Stuhlmannia and Eudri-
loides. The cesophageal diverticulum, however, passes from
end to end of the gland, and expands at its blind extremity into
a network of fine tubes having an intra-cellular lumen and
bearing the strongest possible likeness to nephridial tubes. This
genus is noteworthy from the present point of view as furnish-
ing an intermediate condition between the calciferous glands of
the more typical earthworms and those of the genera Stuhl-
mannia, Notykus, and Eudriloides. The lumen con-
nected with the csophagus is reduced in extent and is not
folded, while at the same time the peritoneal covering is greatly
increased in importance. The next stage is furnished by
Notykus. Stuhlmannia seems to me to have a still more
reduced cesophageal diverticulum. Finally, in Eudriloides
I could not detect any diverticulum at all. In this species the
walls of the esophagus were much folded, so that a short diver-
ticulum, if it exists, would be less conspicuous than in Stuhl-
mannia. As the extent of the epithelial diverticulum of the
cesophagus is lessened there is a corresponding increase in the
amount and also in the specialisation of the peritoneum-like
tissue which surrounds it. Already in Stuhlmannia there
is a commencing conversion of some of these cells into a
definite layer bordering the blood-vessels in certain regions.
In Eudriloides the amount of this specialised tissue is
increased and the specialisation has gone further.. It appears
to me that this remarkable change in the histological characters
of glands, which I cannot but consider to be the homologues
260 FRANK E. BEDDARD.
of the calciferous glands, must indicate a change in func-
tion. |
In the present state of our knowledge, we can do no more
than guess what this change of function can be. We are
helped, however, by certain facts in the histology of the glands,
and by the analogies offered in other animals. The structural
change undergone by the calciferous glands is a reduction of
the lumen, and presumably, therefore, a rapidly decreasing
amount of secretion furnished to the csophagus. I have
never, it should be mentioned, seen the least trace of any cal-
careous particles in the calciferous glands in either Gordio-
drilus, Stuhlmannia, Notykus, or Eudriloides. As the
secreting tissues diminish, the tissues surrounding the glands
increase in amount and in specialisation. They are supplied
with blood from a large vessel which is dilated within the
gland, and by its devious course must prolong the time that
the blood is submitted to the action of the surrounding cells.
The function of these glands must, I believe, have some rela-
tion to the blood. I regard them as analogous to the spleen
of the Vertebrata; and in relation to this matter it may be
pointed out that it has been stated that the spleen is originally
formed as a diverticulum of the gut, thus indicating a conver-
sion from a gland appended to the alimentary tract, and pro-
bably performing the function of a digestive gland to a “‘gland”
concerned in some way with the blood. The instances de-
scribed in the present paper are remarkably analogous. A
series of glands undoubtedly related to the function of diges-
tion are metamorphosed into glands which also appear to have
some relation to the vascular system. In the family Enchy-
treide there is something of the same kind. The genera
Buchholzia and Henlea are furnished with glandular ap-
pendages to the cesophagus, which can hardly be different in
their nature from the calciferous glands of earthworms ; from
these glands (in most cases) the dorsal vessel arises. In the
genus Mesenchytrus there are no such glands, but the
dorsal vessel at its origin from the pericesophageal plexus (or
sinus) contains a cellular rod which has been called the “cardiae
OLIGOCHATA OF TROPICAL EASTERN AFRICA. 261
body.” I agree with Michaelsen in considering this structure as
the reduced equivalent of the cesophageal diverticulum, which,
if it has any function at all, must perform some office in relation
to the blood. In this connection, also, I may refer to my own
description of an analogous organ in the fresh-water Oligo-
chete Phreodrilus. In that worm there are a pair of
perienteric blood-vessels of larger size than the rest which
contain in their interior masses of cells. These, too, may be
the last stage in the conversion of an alimentary gland into a
“blood gland.” Of a different nature are probably the vas-
cular tufts which arise from the dorsal vessel in many Lum-
briculide; though here, too, the error of Grube in terming
these vascular ceca diverticula of the intestine is not unsug-
gestive. Finally there are the “blood glands” of Peri-
cheeta, which I have described in a recent volume of this
Journal. These are hardly referable to cesophageal diver-
ticula which have lost their connection with the gut. Their
existence is interesting as showing the possibility that in the
Annelids we have a group of glands very suggestive of the
spleen, supra-renal bodies, and perhaps some other of the
“ ductless glands”? of the Vertebrata which are not all trace-
able toa common origin. Mr. Weldon has shown how the
supra-renal body derived from the renal epithelium has lost its
renal function and become converted to the interests of the
vascular system. His description and figures of blood-clots
lying in the lumen of the tubes is of particular interest to me
in connection with the structures illustrated in fig. 15 of the
present paper; but in making this comparison, it would be
necessary to assume that the cells which I have regarded as
peritoneal were in reality metamorphosed epithelium of the
calciferous glands.
The only other Eudrilid in which calciferous glands, after
the pattern of those described in the present paper in Stuhl-
manniaand Eudriloides, seems to be Megacheta tenuis.
Michaelsen writes as follows about the matter:—‘‘In den .
folgenden Segmenten erkennt man je ein Paar eigenartiger,
Fettkorper-ahnliche Organe, die zu Seiten des Darmes liegen.
262 FRANK E. BEDDARD.
Die Zellen, aus denen diese Korper hestehen, sind grob granulirt
und erhalten durch Einlagerung zahlreicher, schwarzer Korner
ein Chloragogenzellen-artiger Aussehen. Ein starkes Blut-
gefass geht mitten hindurch. Ich glaube erkannt zu haben,
dass diese Korper mit den Segmentalorganen zusammenhan-
gen, deren in je einem Segment ein Paar vorhanden ist. Es
musste unentschieden bleiden, ob sie durch die ganze Lan ge
des Ko6rpers oder nur am Vorderkorper ausgebildet sind.
Soweit ich das Tier untersuchte, bis zum 20. Segment, sind
sie vorhanden.” It seems to be possible that the structures
which Michaelsen here describes (without any figures) are the
same as the organs regarded by myself as the metamorphosed
equivalents of the calciferous glands of other Eudrilids. The
account of the minute anatomy which Michaelsen gives,
though not very full, agrees so far as it goes with the organs
in question. They have a kind of resemblance to a fat body,
and the cells of which they are composed are distinctly ‘‘ grob
granulirt,” but they do not, so far as my own observations go,
contain any black pigment. Another difficulty, and the most
serious one in the way of comparing the glands of Megacheta
with those of Stuhlmannia, &c., is the fact that in the former
genus they extend back as far as the 20th segment. This is
too far, one would be inclined to suppose, for glands to extend
which are homologous with calciferous glands, connected as
they are in all earthworms with the esophagus. The close
proximity of the nephridia to the glands might easily give the
impression that the two series of organs were connected. The
fact that they are traversed by a strong blood-vessel is another
point of resemblance to the calciferous gland of Stuhl-
mannia, &c.
§ Note on the Substitution of Organs as Illustrated
by the Spermatothece in the Eudrilide.
I do not think that attention has been directed to the ex-
cellent instance which the spermatothece of the Eudrilide
afford of the substitution of one organ for another (physio-
OLIGOCHZTA OF TROPICAL EASTERN AFRICA. 263
logically identical, but morphologicallly different). The princi-
pal feature of interest in the anatomy of the Eudrilide is the
presence of coelomic sacs which do duty as spermatothece. I
have called these sacs consistently “‘ spermatothecal sacs” to
mark their difference in structure from, but their similarity in
function to, the spermatothecee of other Oligocheta. Michael-
sen, it is true, has denied to the spermatothecal sac of the
Eudrilide the function of storing sperm. They have, however,
been proved to contain sperm in Nemertodrilus, Eudrilus,
and Polytoreutus. That these sacs are not homologous
with the spermatothece of other Oligochzta is shown by their
development. I have pointed out that in Libyodrilus the
sac is formed at the expense of the septa, and Rosa has
arrived at the same conclusion by a consideration of the
histological structure of the sacin Paradrilus. Thereis, how-
ever, at least one genus in which the spermatotheca appears
to be of the type general in the Oligocheta. In Heliodrilus
there is a single long and narrow sac which opens externally
on to the 11th segment, and reaches back to the 18th. At
the extreme end the spermatotheca is enclosed by a ccelomic
sac continuous with the ovarian sac, &c. I have no facts
of development to offer in support of my belief that the
spermatotheca in this Annelid is comparable in origin to that
of other earthworms ; I rely upon its structure and relations.
The sac in question is lined by a columnar epithelium, quite
different in appearance from the cells which line the sac
in which it lies, but quite like the cells which are found in the
spermatothece of other worms, in which these organs are
epidermic invaginations. This is, I hope, plainly shown in
the figures illustrating my account of the anatomy of Helio-
drilus. The next stage is seenin Hyperiodrilus. Here
we have the true spermatotheca reduced to very small dimen-
sions, and the sac involving it is greatly increased in size. In
Paradrilus it is possible, whether it actually occurs or not,
for sperm to reach the interior of the large celomic sac; for
this sac communicates directly with the exterior by a short
tube which seems to be an invagination of the epidermis, and
264 FRANK E. BEDDARD.
appears to be the equivalent of a part of the spermatotheca of
Hyperiodrilus. The genus Eudriloides which I have
dealt with in the present paper supplies the next stage. Here
we have a cup-shaped layer of cells which, though they have
lost their connection with the epidermis, are very possibly to
be regarded as a derivative of it. They do not, however, line
the coelomic pouch, but have been, as it were, thrust aside by
the growth of the cells lining the sac, which cells have forced
their way to the exterior. Here, therefore, the original
spermatotheca is entirely relieved of all share in the organ
devoted to the storage of the sperm. Finally, Libyodrilus
and perhaps Polytoreutus have no trace left of the original
epidermic invagination except possibly at the very pore. I
have shown in Libyodrilus that the sac which is formed out
of the tissues of the septa, burrows its way into the thickness
of the body-walls; it is quite likely that it actually reaches
and perforates the epidermis by its own unaided efforts. The
spermatotheca, therefore, of Heliodrilus gradually yields up
its place to the sac developed out of the mesoblastic tissues,
which grow as it diminishes, and finally entirely replace it.
This instance is quite analogous, for example, to the replace-
ment of the notochord by the vertebral column.
§ Classification of the Eudrilide.
This family comprises now so large a number of forms that
it may be desirable to subdivide it. A subdivision has been
attempted by Michaelsen (loc. cit.); but it seems to me that a
further acquaintance with the structure of the various genera
of the family does not tend to confirm the justice of dividing the
family, as Michaelsen does, into two sub-families, Eudrilini and
Teleudrilini. The diagnosis which Michaelsen gives of the
Teleudrilini is as follows:
“ Die Teleudrilinen sind meganephridische, mit 4 Borsten-
paar-Reihen ausgestattete Terricolen, die eine einzige ventral-
mediane minnliche Geschlechtséffuung auf oder am 17
Segment und eine einzige ventral-mediane SamentaschenOff-
nung hinter der Intersegmentalfurche 10/11 besitzen,”
OLIGOCHATA OF TROPICAL EASTERN AFRICA. 265
The two sub-families are, in fact, distinguished by the
median or paired character of the generative apertures alone.
In the absence of any other characters it appears to me that
the paired or unpaired character of the apertures in question is
by no means a difference of first-rate importance. Even if we
follow Michaelsen in separating as a distinct genus the Cryp-
todrilid Fletcherodrilus on account of its median series of
spermatothecz, no one would in all probability consider that
genus to be worthy of being placed in a separate sub-family,
nor, indeed, does Michaelsen propose anything of the kind.
The median spermatotheca of Sutroa does not disguise its
likeness to other Lumbriculide. And in general there are so
many instances in the group of the Oligocheta of structures
which are paired in one genus and unpaired in another, thata
division so pronounced as that which Michaelsen proposes does
not commend itself to me.
There are two characters which seem to me to afford a more
reliable means of subdividing this family into two sub-families,
should such a step be regarded as necessary; as they both
relate to structures which are highly characteristic of the
genera in which they occur, more weight is to be attached to
them. The genus Eudrilus, when it was the only Eudrilid
known, was shown by myself to differ from all other earth-
worms by the possession of ventral median unpaired pouches,
which Michaelsen termed ‘‘ Chylustaschen.” The fact that
these structures are unpaired is not alone a fact that is greatly to
be valued, though it may be pointed out that Gordiodrilus is
at present the only other genus (not a Eudrilid) in which these
so generally present, and with these exceptions paired, struc-
tures are to be found. The most remarkable fact about these
unpaired pouches is that they coexist with one pair of paired
pouches not lying in the same segment as any one of them; this
is more remarkable than if all the calciferous glands were paired
or unpaired, as the case might be. Another (at that time quite
unique) character of Eudrilus is the existence in the epi-
dermis, or rather just below it, of numerous integumental bodies,
which Dr, Horst and I myself have compared to the Pacinian
266 FRANK E. BEDDARD.
bodies of the Vertebrata ; their appearance, at least, is very like
that of the structures mentioned. These same characters are
found in a few other genera of Eudrilide ; they occur in Teleu-
drilus, in Hyperiodrilus, and Heliodrilus (which two
latter Michaelsen unites into a single genus), and finally in
Polytoreutus. I am ina position to state that the integu-
mental sense-organs—if I am justified in applying the term
“‘sense-organs”’ to them—are absent in the following genera:—
Eudriloides, Heliodrilus, Pareudrilus, Nemertodrilus,
and Stuhlmannia. In none of these genera are there cal-
ciferous glands at all like those of the genera mentioned in
the first list. Libyodrilus, Pareudrilus, Alvania, and
Nemertodrilus have no calciferous glands at all—not a trace
of them. In the remaining forms the calciferous glands have
undergone the peculiar modification that has been described on
a previous page. I think that these two characters serve to
distinguish two groups of Eudrilide better than the paired
or unpaired generative apertures. I would furthermore remark
that the condition of the glands, which I believe to be the
representatives of the calciferous glands of other Eudrilide—
in the genus Eudriloides, for example—is not in accord
with the low position among the “ Teleudrilini” to which
Michaelsen assigns it.
These genera are the only ones in which both the points
used for the subdivision of the family are known; in some
others the presence or absence of calciferous glands has been
noted. Thus both Michaelsen and Rosa have shown the ex-
istence of a single pair of calciferous glands in the 12th seg-
ment in the genus Paradrilus. Preussia is said to have a
pair of these glands in the 12th segment; but Michaelsen is
doubtful about their nature, and has stated that they contained
no calcareous particles. Platydrilus, Megacheta, and
Metadrilus have, according to Michaelsen, no calciferous
glands, and the remaining genera are not described in this
respect. It will be noticed that in the two groups into which
I have provisionally divided the family there is yet another
character which divided them, and which may possibly be of
OLIGOCHATA OF TROPICAL EASTERN AFRICA. 267
value. In Eudrilus and the genera which are placed with it,
the sperm-ducts are dilated into a round or oval sac before
they open into the funnel. Michaelsen terms these dilatations
“« Kiweisskapseln.”? These oval dilatations are absent in all the
genera which I have placed in the second group. Paradrilus
has them, and, as it also has at least one pair of calciferous
glands, may perhaps be referred to the first group. Such
dilatations appear to be absent in the genera Platydrilus,
Megacheta, Metadrilus, and Notykus. These genera
have, as has already been pointed out, no calciferous glands; it
remains to be shown whether the integumental sense bodies are
absent. Provisionally, therefore, I group the Eudrilidz! into—
Sub-family 1. HEudrilineze.—Calciferous glands present.
Integumental sense organs generally
present. Funnels of sperm-ducts di-
lated proximally.
Sub-family 2. Pareudrilinz.—Calciferous glands ab-
sent or greatly modified. No integu-
mental sense-organs.* No dilatation of
sperm-ducts.
1 It is very possible that the structure of the nephridia will prove to
separate these two groups. I have to some extent dealt above with the
excretory organs of a few types belonging to the sub-family Pareudriline ;
in these there is either a well-developed integumental plexus of tubules or
traces of such. On the other hand, nothing of this kind occurs in any of
the Eudriline. ,
2 Except in Hudriloides (occasionally).
LS)
for)
ie @)
FRANK E. BEDDARD.
EXPLANATION OF PLATES 16 & 17,
Illustrating Mr. Frank E. Beddard’s paper, ‘“ A Contribution
to our Knowledge of the Oligochzta of Tropical EHastern
Africa.”
Fic. 1.—Transverse section through atrium of Eudriloides Cotterilli.
v. d. Vas deferens, alongside of which runs a blood-vessel.
Fie. 2.—Transverse section through atrium of Eudriloides brunneus,
just at the point where the two atria join. v.d. Vas deferens. gi. Glan-
dular cells containing abundant secretion.
Fic. 3.— Transverse section through atrium of Polytoreutus violaceus.
61. Blood-vessel. m. Special thickenings of muscular coat.
Fic. 4.—Alluroides Pordagei; lateral view of anterior segments. sp.
Spermatothecal pore. . Penis. @. Oviducal pore. The clitellum is in-
dicated by the absence of furrows dividing its segments.
Fie. 5.—Alluroides Pordagei; longitudinal section through atrium
and adjacent structures. The segments are numbered. CV. Anterior, Cl’.
Posterior end of clitellum. ¢. Male pore. p. Penis. o. Ovary. 4?¢. Atrium.
od. Oviduct. ov. A ripe ovum.
Fic. 6.—Polytoreutus Finni. Spermatothecal sacs and atria. s. Sper-
matothecal sacs. 2. Oviducal pores. ©. Terminal muscular sac in which
atria (4¢.) open. sp. Terminal sac through which spermatothecal sacs open.
Fic. 7.—Polytoreutus violaceus. Spermatothecal sacs and atria.
Lettering as in Fig. 6,
Fic. 8.—Polytoreutus kilindinensis. Spermatothecal sacs and atria.
v.d. Sperm-duct. v. d./. Its funnel, opening into interior of s. s., sperm-sac.
Fic. 9.—Reproductive organs of Pareudrilus stagnalis, displayed by
dissection and in siti. @s. Gisophagus. JZ. s. Sac containing ovary, and
connected with spermatothecal sac. Sp. Orifice of latter on to exterior.
Ro. Egg-sac. 9. Oviducal pore. . Terminal sac of 4¢., atrium. yp. s.
Sac containing penial sete.
Fic. 10.—Genital segments of Hudriloides brunneus, from beneath.
9. Spermatothecal pore. @. Male pore. C/., Cl’. Anterior and posterior
boundaries of clitellum.
Fie. 11.—Longitudinal section through a few segments of Stuhlmannia
variabilis in esophageal region. Ws, Qisophagus. Ca. Calciferous glands,
D.v. Dorsal vessel. S. Septum. Zp. Epidermis. m. Muscular layers of
body-wall.
OLIGOCHMTA OF TROPICAL EASTERN AFRICA. 269
Fic. 12.—Transverse section through cesophagus and one of calciferous
glands of the same. ws. (sophagus. d. Duct of gland opening into it.
a. Hxtremity of gland where nuclei are arranged in irregular portions on each
side of a blood-vessel.
Fic. 13.—A portion of a calciferous gland of same, more highly magnified
to show the nuclei (z.), the boundaries of the cells (/.), and the secreted
granules (s.).
Fic. 14.—Extremity of calciferous gland, lettered a in Fig. 12.
Fic. 15.—Section through a calciferous gland of Hudriloides Cotterilli.
ce. Modified cells of gland closely investing the blood-vessels.
Fic, 16.—Section through end of spermatothecal sac of Eudriloides
Cotterilli. Zp. Epidermis. P/. Plug of cells occluding lumen of sac.
m. Mesoblastic cells lining it. #. Layer of epiblastic cells, apparently in-
vaginated to form wall of sac.
Fic. 17.—Genital segments of Polytoreutus Finni. 9. Spermatothecal
pore. ©. Male pore.
Hig. 18.—Terminal male apparatus of Hudriloides Cotterilli. ¢. Atria.
6. Male pore. s. Penial seta. a—e. Muscles referred to in text.
Fic. 19.—Spermatothecal sac of thesame. g/. Its glandular appendices.
Fic. 20.—Hnd of penial seta of the same.
Fic. 21.—Terminal male apparatus of Kudriloides brunneus. v. d.
Sperm-duct. Other letters as in Fig. 18.
Fic. 22.—Spermatothecal sac of the same. wz. End of sac lying in front
of external pore. g/. Glandular appendices.
Fic. 23.—Transverse section through part of spermatothecal sac, lettered
x in last figure.
eqs
tah!
1
ANATOMY OF LIMNOCNIDA TANGANYICZ. 271
A Further Contribution to the Anatomy of
Limnocnida tanganyice.
By
R. T. Gunther, B.A.,
Lecturer of Magdalen College, Oxford.
With Plates 18 and 19.
THE present paper is the result of a more minute examina-
tion of the same material on which my preliminary account of
Limnocnida was based. Having already treated in my
previous paper of the more obvious characters of the Medusa,
I have endeavoured to avoid needless repetition in the present.
The material had been caught and fixed in osmic acid by Mr.
A. Swann on the shores of Lake Tanganyika itself, and reached
me in a good state of preservation in strong alcohol, as already
mentioned in the preliminary notice.
The present work was carried out in the new Morphological
Laboratory at Oxford during Lent term, 1893; and I have to
express my sincere thanks to the President and Fellows of
Magdalen College for enabling me to pursue my studies in
Oxford, and also to Professor Ray Lankester for all the faci-
lities which he afforded me while working in his laboratory.
The following account of Limnocnida is divided into three
sections. The first is devoted to the description of certain
parts of the anatomy of the Medusa, and the arrangement
adopted by O. and R. Hertwig in their ‘ Organismus der
Medusen’ has been followed in the main. The second and
third parts deal with the systematic position of Limnocnida
and the origin of fresh-water Medusz respectively. At the
VOL. 36, PART 2,—NEW SER. T
272 R. T. GUNTHER.
end of the first part is a full description of the process of
reproduction by budding in Limnocnida, which was only
briefly referred to in my preliminary account.
I. Anatomy.—a. THE EctTopERM.
The general limiting epithelium of the body does not
call for particular remark. It consists of a single layer of
polygonal cells, which vary in height at different parts of the
surface. On the more exposed surface of the exumbrella they
are more or less flattened, while on the subumbrella and
velum they are cubical or even slightly columnar in places.
The nucleus is roundish, and usually situated in the middle
of the basal half of the cell.
Thread-cells were only observed on the tentacles, arranged
in groups or batteries, on the margin of the umbrella, thickly
packed in a continuous ring, and on the manubrium, distributed
along the edge. The exumbrellar surface seems to be com-
pletely destitute of any such groups of thread-cells as occur in
the Aiginide or the Geryonide. The thread-cells them-
selves are very small oval bodies, the major axes of which are
about twice as long as the minor axes. A peculiar modifica-
tion of the ectoderm, analogous to the peronium of the
Trachynemide and Geryonide or to the nettle-ring of
Limnocodium, occurs all round the rim of the umbrella, just
outside the circular canal. This ring of modified ectoderm
extends over the roots of the tentacles, and in it are buried
the marginal sense-organs (Pl. 18, figs. 1 and 2). The
minute structure is very similar to that of the peronium of
other Medusz, and, like that organ, apparently consists of
modified thread-cells. Its function is possibly skeletal as well
as defensive.
Muscular System.—Muscle fibres probably exist both on
the exumbrella and on the subumbrella, but they could not be
demonstrated in the material at hand with any degree of cer-
tainty. The muscles of the velum, however, are very con-
spicuous, both in preparations mounted whole and in transverse
sections. The ectoderm of the inner surface of the velum
ANATOMY OF LIMNOOCNIDA TANGANYICA. Kis
gives rise to a number of strong circular muscle-bands, which
may be seen in transverse sections to project into the meso-
gloea. Near the point of attachment of the velum, the muscle-
bands are small with a simple semicircular transverse section,
but as they approach the free margin of the velum, they be-
come larger and their transverse sections lobulated. From
this it appears that the muscle-bands near the free edge of the
velum are those which are chiefly concerned in the contraction
of the organ.
Nervous System.—No trace of a nervous system could
be observed in transverse sections, although special search was
made for a nerve-ring in the neighbourhood of the base of the
velum.
Sense-Organs.—The sense-organs or marginal bodies are
embedded on the velar side of the zone of ectoderm which
constitutes the nettle-ring. In all the individuals examined
no relation could be demonstrated between the arrangement
of the marginal bodies and that of the tentacles, though it is
not improbable that some radially symmetrical disposition may
occur in younger stages. It would seem that new marginal
bodies are continually being developed, as even in the oldest
individual examined quite young sense-buds were observed.
In my former paper I suggested that the marginal bodies
would be shown to be endodermal in origin, from the analogy
of the corresponding organs in Limnocodium. It can now
be definitely affirmed that the axial part of these organs is also
of endodermal origin in Limnocnida. In Pl, 18, figs. 1 and
2, there are figured two sections passing through a young mar-
ginal sense-organ (4). It will be seen that the mesogleea
(ms.), which is present almost everywhere between the ecto-
derm of the nettle-ring and the endoderm of the circular canal,
is broken through just beneath the base of the young sense-
organ, and certain endoderm cells project from the endoderm
into the young sense-bulb. In fig. 1, v. is such an endoderm
cell actually wandering from the endoderm to take up its
place in the axis of the sense-organ. The direction of move-
ment is indicated by the edges of the perforation in the
274 R. T. GUNTHER.
mesogloea being slightly turned up, forming a sort of “ burr”
on the side of the ectoderm. |
The sense-organ consists of an almost spherical sense-bulb
and a containing capsule. The sense-bulb is composed of a
number of clear, refringent nucleated cells, sometimes arranged
in two layers, as in fig. 2, B, where several large central cells
(four are seen in the figure) are surrounded by numerous
flattened cells, but more often the central and sheathing cells
are so mixed up that it is impossible to distinguish between
them. At the base of the sense-bulb there are generally a few
very granular cells, by which the bulb is attached to one wall
of the capsule. The capsule is buried in the nettle-ring and
lined by an endothelium of very thin pavement-cells.
In the young sense-organ the bulb consists of a number of
squarish peripheral cells, all grouped round one or more large
axial granular cells (figs. 1 and 2, 4.). In the young organ
there is not so much space between the wall of the capsule and
the bulb as in the fully-grown organ. The endodermal axis of
the fully-grown organ is completely shut off from the endoderm
of the circular canal, from which it has sprung, by a thick
plate of mesogloea.
Generative Organs.—As is the case with the Ocellata,
the genital organsare developed from the ectoderm of the manu-
brium, In sexually mature animals the ectoderm of the
proximal two thirds of this organ is very much thickened
owing to the prolific growth of the sexual elements in this
region. Inside the area occupied by the genital organs the
endodermal lining of the csophagus is modified (figs. 3
and 4, end.). It is much thickened, and consists of
long cells, the outer ends of which are probably flagellated,
though the flagella were not preserved in the material exa-
mined. The protoplasm contains numerous granules of foreign
matter, and, as in similar cases described by Hertwig (7),
“has a plant cell-like appearance.” The mesoglea becomes
unusually thin in the region of the sexual organs, though
tolerably thick on either side of that region. Thus far, the
structure of the manubrium in the region of the genital organs
ANATOMY OF LIMNOCNIDA TANGANYICA. 275
is the same in both males and females. It is only the ecto-
derm which is different ‘in the two sexes; but in both, the
general type of structure of the germarium is similar.
In both spermarium and ovarium of Medusx the Hertwigs
distinguish three layers, viz. I, a basement layer; II, a layer
of germ-cells in various stages of development ; and III, a cover-
ing epithelium. In Limnocnida, layers I and II are so
intermingled that it is hard to draw a hard-and-fast line
between them (figs. 3 and 4).
The cells which correspond to layer I lie next the mesoglea.
Their nuclei are relatively large and round. Inthe male there
are usually several layers of cells, all of the same nature. In
both sexes these cells may be regarded as the germ mother-
cells, and it is from them that the cells of the next layer
originate.
Layer II is composed of ova or spermatozoa in various
stages of development. On comparing sections taken through
the germaria of the two sexes (figs. 3 and 4) the general
distinguishing features are strikingly brought out. In the
male the germ-cells tend to become as numerous and as small
as possible, while in the female all available material is
employed in the construction of a few gigantic ova. In
the section of a spermarium shown in fig. 4, two stages
of development are seen. The innermost cells have large
round nuclei, nearly twice as large as the nuclei of the parent
cells from which they are derived. These cells eventually give
rise to clusters of numerous cells with very small nuclei, which
stain more deeply than those of the large cells. Ata later
period these small cells will probably grow tails and become
mature spermatozoa, but among the specimens examined none
were found with mature spermatozoa developed.
In the female the cells of the second layer are much larger
and fewer in number than in the male. It is composed of two
kinds of cells (fig. 3, ov. and st.), both of which are derived
from the first layer. The most conspicuous cells in this middle
layer are ova (fig. 3, ov.), and attain to a very considerable size.
They are by far the largest cells in the body of the Medusa,
276 R. T. GUNTHER,
Their shape is rounded, and they are often provided with a
few processes extending inwards between the other cells
towards the cells of the first layer. Their protoplasm is ex-
ceedingly granular, and contains one or more large vacuoles.
The nucleus is a very large clear spherical body in the centre
of the cell, which hardly stains at all, almost all the chromatin
being concentrated in the large nucleolus. The other kind of
cells differ from the ova in their protoplasm being quite clear
and free from granules, and in their not growing to such a
large size. From the way in which they are situated with
regard to the ova, they would seem to be merely packing or
interstitial cells (fig. 3, st.).
The third or outer layer is the same in both sexes. It con-
sists of a covering epithelium of more or less cubical cells with
round nuclei (figs. 3 and 4, ect.). In the female the cells are
somewhat more irregular in shape than in the male.
B. THe ENDODERM.
Epithelium of Gastro-vascular System.—The endo-
dermal lining of the manubrium in the region of the sexual
cells has been described above in the section on the repro-
ductive organs. On either side of this region the endoderm
cells (fig. 5, end. m.) have much the same internal granular
structure, but they are not so long. The epithelium covering
the dorsal wall of the stomach is represented at end. in
fig. 5. It consists of granular columnar cells with a clear space
towards their free ends. No traces of intracellular digestion
were observable in any of the cells of the gastric epithelium.
The radial and circular canals are lined with squarish poly-
gonal cells, many of which are vacuolated. The epithelial
lining of the circular canal on the side nearest the nettle-ring
is thrown into folds (end., fig. 6), the cells of which are some-
times so tightly packed as to obscure the divisions between the
folds. In this way a highly peculiar endodermal organ is
formed in the circular canal, which may grow to such an extent
as to completely fill the lumen of the canal. The cells of this
ANATOMY OF LIMNOCNIDA TANGANYICA. 277
organ are multinucleate and, in some regions, vacuolated.
Moreover, spaces may occur in this hypertrophied mass of cells
within which one or more of the cells may be floating. Similar
cells (fig. 7) occur in the lumen of the circular canal and in
the radial canals. In all, the nucleus is broken up into several
pieces, and in many, vacuoles are present.
The function of this extraordinary organ of the circular canal
is extremely doubtful. Nor are there any means of settling the
matter until living material can be obtained for examination.
The abundant discharge of the cells which are found in the
circular and radial canals, would indicate an excretory function
by which noxious or useless matter is carried to the exterior.
On the other hand, it is possible that the organ is endowed
with the power of secreting some digestive ferment which is
discharged into the circular canal and is thence conveyed into
the shallow gastric cavity by the radial canals; or, again, the
katalysis of the free cells themselves may supply a material of
economic value to the organism.
The endoderm lamella is situated very close to the sub-
umbrellar ectoderm and is composed of a single layer of small
cubical cells which increase in size and become vacuolated near
the gastric cavity (fig. 6, end. /.).
The tentacle axes are hollow and composed of large, clear,
thin-walled cells, each with a small nucleus on one side. The
cell contents seem to be almost entirely fluid or gelatinous, and
hardly any trace of granular protoplasm can be made out. It
is probable that the turgidity of these cells may be the cause
of the rigidity of the tentacles. The lumen of the tentacle is
continuous with that of the circular canal, as seen in the
section figured in fig. 6, which passes through the base of one
of the radial tentacles.
c. REPRODUCTION BY BUDDING.
In certain of the individuals examined, the external wall of
the manubrium was found to be covered with numerous small
Medusa buds in yarious stages of development. The region
278 R. T. GUNTHER.
which is capable of giving rise to buds forms a zone completely
surrounding the manubrium and covering about two thirds of
the surface of that organ, leaving a distal zone free from buds,
as in Sarsia. Thus the area on which Medusa buds may
develop is approximately co-extensive with the germinal area
on which the reproductive elements may arise in sexual in-
dividuals.
The ectoderm cells of the gemmiferous region of the manu-
brium are much vacuolated ; their nuclei for the most part lie
deeply, close to the endoderm, but some have a more super-
ficial position between the vacuoles. Cell boundaries are
either altogether absent or could not be demonstrated. The
cells of the endoderm of this region are arranged in several
layers; in character they are polygonal, very granular, and
provided with definite walls. The mesoglea, which is nowhere
very thick, disappears altogether beneath the larger buds. It
is thickest in the distal third of the manubrium, in the region
which does not give rise to either medusa buds or generative
products.
The young bud first makes its appearance as a small local
outpushing of the wall of the manubrium. This outpushing
(figs. 8 and 8 a) affects all the cell layers and is hollow, its
cavity being a diverticulum of the gastric cavity of the parent.
Thus a part of the digestive epithelium of the parent becomes
the endoderm of the young bud, and it is not unlikely that by
means of this endoderm the bud may obtain the nutriment for
its further growth by the direct absorption of such food matter
as may enter its own “stomach” from the gastric cavity of
the parent.
The young bud does not remain long in this stage. A change
soon appears to come over the ectodermal cells at the apex of
the bud. Their protoplasm instead of being vacuolated be-
comes dense and more granular than before, while the cells
themselves become somewhat thicker, and their nuclei take up
their position in an even row next the endoderm. These
changes foreshadow the next step. The apical ectoderm bodily
invaginates into the endoderm (figs. 9 and 9a) at the tip of the
ANATOMY OF LIMNOCNIDA TANGANYICA. 279
bud to form the “ glockenkern” (g/.) or forecast! of the ecto-
dermic lining of the subumbrella cavity. As the ectodermic
invagination sinks deeper into the interior of the bud, its cells
acquire definite walls and arrange themselves in a single layer
round a central cup-shaped space. The mouth of this space now
closes up, and finally the entire “ glockenkern” becomes covered
over by an overgrowth of ectoderm, and its cells become marked
off from the cells of the peripheral ectoderm (figs. 11 and
lla, gl.). At this stage the “ glockenkern” is a hollow
sphere of ectoderm, one cell thick, enclosing a hollow cavity,
which, as has been shown above, is really a portion of the ex-
terior which has been enclosed during the growth of the bud.
Meanwhile, owing to the invagination of such a comparatively
~ bulky mass of ectoderm, the endoderm thins out so much as to
be reduced to a single cup-shaped layer of cells enveloping the
central “ glockenkern.”
A second ingrowth of ectoderm now occurs. The ectoderm
cells at the apex of the bud become less vacuolated and undergo
changes similar to those which the cells of the “ glockenkern”
passed through, and grow down as a solid plug of cells into the
apex of the “glockenkern.” Before this stage is reached the
“ slockenkern ” was more or less spheroidal or egg-shaped, but
the effect of the second ingrowth of ectoderm is to push one
wall of the “ glockenkern” into the other, with the consequent
result that the “ glockenkern”’ becomes a two-layered cup sur-
rounding a plug of ectoderm cells. A further result is that the
lumen of the “ glockenkern ”—the future subumbrella cavity
—becomes considerably reduced in size and may even, in some
cases, disappear altogether (figs. 12 and 12a). In the latter
case it always reappears at a subsequent stage. The outer
wall of this cup-shaped “ glockenkern”’ will ultimately form
the ectoderm of the subumbrella, while the inpushed wall will
1 T have used this word as the English equivalent of the word “ Anlage,”
which has presented so much difficulty to translators. The use of the terms
“rudiment ” or “fundament ” (Prof. E. L. Mark) is not to be recommended
in this sense. For some considerable time past the members of Professor
Lankester’s “Seminar” at Oxford have been accustomed to use the term
forecast” with the significance of ‘* Anlage,”
280 R. T. GUNTHER.
give rise to the subumbrella side of the velum—the outer side
of the velum being formed from the second ingrowth of ecto-
derm, as will appear in the sequel.
By the next stage the forecasts of most of the important
organs of the Medusa can be readily recognised (figs. 13 and
13a). The subumbrella ectoderm (s. ect.) and velar ecto-
derm (v. ect.) have become much thinner relatively. The sub-
umbrella cavity (s.c.) is very large. The endoderm cells
surrounding the “ glockenkern” have grown further over that
organ, and at certain regions have begun to grow up into the
mass of ectodermic cells at the apex of the bud, thus indicating
the positions of the future tentacles. At this early stage the
forecasts of the tentacles are solid, but in section their endo-
dermal axis (¢. end.) may be shown to be composed of nume-
rous cells arranged round a central axis. A lumen does not
appear until a somewhat later stage. Of the tentacles, four
are greatly in advance of the others, but I was not able to
make out whether two of these make their appearance before
the other two, as is often the case among the Hydrozoa.
Meanwhile a great change takes place in the apical mass of
ectoderm. Its constituent cell nuclei, instead of being more
or less uniformly diffused throughout the mass, now assume
definite positions with regard to the neighbouring organs
(fig. 18). The more deeply seated nuclei range themselves in
a single tier along the outer side of the “ glockenkern,” while
the others surround the forecasts of the endodermal axes of
the tentacles. In the section through a Medusa bud of this
stage in fig. 13, a small space may be seen between the
ectodermic epithelium of the tentacles and the ectoderm over-
lying the ‘“glockenkern.” The subumbrella cavity at this
stage still remains completely closed by a membrane composed
of part of the “ glockenkern” and of the overlying layer of
ectoderm. This membrane is the forecast of the velum and
subsequently will become perforated.
In the most advanced stage observed (fig. 14), the velum no
longer extends right across the mouth of the subumbrella
cavity, but has broken through in the middle. The tentacles
ANATOMY OF LIMNOCNIDA TANGANYICZ#. 281
have increased in number since the previous stage, and the
older ones are relatively much longer, so that they hang
through the velar aperture into the subumbrella cavity.
The tips of some of the tentacles are already provided with
cnidoblasts. Also in some of the older tentacles a consider-
able lumen has made its appearance owing to an axial split in
the endoderm. The gastro-vascular system is well developed
in this stage. The circular canal is the first to appear as a
split in the endoderm at the base of the tentacles, and is soon
placed in communication with the enteric cavity of the parent
by the four radial canals which arise one opposite to each of
the four perradial tentacles. Between the radial canals and all
round the “ glockenkern”’ the endoderm persists as the endo-
derm lamella.
Atavery slightly later period the mesogloea makes its appear-
ance between the endoderm lamella and the ectoderm of the
exumbrella, as seen in the horizontal section, fig. 15, m. s.
An Abnormality in Bud-formation.
As a rule the Medusa buds are situated on the outer surface
of the manubrium of the parent, but on two occasions a bud
with free tentacles was found projecting into the stomach
cavity. This lusus nature in both instances occurred in a
region where the buds were more closely packed than usual,
and, consequently, as there was not sufficient room for the
buds to develop on the outside of the manubrium, they forced
their way through the wall of the manubrium and grew into
the stomach cavity—only in a reversed condition, just as the
young Tenia head grows into the Cysticercus instead of out-
side it.
As a result, the disposition of the cell layers and organs in
such an abnormal bud is completely reversed. The velum
develops outside the circlet of tentacles and the subumbrella
epithelium becomes external, while the exumbrella epithelium
becomes internal. An examination of fig. 16 will make this
clear. The endoderm is continuous with the general endoderm
282 R. T. GUNTHER,
of the manubrium as in a normally developed bud, and so
likewise the exumbrella ectoderm is continuous with the
manubrial ectoderm. But the ectoderm covering the sub-
umbrella has undergone rupture at the point a, and the entire
bud has forced its way through the aperture.
The future of these monstrosities must remain a matter of
doubt. It would be interesting to know whether such abnor-
mal buds are reabsorbed by the parent or whether they are
capable of further growth and are then cast off. In any case
it would seem improbable that a Medusa bud in this condition
should eventually right itself, in spite of the acknowledged
powers of regeneration which are characteristic of the
Hydrozoa.
Bud-formation of Limnocnida contrasted with the
same process in other Hydrozoa.
If we compare the process of development of Medusa buds in
Limnocnida, as detailed above, -with the same process in
other Hydrozoa, several striking points of difference are
noticeable. In the first place the method of invagination of the
ectoderm to form the entocodon or “glockenkern ” in Limno-
cnida differs from the method of formation in most other
Medusz in which the process has been described, and I am in-
clined to think that in this respect Limnocnida exhibits a
more primitive condition.
As a general rule we find that in the Craspedota the earlier
stages of the development of the Medusa buds, whether they
subsequently give rise to free Meduse or whether they grow
into any of the various degenerate modifications of the Medusa,
show considerable shortening of their ontogeny. It would be
superfluous to recapitulate all the arguments used in favour of
this view. Weismann in his monograph on the ‘ Entstehung
der Sexualzellen bei den Hydromedusen’ has brought forward
abundant evidence to indicate the phylogenetic stages by
which the primitive method of the formation of the Medusa-
bell by the outgrowth of a circular fold has become changed
into a process of invagination,
ANATOMY OF LIMNOCNIDA TANGANYICH. 283
Examples of this invagination in a more or less modified
form have been described and figured in many different genera
of Hydroids, but in none does it exhibit quite sucha primitive
condition as in Limnocnida. In the majority of Hydroids
described, the invagination takes the form of an ingrowth of a
solid mass of cells, called the “ glockenkern.” In some cases
however, such as Hydractinia echinata, Weismann (15)
Taf. xxiii, fig. 5, and better in the young buds of Clava
squamata, Taf. v, Galeolaria aurantiaca, Taf. xxi, and
Podocoryne carnea, Taf. xix, the “ glockenkern”’ is seen to
be hollow in the youngest stages figured, but whether a cavity
exists in any Hydroids from the first, or whether it is only sub-
sequently formed as appears in the above cases, remains to be
decided by future investigations. In Limnocnida the
“glockenkern”’ is a hollow invagination from the very first,
and its cavity becomes shut off from the exterior at a later
period, but is typically retained throughout.
The later stages of .Limnocnida buds, and especially the
formation of the velum from two distinct layers of ectoderm
which afterwards break through in the centre, are almost
exactly paralleled by Bougainvillea fruticosa and Peri-
gonimus cidaritis (Weismann, Taf. xii, figs. 10, 12,
and 13).
A noticeable feature in the development of the Medusa
buds of Limnocnida is the entire absence of any trace of
manubrium and mouth in any of the stages examined. It is
true that in the adult the size of the manubrium is very insig-
nificant as compared with the dimensions it assumes in other
Meduse ; but nevertheless its entire absence in the young buds
is, to say the least, remarkable. In the development of all
other Medusa buds in the Hydrozoa, the manubrium is one of
the most conspicuous parts of the young bud, and in the case
of degenerate Medusz its presence is characteristic. The
absence of manubrium and mouth in Limnocnida buds is,
no doubt, to be accounted for by the fact that the young buds
seem to obtain their nutriment direct from the gastric cavity
of the pareut, and therefore have no need of a mouth or manu-
284. R. T. GUNTHER.
brium. The mouth probably breaks through at the date of
the setting free of the young Medusz.
II. Generat CoNcLUSIONS REGARDING SYSTEMATIC
PosITION.
It is perhaps idle to speculate upon the phylogenetic position
of Limnocnida, considering that we are in absolute ignorance
regarding its mode of development from the egg, or of its life-
history, but it is expedient for purposes of reference that all
such newly-discovered organisms should be given a place in
the system, as natural as the state of knowledge will permit.
At the same time any attempt at constructing new groups, or
at arranging old ones, to contain a species about which we
know as little as of Limnocnida, is to be deprecated on the
ground that such a course might add fresh burdens to an
already overladen nomenclature.
According to the preceding account the two most important
characters known in Limnocnida are, firstly, that the gonads
are developed on the walls of the manubrium; and, secondly,
that the sense-organs situated on the margin of the umbrella
have an endodermal axis. The only known Meduse which
possess these two characters in combination are the Narco-
medusz. On the other hand, Limnocnida possesses hollow
tentacles, while those of the Narcomeduse are solid, though
occasionally partly hollow in the adult, and the sense organs
of Limnocnida possess no structure corresponding to the
otoliths of the Narcomeduse.
I regret that my acquaintance with the Narcomeduse does
not justify me in expressing an opinion as to whether the
association of Limnocnida with them is a natural one or
not; but, as Limnocnida has undoubtedly reached the
same grade of development in respect to two of the most
important features of the group, it is convenient to find
a provisional home for Limnocnida in connection with
the Narcomeduse, and to wait until we know more of
ANATOMY OF LIMNOCNIDA TANGANYICA. 285
its life-history before any attempt is made to fix its abode
definitely.
The only fact we have at present about the life-history of
the Medusa is that during part of April, May, and June,
swarms of male, female, and asexual bud-bearing individuals
coexist in Lake Tanganyika. The most important problem
to be solved is, Does Limnocnida pass through a fixed
hydroid stage in any part of its life-history? Is its develop-
ment metagenetic or hypogenetic ?
It is certainly possible that, as in the case of the Narco-
medusz, there is no fixed hydroid stage. It seems to me
conceivable that a kind of alternation of generations may occur
in Limnocnida such as has been described by Brooks (8) in
Willia ornata (ifrightly identified), in which asexual Medusze
produce sexual forms by budding. It is possible that the sexual
individuals of Limuocnida may produce eggs which develop
into free-swimming planula-larve which grow into the asexual
Medusz, which in their turn give rise to sexual Meduse by
budding.
On the other hand, if a fixed hydroid stage form part of the
ontogeny, Limnocnida must be separated from the Narco-
medusz which do not pass through such a stage, and then
would have to be regarded as descended from an Antho-
medusa-like ancestor and as having developed sense-organs
with an endodermal axis, just as Limnocodium is to be
regarded as a descendant of the Leptomedusz with sense-
organs morphologically similar to those of the Tracho-
meduse.
With regard to the striking similarity of the sense-organs in
both the fresh-water Meduse, Limnocodium and Lim-
nocnida, I can only consider it to be another instance of
parallel but independent evolution of similar organs in two
genera which are not related to one another, but which live
under similar conditions (“‘ homoplasy ” of Lankester, cited by
Darwin, ‘ Origin of Species,’ 6th edit., p. 385). If this is
really the case, the similarity of structure is most remarkable.
286 R. T. GUNTHER.
Il].—FresH-water Merpvus2.
The discovery of a new Celenterate living in fresh water
must always be considered an event of no little importance on
account of the fewness of such genera. The chief instances
known at present of fresh-water Coelenterata are the follow-
ing :—
Various species of Hydra are almost cosmopolitan. They
have been described from Egypt (Schmarda), Japan (Hilgen-
dorf), Australia (Bale, von Lendenfeld), and New Zea-
land (Coughtrey). The allied genus Microhydra is an
inhabitant of the New World.
Cordylophora is known from the fresh and brackish
waters of England, and of North Germany as far inland as the
Tegelsee in Prussia; also from America and Australia (v.
Lendenfeld).
Polypodium hydriforme, a parasite of the sturgeon, is
a native of the Volga (Ussow).
The habitat of Limnocodium is as yet unfortunately un-
known, but there is but little doubt that tropical America is
its original home.
Lastly, in 1890 J. von Kennell (9) discovered Halmo-
nises lacustris in Trinidad in a fresh-water lagoon com-
pletely shut off from the sea, and flourishing in the midst of
such truly fresh-water animals as insect and frog larve,
Daphnide, Naids, Holosoma, Dero, Clepsine, Planorbis, &c.
In addition to these truly fresh-water Coelenterates several
marine forms can tolerate brackish water, for example
Aurelia aurita from the Baltic, and Crambessa Tagi
from the mouth of the Tagus, and some others.}
While treating of such isolated Meduse, it may be as well
to mention the supposed salt-water Medusa recently drawn
attention to by Mr. Sclater (13). This Medusa, of which we
have as yet no details, is an inhabitant of Lake Urumiah in
Persia. Lake Urumiah has no communication with the sea,
See also observations by the late Professor Moseley, in his ‘ Naturalist
on the “Challenger,” ’ 2nd edit., p. 236.
ANATOMY OF LIMNOCNIDA TANGANYICA. 287
and consequently its waters are saline. This Medusa (if
Medusa it be) may be descended from fresh-water ancestors.
The first question that naturally occurs to the student of
evolution is, How did the fresh-water Meduse get into fresh
water? It is clear that they must either have originated
there, or they must have immigrated from the sea. The latter
of these alternatives is the only one worthy of consideration,
since we are not acquainted with any fresh-water Celenterate
from which the known lacustrine species could be derived ;
whereas the sea must always be regarded as the true home
and birthplace of Medusz. If this proposition be accepted,
there is no reason why all fresh-water Celenterates known at
present cannot be regarded as having descended from marine
ancestors.
In the first place, the mere fact of the difference of salinity
between fresh and salt water does not seem to be an insuperable
difficulty in the transition from a marine life to a lacustrine
one. The marine forms Aurelia aurita and Crambessa
Tagi, already mentioned, often frequent brackish water. The
only essential condition is, that the change from salt to fresh
water must be very gradual, as any sudden substitution of one
for the other causes almost instantaneous death to a soft
gelatinous creature like a Medusa, owing to the very violent
osmotic action which occurs in animals with a soft skin.
Experiments demonstrating this point were made long ago by
Beudant, and are described by Semper in his ‘ Natural Con-
ditions as they affect Animal Life.’
The circumstances under which the postulated slow change
of environment might occur are of several kinds. All marine
creatures found in fresh water have probably either wandered
up rivers from the sea, or have been living in bays or lagoons
which have become cut off from the sea by the upheaval of
land or by the silting up of the connecting channel, or in some
other way familiar to the geologist. In any of the latter
cases there may perhaps have been an intermediate epoch
when the body of water, completely separated from the sea at
low water, was still in connection with the sea at high tide.
voL. 86, PART 2,—NEW SER. U
288 R. T. GUNTHER.
When such an isolated volume of water has become completely
separated from the sea at all stages of the tide, it is further
necessary that the rainfall received in its basin should be in
excess of the quantity of water lost by evaporation. Given
all these conditions, the salt will gradually be washed out and
the water will become fresher and fresher, and those of the
original inhabitants of the lagoon or bay which could not
accommodate themselves to the changed environment would
die, leaving the rest to survive as a fresh-water fauna.
Such are briefly the changes which have probably occurred
in the case of Halmonises lacustris, which is the inhabitant
of a small fresh-water lagoon removed but a few yards from
the seashore itself. Its ancestors probably wandered into a
small bay or estuary situated where the fresh-water lagoon is
now, but in direct connection with the sea. By the upheaval
of the land or by some other cause the estuary became shut off
from the sea, and the salt water was gradually flushed out by
fresh. In the present case the change must have occurred
with sufficient slowness to allow not only Halmonises, but
also numerous other marine animals, such as several genera of
Polycheta, to become acclimatised to a life in fresh water.
Owing to the proximity of the pool inhabited by Halmo-
nises to the sea, the change is easy to understand. In the
case of Limnocnida it is far otherwise.
At the present day Lake Tanganyika, according to the
description of Mr. E. C. Hore (8), is situated about twice as
far from the west coast of Africa as from the east coast. The
lake lies at an altitude of about 2700 feet above the sea level,
within a mountainous ring fence encompassing a space of,
roughly, 600 miles by 300. The lake itself is situated close
under the mountains on the western side of this ring, and some
2000 to 3000 feet below the higher parts of the range. All
along this western side is an enormous chasm 400 miles long,
with an average width of 20 miles and of great depth, varying
from 500 to 1000 feet in the middle. The superfluous waters
escape by the Lukuga, which runs through a great break in
the mountainous rampart on the west side, and is a tributary
ANATOMY OF LIMNOCNIDA TANGANYICA, 289
of the Congo. It appears that the level of the lake has become
considerably lowered within quite recent times. Such, then, are
the main physiographical facts relating to the habitat of
Limnocnida.
If Limnocnida were the sole representative of a marine
fauna in Lake Tanganyika, it would be a surprising fact
enough, but it is associated with other marine forms. Just
as in the case of the fauna of the Trinidad lagoon, where we
find the sea-sprung Halmonises accompanied by marine
Polychete worms, so in the case of the fauna of Lake
Tanganyika Limnocnida is accompanied by several genera
of molluscs which are perfectly unique in fresh water, and
which would most certainly be described as marine forms were
their true habitat unknown.
Mr. Edgar Smith (14) especially draws attention to
Tiphobia horei and Neothauma tanganyicense, both
of which are quite unlike any genera known from the other
Central African lakes, and both of which differ from any other
molluscs of fluviatile or lacustrine origin.
The question that now arises is, If these organisms are of
marine origin, how did they get into Lake Tanganyika ?
The answer is undoubtedly difficult to find. A possible ex-
planation is afforded us by the analogy of the marine guests of
the Trinidad lagoon. If geologists, when they know more
about the geology of Central Africa than they do at present,
will allow us to imagine a state of things when the level of the
region of the Great Lakes was more than 2700 feet lower than at
the present day, and when the Atlantic Ocean extended over
what is now the plain of the Congo, and Lake Tanganyika
was a fjord communicating with the ocean, then it would have
been possible for the ancestors of Limnocnida, of Tiphobia,
Neothauma, and other molluscs, to have wandered into the
lake, just as at the present day individuals of the Atlantic
fauna wander into the Mediterranean. The lake having thus
received its marine population, was probably gradually separated
from the sea by the uprising of the land and became an inland
salt-water basin. At this stage it was possible for one of two
290 R. T. GUNTHER.
things to occur. It was possible either that the amount of
rainfall received by the lake should be less than the amount
of water lost by evaporation, or that it should be greater. In
the former case the lake would have gradually become salter
and salter, or in the latter case the excess of water would
have overflowed the banks and the water of the lake would
have undergone a gradual process of freshening owing to the
salt water being slowly replaced by fresh. It is possible that
both conditions may have obtained at various periods—indeed,
within our time the lake has been proved to be in both states
of quiescence and of overflow. But on the whole the condition
of overflow must have been the rule, because at the present
day the water of the lake is described as being “ perfectly
potable.”
This hypothesis of the origin of the marine fauna of Lake
Tanganyika may or may not prove to be possible when the
geology of the region is known. It is, at any rate in the present
state of our ignorance, a possible explanation of the presence
of a Medusa in the lake. Should subsequent researches prove
its untenability, it will have served its purpose if it should
stimulate naturalists to discover but a few facts about the past
history of the lake or to discover within its bounds members
of other classes of the animal kingdom with marine character-
istics, from which a new and better theory may be deduced to
account for the remarkable and unique marine character of
the present fauna of Lake Tanganyika.
BIBLIOGRAPHY.
1. Boum, R.—Vide Martens, E. yon, and R. Bohm.
2. Brooxs, W. K.—“ The Life-history of the Hydromeduse,” ‘Mem,
Bost. Nat. Hist. Soc.,’ vol. iii, 1886.
8. Brooxs, W. K.— Budding in Free Meduse,” ‘Amer. Nat.,’ p. 670,
1880.
4, Curzon, —.—‘ Persia and the Persian Question,’ vol. i, p. 533,
5. Guerne, Jutes pE.—“ La Méduse du Lac Tanganyika,” ‘La Nature,’
24th June, 18938, pp. 51, 52, 1893,
ANATOMY OF LIMNOCNIDA TANGANYIC#, 291
6. GintHER, R. T.—“ Preliminary Account of the Fresh-water Medusa of
Lake Tanganyika,” ‘Ann. Mag. Nat. Hist.,’ ser. 6, vol. xi, pp. 269—
275, April, 1893.
. Hertwie, O. and R.—‘ Organismus der Medusen,’ Jena, 1878.
. Hors, E. C.—‘ Proc. Roy. Geogr. Soc.”
9. Kennet, J. von.—‘‘ Uber eine Siisswassermeduse,” ‘S. B. Nat. Ge-
sellsch. Dorpat,’ Bd. ix, p. 282, 1890.
10. LanxestErR, E. R.—“ On Limnocodium Sowerbii,” ‘ Quart. Journ.
Micr. Sci.,’ vol. xx, 1880.
11. Martens, E. von, and R. Boum.—“ Uber eine Qualle im Tanganyika
See, mit Bemerkungen,” ‘ Sitzgsb. naturf. Fr. Berlin,’ pp. 179—200,
1883.
12. MaunsELt, F. R—*“ Kurdistan,” ‘Journ. Roy. Geogr. Soc.,’ June, 1893,
18. Sctater, P. L.— The Jellyfish of Lake Urumiah,” ‘ Nature,’ vol. xlviii,
p. 294, 1893.
14, Smiru, HE, A.—‘ On the Shells of Lake Tanganyika,” ‘ Proc. Zool. Soc.,’
p. 344, 1880.
15. Weismann, A.—‘ Entstehung der Sexualzellen bei den Hydromedusen,’
Jena, 1883.
16. Wissmann, H. von.—‘ Through Equatorial Africa’ (English translation),
p. 253.
os
EXPLANATION OF PLATES 18 & 19,
Illustrating Mr. R. T. Giinther’s paper “A Further Con-
tribution to the Anatomy of Limnocnida tanganyice.”
List or REFERENCE LETTERS.
e.¢, Circular canal. chr. Chromatin. cid. Thread-cells. ¢. Endothelium.
ect. Ectoderm. cf. /. Ectoderm lamella. end. Endoderm. exd. /. Endo-
derm lamella. end. m. Endoderm of manubrium. gl. Glockenkern. 4g. »v.
Gastro-vascular space. ms. Mesoglea. x. r. Nettle-ring. ov. Ovum. 7. ¢.
Radial canal. s. c. Subumbrellar cavity. _s. ect. Ectoderm of sub-umbrella.
sp. Spermatozoa. sé. Interstitial cells. ¢e. Tentacle. v. Velum. vac.
Vacuole.
Fie. 1.—Tangential section of the margin of the umbrella passing through
two of the marginal sense-organs. The knife has just cut off a few of the
peripheral cells of the larger marginal organ B. The smaller, immature
organ A is seen divided in the median plane. At its base is a large triangular
cell (7) thrusting itself through the mesogloa (ms.). The marginal bodies
292 R. T. GUNTHER.
are seen to be situated in a space surrounded by a definite endothelium (e.),
which separates the sense-organs from the mass of modified thread-cells com-
posing the nettle-ring (#.7.). x 1000. Zeiss, apo. 4 mm., comp. oc. 8,
cam. luc.
Fic. 2.—The next section to the one figured in Fig. 1. Sense-organ B is
seen in median section, and shows the difference between the granular basal
cells and the refringent central cells very well. The mesogloea (ms.) forms a
slight cup-like elevation round the base of the sense-capsule. The young
sense-organ A is also cut near the median plane, and the mesoglcea is discon-
tinuous between the organ and the endoderm. X 1000. Zeiss, apo. 4 mm.,
comp. oc. 8, cam. luce.
Fic. 3.—Radial section through the manubrium of a female individual, in
the region of the ovary. Ova (ov.) in various stages of development are
readily distinguishable from the interstitial cells (s¢.) on account of their
granular structure. The largest ova contain vacuoles (vac.), usually situated
on the outer side of the nucleus. Covering all is a limiting epithelium of
irregular cells (ec¢.). 400. Zeiss, apo. 4 mm., comp. oc. 4, cam. luc.
Fic. 4.—Section similar to the one figured in Fig. 3, but from the sper-
marium of a male individual. The layer of ectoderm cells (ect.1) nearest the
mesoglea (ms.) gives rise to the cells of the second layer, the nuclei of
which are alone visible (ect.*), and these in their turn break up to form the
sperm-cells themselves (sp.) No spermatozoa tails have made their appear-
ance in the stage figured. As in the case of the ovary, the entire organ is
enclosed by a limiting membrane (ec/7.3). x 400. Zeiss, apo. 4 mm., comp.
oc. 4, cam. luc.
Fie. 5.—Radial section through corner of gastric cavity (g. c.) to show
the point of junction between the endoderm of the manubrium (end. m.), the
endoderm of the gastric wall of the umbrella (ezd.), and the endoderm
lamella (exd. /.). All cell-walls disappear at the point of junction of the
three layers.
Fic. 6.—Radial section across the circular canal and through the base of an
interradial tentacle (¢e.). The lumen of the circular canal (ec. ¢.) is continuous
with that of the tentacle. The large ‘endodermal organ” fills up most of
the lumen of the circular canal. On the outer side is seen a portion of the
nettle-ring (7. 7.).
Fic. 7.—Multinucleate corpuscles from lumen of the circular and radial
canals. vac. Vacuole. chr. Chromatin. x 1000. Zeiss, apo. 4 mm.,
comp. oc. 8.
Fie. 8.—Longitudinal section through a Medusa bud in the youngest stage.
The bud is a mere hollow outgrowth of the wall of the manubrium, and its
cavity (g. v.) is directly continuous with the gastro-vascular cavity of the
parent.
ANATOMY OF LIMNOCNIDA TANGANYICA. 293
Fie. 9.—Longitudinal section through a later stage. The glockenkern (g/.)
has just commenced to invaginate at the tip of the bud.
Fie. 10.—Oblique longitudinal section through a stage in which the in-
vagination of the glockenkern is considerably advanced, and the lips of the
aperture of the future subumbrella cavity (s. c.) have just commenced to close
over. As the section is cut obliquely, the gastro-vascular space has just been
missed.
Fig. 11.—Oblique section through a Medusa bud, in which the glockenkern
(g/.) has sunk into the body of the bud, and has become closed over by an
overgrowth of ectoderm. Note the persistence of the future subumbrella
cavity in the glockenkern.
Fic. 12.—Longitudinal section through a stage at which the second
ectodermal ingrowth has commenced. This takes the form of a solid plug of
cells (ec¢.), which has pushed into the glockenkern and rendered it cup-shaped.
The lumen of the glockenkern has been squeezed into a mere lamina.
Fic. 13.—Longitudinal section through a considerably later stage. The
forecasts of the tentacles (¢e.) have appeared, and the velum (v.) is well
marked though as yet imperforate. The subumbrella cavity (s. c.) is large.
Fie. 14.—A longitudinal section of a bud passing along one of the four
radial canals (7. c.). The circular canal (c. ¢.) has also appeared in this stage;
the velum (v.) is now perforate, and the tentacles (eight in number) are now
free from one another at the tips, though their bases are still connected by a
bridge of ectoderm (ect.). Note the thread-cells (czéd.) at their tips.
Fie. 15.—Transverse section through a bud at a slightly later stage than
the one figured in Fig. 14. The mesogloea (ms.) has begun to be deposited
between the exumbrella ectoderm and the endoderm lamella (end. /.) Note
the four radial canals (7. c.) The tentacles have been cut more or less
obliquely, and are hanging through the centre of the velum (v.)
Fic. 16.—An abnormal bud, in which the layers are reversed and which
was found growing into the gastro-vascular cavity of the parent. ect. is the
exterior ectoderm. The tentacles (¢e.) and velum (v.) are hanging into the
gastric cavity (g. v.). At # x are the disconnected rims of ectoderm, which in
a normal bud would be joined together to form the centre of the ectodermal
lining of the oral region of the subumbrella cavity, but which have here
become separated by the growth of the bud, which has in fact grown out
through its own mouth.
Figs. 8a—14a are diagrammatic representations of Figs, 8—14 respectively.
{ Ve Cretan ‘ avi : Pr wi IeNTLY fol 2 hig re || A 4
¢ ands TQ F-Pt ry ea eon i
=Gi NGn @ ] (ie Ulia i re hiya h” gees Cat ii Fates
i) Qa) rill wren Teton on aes poy abd wi
, phy maaan y wy ei es fs ? rh 4 Ay iF fuiege ay ye
,
Nit Ab ml wigné Thhheer- oe ole olin s oa ee AL or
riay fon Ly) UP sia i! oi f Pier -2 hos il ijn.) FLA hoe iy
ha Va TI I uy ad yo tg oe ba at Ry fund Wt) iia
4
Miays in] ae ad) we ; 5) noua. be
ha it04 AA ase) ¥ p OL ie DT YIRGs Lait s aeetie ne
. ’ ‘ 4 “
Begu! i i) , of A Le gulice at rr eeit ue sili ne eae
5 ‘ a *
cDore ad tA LY aT thous mie ev nikved and o Hae ay
ea Vaasst itp ASAN Wie AAP ab) Mra etre itl 1a ‘
7 = : Fs : . =~ tare
ny? ee TMLee a ae , Pst ml ; Oz on
liv a L0r San rars i? fa} 7h! + ys Art bs ,
eras uk : Ales}, Wigend rs oy, | es Tae le ah
Tit a7 ‘ i oe ao ~ j ; ‘ mi S idinowens ie ee
j foe 4,9.59) sie VA ii MN A ys
Sia Sid , ral? } Ta) , iS .2iy shia wi Pa,
i ie 4 Nae as i Se
4 FI ’ ) iG A 4 cmt F ‘ j GAMO" ght ip Ja Toth UL. re,
of f LA ta ¢. waits lind oyu A oe Sia HE = th *) ist ols eae
ail ryt i 'y ; té nA i a ews wert rt Vp hwo
ile wht siogel eal. (ive) ambiowiie ait At ity ch Roasts
on tA a oe ; } ia :
we] Lod ehiiee Sap fl ; Dy GL WAG Oren iy le fe 4
- ‘ 7 ~ j ‘ ; f Sa
OE SD SAE Th rw Gerla adam ai, Goat) BRS be tne
Cy Malek Wald i WO Set & Ai) vill ani Siee a vot j
a hb Pe Dea oe hey LP oes AP eae Lactenaiaie ah}
ee ee ae A Cateye (ot Bis Be) i ae it | Nia ube>
Teens \ Bi 4 ie Bios
shes Te phe mo.) 2s & Fol he Oh Bie Gee e pe «
apt Oe OUP eth eS We Ane ae ead navel) oN dk bitin bub)
Ovne RAG oh we dak iwi «Aan til Oe bak Wa ntpaoe Treftes tke ty
OG pS: FB ak ih) ie OU a Ty EP A inal oe .
" 3 “il imily il mtr F '
-_ =
Reiss iI- +6, e604 Ie an al inane uve met ih) ota td rf
hh rea t aN -_, wy ha OPE oh a3 ts os ye “ona
7 » e eo 4 —
:
a ae - i 2 <a
‘ ad _ “oe 0
‘ -
,
'
‘
t
. -
r 7
kK
MINUTE STRUCTURE OF PELOMYXA PALUSTRIS. 295
Notes on the Minute Structure of Pelomyxa
palustris (Greeff).
By
Lilian J. Gould,
Hall Scholar, Somerville Hall, Oxford.
With Plates 20 and 21.
Waite working as a student in the laboratory of the Linacre
Professor at Oxford, during the summer term of 1893, the
opportunity was given me of examining a considerable number
of specimens of the interesting fresh-water Rhizopod Pelomyxa
palustris, originally described by Professor Greeff in 1867 as
Pelobius, and later discussed by him under its present name
in the ‘ Archiv fiir mikroscopische Anatomie,’ Bd. x, 1874. I
was enabled, by the kindness of Professor Lankester (to whom
my best thanks are due for affording me the material for these
observations, as well as for much valuable advice and assist-
ance), to study more than twenty individuals of P. palustris,
both in the living state and by means of sections.
Some points with regard to the minute structure which I was
able to make out have not, as far as | am aware, been observed
before, and others which had been doubtful have received
confirmation. It has therefore seemed desirable to bring
forward the results of my observations, as far as they go, with
the conclusions I have drawn from them, although one or two
of these conclusions must be regarded merely as inferences
until verified by further investigation. Most of the sections
were cut by the skilled hand of my friend Mr. EH. A. Minchin,
B.A. (to whom I am much indebted for his kind help), and
296 LILIAN J. GOULD.
some were also stained by him. Others cut by Mr. Minchin
were prepared by me, and together we obtained a most instruc-
tive series, which I studied very carefully. The figures in the
accompanying plate are from drawings I made with the camera
under high powers, and illustrate the chief points dealt with in
these notes. The stains used were Paul Mayer’s new “ carm-
alum’? and ‘“ paracarmine” (‘Mitth. Neapel, Band x,
1891-3, pp. 489, 491); anilines such as eosin, fuchsin,
fuchsin S., orange G., and gentian violet, either alone or in com-
bination; also picro-carmine and hematoxylin. The sections
varied from 1 to 7 «in thickness. The only observations of mine
which can claim to be new have reference (1) to the appear-
ance of a central mass of doubtful significance in the general
protoplasm, (2) to the staining properties of the refringent
bodies or “ Glanzkérper” of Greeff, and (3) to the perfectly
definite jointing staining-reaction of the rod-like bodies (a
point suspected but hardly established hitherto), which
establish the view that they are bacteria. These points, and
others confirmatory of previous observations, are considered
separately as follows.
General Structure of the Protoplasm.
The protoplasm of Pelomyxa is well known to be highly vacuo-
lated, and in P. viridis Professor A. G. Bourne (‘ Quarterly
Journal of Microscopical Science,’ xxxii, 1891) distinguishes
between the large vacuoles, which are comparatively few, and
the much more numerous smaller ones which he calls “ vesicles,”
and which, in the species described by him, have chlorophyllo-
genous contents. The difference in size, though not in colour,
between ‘‘vacuoles” and “vesicles” obtains in P. palustris,
and the numerical proportion of one to the other is much the
same. Thus the general character of the protoplasm has been
considered as practically identical in the two species. But I
was able to make out that, in P. palustris at least, the proto-
plasm surrounding the vesicles was by no means homogeneous,
as stated by Professor Bourne for P. viridis, but showed
very distinct structure.
MINUTE STRUCTURE OF PELOMYXA PALUSTRIS. 297
The examination of sections under very high powers lent
strong support to the views of Professor Biitschli as to the foam-
like structure of protoplasm (‘ Mikroskopische Schaume,’ 1892).
I was able to confirm the existence of the very fine vacuolisation
described by him (loc. cit., pp. 200 and 216) for P. palustris,
which probably occurs also in P. viridis, but, as he suggests,
may have escaped Professor Bourne’s notice. The vesicles in P.
palustris, which corresponded to the smallest seen by Professor
Bourne in P. viridis, were easily known by their size as com-
pared with that of the nuclei in both forms; and a foam-like
structure, consisting of alveoli of infinitely smaller diameter
(4—1 ) than these vesicles, was most distinct in many of the
sections. Fig. 11 shows this fine alveolar structure with a
nucleus and two vesicles for comparison; in fig. 10 it is seen
on a larger scale. The strands of protoplasm bounding these
smallest alveoli, and also some of the thin strands between the
large vacuoles, appeared to me homogeneous, and comparable
with the finest pseudopodia of Rhizopods and the finest strands
of vegetable protoplasm described by Biitschli (loc. cit., pp.
67, 79). The peripheral radiate alveolar border of Biitschli,
said to be characteristic of froths, was clearly distinguishable
in some sections (fig. 2), and the radiate alveolar layer de-
scribed by him is seen in fig. 9, (2) round a nucleus, and (6)
round a refringent body.
Appearance of a Central Mass.
In one individual of Pelomyxa, of which a consecutive
series of sections had been cut, a very curious appearance was
observable. The animal had been killed with osmic acid, and
stained in bulk with carm-alum, a delicate protoplasmic stain.
_ The sections presented a central, more deeply staining, irre-
gularly oval ring of apparently denser protoplasm, which was
traceable through many sections. ‘The ring diminished at
each end of the series to a small central solid patch of denser
protoplasm, and hence apparently represented a more or less
spherical or oval mass, reminding one somewhat of a central
298 LILIAN J. GOULD.
capsule. The character of the protoplasm was somewhat
different without the ring from that within. External to the
ring it was highly vacuolated; from the periphery of the
animal inwards areas containing very few nuclei or vacuoles,
but consisting almost entirely of small vesicles, extended
nearly to the boundary of the ring, in some places running
quite up to it. Fig. 1 shows a whole section about the middle
of the series, drawn without details, showing merely the gene-
ral appearance and size of the ring, with the nuclei and largest
vacuoles. The dotted regions represent the vesicular areas.
The ring itself exhibited the extremely fine alveolar structure
above mentioned. Internal to the ring I could make out
nothing more than a very finely granular appearance of the
protoplasm; alveoli, if present, must have been infinitely
small in diameter. Fig. 8 is a high-power view of a portion
of the ring; the area marked out by dotted lines is represented
on a much larger scale in fig. 10, to show more accurately the
character of the alveoli. The capsular appearance was not due
to the effect of semi-penetration of the osmic acid used in kill-
ing, since an individual treated with the same reagent for the
same length of time, but stained with picro-carmine, showed
no trace of the structure. Some specimens of P. palustris
obtained later, which were killed with alcohol and stained in
bulk with carm-alum, showed a slight tendency to the same
appearance; there seemed to be a drawing together or central
concentration of the protoplasm, though there was no definite
ring formed. But these individuals presented, in the living
state, quite a different appearance from those examined pre-
viously. They were shrunk up into a globular shape, were
brownish in colour, and perfectly quiescent, so that on first
examination they seemed to be dead. But after they had
been under the microscope for a long time, they gradually
began to assume a more normal appearance, and very slowly
put forth pseudopodia. It was a condition suggesting en-
cystment, but no definite cyst-wall was found. It appeared
to me possible that these individuals were in a stage leading
on to, or nearly connected with, that seen in the first-mentioned
MINUTE STRUCTURE OF PELOMYXA PALUSTRIS. 299
specimen, and that the quiescent condition may have been pre-
paratory to the production of swarm-spores.
The outward appearance of quiescent individuals certainly
seemed somewhat to resemble that described by Greeff (loc.
cit.) as preceding reproduction. The protoplasm of all quies-
cent individuals was filled with sand-particles to such an
unusual extent that the cutting of sections was a matter of
extreme difficulty.
The “ Glanzkérper.”
The refringent bodies, or “ Glanzkérper” of Greeff, were very
numerous in P. palustris. Professor Bourne says he never
saw “anything in the protoplasm of P. viridis resembling ”
them, but it seems to me possible that he might have con-
founded them with vacuoles of the same size, as in life they
were often not easily distinguishable (except by the fact of their
extrusion from the body) from these.1_ Nor were they always
recognisable in sections when stains such ascarm-alum and picro-
carmine were used. Alum-carmine, used by Professor Bourne,
probably would not show them up either. But I found that
they stained very readily and deeply with fuchsin, eosin, dahlia,
solution of iodine in potassium iodide, picric acid dissolved in
turpentine, and some other stains, and could thus be picked out
with beautiful distinctness. With all stains except picric acid in
turpentine they appeared perfectly homogeneous, but with the
latter reagent they showed plainly a fine granulation, and
sometimes contained a small bright crescentic area which
might represent a space or cavity in the interior. From my
observations I concluded that they were almost certainly either
solid structures or filled with a coagulable fluid. Greeff
observed the granulation with acetic acid, and found these
bodies to become deep brown with iodine. I found a solution
of iodine in potassium iodide to have the same effect. He also
thought he could sometimes distinguish in the refringent bodies
1 IT am inclined to regard the chlorophyll-bearing “ vesicles” of Bourne’s
P. viridis as equivalent to the “ glanzkorper ” of P. palustris—E. R, L.
300 LILIAN J. GOULD.
some kind of contents and a nucleus, but I found no appear-
ance of internal structure other than that described above.
I am inclined to think that he, too, sometimes confounded
refringent bodies with vacuoles, since he describes a falling in
of the walls of the former which occurred with some reagents,
and I found a crenellation of this kind to be very character-
istic of the food-vacuoles under certain conditions. The re-
fringent bodies divide by constriction, and in fig. 6 the process
is seen near completion. M. Pénard (‘ Archives des Sciences
physiques et naturelles,’ tome xix, 1893) says that colouring-
matters have little or no effect upon them, and that they are
either structureless or coutain vacuoles. Possibly he looked
upon the crescentic areas as vacuoles, or, as he evidently used
different reagents, he may have failed to distinguish the
refringent bodies from the food-vacuoles, which generally have
contents of some sort.
The Vacuoles.
The vacuoles proper were of two kinds, viz. (1) large non-
contractile vacuoles, which did not stain (figs. 2 and 8), and (2)
food vacuoles of varying size, which were found with and
without contents. These contents in all cases stained freely
with carm-alum and picro-carmine, and the vacuoles were
further distinguishable by the greater thickness, and often by
the crenellation of their walls (fig. 8). The “ vesicles” of
Professor Bourne, which greatly outnumber the vacuoles, and
must be placed apart on account of their having in P. viridis
chlorophyllogenous contents, are perhaps, as suggested by
Professor Lankester, not to be regarded as “ vacuoles,” but
as corresponding to the glanzkorper of P. palustris.
The Nuclei.
With regard to these I have nothing new to add. The
nuclei were lodged in the nodes of the protoplasmic network,
and presented, as described by Greeff and others, a finely
granular structure with several nucleoli in the middle, and
deeply-staining chromatin granules arranged peripherally in
MINUTE STRUCTURE OF PELOMYXA PALUSTRIS. 301
a more or less ring-like fashion. I was not able to detect any
appearance of nuclear division. The radiate alveolar layer
round the nuclei and refringent bodies has been noticed
above.
The Bacteria.
The rod-like bodies found in profusion in Pelomyxa were
originally described by Greeff as “crystals,” but later observers,
firstly Bourne, and after him Pénard (loc. cit.), have been
inclined to regard them as symbiotic bacteria. Leidy (‘ Fresh-
water Rhizopods of N. America, 1879’) and Greeff both
thought they distinguished a “ transverse striation’’ of the
rods.
M. Pénard considered the rods as “ véritables bactéries.”
He saw in them “une striation transversale, ou plutdt une
apparence de divisions 4 intervalles réguliers, telles qu’on
les trouve dans les algues filamenteuses inférieures.’’ With
reagents the rods appeared “‘nettement divisées en plusieurs
parties par des étranglements,” or “ reduites en quelque sorte
a des chapelets, dont les nombres respectifs de grains étaient
de 2, 3, ou 4.” But he did not make it very clear whether
the appearance he saw was one of constriction merely and due
to reagents, or jointing, which is rather a different thing.
Nor did he state definitely the largest number of divisions
seen in a single rod, while his description of some of the
longest as “ filaments ondulés ou recourbés” rather inclines one
to think that some of those he saw might have been really the
“algues filamenteuses” to which he compared them, and
which are not uncommonly found in the protoplasm of
Pelomyxa. The rods, as figured in the plate accompanying
M. Pénard’s paper, show either transverse striation or a
moniliform appearance ; the former does not of course amount
to jointing, and the latter is very different from anything seen
by me.
In a Pelomyxa killed with osmic acid, stained in bulk with
carm-alum, and teased up in glycerine, I found that the
rods were not constricted, but very distinctly jointed
302 LILIAN J. GOULD.
(fig. 4). I obtained the same result with a specimen treated
in the same way, but stained with picro-carmine. In the sec-
tions, which were mostly double- or treble-stained with anilines
such as eosin, fuchsin, fuchsin S., orange G., and dahlia, the
rods were deeply stained, and the jointing could be well seen ;
in some cases the terminal joint could be seen separating off
(fig. 7). The rods were always straight and made up of 2, 3, 4,
6, and sometimes even 9 joints. They were highly refringent,
and their refractive index seemed to be nearly the same as
that of Canada balsam, since in preparations mounted in the
latter medium the rods were difficult to see, while in glycerine
they were most distinct. The joints had the shape of long
prisms, and in a few of the 6-jointed rods a further transverse
division of each joint into two was apparent.
In the living Pelomyxa the rods were frequently thrown
out into the water, together with refringent bodies and nuclei,
a process evidently abnormal and the result of unfavorable
conditions. When thus thrown out, the rods exhibited active
movement of a kind which has been considered as possibly
molecular (Bourne, loc. cit.), but they also travelled round
the periphery of the animal. I could not absolutely satisfy
myself that the latter movement might not have been due to
currents created in the water by the activity of the pseudo-
podia. Still, taking all the appearances together, it seemed
impossible to doubt that the rods were really bacteria.
My friend and fellow-student Mr. M. D. Hill, of New
College, Oxford, undertook to prove this by cultivation of the
bacteria in suitable media, and some account of his prelimi-
nary investigations, which are not yet completed, are appended
here.
With regard to the situation of the bacteria, they were
scattered more or less throughout the protoplasm, but were,
as stated by Greeff, especially abundant around and adhered
thickly to the walls of the refringent bodies. Fig. 7 shows
some of the latter cut through at different levels, and here the
rods are plainly seen in situ. The rods alone are represented
in fig. 12.
MINUTE STRUCTURE OF PELOMYXA PALUSTRIS. 308
New Species.
If further criticism of M. Pénard’s observations on Pelo-
myxa be permissible, it would seem that he has scarcely
sufficient grounds for the establishment of a new species
(P. beleosteii). He does not mention any definite feature
which is not equally characteristic of P. palustris. Size
is no criterion, since individuals of P. palustris vary very
much in this particular, as also in the presence or absence of
sand débris in the protoplasm. The sole real difference appears
to be in the structure of the nuclei; but as this also differs in
two nuclei from the same animal (according to the figure),
and both of these, from their thick walls, size relative to the
vesicles, and general appearance, bear far more resemblance
to food-vacuoles with contents than to the nuclei of any
amceba, it would seem doubtful whether they were really nuclei
at all. In a paper published several years ago in the ‘Archives
Exp. de Zoologie,’ Korotneff has given reasons for recognising
a second species of Pelomyxa. I have no doubt that the form
studied by me is the P. palustris of Greeff.
ADDENDUM.
November 10th, 1893.
In order to furnish, if possible, a conclusive proof of the
organic nature of the so-called “ rods” of Pelomyxa, Professor
Lankester suggested to me that I should attempt to obtain a
cultivation of these organisms by means of the usual bacterio-
logical methods. This I did during June, 1893, and the
following is a short account of the work.
The investigation was carried on in the laboratory of the
Regius Professor of Medicine in the University Museum, with
the kind collaboration of Dr. Ritchie, who was then, and is at
present, engaged in bacteriological research.
We made a large number of cultivations by teasing up fresh
specimens of Pelomyxa in sterilised water, after removing
them straight from the pond water in which they were kept,
VOL. 36, PART 2,—NEW SER. x
304 LILIAN J. GOULD.
and “sowing” them in test-tubes containing various media,
e.g. blood-serum, beef jelly, bovril, &e. The tubes were kept
in the dark at the ordinary temperature of the room, except
in some cases where artificial warmth was applied. In all
these cases a very mixed culture was obtained, and it was im-
possible to say whether the growth was derived from the
“rods.” In one instance, however, we were able to detect a
few long bacteria, which probably were those for which we
were seeking, but it was found impossible to convince ourselves
that this was the case.
Furthermore, at Professor Lankester’s suggestion, we tried
immersing the animals for a moment, some in dilute corrosive
sublimate, others in strong alcohol, in order to kill, if possible,
the foreign micro-organisms which would naturally be clinging
to the surface of the protoplasm, without injuring those in the
interior. Cultures were made in the same way as in the pre-
vious experiments, but only one was successful. Here we
obtained a pure colony of short rod-like forms, which may
have been the “rods” in a more finely divided state than
that in which they appear normally in the living Pelomyxa.
It is probable that, when supplied with abundant nutriment,
the ‘‘rods”’ would break down and multiply so rapidly as not to
allow themselves to assume the many-jointed condition which
Miss Gould has described.
Other methods were tried, such as the “ hanging drop” and
“ fractional ”’? methods, but where colonies were produced they
were of too mixed a character to give conclusive results.
At present, therefore, our results are negative as regards
having obtained a demonstrably pure cultivation.
My best thanks are due to Dr. Ritchie for his kindness in
putting his apparatus at my disposal, and for his help during
the work.
M. D. Hi.
MINUTE STRUCTURE OF PELOMYXA PALUSTRIS. 305
DESCRIPTION OF PLATES 20 & 21,
Illustrating Miss Lilian J. Gould’s paper, ‘‘ Notes on the
Minute Structure of Pelomyxa palustris” (Greeff).
Fic. 1.—General view of complete section of Pelomyxa palustris
without details. Killed osmic acid, stained in bulk carm-alum. Zeiss,
obj. B, comp. oc. 4; camera. 7. Ring of denser protoplasm. 2. Nuclei.
v. Vacuoles. v.a. Vesicular area.
Fic. 2.—Portion of section. Killed osmic acid, stained in bulk carm-
alum. Zeiss, obj. E, comp. oc. 4; tube at 17 mm. z. Nuclei. v. Non-staining
large vacuoles. fv. Food-vacuoles. v.v. Vesicles. yp. 1.4. Peripheral
radiate border.
Fie. 3.—Portion of section. Killed osmic, stained carm-alum, fuchsin,
and eosin. Zeiss, obj. E, comp. oc. 4; tube not drawn out; camera. g. Re-
fringent bodies. «. Nuclei. 4. Bacteria. s. Sand and débris. p.7. 2. Peri-
pheral radiate border. d. Diatom.
Fic. 4.—Teased-up portion of P. palustris in glycerine. Killed osmic,
stained carm-alum. Zeiss, obj. E, comp. oc. 4; tube not drawn out; camera ;
b. Bacteria. g. Refringent bodies. 2. Nuclei. pr. Droplets of protoplasm.
Fig. 5.—Teased-up portion in glycerine. Killed osmic, stained picro-
carmine. Zeiss, obj. E, comp. oc. 4; tube not drawn out; camera. 4. Bac-
teria. g. Refringent bodies. x. Nuclei.
Fic. 6.—Portion of section. Killed osmic, stained picro-carmine and
picric turpentine. Zeiss, obj. E, comp. oc. 4; tube not drawn out; camera.
g. Refringent bodies. gg. Dividing refringent body. x. Nuclei.
Fic. 7.—Refringent bodies, to show rods in situ. Killed osmic, stained
picro-carmine and dahlia. Zeiss, oil imm. 5, comp. oc. 4; tube at 17 mm. ;
camera. z. Nuclei. 4. Bacteria. g. Refringent bodies. w. Wall of refringent
body.
Fic. 8.—Small portion of section, including bit of capsular region. 77.
Width of ring. 7. pr. Protoplasm within ring. e. pr. Protoplasm external
to ring. fv. Food-vacuoles. sg. Alveolar area represented in Fig. 10.
Killed osmic, stained carm-alum and dahlia. Zeiss, oil imm. 3, comp. oc. 4;
tube at 19 mm., finished with comp. oc. 8; camera.
Fic. 9.—A. Nucleus, with surrounding radiate alveolar layer. Killed
osmic, stained carm-alum. 3B. Refringent body, with surrounding radiate
306 LILIAN J. GOULD.
alveolar layer. Killed osmic, stained picro-carmine and picric turpentine. Zeiss,
oil imm. 4, comp. oc. 8; tube at 19 mm.; camera.
Fie. 10.—Protoplasm between vacuoles, to show finest alveolar structure.
v. Vacuoles. a. Alveoli.
Fic. 11.—Small portion of section. Killed osmic, stained picro-carmine
and hematoxylin. Zeiss, oil imm., comp. oc. 8. 2. Nucleus. v. Vacuoles.
v,v. Vesicles corresponding to the chlorophyllogenous vesicles of P. viridis.
Fic. 12.—Bacteria. Zeiss, oil imm. 3, comp. oc. 4; tube not drawn out.
MONILIGASTER GRANDIS, A. G. B. 307
On Moniligaster grandis, A.G.B., from the
Nilgiris, 8. India; together with Descriptions
of other Species of the Genus Moniligaster.
By
Alfred Gibbs Bourne, D.Sc.Lond.,
Professor of Biology in the Presidency College, Madras.
With Plates 22—28.
EXTERNAL CHARACTERS.
Colour and Size.—The general appearance, colour, and
size may be judged from figs. 1 and 2. Anaverage-sized worm
which I measured was 590 mm. long when living and at rest,
270 mm. when fully contracted, and 1080 mm. when stretched
out after having been (purposely) badly preserved.
The most striking feature in the living worm is the great
activity of the twelve or fourteen most anterior somites as
compared with the extreme sluggishness of the greater portion
of the body. When removed from the ground the animal
seems to have very little control over all the hinder portion of
its body. The body-wall is here very thin and weak, and often
becomes ruptured in specimens which are kept in captivity.
When this happens the gut instantly bulges out at the point of
rupture.!
Great differences exist among earthworms with regard to what happens
when the body-wall is injured. Some worms, like M. grandis, will live on
for a long time with a portion of the gut bulging out, while others, e. g.
Perionyx saltans, will, on receiving the slightest injury to the body-wall,
breek at once into two pieces. The body-wall behaves in some cases as though
VOL. 36, PART 3.—NEW SER. Y
308 ALFRED GIBBS BOURNE.
There is very little pigment in the body-wall, which in the
greater portion of the animal’s length is very transparent, but
which becomes opaque at the two extremities. The opacity is
due even here rather to the greater amount of muscle present
than to the presence of pigment. What little pigment there is
occurs in the connective tissue which lies below the epidermis
and between the muscle-fibres. There is in Segments x—x111
(the region where the clitellum develops) in all fully-grown
worms an orange tint which is due to pigment, and the dorsal
region of the anterior segments of a worm which has been
killed in spirit presents a greenish appearance which must be
due to the alteration in colour of a pigment. When the worm
has been preserved in spirit for some time, this coloration passes
away.
The coloration of M. grandis depends chiefly upon the
earth which it takes into the gut. It almost always occurs
here in a light-coloured soil, and is consequently itself light in
colour.
Number and Character of the Segments. — The
largest number of segments which I have met with is 480, the
smallest 266, this in a worm 3% inches long. One of the most
striking features in the external appearance of the worm is the
great variation in size of the segments (figs. 1,2, 15,16). They
increase in both length and circumference from before backwards
down to Segment vit1, which is the longest in the body. Seg-
ments 1x and x are of about the same length as one another, but
rather shorter than Segment vi1r. Segments x1, x11, and x1Ir
are each rather shorter than the preceding segment. From
Segment x111 there is but very little decrease in the length of the
segments, which are all, as compared with most worms, very
short relatively to their circumference. The circumference of
the body varies so much with the amount of earth in the
alimentary canal that one must be very guarded in making any
statement respecting it (fig. 1 gives the natural size of a big
it: were tough, and in others as though it were brittle, and the end of the
worm breaks off on receiving the slightest injury—like the end of a lizard’s
tail.
MONILIGASTER GRANDIS, A. G B. 309
worm full of earth), but if a starved worm, from which most of
the earth has been. voided, is killed by immersion for a few
minutes in strong spirit and then stretched, it may be very
clearly seen that the body tapers from Segment vuit, gradually
but regularly, right away down to the tail end, which is then a
blunt point. If such a worm is left in strong spirit until all the
muscles have been contracted and become firm and hard, the
region immediately in front of the anus becomes swollen out _
into a knob. I think that this is due to the strong contraction
of the thick layer of longitudinal rectal muscles.
This particular shape and this great variation in the size of
the segments is not by any means common among earthworms,
and does not obtain, for instance, in the majority of Monili-
gasters.
As a rule the segment is clearly divided into two anuuli,
while each of these is again obscurely divided (at any rate in
about the anterior half of the worm) into two, making in all
four annuli (see fig. 15, Segment xv). The sete are placed
upon the second of these four annuli. The fourteen anterior
segments are peculiar in their annulation. Segments 1, 11, and
111 consist of one annulus only, and Segments tv and v of
two annuli only, while in Segments vi—x there are more than
the usual number of annuli, but the additional annuli are very
irregularly marked and frequently do not extend round the
whole circumference. Fig. 3 is accurately drawn from one in-
dividual. When the clitellum is developed the annulation
in that region disappears, but I have never seen the demarca-
tion between the segments disappear or even become to any
great extent obscured.
Prostomium.—This is not dovetailed into the peristomial
segment (Segment 1), but is very definitely marked off from it
by a groove. The anterior edge of the peristomial segment is
almost always turned in, so that the prostomium appears to
protrude from the buccal cavity. When the prostomium is
fully protruded a portion of the buccal membrane is protruded
also (see fig. 4). It can be entirely withdrawn, and in spirit
specimens is usually invisible (in fig. 15 the prostomium is
310 ALFRED GIBBS BOURNE.
entirely withdrawn). It is constantly changing its shape
during life. It is hollow, the body-cavity being prolonged
into it. Muscles attached to its inner wall and passing back-
wards to the body-wall serve to withdraw it (see fig. 17),
while if coelomic fluid is forced into it, it becomes much dis-
tended.
Setz.—The setz are very ordinary in shape (fig. 26), and
bluntly pointed at either end. They present no remarkable
features, nor do they differ, so far as I can see, to any extent
from one another. They are as nearly as possible 0°5 mm. in
length ; that is to say, they are very small as compared with
the setze of most worms—as small as those of Microcheta.
They are arranged in couples, the two setz in each couple being
very near together. In a large-sized specimen which I measured,
the circumference at about Segment xxv was 36°5 mm.,
the dorsal gap 18 mm., the lateral gaps 6°5 mm. each, and
the ventral gap 5°55 mm. More anteriorly the dorsal gap
increases very slightly at the expense of the lateral gaps.!
Both seta couples are, as usual, absent from Segment 1; they
are also absent in this worm from Segment 11, while in Seg-
ment 111 the inner couples alone are present. I have never
found any instance of any of the sete having accidentally fallen
out or been lost ; and young sete (soies de remplacement),
sO common in many worms, do not often occur; I think that
they occur at certain particular seasons, when a new set of
setee develop. There are no modified sete of any kind. The
setee are so small that the position of neither the individual
seta nor of the couple can be observed on an inspection of the
inner surface of the body-wall. The position of the rows of
seta couples is, however, clearly marked by the arrangement
of the muscles (see below, Muscular system).
1 The relative positions of the seta rows are of importance for systematic
purposes in this group. My measurements are taken in the following
manner: the worm is killed in-strong spirit, a ring of the body-wall, con-
sisting of one or two somites, is cut off at about Segment xxv, cut through
along one of the seta rows, and then flattened out, avoiding any stretching,
and measured.
MONILIGASTER GRANDIS, A. G. B. 311
Clitellum.—A clitellum undoubtedly develops—at a cer-
tain restricted time of year—in connection with Segments x,
XI, x11, and x111. It commences, even in the most fully de-
veloped specimens which I have seen, on the second annulus
of Segment x. It is not developed at all upon the ventral area
of Segments x, x1, and x11, and so does not actually include
either the male or the oviducal pores, but it is developed over
the whole of Segment x111._ Before it becomes recognisable in
living specimens, it may be distinguished in a worm which
has been placed in spirit for half an hour, as the future clitellar
area then becomes orange-coloured.
Mouth.—The mouth is ordinarily completely occluded by
the retracted prostomium, but it is capable of great distension.
From time to time the worm protrudes the prostomium and
the whole of the buccal membrane, and, indeed, the whole of
the dorsal wall of the pharynx.
Anus.—The anus is a narrow vertical slit.
Dorsal pores.—There are no dorsal pores.
Nephridiopores.—These are very easily seen. There is a
pair in every segment, except Segments 1 and 11 and the two
most posterior segments. They are all placed close to the in-
tersegmental groove, but upon the auterior portion of the
segment in which the nephridium les. The seven most
anterior pairs of nephridiopores, 1. e. those of the nephridia of
Segments 111—1x, lie to an equal extent dorsad of the outer
seta row. The remaining pores lie exactly in the outer seta
row.!
Genital Apertures.—There are three pairs of genital
apertures, all strictly intersegmental, and lying between Seg-
ments VII-VIII, X-XI, and x1-x11 respectively.
The male pores lie between Segments x-x1i, and slightly
dorsad of the inner seta couple. They are large oval apertures,
and the immediately adjacent annuli form concentric lips (fig.
15), and the surface is here and in the immediate neighbour-
1 This peculiar arrangement shows how careful it is necessary to be in de-
scribing the position of nephridiopores, especially in spirit specimens, where
it is often very difficult to see them at all,
312 ALFRED GIBBS BOURNE.
hood much corrugated. Sometimes after the animal has been
killed in spirit a conical papilla is formed, at the top of which
lies the actual pore. No other papille are ever found in this
region.
The oviducal pores lie between Segments x1 and xu, slightly
nearer the median ventral line than the male pores, and there-
fore almost exactly in the direction of the inner seta couples
(fig. 15). The pores are exceedingly minute, and are indeed
hardly recognisable. The corrugations of the surface above
mentioned extend as far as the oviducal pores (fig. 15).
The spermathecal pores lie between Segments vii and vit,
and very slightly ventrad of the outer seta row. They are
easily recognisable, and appear to lie upon a little eye-shaped
piece of tissue which is inserted between the two segments
(figs. 15 and16). Here, as in the case of the male pores, little
papillz may stand out with the pore upon the apex.
Bopy-watt, Bopy-cavity, anD Septa—Muscutar System.
Body-wall.—The body-wall is very thick in the anterior
region, but is everywhere else very thin compared to what it is
in many worms.
The cuticle presents the usual characters.
The structure of the epidermis is very clear and distinct.
Five kinds of cells occur in the epidermis—ordinary epidermic
cells, sensory cells, and three kinds of gland-cells, viz. (1)
goblet-cells, (2) short club-shaped cells, and (3) long club-
shaped cells.
Setting aside the sensory cells, it is only necessary to dis-
tinguish between the clitellar segments and the non-clitellar
segments or main portion of the worm.
In a non-clitellar segment the only glands present are the
goblet-cells (fig. 18, gob.). These occur in great numbers in
the middle of the segment, but are absent from the interseg-
mental grooves (fig. 59). They are as numerous in the inter-
annular grooves as elsewhere. The ordinary cells are rather
shorter in the intersegmental grooves than elsewhere, and
MONILIGASTER GRANDIS, A.G. B. 313
strictly columnar in shape. Those cells which are packed be-
tween the goblet-cells are—at any rate when the goblet-cells
are full of secretion—much pressed out of shape. In many of
my sections the goblet-cells are extremely distended, but they
never dip down below the ordinary epidermic layer. The
nucleus is pushed right down to the base of the cell. The
secretion does not become stained on treatment with alum
carmine. These cells correspond apparently to the finely
granulated cells of Lumbricus,! and they are the only gland-
cells present, there being no cells in a non-clitellar segment
corresponding to the coarsely granulated cells of Lumbricus.
I have recognised these two kinds of gland-cells in a good
many genera, including Lumbricus, but it is very clear that
only one kind is present in M. grandis, or, I believe, any
Moniligaster.
The clitellum, as stated above, develops upon Segments x, x1,
x11, and x11, but not upon the whole of Segments x, x1, and
x11, so that in the clitellar region we have to distinguish
between clitellar epidermis and non-clitellar epidermis, and
there is further a difference between the epidermis in the
neighbourhood of the genital apertures and the rest of the
non-clitellar epidermis.
In the clitellar epidermis (fig. 19) ordinary epidermic-cells
and goblet-cells occur, and in addition to these the two other
kinds of gland-cells mentioned above. The ordinary cells and
the goblet-cells form a surface layer through which pass the
necks of the other gland-cells. The main portions of what I
have called above the short club-shaped cells form a distinct
layer of more than twice the thickness of the superficial layer.
These short club-shaped cells (figs. 19, s. cl.) are filled with
large granules which stain deeply with alum carmine, and
correspond, I believe, to the coarsely granulated cells of Lum-
bricus. The innermost layer of the clitellum, a layer which
is thicker than the other two layers together, is made up
of the main portions of the long club-shaped cells (fig. 19,
1. cl.), so that some of these, the ones that dip furthest down,
Cf. Cerfontaine, ‘Arch, de Biologie,’ t. x, 1890,
314 ALFRED GIBBS BOURNE.
are very long indeed. Their contents are finely granular, and
do not stain with alum carmine. Connective tissue runs up
between these cells, and tends to separate them into groups.
Such cells appear to exist in the clitellum of all worms, and
not elsewhere, so that they are probably the true clitellar
glands, and secrete the egg-capsule.
The epidermis in the neighbourhood of the genital apertures
consists of columnar cells which are rather thinner and longer
than those elsewhere, but none of them have become glandular
(fig. 20). In the other non-clitellar portions of the clitellar
region, the epidermis is similar to that found in the clitellum
proper, except that the long club-shaped cells are entirely
absent, and the short club-shaped cells are rather shorter.
These cells, which are very clearly confined to the clitellar
region of M. grandis, are found all over the body in many
other worms, and it is difficult to assign a function to them.
As our Moniligaster needs to have the surface of its body kept
moist with mucus (it lives in what is often very dry earth),
and the only glands present all over are the goblet-cells, we
may assign to them a mucus-secreting function, a theory which
is borne out by their appearance, and as the only glands which
seem to be always confined to the clitellum are the long club-
shaped glands, we may assign to these a capsule-secreting
function—but what is the function of these coarsely granulated
glands ?
Capillary blood-vessels are remarkably abundant in all
Moniligasters throughout the epidermis.
The usual circular and longitudinal muscular layers of the
body-wall are present. The circular muscles form, as Cerfon-
taine has pointed out for Lumbricus, an almost continuous
sheath, thickest towards the middle of a segment and thinnest
in the region of the intersegmental grooves, and interrupted
only for the passage of the sete, excretory ducts. of the
nephridia and generative organs, and the muscles of the
sete, &c.
The muscle-fibres are unlike those of Lumbricus and
most worms, They are like those of a leech, each cell having
MONILIGASTER GRANDIS, A. G. B. 315
a superficial layer of contractile substance, while the medullary
portion consists of unmodified protoplasm and contains the
nucleus.
The longitudinal muscles form four longitudinal bands run-
ning the whole length of the body. The dorsal band is semi-
cylindrical, extending over, as nearly as possible, half the body-
wall. The ventral band is the narrowest; it lies between the
inner rows of seta couples, while the lateral bands lie on each
side between the inner and outer rows of setacouples. As the
setze are so small that they do not project through the body-
wall, and the two setz in a couple so near together, there are
no intersetal or accessory bands such as are present in Lum-
bricus. Each band is built up of a number of bundles of
cells. There is no bipinnate arrangement, the cells in each
bundle form a solid block. The bundles lie very close to one
another, there being only a very small amount of connective
tissue present, and this has the nature of a granular stroma.
The parietal layer of coelomic epithelium has the usual
character.
Septa.—Four of the septa are very thick, viz. v-v1, vi-vn,
VII-VIII, and vitl-1x; the remainder are all very thin and do
not vary! in thickness, excepting that towards the anal end of
the body they become gradually thicker.
The thin septa are very thin and absolutely transparent
when fresh, and remain fairly so even in spirit.
The arrangement of the septa in the anterior region and the
consequent location of various organs has given me a good
deal of trouble. The safest guide to the numbering of the
various segments is to be found in the efferent ducts of the
nephridia, which can with care be always traced with certainty
up to the point where they penetrate the body-wall. The
nerves arising from the ventral nerve-cord are also useful for
purposes of enumeration, as they come off with absolute
1 IT note this in contradistinction to the condition of many worms, where
the septa which follow the thickened septa become gradually thinner and
thinner. Here the transition is perfectly sudden, Septum vitI-rx is the
thickest, and Septum 1x=x is as thin as any other, in the body.
316 ALFRED GIBBS BOURNE.
regularity (figs. 27 and 49). Without careful attention to
these points one might be very easily led into error. The
sete are so small that unless well protruded they are often not
visible in the anterior region, where the body-wall is so thick,
and the annular grooves are in places so pronounced, that
from an inspection of the cut edge of the body-wall it is
impossible to say with certainty which grooves serve as
demarcations between neighbouring segments and which
between neighbouring anunuli only.
It is well known that in some worms certain of the septa
may be absent, but here all the septa are present, only some
become displaced; and I need hardly say that in all cases the
septa should be numbered with regard to their relation to such
organs as occur in continuous series (e. g. nerve-branches,
nephridial ducts), rather than with regard to the position of
their attachment to the body-wall.
The four thickened septa above mentioned join the body-
wall in the region of the intersegmental grooves, the most
anterior, v-v1, showing, however, a tendency to run a little
forward along the body-wall.
I can recognise three structures in front of these which
might be the septa 11-111, 111-Iv, and 1v-v, but these are attached
(see fig. 17) to Septum v-v1, and never reach the body-wall,
and are therefore not septa, but special mesenteries supporting
the nephridia, such as are connected with the other nephridia
(see below).
The only septa with regard to which there is any real
difficulty are 1x-x, X-x1, and XI-xII.
Septum Ix-x is attached to the body-wall in the dorsal
region near the intersegmental groove x-x1, while in the
ventral region it is attached in its normal position. This is
the septum which supports the seminal capsule, and through
which the vas deferens passes. At the spot where the excretory
duct of the nephridium of Segment x penctrates the body-wall,
this segment is carried forward so as to leave nephridium 10—
i. e. the nephridium of Segment x—entirely behind it.
Septum x-x1 is the most modified; the greater portion of it
MONILIGASTER GRANDIS, A. G. B. 317
is in an adult worm so arranged as to reduce the (body) cavity
of Segment x1 to two sacs; one of these contains the ovaries
aud nephridia of the segment, and into it the oviducts open
internally (the oviducts are, as a matter of fact, mere modifica-
tions of a portion of the wall of this segment) ; the other sac
contains the portions of the dorsal vessel, the alimentary canal,
and the ventral vessel, which belong properly to the segment.
The portion of this septum which forms the latter of the above-
mentioned sacs does not touch the body-wall at all; the portion
forming the sac containing the ovaries, &c., joins the body-
wall along a small portion of the ventral region only; else-
where it is attached to the septum which follows, viz. x1-x11,
The nerve-cord does not lie in any portion of its course in the
cavity of Segment x1; there is a free passage through from
Segment x to Segment x11. (See the arrow in fig. 27, and note
also in this figure the arrangement of this septum in relation
to the duct of the nephridium of this segment.) The cavity of
Segment xt only passes forward to its proper limit as a cecal
prolongation surrounding the nephridial duct.
Septum xi-x11 is attached to the body-wall in the normal
position, and is only peculiar in being prolonged backwards
to form the wall of the receptaculum ovorum (figs. 17
and 59),
The remaining septa are normally attached ; those of the
three or more segments following x1-x11 get pushed back by the
receptaculum ovorum on each side, the number of septa thus
displaced depending on the time of year, i.e. the state of the
development of the receptaculum ovorum; but when this is
fully developed it may rupture some of the septa; further,
those septa in the neighbourhood of the gizzard become much
displaced where they surround the alimentary canal, as they
must be, considering that the gizzard region when extended
occupies 35 mm., while the five segments containing it measure
when contracted only about 5 mm. in length.
The relation of the septum to the chief organs which run a
longitudinal course from segment to segment is indicated in
fig. 13, from which it will be seen that the septum is in actual
318 ALFRED GIBBS BOURNE.
contact with the dorsal vessel, intestine, ventral vessel, and the
lateral longitudinal vessels (only in the region where these exist
of course), the nerve-cord itself runs freely through the septa,
there being a space left all round through which the cavity of
one segment communicates with that of the adjacent segments.
Mesenteries.—I apply this term (it is using the word in
a forced sense) to a set of mesentery-like structures which
attach various organs to one another or to the body-wall
(fig. 41), In the interseptal regions the dorsal vessel and the
ventral vessel are each attached by a longitudinal mesentery
to the intestinal wall, so that these three structures run through
the segment attached to one another, but not attached to any
other structures. The lateral longitudinal vessels are attached,
along their whole course through each segment in which they
occur, to the body-wall in the ventral region by longitudinal
mesenteries, but are otherwise unattached. The line of
attachment of these last-mentioned mesenteries begins on the
hinder surface of the septum near the middle line, at the level
of the dorsal vessel, and passes along the septum outwards and
downwards and then inwards, curving round to the inner seta
line, and then back along the wall of the segment just inside the
inner seta line, and then up on the anterior face of the hinder
septum of the segment in which they lie towards the dorsal
vessel. Prolongations from these mesenteries hold the ne-
phridia, the hearts, and where necessary the spermathecz
and their ducts, and the vasa deferentia.
Behind the region of the lateral longitudinal vessels, and
therefore of these special mesenteries, there are much smaller
mesenteries attaching the nephridia to the posterior face of
each septum.
All the septa and these mesenteries are lined on both sides
with irregularly-shaped pavement-epithelium cells. The thin
septa consist of little else, a few straggling muscular fibres and
a little connective tissue, but the four specially thickened septa
and the septa in the posterior region of the body contain
a much larger amount of muscle; on each surface there is a
layer of centripetally placed fibres, while the main mass of the
MONILIGASTER GRANDIS, A. G. B. 319
septum is made up of fibres crossing one another in all direc-
tions. Where the septum joins the body-wall the centripetally
placed fibres spread out and penetrate to a certain extent the
longitudinal muscular layer of the body-wall. In the case of
the septa in the hinder region the same sort of thing occurs at
the junction of the septum with the intestinal wall, indeed,
close to the intestinal wall the septa become so thick as to
almost join one another along the intestinal wall.
ALIMENTARY TRACT.
It is, so far as I can see, impossible to say with certainty
from an examination of the adult worm, to what segments the
anterior portions of the alimentary canal really belong. In
front of Septum v-vi, i. e. in the five most anterior segments,
lie the buccal region, the pharynx, and a considerable length
of csophagus. From the arrangement of the nephridia and
blood-vessels I should say that this latter occupies, at any
rate, Segments 111, Iv, and v; this leaves Segments 1 and 11
only for the buccal region and pharynx. Longitudinal sec-
tions show a slight change in the character of the alimentary
epithelium at the posterior limit of Segment 1, so that I think
we may assume that the buccal region occupies this segment
while the pharynx is confined to Segment 11. The pharynx
extends backwards, of course, far beyond the limits of this
segment as marked by the grooves on the body-wall, but these,
as we know, by no means determine the true anatomical posi-
tion of an organ.
The buccal region is completely eversible and can be pro-
truded with the prostomium (fig. 16), or it can even be pro-
truded while the prostomium remains completely withdrawn.
The pharynx is comparatively thin-walled except in the
dorsal region. When the pharynx is cut open ventrally a
circular portion of its dorsal wall appears as a thick pad, while
a circular wall stands up all round this. This pad can be pro-
truded, and forms the piston which is used in sucking in
the earth and in excavating the burrows (fig. 17, m.). Numer-
ous strong muscles run from the outer wall of the pharynx to
320 ALFRED GIBBS BOURNE.
the body-wall, but those which serve to retract this pad are
much stronger than any of the others, and are attached as far
back as the wall of Segment vir. The cesophagus really opens
into the ventral wall of the cavity of the pharynx, so that this
latter might be regarded as a diverticulum of the dorsal wall of
the cesophagus. The cesophagus is very straight and narrow,
slightly constricted at the septal regions, and does not change
its character in any way until Segment xv. The walls through-
out are much firmer and stronger, although no thicker, than the
cesophageal walls in, for instance, Megascolex ceruleus.
There is really more muscle present in them.
When the other organs and the greater part of the septa are
dissected away, it becomes clear that there is a segment of
oesophagus belonging to each segment of the body, although,
owing to the peculiar arrangement of Septum x and x1, the
portion of cesophagus belonging to Segment x1 might be easily
overlooked, as it is completely enclosed. In Segments xv, xv1,
and xvi the cesophagus gradually dilates and its walls become
a little thicker. In Segment xviit is the first of the series of
gizzards. There are five gizzards belonging to as many succes-
sive segments, i.e. the most anterior gizzard lies between Septa
XVII-XVIII and xvi1I-xIx and the most posterior between Septa
XXI-XX1I and xx1I-xx111 (fig. 22).! The gizzards differ very
little from one another, that in Segment xviii is slightly
smaller than the other four. The arrangement which obtains
in each of these gizzard segments is as follows :—Immediately
following the septum comes the gizzard itself; this is fairly
globular in form, the greater part of its wall consisting of a
ring of circularly-disposed muscle ; this ring of muscle is oval
in longitudinal section, so that the lumen of the gizzard is
smallest in the centre. Following the gizzard itself is a cylin-
drical tubular portion of the cesophagus which joins that
gizzard to the next following gizzard. Lastly, there are a
1 It is especially necessary to insist upon this point, because Benham’s
diagram (11), p. 295, would lead one to suspect a different arrangement. The
gizzards should have been shown as they are in Perissogaster, Trigaster,
or Hormogaster. :
MONILIGASTER GRANDIS, A. G. B. 321
number of bands of muscle attached round the outside of the
anterior portion of the gizzard, free along their length, and in-
serted into the cylindrical portion of the csophagus in the
region of the septum (fig. 23).
In Segment xx111 the wall still remains firm and tough. In
the ten segments, xxIv to xxx1tt, lies the portion of the canal
which corresponds to the tubular intestine of other worms. In
the whole of this region there lies embedded in the thickness of
the mucous membrane an enormous number of tubular glands,
the apertures of which may be seen with a strong lens as little
round holes on the surface of the mucous membrane. I have
generally found this portion of the canal empty and much
shrunk, and consequently immediately obvious, being smaller
than the region which follows, but I have found it full of earth
and distended, in which case it looks from the outside just like
the rest of the intestine.
The “saccular” intestine is very slightly constricted by the
septa, and changes very little in character throughout its whole
length. In about the last eighty segments its walls become
rather stronger, and its blood-supply becomes much diminished;
we may term this the rectal region.
There is no typhlosole.
The alimentary epithelium consists in the buccal region of
small columnar cells very closely set, and there are no glands.
In the pharynx, in the epithelial layer covering the protrusible
dorsal wall there are no glands, but between the ordinary cells
the salivary glands open in immense numbers. ‘These salivary
glands are obviously masses of epithelial cells which have taken
up a deep-lying position and a grouped arrangement, but each
cell retains its connection with the surface, and pours out its
secretion into the pharynx. These salivary glands, which are
thin masses of unicellular glands, lie upon the bands of muscle
which connect the pharyngeal wall with the body-wall. There
are four pairs of such masses lying immediately anterior to
such muscle bands, and a much larger median mass lying im-
mediately behind the pharynx (fig. 17). In the epithelium of
the ventral wall of the pharynx there is an enormous number
322 ALFRED GIBBS BOURNE.
of goblet-cells, looking very like those in the epidermis and
probably secreting mucus.!
The epithelial cells of the cesophagus are also many of them
glandular, and the layer remains unchanged until we reach the
gizzard. There the cells become very short, small, cylindrical
cells, very closely set together; and here there is a cuticle and
no glands. This cuticle is continuous from one end of the
gizzard region to the other, but varies in thickness, becoming
very thick in the middle of each gizzard and thinning out at
either end, being thinnest in the cylindrical regions between
the gizzards.
In one segment behind the gizzard there are no glands in the
epithelium, and then comes the region of tubular glands. In
this region the epithelium on the surface appears only slightly
glandular. In a section passing through the aperture of a
tubular gland (fig. 24) the ordinary epithelium is seen to be
continued some way down the follicle, and then its character
gradually changes, the cells become broader and evidently
glandular, but they do not become goblet-shaped, and they are
all alike; their nuclei are smaller than those of the ordinary
epithelial cells. ‘The follicles run straight down to the base-
ment membrane, and then branch and coil.
In the saccular region of the intestine the mucous membrane
becomes thin again, and consists of a single layer of epithelial
cells only. Most of these have granular contents and appear
to be glandular, but they do not become much swollen, or at
any rate are not so in any of my sections, which come from
several different worms.
In the rectal region the cells are all narrow and non-
glandular. Longitudinal sections passing through the anus
show that there is no sharp demarcation between rectal epi-
thelium and epidermis.
Of the special muscles of the pharynx and gizzards I have
already spoken. In the wall of the rest of the cesophagus and
of the intestine the usual muscles are present, a very thin layer
1 These glands show best when the tissue has been prepared in Flemming’s
mixture of chromic, osmic, and acetic acids.
MONILIGASTER GRANDIS, A. G. B. 323
of longitudinal muscle, and below that an equally thin layer of
circular muscle; in both cases there are scattered fibres rather
than a regular layer. In the intestinal wall there are fairly
strong longitudinal bands in the median dorsal and ventral
walls. Further, in this region, to an extent which increases
from before backwards, the muscle fibres of the septa spread
out circularly round the outside of the intestinal wall. These
fibres lie outside the ccelomic epithelium, and in the rectal
region the circular muscular bands so formed become so broad
as to be almost continuous (figs. 25).
The coelomic epithelium of the intestinal region presents the
usual characters.
VASCULAR SYSTEM.
Having had a practically unlimited supply of fresh speci-
mens, I have been able to work out this system in considerable
detail. There is no doubt but that the system has become
very fully developed ; in dissections, in sections, and in other
microscopic preparations I have again and again seen most ex-
quisite “capillary” networks. I have also procured admirable
injections for dissecting purposes by injecting a mixture of any
very soluble oil paint dissolved in turpentine. This mixture is
very excellent in one way, viz. that when the injected speci-
men is plunged into spirit loose drops of injection at once
separate, leaving the specimen quite clean.
Dorsal Vessel.—The dorsal vessel extends from the an-
terior portion of the pharynx to the last segment of the body.
It lies above the alimentary canal, and is adherent along its
whole length to the wall of this organ. For the greater part
of its length it appears to be actually attached to the intes-
tinal wall, but anteriorly (in the region of the gizzards and in
front of them) the nature of the attachment becomes clear,
and the latter is seen to be due to the presence of a longi-
tudinal double fold of mesentery (fig. 41), so that it is nowhere
possible to put a needle under the dorsal vessel without
piercing this mesentery. There is no mesentery connecting it
with the dorsal region of the body-wall.
VOL. 36, PART 3.—NEW SER. Z
324 ALFRED GIBBS BOURNE.
Its walls are muscular throughout its length, and in the
anterior segments the ccelomic epithelium covering it becomes
pigmented; this is more especially the case near the septa
and at the sides.
It is nowhere much dilated, but is largest in Segments
vi—1x, and constricted in the region of Septa v-v1, vi-vII,
vil-vil1, and viti-1x (the thick septa). In front of Septum
v-v1 it very rapidly narrows, and becomes a mere thread on
the surface of the pharynx where it dips down among the
salivary glands. It does not appear to have any special mode
of termination at this anterior end, but simply loses itself in
a fine capillary network.
It presents a series of valves (fig. 39) as in Megascolex
ceeruleus.! The valves are placed in the septal regions and.
prevent any backward flow. There are also valves (mentioned
below) at the orifices of the dorso-intestinal vessels.
It is never double in any part of its course.
Posteriorly it ends by dividing into a pair of dorso-tegu-
mentary branches.
Ventral Vessel.—(Fig. 27, V. V.) This is also known as
the subintestinal or supra-neural vessel.
Its walls are thick and strong, but are formed of connective
tissue and not muscle.
There are no valves in the ventral vessel.
It is of uniform calibre for the greater portion of its length,
but posteriorly it gets very gradually smaller and smailer,
while anteriorly it narrows rapidly in front of the hearts.
It comes to an end anteriorly and posteriorly by dividing
into a pair of branches in each case, those at its anterior
extremity lying on the subesophageal ganglion.
Supra-intestinal vessels are absent.
Ty phlosolar Vessel.—No such vessel exists, there being,
as I have stated above, no typhlosole.
Latero-longitudinal and Subneural System.—In-
testino-tegumentary vessels, i. e. vessels carrying blood between
the peripheral capillary networks and the intestinal capillary
Cf. Bourne, ‘Quart. Journ. Micr. Sci.,’ vol. xxxii, Pl. ix, fig. 12, A,
MONILIGASTER GRANDIS, A. G. B. 325
networks, are not present in any region of the body. ‘‘ Latero-
longitudinal ” vessels are very obvious in the anterior segments,
from Segment x forwards to Segment v. They are the main
trunks, in the anterior region of the body, of a system of vessels
of which the subneural vessel is the main trunk in the pos-
terior region of the body. This system must be dealt with as
a whole.
This system is connected on the one hand with capillary
networks, and on the other with the hearts of Segment 1x.
It is, in fact, a venous (using the word in an anatomical sense)
system for the greater portion of the body. The only capil-
lary networks with which vessels of this system do not com-
municate, from which, in fact, they do not carry blood back
to the hearts, are the capillary networks of the intestinal wall
from Segment x11 backwards.
Fach latero-longitudinal trunk may be divided, for purposes
of description, into three parts, an anterior trunk and a pair
of posterior trunks. These come together in the anterior
portion of Segment x, unite into a common trunk which
passes through Septum 1x-x and opens on each side into the
heart, or, rather, the neck of the heart of Segment 1x. These
trunks belong to the cephalised region. The subneural vessel
exists along the greater portion of the length of the body. It
is the longitudinal vessel into which open all the vessels
coming from the capillary networks of the body-wall, septa,
and nephridia of all the segments posterior to the cephalised
region. It comes to an end anteriorly in about Segment xiv
in a very indefinite manner, but all the blood brought for-
wards by it finds its way into one of the posterior trunks of
the latero-longitudinal vessel on each side by such vessels as
a., 6., fig. 27; the exact arrangement is liable to individual
variation. This strikes one as a very inefficient way of dis-
posing of so much blood, but the circulation is here, as in all
veins, comparatively slow, and there can be very little blood-
pressure; and, besides, the whole circulation in Monili-
gaster grandis is but feebly carried on in the hinder portion
of the worm. It is interesting to recall in this connection
326 ALFRED GIBBS BOURNE.
the extreme sluggishness of the hinder end of the worm, and
to contrast this state of things with that which obtains in,
for instance, M. ophidioides, M. robustus, and M. sap-
phirinaoides, where, coexisting with a much more active
body throughout its entire length, there is a much more
definite connection between the subneural vessel and one of
the posterior longitudinal trunks (fig. 29).
There are no latero-neural vessels.
It will be convenient to defer for the present the description
of the various branches of this latero-longitudinal and subneural
system.
Hearts (figs. 27, 32, and 33).—There are four pairs of
hearts, and these belong to Segments vi, vii, viti, and Ix.
There is in some worms a little. difficulty in determining
whether certain particular branches of the dorsal vessel should
be regarded as hearts or not. In Moniligaster there is no such
difficulty. In all the species which I have examined there are
four pairs of hearts and four only.
Their walls are so muscular that when spirit is poured upon
a freshly-opened worm the walls of the hearts become so opaque
that their red contents can no longer be seen, while the well-
developed network of vasa vasorum (fig. 32) supplying their
walls comes very clearly into view. The walls of the dorsal
vessel also become opaque under similar conditions, only to a
very much less degree. In no other vessels of the body are
the contents in the least degree obscured by the action of the
spirit, nor do I know any other worm in which this pheno-
menon occurs to anything like the same degree, even in the
hearts, as in Moniligaster grandis.
The three anterior pairs of hearts are very simply connected
with the dorsal vessel (fig. 32). There is a short narrow neck,
and at the junction of this with the dilated heart there is a
sphincter muscle. The hearts stand out well from the walls
of the cesophagus, but each is attached to the wall along its
whole length by a double fold of mesentery.
The hearts of the most posterior pairs have necks at their
upper extremities similar to those of the more anterior pairs,
MONILIGASTER GRANDIS, A. G. B. 327
but these are connected one with the other across the middle
line, and a short median vessel puts them in connection with
the dorsal vessel, while into the necks open the two main
trunks of the latero-longitudinal subneural system (fig. 33).
At the lower (or distal) extremity each heart opens into a
bulb, between which and the heart proper is a sphincter
muscle. This bulb opens directly into the ventral vessel in
the case of the three anterior pairs of hearts, but with the
intervention of a short neck in the case of the most posterior
pair of hearts (fig. 33).
Capillary Networks.—These may be spoken of as peri-
pheral networks and intestinal networks.
Peripheral Networks.—Under this term I include not
only the capillary networks of the body-wall and septa, but
those of the nephridia, generative organs, walls of the large
blood-vessels (dorsal vessel and hearts), and, indeed, those of
all the viscera with the exception of the intestinal walls; the
capillaries of these latter are really central as compared with
the others, but are conveniently termed intestinal networks.
The peripheral networks are extremely well developed
throughout, the capillaries being always perfectly clear and
distinct. The smallest capillaries measure not more than —,.
of an inch in diameter. There is no trace of any lacunar
structure.
There is a sort of fascination in following out these exquisite
networks, and I have spent much time in so doing, but it
would serve no purpose to describe them in detail. I merely
mention this to emphasise the fact that figs. 35—37 are no
imaginary diagrams, but accurate drawings of actual prepara-
tions. Perhaps the most striking feature of these networks is
the strict parallelism which obtains throughout between
“artery ’’ and “vein.” They are not, indeed, strictly speak-
ing, networks; each small “ artery ”’ loops round and becomes
a small “ vein.”
In dealing with Megascolex ceruleus I had only ascanty
supply of specimens, and was consequently unable to study
these capillary networks in such detail, I used the term
328 ALFRED GIBBS BOURNE.
“triple connections”’ and imagined that networks of capillaries
existed in certain places, to which blood was brought by
branches of both the dorsal and ventral vessels, and from
which blood was carried away by branches of intestino-tegu-
mentary vessels. I am now convinced (from a re-examination
of my preparations of Megascolex ceruleus and a study of
fresh preparations made for the purpose from another large
Perichete) that this is not the case. I still believe it to be
perfectly true that blood carried outwards from the dorsal and
ventral vessels passes into intestino-tegumentary vessels, but in
one place small capillaries from the dorsal vessel loop over into
some branches of an intestino-tegumentary vessel, and in other
places small capillaries from the ventral vessel loop over into
other branches of the same vessel.!
Intestinal Networks.—There is not quite so elaborate a
system of blood-vessels in the walls of the alimentary canal as
in Megascolex ceruleus, and there is not in a sense the
same interest attaching to them as in that form, because their
relations to the main trunks are muchsimpler. The intestinal
networks are in the greater portion of the canal (from Seg-
ment xiv backwards) directly connected with the dorsal vessel
on the one hand and the ventral vessel on the other. There
are no vessels having the relations of the intestino-tegumentary
vessels of Megascolex ; there are, indeed, no vessels communi-
cating at both extremities with capillary networks.
In the anterior region, as far down as Segment x111, blood
is carried to the canal wall by branches of the ventral vessel
(ventro-intestinal vessels), and taken away by vessels which
run into the anterior longitudinal vessels in the rest of the
worm; there are similar afferent vessels, but the efferent
vessels run into the dorsal vessel direct (dorso-intestinal
vessels), The intestinal networks are best developed in the
absorbent region. There the main vessels of the network run
1 I have been unable to find any account in the literature at my command
of the exact connections between the ultimate ramifications of the hepatic
artery, the portal vein, and the hepatic vein, in the Vertebrate liver. It would
be interesting to know if we have to do here with true “triple connections.”
MONILIGASTER GRANDIS, A. G. B. 329
circularly round the intestinal wall. These circular vessels lie
very close together, and are connected with one another at
short intervals by longitudinally-running junctions, the whole
forming a rectangular meshwork. In the dorsal region these
circular vessels join together to form the dorso-intestinal
vessel in each segment. They take their origin from a net-
work which runs longitudinally along the ventral surface of the
intestinal wall on each side of the ventral median muscular
band. As part of this network there is a small longitudinal
vessel on each side which runs right on from segment to seg-
ment; from this there runs in the region of each septum a
small circular vessel, lying on the outer wall of the intestine,
to the dorsal region. When nearly arrived there it bifurcates,
and one branch runs over the intestine and under the dorsal
vessel and joins its fellow of the opposite side, while the other
is connected with a longitudinal network which lies on either
side of the dorsal median muscular band, and which is con-
tinuous from segment to segment. In this way provision is
made in every segment for communication between the net-
works of the two sides, and by means of the networks, for com-
munication between the blood supply of one segment and that
of another. The ventro-intestinal vessels open into dorsal
longitudinal networks on each side.
Dorso-intestinal Vessels.—These are the branches of
the dorsal vessel mentioned above which place it in connection
with the intestinal capillary networks. They are the efferent
vessels of these networks, and the only vessels which bring
blood into the dorsal vessel. The most anterior pair are in
Segment x111, and behind these there is one pair in each seg-
ment. They are joined to the dorsal vessel immediately be-
hind the septum, i.e. in the most anterior portion of the
segment, and each is formed of from ten to twenty branches,
all of which spring from the intestinal wall. At the junction
of each with the dorsal vessel is a valve precisely similar to
that of Megascolex ceruleus, which must effectually pre-
vent any blood passing outwards into it from the dorsal vessel
(fig. 39).
300 ALFRED GIBBS BOURNE.
These vessels and their main branches are covered, as is
frequently the case, with yellowish-brown celomic epithelium
cells, which make them look very large and very different
from the other branches of the dorsal vessel (dorso-tegumen-
tary vessels) which are not covered by pigmented cells.
The connection of these dorso-intestinal vessels with the
intestinal capillaries has been described above.
Ventro-intestinal Vessels.—These are branches of the
ventral vessel running from it to the intestinal walls. There
is one pair in every segment from Segment vi (at any rate)
backwards, arising immediately posterior to the septum, ex-
cept in Segments vi, viI, vii1, and 1x, where they arise imme-
diately in front of the hearts. They are the sole afferent
vessels of the intestinal walls. There are no such vessels in
Megascolex ceruleus, their function is performed by the
intestino-tegumentary vessels.
Ventro-tegumentary Vessels.—These are branches of
the ventral vessel connecting it with the peripheral networks.
There is a pair of these vessels in every segment except the
first, in which there is a branch of that belonging to the 2nd
segment on each side, and except in those segments in which
the ventral vessel is joined by hearts, viz. Segments vi, vil,
vull, and rx, and in Segment x.
In Segments vi, viz, and vii1 the place of ventro-tegumen-
tary vessels is taken by vessels which come off from a bulbous
dilatation, which is interposed between the heart proper and
the ventral vessel. The limit of the heart proper is marked
by a sphincter, and between this sphincter and the ventral
vessel is the bulbous dilatation.
In Segment viii, in addition to the ventro-tegumentary
vessels belonging to that segment, there is a second pair of
vessels arising from the bulbous dilatations which run back-
wards through Septum vitt-1x, and form the ventro-tegumen-
tary vessels, or rather the vessels which correspond to them, of
Segments 1x and x,
Ordinarily these ventro-tegumentary vessels come off from
the ventral vessel immediately in front of the septum which
MONILIGASTER GRANDIS, A. G. B. 331
bounds posteriorly the segment to which they belong, so that
they lie in the posterior portion of their segment.
These ventro-tegumentary vessels supply not only the body-
wall, but all the viscera excepting only the walls of the
alimentary canal, and they and their branches almost always
run strictly side by side with vessels belonging to the latero-
longitudinal and subneural system; their finest branches
(capillaries) are in every case continuous with the finest
branches of this system—they are, in fact, the afferent vessels
while the branches of the latero-longitudinal and subneural
system are the efferent vessels of the tissues. So closely are
these afferent and efferent vessels associated throughout the
greater part of their course that I have again and again ima-
gined that I was dealing with a single vessel, but when I
knew what to expect, more careful examination always re-
vealed two vessels lying side by side and branching at the
same spots.
As atype of what occurs with regard to the majority of
these ventro-tegumentary vessels we may take that of Segment
x11. The vessel lies on the mesentery and runs towards the
dorsal region : it sends outwards towards the body-wall a series
of branches; the first of these supplies the region of the body-
wall which lies ventrad of the inner row of setz following this ;
then there are three branches going to the region of the body-
wall which lies between the two rows of set, and near the first
of these a branch to the nephridium of the next following seg-
ment (tothe nephridium, that is to say, all of which except the
funnel lies in the next following segment). Up to the point
where the last of these branches is given off, the vessel is closely
accompanied by a branch of the subneural vessel, but at this
point the ventro-tegumentary vessel leaves for a short time
the subneural-tegumentary branch and runs more closely
along the body-wall, giving off a series of five or six small
vessels in its course towards the dorsal region. The subdivi-
sions of all these branches to the body-wall run backwards as
well as forwards, so that the vessel is not distributed to its
own segment only, but to the anterior portion of the next
Doe ALFRED GIBBS BOURNE.
following segment. There is, however, so far as I can see, no
anastomosis between the branches of one ventro-tegumentary
vessel and those of another.
In Segments vi, vii, and vir each ventro-tegumentary
vessel, very soon after its origin from the bulbous enlargement
at the base of the heart, gives off a small branch which supplies
the wall of the heart itself and the wall of the dorsal vessel ;
the next branch runs along the anterior surface of the septum
which bounds posteriorly the segment to which it belongs,
gives off some small branches, and then passes up towards the
dorsal region of the posterior portion of the body-wall of the
segment, giving off numerous branches, some of which run
backwards and some forwards, while all supply the various
layers of the body-wall. The ventro-tegumentary vessel itself
runs forward in the segment and gives off a branch to the
nephridium, and, lastly, supplies the posterior side of the
septum which bounds the segment anteriorly, and, passing
towards the dorsal region, supplies the body-wall of the ante-
rior portion of the segment.
The ventro-tegumentary vessel, which originates in Segment
vim but passes backwards to supply Segments 1x, x, and, to
a certain extent, x1, is the largest of the whole series. It
penetrates the Septum vui1-1x close alongside the latero-
longitudinal vessel and runs backwards with this. It gives
off a branch to the posterior side of Septum viit-1x and the
anterior region of the body-wall of 1x, one of the subdivi-
sions of which branch forms the afferent vessel of the ne-
phridium of Segment 1x, then a branch to the posterior portion
of the body-wall of Segment 1x, which will be seen from
fig. 27 to pass just in front of the external aperture of the
nephridium of Segment x, and close to this last branch
another, which, after giving off a small branch to the wall of
the heart of Segment 1x, passes to the seminal capsule ; its next
branch supplies the nephridium of Segment x; its next branch,
after giving off a vessel to the prostate, supplies the ventral
region of the body-wall in the anterior portion of Segment x,
and then, after giving off a branch which goes to the nephridium
MONILIGASTER GRANDIS, A. G. B. 333
of Segment x1 and also to the ovary, this vessel supplies the
dorsal region of the body-wall of Segment x.
The ventro-tegumentary vessel of Segment v calls for special
remark. After leaving the ventral vessel it runs outwards and
forwards along the body-wall towards the dorsal region. The
first branch runs to the salivary glands, the second penetrates
the muscular layers of the body-wall; the next two branches
come off at almost exactly the same spot; the one supplies
the nephridium of Segment v, the other runs forwards, and
on its way supplies the nephridia of Segments rv and 111.
There are seven other branches, which all run forward and
penetrate the body-wall. There are thus eight branches from
each of these ventro-tegmentary vessels (those of Segment v),
which run forwards in the body-wall of the anterior extremity.
The ventro-tegumentary vessels of Segments Iv, 111, and 11
are all small. |
The ventral vessel comes to an end anteriorly in the region
of the subcesophageal ganglion by giving off the ventro-
tegumentary vessels of Segment 11, each of which gives off
three main branches, supplying, doubtless, the prostomial
and peristomial regions, and of course Segment 11.
Dorso-tegumentary Vessels.—These are branches of
the dorsal vessel connecting it with the peripheral networks.
All the branches of the dorsal vessel anterior to the hearts,
and one pair of those branches in each segment posterior to
them, belong to this category.
In the most anterior portion of the dorsal vessel they arise
from this slightly irregularly, i.e. asymmetrically.
In all other segments they arise regularly from the dorsal
vessel immediately in front of the septum, which forms the
anterior boundary of the segment in which they lie.
All these dorso-tegumentary vessels which arise behind the
hearts have similar relations. Each runs boldly outwards,
unattached to any other structure, and appearing always as
though the blood in it was under considerable pressure, and at
about the outer seta line it bifurcates ; the two branches are
really portions of the subneural system, and are coloured blue in
334 ALFRED GIBBS BOURNE.
fig. 27. One branch—the “ branche tegumentaire” of Jaquet!
—runs on to the body, comes into the neighbourhood of the
distal portion of the ventro-tegumentary vessel of the segment,
and runs along parallel to this towards the dorsal median line ;
its branches run parallel to those of the ventro-tegumentary
vessel, and their capillaries communicate. It is evident that,
as the ventro-tegumentary vessel is the afferent vessel for these
capillaries, this branch of the dorso-tegumentary vessel is the
efferent vessel. The other branch (the “ branche dorso-sous-
nervien”’? of Jaquet) loops round and becomes directly con-
tinuous with the branch of the subneural vessel belonging to
the segment; from about the outer seta line inwards it runs
parallel to and close alongside of the proximal portion of the
ventro-tegumentary vessel, and there are branches of this dorso-
subneural vessel corresponding to each of the branches of this
portion of the ventro-tegumentary vessel; there are, again,
efferent vessels of the capillary networks in the body-wall and
nephridia, supplied by the ventro-tegumentary vessel.
There is thus a loop or commissural vessel which is connected
on the one hand with the dorsal vessel, and on the other with
the subneural vessel, all the branches of which bring blood
back from capillaries. Does this blood pass into the dorsal
vessel or into the subneural vessel, or into both? I believe
that the blood passes into the subneural vessel only, and that
no blood enters the dorsal vessel by means of the dorso-
tegumentary vessel. I have already discussed this question
when speaking of Megascolex, and although we are now dealing
with a very different genus, it is not, I think, on a priori
grounds, likely that the dorso-tegumentary vessel would carry
blood out of the dorsal vessel in the one case and into it in the
other ; moreover, in Moniligaster as in Megascolex, while there
are valves which would mechanically prevent blood flowing
into the dorso-intestinal vessels from the dorsal vessel, there
are no such valves where the dorso-tegumentary vessels join
the dorsal vessel. I have, however, observed in Moniligaster
and some other worms a sphincter muscle in the wall of the
1 «Mitth. Zool. Stat. Neap.,’ Bd. vi, 1885-6.
MONILIGASTER GRANDIS, A. G. B. 330
dorso-tegumentary vessel close to its origin. Now if it were
the usual thing for blood to enter the dorsal vessel from the
dorso-tegumentary vessels, there would, judging from analogy,
be valves which, acting mechanically, would prevent blood
taking the reverse direction during the powerful contraction of
the dorsal vessel ; if, on the other hand, the normal course is
for the blood to pass out into these dorso-tegumentary vessels,
the sphincter muscles on the latter would regulate the amount
so passed out. I do not believe that any great quantity of
blood flows through these vessels at all, but that what flow
there is is an outward one, and one which would serve to
increase the pressure in the peripheral capillaries, and could
be varied in amount from time to time and regulated.!
1 Benham thinks that the dorso-tegumentary vessels bring blood to the
dorsal vessel. He is speaking of Lumbricus, but, as far as I can see,
there is very little difference between Lumbricus and Moniligaster in
the matter of the vascular system. The diagram (fig. 34), represents the
state of things in an ordinary segment of Moniligaster. In this figure I
have inserted various arrows to indicate the possible direction of the blood-
flow. Now I think that there is no question as the correctness of these arrows
>>——, and I believe that these arrows 3>—— are also rightly placed ;
Benham considers these arrow -+>——> to be correct. I consider the
question settled in Moniligaster by the valves; Benham thinks that his
view is confirmed by the arrangement of valves in Lumbricus. Setting
aside for a moment the question of valves, let us follow out his view. He
says, that in all these places blood is flowing out from the subneural vessel ;
very well, where does it flow into the subneural ? The subneural is, as far
as I can make out, always a part of a system of vessels of which the latero-
longitudinals form another part; now this system is known to be connected in
Moniligaster and Lumbricus with the dorsal vessel (or a pair of hearts).
It is from the dorsal vessel (or hearts), then, that the subneural must get its
blood, which means that blood must flow, according to Benham’s theory, from
the dorsal vessel into not only the subneural but into the whole system,
including the latero-longitudinals ; but branches of these latero-longitudinals
evidently play an opposite réJe to the branches of the ventro-tegumentary
vessels in the anterior part of the body, and if the latero-longitudinal vessels
serve as arteries, the ventro-tegumentary vessels must serve as veins, and
carry blood into the ventral vessel (in the anterior region of the body). I
have already given my reasons for believing that this is not the case in Me-
gascolex (I. c., p. 77), and I think that they hold good, mutatis mutandis,
336 ALFRED GIBBS BOURNE.
Latero-longitudinal and Subneural System.—Hav-
ing already spoken of the main vessels of this system, and
made various references to their branches and detailed the
distribution of the ventro-tegumentary vessels, which run for
the most part strictly parallel to the branches of this system,
it is not necessary to do more than refer the reader to fig. 27,
where this system is shown throughout in blue.
CouRSE OF THE BLoop.
I have found it convenient to refer to the course taken by
the blood in describing the anatomy of the vascular system,
and I have, in writing of Megascolex ceruleus, dealt at
length with certain general problems relating to the blood-flow
in earthworms (see also below, p. 42), so that I need only sum-
marise here the chief facts with regard to Moniligaster
grandis.
The blood flows forwards in the dorsal vessel, and while a
little passes on in the anterior portion of the dorsal vessel,
most of it is forced down into the ventral vessel by means of
the hearts (chiefly the three anterior pairs) and is distributed
by the ventro-tegumentary vessels on the one hand, to peri-
pheral capillaries, and by the ventro-intestinal vessels on the
other hand, to the intestinal capillaries.
The peripheral capillaries in the region of the body behind
the hearts are also supplied, to an extent which probably varies
from time to time and is, I expect, never very great, from the
dorsal vessel by means of the dorso-tegumentary vessels.
The blood sent to peripheral capillaries is returned by the
latero-longitudinal and subneural system to the most posterior
pair of hearts, and the blood sent to the intestinal walls is
returned by the dorso-intestinal vessels to the dorsal vessel.
for Lumbricus. Upon my premises I do not think that the logical out-
come of Benham’s view is tenable.
But, apart from all this, 1 have investigated the arrangement of valves
connected with these vessels in Lumbricus, and 1 think that it resembles
that which obtains in Moniligaster.
MONILIGASTER GRANDIS, A.G. B. 337
NEPHRIDIA.
There is in Moniligaster grandis a pair of large ne-
phridia in each segment except Segments 1 and 11 and (at
any rate in the worm from which I cut sections of the anal
end of the body) the two most posterior segments.
The nephridium consists of (1) a nephrostome or funnel
which, with a small portion of the duct, is preseptal, (2) the
nephrostomial duct passing from this to (3) the glandular lobe,
(4) the vesicle duct leading to (5) the large vesicle, and (6) the
terminal duct leading to the nephridiopore. Regions 2—6 are
post-septal.
The general relations of these regions are shown in fig. 42.
On the outside of the greater part of the nephridium there is
a layer of connective tissue (fibrous, not vesicular), and outside
this, again, a layer of coelomic epithelium cells.! There is a
continuous lumen passing from the nephrostome to the
nephridiopore which is intra-cellular, at any rate as far as the
entrance to the vesicle, and possibly even in the vesicle and ter-
minal duct.
The lumen consists of the following regions :—(“S¢. a”) A
narrow ciliated tube leading from the nephrostome to the
glandular lobe; here it communicates with (‘‘ab”’) a branch-
ing system of ductules, the cells pierced by which lie on the
outside of the first third of the glandular lobe; these branch-
ing ductules collect into a narrow duct (“‘dc” and “cd”), which
passes on to the apex of the glandular lobe and returns upon
itself and passes right back to the first bend of the glandular
lobe. Where there are two portions of this narrow duct
running side by side (as there are in the distal two thirds of
the glandular lobe), there are transverse ducts every here and
there placing one of these limbs of the narrow duct in com-
munication with the other (fully shown in the figure in a little
1 These cells, as may be proved by the careful use of silver nitrate, occur
on the wall of the vesicle, as well as elsewhere.
338 ALFRED GIBBS BOURNE.
piece only of the spirally-twisted portion of the glandular lobe,
diagrammatically shown in the apical position). Where the
narrow tube ends (d) what may be called the middle tube
commences; this (‘‘de,” “ef,” and “‘fg”’) runs back along the
whole length of the glandular lobe to the apex, and then returns
upon itself and passes on to the proximal extremity of the
glandular lobe (g), close to the spot where the latter was joined
by the nephrostomial duct. The middle tube is followed by the
wide tube (the vesicle duct, marked ‘gh’’) which enters the
vesicle (at the point marked h). The shape of the vesicle
is shown in the figure; there is a sphincter muscle (7) at the
spot where the terminal duct is given off. The terminal duct
is a wide duct, and is much longer in some of the nephridia
(e. g. those of Segment x1) than in others.
The nephrostome is very like that of Lumbricus. Ihave .
studied a very large number of these organs and have counted
between forty and fifty marginal cells, and have satisfied my-
self as to the existence of one or perhaps two central cells (in
fig. 42 I have drawn one nucleus of such a cell, but in one
preparation [fig. 43] at any rate there were two similar nuclei
side by side). I have been unable to ascertain anything clearly
with regard to ‘ gutter ” cells.
The nephrostomial duct penetrates the septum at a spot
somewhat removed from the body-wall and opposite the inner
seta rows. The nephrostome is about ;4,; inch across at the
widest part.
The nephrostomial duct has very thin walls, richly ciliated
(I do not think that the cilia are arranged in spiral lines, there
is no appearance of such an arrangement) and covered with
connective tissue in which there is a little pigment, rendering
the duct rather easy to follow.
In the branching system of ductules the lumen is always
- very irregular; it can be well seen in quite fresh prepara-
tions, the finer ramifications show only when distended with
fluid. Fig. 46 accurately represents the appearance obtained
by focussing just below the surface of a portion of this region
of the glandular lobe, but I have seen in preparations treated
MONILIGASTER GRANDIS, A.G.B. 339
with a chromo-osmo-acetic acid mixture much finer ramifica-
tions which terminate in the periphery of cells. The limit
between one cell and another does not show clearly, as it does
in the case of similar cells in Hirudo.
The walls of the large ductules of this system are here and
there set with cilia arranged very irregularly, and there are also
numerous undulating structures composed of bundles of long
cilia-like filaments running along, or very obliquely across, the
lumen (fig. 45). At each end the filaments of the bundle are
gathered together and attached to the wall of the ductule. The
bundles vary in length from about ;1, inch to 5,55 inch, and
each bundle consists of some ten or twelve filaments. When
the movement is rapid the bundle undulates as a whole, the
wave always starting from one end—the end nearest the
nephrostome—but as the movement becomes slacker it con-
stantly happens that the waves start at slightly different mo-
ments in the individual filaments, with the result that they
separate from one another ; in fact, a wave of separation passes
along the bundle. Except for the fact that these filaments
are attached at both ends, they are remarkably like the tufts
of fine vibratile cilia which constitute the so-called “flame”
of the excretory canals of the Platodes. I have made a very
careful study of these structures, and am quite convinced of
the accuracy of the above statements. ‘They are quite clear
in some of my chromo-osmo-acetic acid preparations, and on
one occasion I watched them in a fresh preparation on and off
for nine hours as the movement became gradually slower and
slower. ae
The walls of the narrow duct are very like those of the nar-
row duct of Lumbricus in structure. I have never been
able to quite satisfy myself that the branching tubules collect
into one limb of the narrow duct only, as shown in fig. 42 at
6, but I think that they do, and that the transverse tubules
joining the two limbs of the narrow duct are secondary forma-
tions. I have not found cilia in any part of this narrow tube.
The walls of the middle tube are much thicker, and they pre-
sent a striation similar to that which I described as existing
VoL. 36, PART 3.—NEW SER. AUK
340 ALFRED GIBBS BOURNE.
in many of the nephridial cells of Hirudo. It is not ciliated
in any part of its course. It frequently loops back upon itself ;
one such loop only is shown in fig.42. The lumen presents a
somewhat irregular contour.
The cells of the wide tube are similar to those of the middle
tube, but the contour of the lumen is more regular. The wall
of the vesicle is not lined inside by any layer of visible cells,
but sections show scattered nuclei on the inner wall, and I
think that the lumen must be intracellular, the cells being
very much flattened out. There is a well-developed muscular
layer which consists of an inner set of longitudinally placed
fibres, and an outer set of more or less obliquely placed cir-
cular fibres. Outside the muscle is a little connective tissue,
and outside all a layer of pavement (ccelomic) epithelium cells ;
these are very delicate, and, owing to the contraction of the
muscle, somewhat difficult to demonstrate, but in a fortunate
silver preparation they are very clear indeed. The vesicle is
a tube closed at either end, and has a short prolongation on
one side leading to the terminal duct. At the junction there
is a well-developed sphincter, and the lining epithelium is
ciliated all round the aperture. The terminal duct has very
thin walls; these are like those of the vesicle, except that there
is no muscle. The connective-tissue layer is very thin, but
very strong, and appears structureless.
I have already spoken of the very elaborate vascular supply
of the nephridia.
Nervous System.
The cerebral ganglion lies as usual far forward on the dorsal
wall of the pharynx. It is a single, somewhat square-shaped
lobe (fig. 27). Arising from its anterior outer corners are the
cesophageal commissures. These join together to form the
subeesophageal ganglion, whence arise the ventral cords,
which are united at intervals by ganglia. These cords are
firmly bound up together, by a sheath composed of muscular
and connective tissue, into a single cord; but as nerves they
are really as separate and distinct from one another in the
MONILIGASTER GRANDIS, A. G. RB. 341
interganglionic regions as are the nerve cords of a leech. The
ganglia, however, spread along the cords so that the inter-
ganglionic regions become very short, and two or three other
post-oral ganglia fuse with the subceesophageal ganglion itself.
The ganglion lies in the anterior region of a segment.
A nerve arises on each side from the anterior edge of the
cerebral ganglion, the branches of which clearly supply the
prostomium.
Five nerves arise on each side from the inner surface of the
commissure, and pass on to the pharyngeal wall to form a
stomato-gastric system. I find no nerve ring on the pharyn-
geal wall. Ordinarily three nerves arise on each side from
each ganglion. One of these arises from the anterior portion
of the ganglion, and this passes at once into the septum (see
this branch, of the right hand side, in Segment vii in fig. 27.
The septa are supposed to have been dissected entirely away
in Segments 1x and x). In the case of the thickened septa
this nerve can be very clearly seen running in between the
muscular layers of the septum; it passes into the body-wall,
and runs round towards the dorsal region, behind the inter-
segmental groove.
The other two nerves arise very near together from the
middle region of the ganglion. They run outwards near the
inner surface of the body-wall until they reach the inner seta
row. Here they penetrate the body-wall, and run round
towards the dorsal region between the muscle and the epi-
dermis. The more posterior one of the two gives off a large
branch soon after penetrating the body-wall, which also runs
round towards the dorsal region.
These three nerves may be called the first, second, and third
nerves of the segment respectively.
The first nerve gives off a small branch, which bifurcates
and is supplied to the septum.
The second and third nerves each give off a small branch
before arriving at the inner seta line; these branches join and
supply, I believe, the viscera of the segment (except the
alimentary canal),
342 ALFRED GIBBS BOURNE.
The third nerve also gives off another branch before it
arrives at the inner seta line, which runs backwards and joins
the branch of the first nerve of the next following segment.
The three nerves of the lst segment which really arise
from the subcesophageal ganglion appear to come off from
the commissure.
From about Segment xx onwards the second and third
nerves arise so close together as to be indistinguishable one
from the other, but the microscope shows distinct bands of
fibres, and they separate after penetrating the body-wall.
The histology of the ventral cord is dealt with in figs. 47, 48.
The cord is covered throughout by flattened ccelomic epi-
thelium cells. Under these is a layer of connective tissue
in which occur large oval nuclei. Then follows a very thick
sheath of almost hyaline substance in which are embedded the
muscle-fibres; this shrinks away from the outer connective-
tissue sheath in preserved specimens, as shown in the draw-
ings ; a few scattered histological elements occur in it.
' The muscle is present in enormous quantity as compared
with other worms. It consists of longitudinally running cells
placed either singly or in bundles. At the origins of the nerves
these muscle cells run outwards to form a sheath to the nerve
(fig. 48).
With regard to the nerve tissue proper, there is a single
giant-fibre only ; in the inter-ganglionic regions, the two nerve-
cords are quite distinct from one another and from the giant-
fibre; the inner and dorsal portion alone of each cord is
occupied by longitudinally running nerve-fibres arranged
more or less definitely in bundles; in the ganglia the two
cords and the giant-fibre all form part of the same mass, the
nerve-fibres retain here a similar position while the ganglion
cells lie on the outer and ventral sides.
Both the sections figured run slightly obliquely across the
cord, which accounts for the facts that in fig. 47 one cord is
slightly larger than the other, and that in fig. 48 the origin of
a nerve is seen on one side only.
MONILIGASTER GRANDIS, A. G. B. 343
GENERATIVE SYsremM.
This consists of a pair of testes, each testis lying in a sperm- |
sac, a portion of the wall of which forms the “ ciliated rosette,”
and is continued as sperm-duct to open into the atrium on
each side, the prostate glands, a pair of ovaries, a pair of ovi-
ducts formed by a modification of the septal wall, a pair of
ovisacs, and a single pair of spermathecz provided with long
ducts which present special enlargements, copulatory pouches,
close to their external apertures.
The testes belong to Segment 1x, the ovaries to Begpient XI,
the spermathece to Segment vi11.
Sperm-sacs, Testes, Ciliated Rosettes, Sperm-
ducts, Atria, and Prostates.—The sperm-sacs are sus-
pended in Septum 1x-x. Each appears in the adult to pierce
the septum so that half lies in one segment and half in the
other, while the septum is attached all round the equator, so
to speak, of the sac. The sac can be pushed forwards or back-
wards, carrying the septum along its attachment with it, and
when this is done, the septum, owing to its contractility, shrinks
behind or in front of them so that they appear then to lie
wholly in one segment or the other. Sections of a young
worm in which the sperm-sacs were just visible to the naked
eye show the sperm-sac lying in front of, although in contact
with, Septum r1x-x. The sperm-sac is oval in shape; it con-
tains the testis, and in the adult, developing spermatozoa,
while a portion of its wall forms the “ciliated rosette.” The
cavity is traversed by trabecule which consist of blood-vessels
with a minute amount of connective tissue and muscle.
The testis is attached to the inner wall of the sac on the
ventral side, in front of the “ ciliated rosette.” It is a small,
white, flattened body, divided up into lobes at its free edge,
and resembles in structure the testis of other earthworms,
The sperm-duct joins the sperm-sac just behind the testis,
but still in front of the equatorial attachment of the septum—
i. e. in Segment 1x—and its wall spreads out so as to become,
I think, the walls ofthe sac. The epithelium lining the sperm-
344 ALFRED GIBBS BOURNE.
sac is thick all round the entrance of the sperm-duct, and
ciliated; it forms, in fact, the “ciliated rosette,’’ which does
not stand out freely into the cavity of the sac at all: indeed,
the ciliated epithelium is directly continuous with the rest of
the epithelium of the sperm-sac. The “ ciliated rosette ” does
not even present any convolutions of its surface.
The sperm-duct is much convoluted, and the whole is
supported and its convolutions bound together by a special
development of the mesentery, which in other segments
supports the nephridia. A mass of convolutions lies anterior
to Septum rx-x, the duct then penetrates the septum, and
another mass lies posterior to the septum; finally the duct
runs down and enters the prostate close to the body-wall, and
at its inner (median) side. The sperm-duct is enormously
long; I have completely unravelled it under a dissecting
microscope, and found that when straightened out, without
stretching, it measures as much as 94 inches.
I cannot suggest what may be the use of this great length.
Its walls are ciliated and not glandular, and the spermatozoa
are not, I believe, built up into spermatophores.
The wall consists of a layer of slightly flattened cells, ciliated
upon apparently a portion of their surface only, while outside
is a thick layer of connective tissue. There is no muscle in
the wall (figs. 53, 54).
The sperm-duct opens into the atrium. The shape of the
atrium varies according as the papilla bearing the male pore
is protruded or withdrawn, and as the muscles of its own walls
are contracted or relaxed. When the papilla is protruded the
atrium is a mere tube; when the papilla is withdrawn, if the
muscles in the atrial wall are relaxed, the cavity is approxi-
mately spherical; but if the muscles are contracted the epi-
thelial lining becomes much plicated and the cavity quite
irregular,
The atrial epithelium consists of two kinds of cells—non-
glandular columnar cells and greatly elongated gland-cells,
which dip down a great distance below the epithelial layer,
and are arranged in groups. Hach, however, sends its duct up
MONILIGASTER GRANDIS, A. G. B. 345
to take its place among the columnar cells. These gland-
cells do not occur near the external aperture; the glandular
mass formed by them constitutes the “ prostate ;’? a layer of
muscle, the fibres of which run in various directions, overlies
the epithelial layer ; outside this comes the thick layer formed
by the deep-lying glandular portions of the prostatic cells; out-
side this is another layer of muscle, and outside this the ordi-
nary layer of coelomic epithelium. The nuclei of these
epithelium-cells can be seen with difficulty in sections, but the
presence of the cells themselves may be very clearly demon-
strated by the use of silver nitrate.
The sperm-duct penetrates the prostate, and then opens into
the atrium at the point furthest from the male pores (fig. 55).
Ovaries, Oviducts, and Ovisacs.—The ovaries lie in
Segment x1, or rather in that closed portion of the segment
which contains also the nephridium ; they are not therefore
exposed in an ordinary dorsal dissection. When the septum
is cut so as to expose them, they are seen as large, brilliantly
white, frill-like organs ; they really lie above.and at the side of
the csophagus, and not near the ventral wall. They are as
much as 3 inch long and 4 inch deep. Each is really a very
closely-set zigzag, the free edge of which is slightly lobed and
thicker than the attached edge owing tothe greater development
of the ova which are set free from the free edge (figs. 50 and 57).
The oviduct is merely a modification of a portion of Septum
x1-xi1. There is nothing standing out freely into the segment.
The epithelium is thickened and thrown into folds and ciliated,
and as the wall is very thin in the intersegmental region, there
is a very short little duct running to the exterior ; the aperture,
as mentioned above, is strictly intersegmental. The ovisacs
are diverticula of that portion of Septum x1-x1 which forms
part of the wall of the sac containing the ovaries and nephridia
of Segment x1; they, and indeed the whole sac, become packed
with ova; the pair of ovisacs and this portion of Segment x1
thus packed are shown of natural size in fig. 51. A young
ovisac just developing is shown in fig. 59. The wall becomes
thick and very vascular, and when packed with ova there is an
346 ALFRED GIBBS BOURNE.
exceedingly rich network of capillaries forming a perfect net-
work throughout the lumen; very little, if any, connective
tissue accompanies these capillaries. The ova lie in the
meshes, and the whole thing becomes like an exceedingly vas-
cular solid gland ; the ova, however, remain perfectly free and
imbibe nutriment, and attain three or four times the size they
were when they entered the sac. A ripe ovum is about >,°5
inch in diameter.
Spermathece, Spermathecal Ducts, and Copulatory
Pouches.—The spermathece are pear-shaped, and the duct is
continued from the thin end. They lie attached by mesentery
to the posterior face of Septum vii-vi11 at the level of the
cesophagus. The duct held by the mesentery coils a good
deal, and then penetrates the septum and runs along for a
considerable distance embedded in the muscle of the septum.
It never actually passes through into Segment vir. It opens
into a very small oval pouch, which is itself embedded in the
body-wall and opens to the exterior. The pouch must serve
as a copulatory pouch, and may be so styled.
The spermathecal wall presents a muscular layer immedi-
ately under-the celomic epithelium, and below this comes the
spermathecal epithelium. Near the opening of the duct and
for some little distance inwards the epithelium is composed
of small columnar cells, but the rest of the epithelium is
glandular. All the cells appear to be gland-cells, but they are
long narrow cells; the epithelium in this glandular region is
about three times as thick as elsewhere. The epithelium of
the duct is composed of very small cells, outside this is a
muscular layer. The epithelium of the copulatory pouch is
composed of rather larger cells, but they are still small and
not glandular, and it must be remembered that a papilla can
be protruded here, and when this is done the pouch must
become a tube as is the case with the atrium.
I have found in the spermatheca masses of spermatozoa
and an albuminous-looking mass, secreted doubtless by the
glandular cells, as there is generally a mass of it up at that end
of the sac.
MONILIGASTER GRANDIS, A. G. B. 347
The Vascular System in some other Species of
Moniligaster.
The two main variations from the arrangement of this
system in M. grandis which obtain among my species are:
1. A supra-intestinal vessel may be present.
2. The latero-longitudinal and subneural system may com-
municate with two pairs of hearts instead of only one.
In M. pellucida there is a supra-intestinal vessel (fig. 30).
This means that the intestinal veins of certain segments—in
this case Segments x—xvi11—instead of opening into the dorsal
vessel by means of dorso-intestinal vessels, collect into’a special
trunk—the supra-intestinal—which, after receiving all the
blood from the latero-longitudinal and subneural system, com-
municates with the most posterior pair of hearts; the dorsal
vessel has no connection with these, they are therefore intes-
tinal hearts.!
The supra-intestinal vessel is practically a specialisation of
the latero-longitudinal and subneural system.
In M. sapphirinaoides, M. ophidioides, and M. ro-
busta the latero-longitudinal and subneural system communi-
cates with hearts as in M. grandis, but in addition to vessels
passing up to the hearts of Segment rx, as in the latter species,
there is a similar pair opening into the hearts of Segment viit.
This entails no other variations.
In fig. 29 I have shown the way in which the subneural
trunk communicates with the latero-longitudinals. There is
always asymmetry in this region, the connecting trunk being
always larger on the right hand side than on the left. The
communication here is better developed than in M. grandis,
but in neither case does it seem perhaps to be a very ample
way of providing for the return of all the blood from the peri-
pheral capillaries in the posterior region of the body to the dorsal
vessel or hearts, but it must suffice for the purpose, and it
1 Cf. Bourne on Megascolex coeruleus, in ‘Quart. Journ. Micr.
Sci.,’ vol. xxxii, p. 64, footnote.
348 ALFRED GIBBS BOURNE.
must be borne in mind that the flow is here a continuous one
while in the arteries it is intermittent.
The arrangement of the vessels in a segment of M. ophi-
dioides shown in fig. 28 emphasises my view as to the direc-
tion of the blood flow, such as it is, in the dorso-tegumentary
vessels. The dorso-intestinal vessels on the gizzard wall have
a characteristic arrangement which is here shown.
Generalisations with regard to the Vascular System
in Earthworms.
IT am now in a position to offer some wider generalisations
with regard to this system than I was when writing on Megas-
colex ceruleus, but, considering the large number of genera
known, our information on the subject is still very scanty. If
the anatomical relations of their afferent and efferent vessels
be taken into account, we may conveniently distinguish two
sets of capillary “ networks ” and term them respectively peri-
pheral and intestinal networks. To the former category belong
all capillaries in the body-wall, septa, nerve-cord, nephridia,
and, indeed, all the viscera except the intestine. To the latter
category belong the capillaries in the intestinal wall; and here
we must distinguish in some worms between those in the
anterior region and those in the region behind the gizzard,
those in the anterior region having in some worms relations
similar to those of peripheral networks,
It is, I think, desirable to use the terms artery and vein for
the vessels which have relations with the capillary networks,
We can recognise :
1. Arteries which carry blood from the main trunks towards
capillary networks ; these arise from the main trunks and then
repeatedly subdivide.
2. Veins which carry blood from capillary networks to main
trunks; these are formed by the repeated junction of smaller
vessels, and finally open as single large vessels into a main
trunk.
3. Veins which, like the vertebrate portal vein, carry blood
from one capillary network to another. These are vessels or
MONILIGASTER GRANDIS, A. G.B. 349
networks of vessels, the branches of which all repeatedly sub-
divide until they become capillaries, some of which, i. e. those
in one particular region of the body, are continuous with the
capillaries formed by an artery, and others with those formed
by an ordinary vein.
I shall speak of these last-mentioned veins collectively as a
portal system.}
The most important distinction which I find between the
vascular system of one worm and that of another lies in the
presence or absence of a “ portal system.’ Where a portal
system is present there are no intestinal arteries; the blood is
brought to the capillaries of the intestine from the peripheral
capillaries by means of the portal system.
Where there is no portal system the intestine receives intes-
tinal arteries (ventro-intestinal vessels), while the blood in the
peripheral networks is carried by ordinary veins (latero-longi-
tudinal and subneural vessels) direct to the main trunks (dorsal
vessel, supra-intestinal vessel, or hearts).
There is no portal system in Moniligaster or Lumbri-
cus, while on the other hand I have observed such a system in
all the Perichzetide which I have examined, in Perionyx and
Acanthodrilus, and from facts recorded by Perrier and
Benham, I have no doubt but that such is present in Uro-
cheta, Pontodrilus, Microcheta, Rhinodrilus, Plu-
tellus, Digaster, and Titanus.
The character which seems to me to come next to the presence
or absence of a portal system in order of importance is the
presence or absence of a subneural trunk.
Where a subneural is present in addition to a portal system,
as in Urocheta and Pericheta (sensu restrictu) it appears
to take some of the blood more directly into a main trunk (a
conclusion to which I am led by a study of Perichzta).
Where a subneural is present without a portal system it
1 As the term “portal” is now in general use to designate veins other
than that to which it was originally applied, there seems to be no objection
to its use for all vessels having similar relationships with regard to capillary
networks,
350 ALFRED GIBBS BOURNE.
forms the main trunk of the venous system in the posterior
region of the body, while the latero-longitudinals form the
main trunks in the anterior region, and they join together to
open into the dorsal vessel or a pair or pairs of hearts.
Megascolides appears to have no portal system, because,
according to Spencer, the latero-longitudinals open into the
dorsal vessel; but no subneural is present, and no mention is
made of ventro-intestinal vessels; so that unless the blood’
carried outwards by ventro-tegumentary vessels is returned to
the dorsal vessel by the dorso-tegumentary vessels, and blood
is supplied to the intestinal wall by dorso-intestinal vessels
and returned—where and how we do not know—I am at a loss
to explain Spencer’s account. But I am not prepared to base
any arguments upon it, as one might imagine from the account
that there was a perpetual supply of blood to the various
capillaries which never returned thence. =
Other variations in the vascular system are of apparently less
importance, e. g. a supra-intestinal vessel is present in some
species of a genus and absent in others.
With regard to the course of the blood, I have little to add
to what I have already said with regard to Megascolex
ceruleus and Moniligaster grandis. These two types
cover, with slight variations, all cases which I expect exist,
except a case where there is no portal system and no subneural
vessel; in such a case—and, as Spencer’s description stands,
Megascolides presents this arrangement—the dorso-tegu-
mentary vessels may be the veins of the peripheral capillaries.
I have again and again returned to the course taken by the
blood in these vessels. I cannot help thinking that primitively
they are efferent vessels, and that both they and the dorso-
intestinal vessels bring blood to the dorsal vessel. In this case
they can only have, in worms otherwise well provided with a
venous system, the function suggested above for M. grandis
of regulating the pressure in the peripheral capillaries, and
have practically no flow in them in one direction or the other,
In this case it would not be surprising, in cases where there is
no portal system and no subneural vessel or other main venous
MONILIGASTER GRANDIS, A. G. B. 351
trunk, to find them acting as veins. This theory is borne out
by observations on the development of a large Perichete. I
do not find any stage in which commissural vessels run in
each segment from the dorsal vessel to the ventral ; the hearts
develop as such vessels, but in all the segments, those contain-
ing the hearts as well as the others, a pair of dorso-tegumentary
and a pair of ventro-tegumentary vessels develop and com-
municate only through capillaries, so that at this stage the
ventro-tegumentary vessels are the arteries of the peripheral
tissues, and the dorso-tegumentary vessels the veins. This is
a worm which, like all Perichetez, subsequently develops
intestino-tegumentary vessels.
Notes on the Clitellum and Generative Organs.
Clitellum.—Beddard (13) has figured the structure of the
clitellum and insists upon its being composed of ‘‘a single
layer of cells,” by which I presume he means that none of the
-cells dip down into the body-wall further than others, but
they do so to a certain extent in his figure, and they do so to
a great extent in all the clitella which I have examined, and
I expect that the clitellum of his specimen of M. bahamensis
was not fully developed.
Even in the most fully-developed clitella which I have exa-
mined, a number of the cells remain unglandular between those
which have become glandular, a state of things which does not
appear to occur in other earthworms.
Generative Organs.—The resemblance of these organs
to those of various lower Oligocheta is, as has been fully
pointed out, very striking. My observations all tend to con-
firm the resemblance.
The ciliation of the speym-ducts is a very striking feature.
The testes undoubte: y belong to Segment 1x. I may
anticipate my observations on the development of the genus
by stating that these organs develop as proliferations of the
celomic epithelium of the anterior face of Septum 1x-x, and
the ovaries as precisely similar structures on the posterior
352 ALFRED GIBBS BOURNE.
face of Septum x-x1, while no rudimentary genital organs
form in any other segment. The “ ciliated rosette” develops,
I believe, from the epithelium of this septum just ventrad of
the testis, and afterwards, together with some of the muscles
of the septum, grows round to enclose the testis and form the
sperm-sac. The development of the sperm-duct I have not
yet been able to make out. The wide mouth of the oviduct is
in the adult nothing but a specialised portion of the septal
wall. So that there seems to be in Moniligaster no connec-
tion between genital ducts and nephridia.
With regard to the relation of prostate to atrium I agree
entirely with Benham (14), but I find that each gland-cell in
the prostate opens by its own duct into the atrium, that it is,
in fact, merely an enlarged cell of the atrial epithelium, so
that I do not see why we should use the term ‘ multicellular
gland.” The ccelomic epithelium is always to be found outside
the other structures, in the form of pavement cells easily
demonstrated by the use of silver nitrate.
Systematic Account of the Genus Moniligaster.
Before describing any other species! I shall discuss the value
of the various characters for systematic purposes, taking ac-
count of my own observations only.
Colour has a certain value ; there are groups of species with
pigment in the body-wall (figs. 6—10) and groups with little
or no pigment (figs. 3,4), but the exact colouring varies much
in some cases within the species (figs. 8—10).
Size has some importance as there is great variation, but
the condition of the worm when measured must be estimated
(see above, M. grandis—size, p.307),and the specimen must be
1 Out of the thirteen species described below, I place three (M. grandis,
M. naduvatamensis, and M. nilamburensis) in one group and call it
in the following account the M. grandis group; three (M. ophidioides,
M. robusta, and M. sapphirinaoides) in another group and call this
the ophidioid group; two (M. pellucida and M. uniqua) in another
group and call this the pellucid group; the other species I do not group
together in any way.
MONILIGASTER GRANDIS, A. G. B. 353
adult as shown by the generative organs; even then there is a
variation of at least 25 per cent. of the total length between
adult individuals of the same species. ‘This variation does not
depend upon age, there are long individuals and short indivi-
duals as in the human species. I have proved this by observa-
tions taken at the time of hatching; at this time and indeed
long before, growth by the addition of new segments has
ceased while there is considerable variation in the number of
segments—and the adult length depends, given an equal amount
of contraction, solely upon the number of segments. The
well-known phenomenon of regeneration of tail segments must
be mentioned in this connection ; where the original tail has
been injured the number of new segments which may form
seems to be about the same as the number lost, as in any
species the total number varies only to the extent of the usual
percentage, whether the individuals have regenerated tails or
not. I give a tabular statement for M. sapphirinaoides.
The specimens were part of the same collection, from the same
spot, and it was from this same collection that the nine speci-
mens described below as varieties of this species were taken.
The specimens in both cases were taken at random.
354. ALFRED GIBBS BOURNE.
Specimen. | Length. | N maa ig | Gamntrasemee Total. Remarks.
1 170 mm, 140 | 15 155
2 UD Dita, 168 -— 168
3 180"; 148 | 12 160
4 ie reer 91 26 117 | Abnormal.
5 920 ,, 143 56 199
6 i Ui 172 | _— 172
7 15017, 170 | — 170
8 il See 133 | 44, 177
9 150°" ;; 140 | — 140
10 1D ke 115 | 114+ 6 132 | Twice injured.
ll 05 asf 146 | 10 156 | Apparently still
growing.
12 190 _,, 163 | 36 199
13 95 vis 84. | 40 124 | Still growing.
caves, 109 6 115 | Still growing.
15 120°-;, 160 | — 160 | -
16 140 ,, 140 24. 164
17 160"; 139 37 176
18 120) 559 164 ~- 164.
19 IBD F5,. | 164. _ 164.
20 145 ,, | 157 — 157
21 Dp? sa 150 — 150
22 ay aay 172 -- 172
We get an average number of segments in the specimens
which had not been injured of 162, and in those which had
grown new tails of 156, or, omitting Nos. 13 and 14 which
were still growing, and No. 10 which had been twice injured,
-and No. 4 which was abnormal, the half segments on one side
not always corresponding with the halves on the other side, we
get an average of 16] segments in an injured worm to corre-
spond with the average of 162 segments in an uninjured one.
Segments.—I have spoken above of their number. Some
importance attaches to the variation in size of the anterior
segments (compare figs. 1 and 6 with figs. 5 and 12). Annu-
lation of the segments, which depends upon a particular
arrangement of the circular muscles, is characteristic of groups
of species.
Prostomium.—This presents no easily recognisable varia
tion in the group (see M. grandis—prostomium, p. 309).
Setx.—The length varies in the group (my species) from
MONILIGASTER GRANDIS, A. G. B, 355
0°24 mm. to 0°63 mm., but in variable species like M. sap-
phirinaoides it varies in different individuals from 0'375 mm.
to 0625 mm., so that I have judged of the normal size for the
species after examining several individuals only. There is no
appreciable variation in the size of different sete (full grown
and unbroken) from any one individual.
The presence or absence of all or one couple from Segment
I1 is a good specific character, but I have found them present
in individuals of a species where they are as a rule absent.
The two setz of a couple are always very close together.
The position of the rows of couples, i.e. the extent of the
dorsal, lateral, and ventral gaps, is a very good specific cha-
racter, and only fails, to my knowledge, in the case of what I
consider to be hybrids between M. ophidioides, M. ro-
busta, and M. sapphirinaoides, and between M. uniqua
and M. pellucida. The position of the seta rows is only
quite obvious (without mounting) in most species behind
about Segment x111, but from that point onwards the relative
distance of the rows to one another does not vary in the in-
dividual.
Occasionally in a species the distance between the two sete
of a couple is greater than usual. My only instance of this is
M. nilamburensis.
Clitellum.—This, owing to its transient nature, would not
in any case be a good character (it does not show externally for
more than a month or two in the year, but I doubt very much
whether, having been once developed, it would ever become
unrecognisable in sections ; I have never found it so). So far
as my observations go it always develops, and that upon Seg-
ments x—x111, strictly confined to those segments, never com-
plete ventrally (see M. grandis—clitellum), and never obscur-
ing, or developing over, the intersegmental grooves. All the
egg-capsules of Moniligasters I know (I have never found those
of M. grandis!) are perfectly neatly formed and very globular.
1 The small worm which so often occurs crawling in the mucus on M.
grandis, is not, as I at first thought, the young of this animal; nor is it a
Moniligaster at all.
You. 36, PART 3,—NEW SER. BB
356 ALFRED GIBBS BOURNE.
In a M. sapphirinaoides where the clitellum is well de-
veloped, the worm, when suddenly killed, makes this region of
the body very globular by means of certain muscles running
from the dorsal to the ventral part of the segment on each
side, and it no doubt assumes this shape when forming the
ege-capsule, and allows the semi-fluid secretion to set before
resuming its ordinary shape and crawling out.
Dorsal pores are never, to my knowledge, present.
Nephridiopores are usually placed on the very anterior
margin of the segment in the direction of the outer seta rows,
but the more dorsal position of the pores of Segments 111—1x,
and occasionally others, forms a good specific character. It
serves, for instance, to distinguish between M. grandis and
M. naduvatamensis, but in the “ ophidioid” group (see
below) is liable to variation.
Genital Apertures.—These always lie between the same
segments, Viz. VII-VIII, X-XI, xI-x11. There is a slight specific
variation as to their relative distance from the median line,
but they always lie somewhere between the direction of the
outer and inner seta rows or precisely in the direction of one
of those rows. With regard to their character, I can only say
that in the M. grandis group and in the pellucid group
papillz commonly protrude from the male pores when killed,
while in the ophidioid group they do not.
Body-wall.—This is thickened at the sides in the ophidioid
group, enabling them to move very rapidly in a serpentine
manner.
Septa.—The only septa which thicken are v-v1 to viI-1x,
and these vary in different species in the extent to which they
are thickened, the most anterior being in some cases hardly
any thicker than the majority of septa in the body. The
peculiar arrangement of some of the septa such as is described
in M. grandis obtains to a greater or less extent in all
species. One variation deserves special notice: in some cases,
asin M. grandis, the arrangement of Septum x-x1 is such as
to completely shut off the ovaries and oviducts and the aper-
tures of the ovisacs from the rest of the cavity of the segment ;
MONILIGASTER GRANDIS, A. G. B. 357
this modification obtains only when the worm is fully adult ;
in other species even in the adult this modification does not
obtain. I express these two conditions by saying “ ovaries
enclosed ” or ‘ ovaries free” as the case may be.
Muscles.—The longitudinal muscles have in some and not
in other species the “ feathered ” arrangement.
Setal muscle-bands are present or absent according as the
setz project or not into the celom.
Alimentary Canal.—The number of the gizzards and the
particular segments in which they occur, although subject to a
considerable amount of individual variation, is to a certain
extent a good specific character. It is not difficult by examin-
ing a number of individuals to determine the number of
gizzards normal for the species, and it is possible, by examin-
ing a still larger number, in some cases to fix upon a normal
position. It is very common to find a slight gizzard develop-
ment in front of or behind the normal gizzards, and sometimes
these additional gizzards are as large as the others. The giz-
zards which do develop are always in contiguous segments.
There are rarely if ever more than one or two gizzards in
excess of the normal number. The variation in position de-
pends of course for one thing upon the number of gizzards
developed, but given the normal number the position may still
vary, especially in some species, by two or even three seg-
ments. It is evident that the cesophageal wall in the gizzard
neighbourhood can very easily produce these muscular deve-
lopments, and the variation in position is fully accounted for
by the fact that the mesenteron is not primarily metameric-
ally segmented, and that the portion of the gut bounded by any
particular septa is not necessarily homogenous with the portion
bounded by the homogenous septa, in another individual.
A typhlosole is never present.
Vascular System.—This presents slight individual varia-
tions only, but certain well-marked specific variations, some of
which I describe elsewhere in this paper.
Nephridia.—I have not specially studied the variations in
the structure of the nephridium ; so far as my observations
358 ALFRED GIBBS BOURNE.
go, it is always built upon the same general type as in
M. grandis.
Nerve-cord.—This presents no marked variations,
Generative Organs.—It is with regard to these organs
that the greatest variations are usually looked for, but there is
in this genus, so far as it is known to me, the most striking
anatomical, and indeed histological, uniformity in this matter.
The testes may vary in the position of their attachment to the
wall of the sperm-sac. The inner wall of the prostate may or
may not present a layer of muscle between the gland-cells and
the coelomic epithelium; in the former case these organs
present on opening the worm a shining appearance, and in the
latter their glandular nature is very apparent. I express the
latter condition by saying “ prostates glandular in appearance.”
I have not examined them all histologically, and there may be
in some species a few muscle-fibres present, even when I have
called them “ glandular ;”’ but anyone who has compared the
“slandular” with the “non-glandular” appearance cannot
fail to distinguish between them, and I have never found
individual variation in the matter.
The precise point of penetration of the sperm-duct into the
prostate varies in different species. The prostates vary in
shape to a limited extent in different species, but they also vary
according as the glands are full or empty of secretion.
The spermathece and sperm-sacs vary in shape in different
species, but they also vary according as they are full or empty,
a pyriform spermatheca becomes almost globular when quite
full, and an oval sperm-sac becomes to a slight extent broader
along its shortest axes.
The most valuable character is to be found in the copulatory
pouch ; this is in some species small and completely embedded
in the body-wall, and at first sight would be overlooked ; in
others it is large, but not bilobed; in others, again, it is bilobed.
I have never found individual variations in respect of this
matter except in the ophidioid group of species, where I
think hybrids occur ; this matter is dealt with elsewhere in this
paper, It is not easy always to see, but I believe that the
MONILIGASTER GRANDIS, A. G. B. 359
spermathecal duct sometimes does and sometimes does not
become embedded in Septum vii-viit before penetrating the
body-wall.
4
Genus Moniligaster, Perrier, 1872.
Prostomium never dovetailed into peristomium.
Sete in four couples.
Clitellum develops on Segments x, xI, X11, and XIJI; it is
a transient structure, and does not extend over the interseg-
mental grooves.
Dorsal pores absent (? M. barwelli, M. beddardi).
Nephridiopores usually in the direction of the outer seta
couple, but in some cases more dorsally placed in certain of the
anterior segments.
Genital apertures, three pairs, all strictly intersegmental.
Male pores between Segments x and x1.
Oviducal pores between Segments x1 and XII.
Spermathecopores between Segments vil and VIII.
Testes, one pair, each enclosed in a sperm-sac; they belong
to Segment Ix.
Sperm-sacs, one pair, provided with vascular trabecule ;
held equatorially by Septum 1x-x.
Ciliated rosettes are specialised portions of the walls of
the sperm-sacs.
S perm-ducts very long and convoluted ; ciliated along their
whole length.
Atria, large muscular sacs.
Prostates developed as glandular walls of the atria.
Ovaries, one pair of frill-like bands in Segment x1, produc-
ing enormous numbers of eggs, which, after further growth in
the ovisacs, are placed in enormous numbers in the egg-capsule,
and take the place of albumen in nourishing the one or two
which develop.!
1 [ have added this very interesting fact here because it is, in my experience,
very characteristic of the genus; and I have examined the egg-capsules of
three species of Moniligaster. I hope to give a fuller account of the phe-
nomenon in a subsequent paper.
360 ALFRED GIBBS BOURNE.
Ovisacs, one pair, developed as diverticula of Segment x1;
provided with vascular walls and trabecule, capable of great
extension.
Spermathece, one pair, with muscular ducts, the terminal
portions of which form copulatory pouches in Segment VIut.
Nephridia paired, provided with a muscular-we'led bladder ;
most anterior pair in Segment 111 ; anterior pairs not modified.
Gizzard moniliform, two- to six-lobed; each lobe confined
to one segment; lies somewhere between Segments x1 and
06:6. 6006
Ty phlosole absent.
Vascular System.—No portal system, i. e. peripheral veins
go direct to the dorsal vessel (or hearts), while the intestinal
wall is supplied by ventro-intestinal vessels ; subneural present ;
supra-intestinal vessel may be present or absent ; four pairs of
hearts in Segments vi—1x.!
M. grandis, A. G. B., 1886.
Prostomium broad.
Size, 590 mm. long; 36°5 mm. in circumference.
Segments, 480, annulated.
Pigment, very little present.
Sete 0°5 mm. long, dorsal gap 18 mm., lateral gaps 6°95
mm., ventral gap 5° mm.
Nephridiopores in Segments 111 to 1x dorsal of the outer
seta rows; elsewhere in the direction of the outer seta rows.
Male pores nearer the inner seta rows than the outer.
Oviducal pores in the direction of the inner seta rows.
Spermathecopores nearer the outer seta rows than the
inner.
Septa v-vi to vili-1x very thick.
Gizzard in five lobes, occupying Segments XVIII—XxXII.
Sperm-sacs oval.
1 [ have included in this diagnosis of the genus many characters which, no
doubt, properly belong to the family. The family contains at present one
other genus only, viz. Desmogaster, Rosa, 1890.
MONILIGASTER GRANDIS, -A. G. B. 361
Prostates hemispherical, not glandular in appearance (see
p- 398).
Ovaries enclosed (see p. 357).
Spermathece pyriform to globular.
. Copulatory pouches small, simple, embedded in the wre
wall.
Vascular System.—Latero-longitudinal vessels connected
with the hearts of Segment 1x. No supra-intestinal vessel.
Habitat.—Nilgiris, S. India. Widely spread at elevations
from 5000—8000 feet, in dry grass land; burrows to 9—10
feet, and remains coiled up at the bottom of the burrow during
the dry season.
M. naduvatamensis, sp. n.
Prostomium small and pointed.
Size, 500 mm. long, 16 mm. in circumference.
Segments, 400.
Pigment absent.
Setz 0°3 mm. long, dorsal gap 9 mm., lateral gaps 2°5 mm.,
ventral gap 2 mm.
Nephridiopores all almost exactly in the direction of the
outer seta rows.
Male pores rather nearer the outer seta rows than the
inner.
Oviducal pores in the direction of the inner seta rows.
Spermathecopores midway between the outer and inner
seta rows.
Septa v-vi to vi1-1x thickened, but not so much so as in
M. grandis.
Gizzard in three lobes, occupying Segments xv—xvu.
Sperm-sacs slightly kidney-shaped.
Prostates glandular in appearance.
Ovaries not enclosed.
Spermathece pyriform to globular.
Copulatory pouches small, simple, embedded in the body-
wall.
Habitat.—Naduvatam, Nilgiris, at about 6000 feet.
362 ALFRED GIBBS BOURNE.
Remarks.—I have only had a few specimens of this worm.
It is very like M. grandis, and occurs along with that species.
It is easily to be distinguished from M. grandis by its small,
pointed prostomium. The whole anterior end is pointed, and
when killed in spirit it is flaccid, while M. grandis is for a
time very rigid.
M. nilamburensis, sp. n.
Prostomium broad.
Size, 760 mm. long, 23°5 mm. in circumference.
Segments, 566, annulated.
Pigment, very little present.
Setz 0°63 mm. long. They appear black to the naked eye.
Dorsal gap 14 mm., lateral gaps 3 mm., ventral gap 3-5 mm.
Each couple is separated by 0°5 mm., and are therefore more
widely separated than usual.
Nephridiopores all, I think, in the direction of the outer
seta rows.
Male pores nearer the inner seta rows than the outer.
Oviducal pores in the direction of the inner seta rows.
Spermathecopores in the direction of the outer seta
rows.
Septa v-vi to vii1-1x very thick.
Gizzard in five (or six) lobes, occupying about Segments
XXVIII—XXXIII (normal for five lobes, the sixth lobe may be in
XXVII OF XXXIV).
Sperm-sacs oval.
Prostates 7
Ovaries :
Spermathece eas in M. grandis,
Copulatory pouches
Habitat.—I have had preserved specimens only from
Nilambur, near the sea-level, which were kindly sent me by
Mr. Hadfield, of the Forest Department, who informs me that
the worm burrows along in the wet season just underneath the
surface.
Remarks.—This is another worm like M. grandis. It is
MONILIGASTER GRANDIS, A. G. B. 363
the only species in which I have found the gizzard so far back ;
in Desmogaster the gizzard occupies, according to Rosa, Seg-
ments XX—XXIX.
M. pellucida, sp. n.
Size, 90 mm. long (presents great variations, many speci-
mens as long as 190 mm., as in the specimen figured), 9°25 mm.
in circumference.
Segments, 130, not annulated.
Pigment absent.
Setz 0:24 mm. long, dorsal gap 6 mm., lateral gaps 1 mm.,
ventral gap 1:25 mm. Setz absent from Segment 11.
Nephridiopores in the direction of the outer seta rows.
Male pores nearer the inner seta rows than the outer.
Oviduca) pores slightly dorsad of inner seta rows.
Spermathecopores slightly ventrad of outer seta rows.
Septa v-vi—vitl-tx thickened.
Gizzard in four (or five) lobes occupying Segments xv—
XVIII (XVIIMXXI, XVI—XX, XV—XIX, XIV—XVII, XIV—XVIII, are
not uncommon variations).
Sperm-sacs lemon-shaped.
Prostates, flattened hemispheres.
Ovaries not enclosed.
Spermathece lemon-shaped, in one or two specimens oval.
Copulatory pouches simple, embedded in the body-wall.
Vascular System.—Supra-intestinal vessel present, con-
nected anteriorly with the hearts of Segment 1x, which are
therefore not “dorsal hearts.”
Habitat.—Ootacamund and Naduvatam, Nilgiris.
Remarks.—The habit of this worm of contracting its ante-
rior end into the shape shown in fig. 3 is very characteristic ;
this is the shape usually found in spirit specimens. The body-
wall is very transparent.
M. uniqua, A. G. B., 1886. (Syn. M. papillatus, A. G. B.,
1886.)
Size, 220 mm. long, 15°6 mm. in circumference.
Segments, 316, faintly annulated.
364 ALFRED GIBBS BOURNE.
Pigment absent.
Sete 0°32 mm. long, dorsal gap 8 mm., lateral gaps 2°6
mm., ventral gap 2°4 mm.; present on Segment 11.
Nephridiopores. closer even than usual, to the interseg-
mental grooves, in the direction of the outer seta rows.
Male pores between outer and inner seta rows, papilla
very frequently protruded (this is the character I first noted
in M. papillatus, but it is by no means specific, and this does
not form a distinct species).
Oviducal pores in the direction of the inner seta rows.
Spermathecopores in the direction of the outer seta rows
(between setz 1 and 2).
Septa v-vi to vitl-1x thickened.
Gizzard in four (or five) lobes occupying Segments xvi1—xx
(XV—XIX, XVI—XX, XVII—XxXI).
Sperm-sacs somewhat lemon-shaped.
Prostates teat-shaped, projecting rather backwards.
Ovaries not enclosed.
Spermathece pyriform and oval or lemon-shaped.
Copulatory pouches simple.
Habitat.—Ootacamund, Coonoor.
Remarks.—This species and M. pellucida occur together
and form, I believe, hybrids,—I have found so many specimens
with an intermixture of characters,—but as I dwell at length
below upon hybrids in the ophidioid group, it is not necessary
to do so in connection with these species.
M. chlorina, sp. n.
Size, 130 mm. long, 11°5 mm. in circumference.
Segments, 135, not annulated.
Pigment, a little present, the worm becomes greenish when
put in spirit.
Setz 0'5 mm. long, dorsal gap 6 mm., lateral gaps 2 mm.,
ventral gap 15 mm. Setz absent from Segment 11.
Nephridiopores in the direction of the outer seta rows.
Male pores nearer outer seta rows than inner.
Oviducal pores in the direction of the inner seta rows.
MONILIGASTER GRANDIS, A. G. B. 365
S permathecopores in the direction of the outer seta rows.
Septa vi-vi to vil1-1x slightly thickened.
Gizzard in four lobes, occupying Segments x1v—xvII.
Sperm-sacs oval, rather pointed at the ends.
Prostates hemispherical, appear glandular.
Ovaries not enclosed.
Spermathece pyriform to oval.
Copulatory pouches small and simple.
Habitat.—Ootacamund.
Remarks.—This occurs with the two preceding species,
but is very easily picked out owing to its pigment.
M. ophidioides, sp. n.
Size, 310 mm. long, 21:5 mm. in circumference.
Segments, 200, not annulated.
Pigment present in quantity, bluish to olive green.
Sete 0°35 mm. long, dorsal gap 12 mm., lateral gaps
3°25 mm., ventral gap 3 mm. Couples very closely paired.
Nephridiopores mostly in the direction of the outer seta
rows, those of Segments vil, vill, and x11 usually, and others
occasionally, 3 mm. nearer the dorsal median line; occasion-
ally some of the pores are in the direction of the inner seta rows.
Male pores nearer the outer seta rows than the inner.
Oviducal pores in the direction of the inner seta rows.
Spermathecopores in the direction of the outer setz.
Septa v-vr to viit-1x thickened but only slightly so, the
most anterior of the four hardly so at all.
Gizzard in three lobes occupying Segments x1v—xvi (see
remarks below).
Sperm-sacs very rounded, oval.
Prostates oblately hemispherical or, including a little
more than the hemisphere, glandular in appearance.
~ Ovaries enclosed.
Spermathece pyriform and globular.
Capillary pouches bilobed, a short teat-shaped lobe in
front of Septum vii-vi11, and a longer one behind the septum.
Vascular System.—The latero-longitudinal vessels are
366 ALFRED GIBBS BOURNE.
connected with the hearts of Segment 1x only; no supra-
intestinal vessel.
Habitat: Ootacamund and Coonoor.
M. robusta, A.G.B., 1886. (Syn. M. indicus, Benham, 1893.)
Size, 200 mm. long, 19 mm. in circumference.
Segments 160.
Pigment present in quantity, bluish to greenish brown.
Setz 0°3 mm. long, dorsal gap 11 mm., lateral gaps 2°5 mm.,
ventral gap 3 mm.
Nephridiopores, perhaps normally, all in the direction of
the outer seta rows (see remarks below).
Male pores nearer the outer seta rows than the inner.
Oviducal pores in the direction of the inner seta rows.
Spermathecopores in the direction of the outer seta rows.
Septa vi-vir to vii1-1x slightly thickened, v-v1 not to any
appreciable extent.
Gizzard in four lobes, occupying segments XII—Xv.
S perm-sacs broadly oval.
Prostates differ in shape from those of the preceding species
only in the fact that they usually lap over a little towards the
middle line ; glandular in appearance.
Ovaries enclosed.
Spermathece pyriform to oval.
Copulatory pouches bilobed, the anterior lobe larger than
the posterior, and its extremity usually folded over, so as to
point backwards.
Vascular System.—The latero-longitudinal vessels are
connected with the hearts of Segments vill and 1x. No supra-
intestinal vessel.
Habitat.—Widely spread on the Nilgiris.
M. sapphirinaoides, A.G.B., 1886.
Size, 140 mm. long, 17 mm. in circumference,
Segments, 160.
Pigment present in quantity, bluish-red.
MONILIGASTER GRANDIS, A. G. B. 367
Setz 0°5 mm. long, dorsal gap 10 mm., lateral gaps 2°4
mm,, ventral gap 2:1 mm.
Nephridiopores in the direction of the outer seta rows.
Male pores midway between the outer and inner seta
rows.
Oviducal pores in the direction of the inner seta rows.
Spermathecopores in the direction of the outer seta rows.
The nephridiopores of Segment vii are just beyond (dorsal)
the spermathecopores, and, of course, being just on the seg-
ment and not on the intersegmental groove, just behind them.
Septa vi-vi1 to vi11-1x very slightly thickened, v-v1 not to
any appreciable extent.
Gizzard in four lobes, occupying Segments xv1I—xx.
Sperm-sacs oval,
Prostates, flattened hemispheres, glandular in appearance.
Ovaries enclosed,
Spermathece pyriform to globular,
Copulatory pouches not bilobed, but large (actually
larger than they are, for instance, in M. grandis).
Vascular System.—The latero-longitudinal vessels are
connected with the hearts of Segments vi1r and 1x. No supra-
intestinal vessel.
Habitat.— Widely spread on the Nilgiris,
Remarks on M. ophidioides, M. robusta, and M, sap-
phirinaoides,
This group of species, which I have alluded to above as the
ophidioid group, is, as a group, well characterised. The body-
wall is always absolutely opaque, its muscles, especially the
longitudinal muscles at the sides of the body, are well deve-
loped. They are all strong active worms which, when excited,
exhibit a serpentine mode of progression ; it is almost impossible
to keep hold of one when excited without injuring it. Owing
to the arrangement of the longitudinal muscles the body when
at rest is somewhat flattened, but when in active movement
becomes quite cylindrical; the tail is always capable of becom-
368 ALFRED GIBBS BOURNE.
ing very pointed and is generally pink at the tip. The worms
live in swamps and wet ground, hence no doubt the feeble
thickening of the anterior septa (note in fig. 5 how the strong
pharyngeal region is immediately followed in M. grandis by
the four thickened septa). The intersegmental grooves are
always most marked at the sides of the body, from about Seg-
ment xiv onwards. The dorsal region often has a very smooth
appearance ; the darker colour is, in spirit specimens (as noted
by Benham), very sharply confined in this part of the body to
this region ; papillz never protrude from the male pores. But
when we come to diagnose species the matter becomes almost
hopeless, and therefore most interesting. The trouble which I
have had with this group is mainly responsible for the delay
in the publication of this paper. I have at last determined to
defer further examination of the matter till a future occasion.
I have for the present made three species, and the diagnoses
given above are the outcome of the examination of an enor-
mous number of specimens, but the chances are about equal
that any particular specimen which by external appearance I
should refer to one of the three species, will or will not present
an intermixture of characters. The three species are rather
three types around which I can group the variations.
In connection with these variations I have considered three
factors,—age, habitat, and the production of hybrids. It is diffi-
cult to altogether eliminate variations due to age. The condi-
tion of the generative organs is of course the great criterion,
but the transient nature of the clitellum introduces one diffi-
culty, and I often find specimens in which some portion of the
generative apparatus, particularly the ovisacs, has a very
immature look while other parts are well developed.
There is an undoubted tendency among worms to vary with
even a slight difference of habitat (I have such variations re-
corded for future publication in connection with a Perichete),
but the variations partake in this case so markedly of an inter-
mixture of the characters of the three types that they can, I
think, only be explained by the fact that hybrids are produced,
and although to make it certain the matter needs further
MONILIGASTER GRANDIS, A. G. B. 369
investigation, it has been most forcibly borne in upon me from
time to time that in the specimens presenting intermixed cha-
racters there was something wrong with generative apparatus,
e.g. ova not found separated from the ovary and the ovisacs in a
particularly shrivelled condition, while in more typical speci-
mens at the same time of year the ovisacs are full of ripen-
ing ova.
I have given what I consider to be the typical distinctive
characters of the three species in the above diagnoses, and I
now give, in the annexed table, a few particulars of some speci-
mens of what I should call variations of M. sapphirinaoides
among worms taken at random from one and the same locality.
To carry the subject further it would be necessary to do this
for each species in groups from various localities, but this
would prolong this paper beyond reasonable limits, and I should
like to do this at some future time on a really large scale.
Specimens 8 and 9 at any rate are immature, and they may
pass as young M. sapphirinaoides; the sete are rathersmall,
but the setz, like the exoskeleton of an arthropod, are shed at
intervals and replaced by larger ones; the seta has not a per-
manent growth, and the full-grown embryonic seta of one of
these species is 0'1 mm. long. The gizzard is far forward in
them and, indeed, in all these specimens, but the normal posi-
tion for M. sapphirinaoides has been determined by the
examination of a much greater number. The condition of the
ovaries and the spermathece in 4, 8, and 9 is due, I think, to
immaturity.
Specimen 2 is what I should call a hybrid between M.
sapphirinaoides and M. robusta,—the sete, gizzard, and
copulatory pouches resemble the latter species ; the size, colour-
ing (as in all the specimens), relative distance between the seta
rows are as in the former.
Specimen 1 also resembles a M. sapphirinaoides, but its
seta gaps have the peculiar arrangement characteristic of M.
robusta.
The nephridiopores are not mentioned above ; by the time I
could have made certain of the position of all of them the worm
ALFRED GIBBS BOURNE.
370
*|eurs Ai9a
pue ‘ayrya Sus0jtsh 7
*]yews <19A
pue ‘ayy ‘urrosas
*moTfad pue repngoly
*MoO][a4 pue Je[NgoTy
“morjak pu reyngopy
“aq pue wr0ytIs |
*MoT[as pue TelNgoTD
‘aqiqa pue wot gy
“morjad pue aejnqoyy
‘moaqyeuieds
Tews pasojoua
AxaA “paqoriq JON) JON
[jews pasojoua
£J9A *paqgoliq JON | YON
P2qQoi!q FON =| PosopouT
P2Q0TtG JON =| Posopoug
P2QO!G ON =| pasopougy
pasojous
peqong JON FON
PEqOTG FON =| Pesopoug
JOAO papjoy
pure odie] oqoy
Jolayue ‘pagorig | pasojoug
P2qoT!g JON | pasoyougy
*saqonog Ar0jeyndog *SOIIBAQ
IIIAX—AIX aes 5) |
IIIAX—AIX sale
XIX—AX ey PSL
XIX—AX fe 6 oT
XX—AX se V4
ITIAX—AIX ee PA
jews f1aa xx
XX—IAX << GhT
[jews A194 TAX ‘[]euMs ITX
IAX—IIX ‘a g
XIX—AX ‘WU G).Z
*pxezz19 ‘dey
Zuiute} 109 syusuZeg [21309 A,
© GL-T
* GLB
“UU ¢.g
‘dey
yeqaqe'T
“82-0
“WU SBF-0
* 00L
© OFT
* 807
sma i
Sma ot
= ORT
“OCT
“WU CFT
mo Oo eS
‘wamoads
MONILIGASTER GRANDIS, A. G. B. 371
would have been ruined for dissection. I quote one or two
instances in which M. robusta approaches M. ophidioides.
M. robusta specimen—
Nephridiopores in Segments 111, Iv, v, and vi . . Outer seta rows.
Pr, Pr vit and vir. . . . More dorsal.
=f ay PX, &, ET 4) oy bs yon, Outerseta rows.
Ss Pr XIk ; scm 6 <ec-e = More dorsal:
3 $5 XG s,s e “Outer seta rows.
Ss 3 MV, XV « .« «4* sf ol) Moré dorsal:
46 Pr XVI, XVII. . «yj »° Outériseta rows:
2 KYAT, TAs |. a vant egy e pore, dorsal.
Pr ss x1x onwards. . . . Outer seta rows.
M. robusta specimen—
Nephridiopores in Segments 11—vil1 . . . . . Dorsad of outer seta
rows.
‘ Fe Ix onwards . . . . Outer seta rows.
My original specimens of M. sapphirinaoides were very
iridescent, this extreme iridescence is a local variation. The
clitellum in all these ophidioid species becomes pinkish in
spirit.
‘Benham’s M. indicus is probably M. robusta, or
perhaps a hybrid between that and M. sapphirinaoides,
having most of the characters of the former, and a few (includ-
ing size) of the latter species. Benham says the setz are
2mm. long; this is surely a misprint, if not it must be a
different species, or possibly a very strongly marked local
variety. The specimen came, I think, from a part of the
Nilgiris which I have not visited.
M. parva, sp. n.
Size, 75 mm. long, 4°25 mm. in circumference.
Segments, 115.
Pigment present in small quantity.
Setz 0°4 mm. long, dorsal gap 2°75 mm., lateral gaps 0°75
mm., ventral gap 0°75 mm. ; absent from Segment 11.
Nephridiopores in the direction of seta 4, in Segment vi11
just dorsal of (and behind, of course) the spermathecopore.
Male pores between the outer and inner seta rows; papille
often protrude.
VOL. 36, PART 3.—NEW SER. cc
372 ALFRED GIBBS BOURNE.
Oviducal pores in the direction of the inner seta rows.
Spermathecopores in the direction of the outer seta row.
Septa v-vi—viii-1x thickened.
Gizzard in three (four) lobes, occupying Segments x1v—xvI
(XIV—XVII).
Sperm-sacs slightly kidney-shaped.
Prostates hemispherical, glandular in appearance.
Ovaries not enclosed.
Spermathece oval.
Copulatory pouches simple, rounded.
Vascular System.—A supra-intestinal vessel present, the
hearts of Segment 1x are connected with it.
Habitat.—Ootacamund.
M. minuta, A. G. B., 1886.
Size, 47 mm. long, 5°7 mm. in circumference.
Segments, 150.
Pigment, a considerable amount present.
Setz 0:25 (to 0:15) mm. long, dorsal gap 3°6 mm., lateral
gaps 1:1 mm., ventral gap 1 mm.; present in Segment 11.
Nephridiopores in the direction of the outer seta row.
_ Male pores in the direction of the inner seta rows (a very
unusual position).
Oviducal pores in the direction of the inner seta rows.
Spermathecopores in the direction of the inner seta rows
(a very unusual position).
Septa v-vi to vi1t-1x very slightly thickened.
Gizzard in two (three) lobes, occupying Segments x111 and
XIV (XII—xIV).
Sperm-sacs oval.
Prostates hemispherical.
Ovaries not enclosed.
Spermathece oval.
Copulatory pouches bilobed, lobes about equal size.
Habitat.—Widely spread in the Madras Presidency at sea
level and up to about 6000 feet.
MONILIGASTER GRANDIS, A. G. B. 373
M. rubra, A. G. B., 1886.
I can, unfortunately, give no further details here with
regard to this species. For the present I merely place a figure
of it on record.
I now give, as far as information is available, diagnoses in
the form I have adopted above of the species which have
been previously described by others.
M. deshayesi, Perrier (1), 1872.
Size, 150 mm. long (Perrier), “‘ largeur 6 mm.”
Nephridiopores in the direction of the outer seta row.
Male pores between the outer and inner seta rows.
Spermathecopores in the direction of the inner seta rows.
Gizzard in three or four lobes, apparently occupying about
Segments x1v—xvi._ Perrier describes also an anterior gizzard.
Habitat.—Ceylon.
M. barwelli, Beddard (2), 1886.
Size, about 40—50 mm. long.
- Pigment, very little present in the body-wall.
Spermathecopores in the direction of the outer seta row.
Prostates oval.
Gizzard in four lobes.
Habitat.—Manila.
M. houteni, Horst (4), 1887.
Size! probably about 820 mm. long, about 55 mm. in cir-
cumference.
Segments, 443.
Setz 1 mm. long ; ventral gap twice the length of a lateral .
Sap.
1 Horst gives the length as 1500 mm., and the greatest circumference as
about 55 mm. From experiments with poorly-preserved M. grandis I
calculate that the natural length of M. houteni, when living and at rest,
would not be less than 820 mm., a greater length than M. grandis ever
attains. The circumference, where it is greatest, viz. in the anterior
region, does not vary to any appreciable extent with the state of preservation
and extension, as it depends upon the size of the sperm-sacs, &c.
374. ALFRED GIBBS BOURNE.
Pigment: There is probably little pigment in the body-
wall, or Horst would have mentioned the colour.
Prostates long and tubular.
Habitat.—Sumatra (Tapanoeli).
M. beddardi, Rosa (12), 1890.
Size, probably about 40 mm. long (Rosa says 30 mm., “ es.
piuttosto contratto’’).
Segments, about 115.
Setz: Ventral gap equal in length to a lateral gap.
Pigment present.
Nephridiopores in the direction of the outer seta rows.
Male pores between the outer and inner seta rows.
Oviducal pores in the direction of Seta 2.
Spermathecopores in the direction of Seta 3.
Dorsal pores said to be present.
Septa v-vi—vill-1x much thickened. a
Gilzzard in three lobes, occupying Segments x1lI—xv.
Prostates pyriform.
Ovaries probably not enclosed.
Habitat—Burmah (Chiala), at an elevation of about
4500—5500 feet.
M. japonicus, Michaelsen, 1892.
I have not seen the account of this worm.
M. bahamensis, Beddard (18), 1892.
Size, 25 mm. long.
Sete: Judging from the figure, the lateral gaps are larger
than the ventral gap; the sete are absent from Segment 11.
Male pores between the outer and inner seta rows.
Spermathecopores between the outer and inner seta rows.
Septa v-vi—vi11-1x thick.
Gizzard in three lobes, occupying Segments x11I—xv.
Prostates glandular in appearance.
Spermathece pyriform.
Copulatory pouches simple.
Habitat.—Bahamas.
MONILIGASTER GRANDIS, A. G. B. 375
Remarks upon previously described Species.
M. houteni differs from all my species in the long tubular
prostate, about which there can be no mistake ; it is evidently
more like one of the prostates of Desmogaster. With regard
to the numbering of the segments I say nothing.
M. barwelli has been described piecemeal, and even now I
am unable to be quite sure whether it is the same as M.
minuta. I presume not, as the spermathecopores are in a
different position.
M. beddardi is obviously not M. parva, and it could be
no other known species.
M. deshayesi is lost in obscurity, and unless by means of
the type specimen there will be little chance of identifying it.
1 have found one species of Moniligaster in Ceylon among some
thirty species of other genera, so that Moniligaster is not the
dominant genus in Ceylon, even in the hills, that it is in
S. India (at any rate on the Nilgiris). My notes of this
Ceylon specimen are insufficient. I refrain, therefore, from
naming it at present, and give an external figure only, for future
reference. Its colouring distinguishes it from all my other
species. The gizzard occupies Segments xv—xvitt. I obtained
one specimen only at Kandy. ao ets
M. bahamensis is at present very insufficiently charac-
terised. I have gathered most of the characters mentioned
above from the figures. :
BIBLIOGRAPHY.
The Moniligastride.
1. PERRIER.—“ Recherches pour servir a Vhistoire des Lombriciens ter-
restres,” ‘ Nouvelles Archives du Muséum d’Histoire Naturelle de
Paris,’ t. viii, 1872.
2. Bepparp.—“ Notes on some Karthworms from Ceylon and the Philip-
pine Islands, including a Description of Two New Species,” ‘ Annals
and Magazine of Natural History,’ ser. 5, vol. xvii, 1886.
376 ALFRED GIBBS BOURNE.
14
. Bourne.—“ Preliminary Notice of Earthworms from the Nilgiris and
Shevarays,” ‘ Proceedings of the Zoological Society,’ 1856.
. Horst.—“ Descriptions of Earthworms,” ‘Notes from the Leyde
Museun,’ vol. ix, 1887.
. Bepparp.—“ Note on the Reproductive Organs of Moniligaster,”
* Zoologischen Anzeiger,’ 1887.
. Bepparp.—“On the So-called Prostate Glands of the Oligochzta,”
‘Zoologischen Anzeiger,’ 1887.
. BEpparp.—“ On the Structure of Three New Species of Earthworms,
with Remarks on Certain Points in the Morphology of the Oligo-
cheta,” ‘Quart. Journ. Micr. Sci.,’ vol. xxix, 1888.
. Rosa.— Nuova classificazione dei Terricoli,’’ ‘ Bolletino dei Musei di
Zoologica ed Anatomia Comparata di Torino,’ vol. iii, 1888.
. Bepparp.—‘“‘ Preliminary Notes on some Oligocheta,” ‘ Zoologischen
Anzeiger,’ 1889.
. Bennam.—“ ‘Atrium’ or ‘ Prostate’? ” ‘ Zoologischen Anzeiger,’ 1890.
. Bennam.— An Attempt to Classify Earthworms,” ‘ Quart. Journ. Micr.
Sci.,’ vol. xxxi, 1890.
. Rosa.—* Moniligastridi, Geoscolecidi, ed Eudrilidi,” ‘ Annali del Museo
Civico di Storia Naturale di Genova,’ ser. 2, vol. ix (xxix), 1890.
. Bepparp.—“ On some New Species of Earthworms from various Parts
of the World,” ‘ Proceedings of the Zoological Society,’ 1892.
. Bennam.— Description of a New Species of Moniligaster from India,”
‘Quart. Journ. Micr. Sci.,’ vol. xxxiv, 1893.
Also which I have not seen—
15
16
. Bepparp.—“ Observations upon the Structure of a Genus of Oligocheta
belonging to the Limicoline Section,” ‘Transactions of the Royal
Society of Edinburgh,’ vol. xxxvi.
. Micusartsen.— Terricolen d. Berliner Zool. Sammlung,” ‘Arch. fiir
Naturgesch.,’ 1892,
MONILIGASTER GRANDIS, A. G. B. 377
EXPLANATION OF PLATES 22—28,
Illustrating Professor A. G. Bourne’s memoir “ On Monili-
gaster grandis, A. G. B., from the Nilgiris, S. India ;
together with descriptions of other species of the genus
Moniligaster.”
PLATE 22.
Fie. 1.—Moniligaster grandis. The entire worm drawn from life,
natural size. This was the largest specimen I ever found. The clitellum
extending over Segments X, XI, XII, and x111 is much better marked than in
the majority of individuals during life. The tail had been regenerated.
Fic. 2.—Moniligaster grandis. Ventral view of the thirty-six most
anterior segments drawn from life from a smaller specimen. sp. Position of
the spermathecopores. @, 9. Position of the male and oviducal pores.
i. s. Letters with reference lines, inserted to show that the structures so
marked represent inner seta couples.
PLATE 23.
Entire worms drawn from life, natural size. c/. Clitellum.
Fic. 3.—Moniligaster pellucida. The anterior end is shown in the
characteristic form it assumes when contracted.
Fic. 4.—Moniligaster uniqua.
Fic. 5.—Moniligaster chlorina.
Fie. 6.—Moniligaster ophidioides.
Fic. 7.—Moniligaster robusta. From r. onwards the tail had been
regenerated.
Fic. 8.—Moniligaster sapphirinaoides.—A typical specimen from
very wet ground at Naduvatam showing the iridescence.
Fic. 9.—Moniligaster sapphirinaoides-robusta, An example of
a worm with external and internal characters intermediate between those of
the two species, probably a hybrid.
Fic. 10.—Moniligaster sapphirinaoides-ophidioides,—An example
of a worm with external and internal characters intermediate between those
of the two species, probably a hybrid.
Fie. 11.—Moniligaster parva.
Fie. 12.—Moniligaster minuta.
Fie. 13.—Moniligaster rubra.
Ftc. 14,—Moniligaster sy. From Ceylon,
378 ALFRED GIBBS BOURNE.
PLATE 24.
Moniligaster grandis.
Fic. 15.—Ventral view of the head end. Segments numbered 1—xv.
7. s. Inner setacouples. cl. Clitellum. Sp. apt. Spermathecopore. ¢. Male
pore. 9%. Oviducal pore. The prostomium entirely withdrawn.
Fie. 16.—Lateral view of the head end. Segments numbered 1—x!.
a. & to 2. 11. Nephridiopores. o. s. Outer seta couples. pr. Prostomium
fully extended along with the buccal membrane. Sp. apt. and ¢ as before.
Fic. 17.—Diagrammatic vertical longitudinal section of the anterior Seg-
ments I—xIv. The italic roman numerals ///—XJ/J are placed in the
cavities of the several segments. 7/8—12/13. Septa bounding various seg-
ments. pr. Prostomium. m. Muscular pad in the dorsal wall of the pharynx.
Sal. gl. Salivary glands. Sp. Spermatheca. Cop. Copulatory pouch. Sp. apt.
Spermathecopore. Sp. sac. Sperm-sac. 7. Testis. f. “Ciliated rosette.”
V.d. Seminal duct. @. Male pore with the prostate and atrium. Ov.
Ovary. 9. Oviducal pore with the oviduct. Ovs. Ovisac.
Fic. 18.—Ordinary epidermis. Cz. Cuticle. Zp. Unmodified epidermis
cells. god. Gland-cells.
Fie. 19.—Clitellar epidermis. §. c/. Short club-shaped cells. 7. e/. Long
club-shaped cells.
Fie. 20.—Epidermis cells from the neighbourhood of the genital apertures.
Fic. 21.—Epidermis cells from an intersegmental groove.
Fic. 22.—The gizzard in longitudinal section. The septa bounding Seg-
ments XVIII—xxII are indicated. giz. The annular muscular bands. muse.
The outstanding muscle fibres. Oe. The soft-walled portions between the
annular muscular bands.
Fic, 23.—Surface view of a portion of the gizzard to show the outstanding
muscle fibres. Letters as before.
Fic. 24.—From a longitudinal section of the intestinal wall in the region
of the tubular glands. g/. The glands; the section passes through the aper-
ture of one into the intestine. Cal. ep. Ceelomic epithelium. A/. ep. Intes-
tinal epithelium. V. Blood-vessels. (Leitz, ocular 3, obj. 3, cam. luc.)
Fic. 25.—Longitudinal section at the anal extremity to show the thick-
ened septa. Hp. Epidermis. /. ep. Rectal epithelium. Sept. Septum.
At aa. is the anus.
Fic. 26.—A seta couple. One being a little more bent over than thie
other, appears a little shorter,
MONILIGASTER GRANDIS, A. G. B. 379
PLATE 25,
Fic. 27.—M. grandis. A dissection of the anterior region. The seg-
ments as defined by the intersegmental grooves are marked vi to xiv. Certain
of the segments as defined by the septa are marked rx, x, xz, x1. The cut
edges of certain septa are marked 9/10, 10/11, 11/12. 0. s., 7.s. Outer and
inner seta lines.
Nervous System.—The cerebral ganglion, cesophageal commissures,
the ventral nerve-cord, and some of the nerves are shown. In the cerebral
ganglion the bilobed ganglionic mass is shown surrounded by the sheath.
N. Pro. Prostomial nerves. . N. 1, N. 11, &c. Nerves belonging to the
Segments 1, 11, &c. The five pairs of nerves to the cesophageal wall are
shown. A small portion of the thick septum 7/8 is left undissected away to
show the anterior nerve of Segment vir passing into its substance. The
nerves in most segments are shown only up to the inner seta line where they
pass into the muscular layers, but in Segment 1x and a portion of Segment x
these are dissected away to show the distribution of the nerves. 2. 1, x. 2,
n.3. The three nerves of the segment. ~z. sept. Branches supplying the
septum. .v. Nerve to the viscera of the segment. 2. 2. Nerve joining
a. 3 with the septal nerve of the next following segment. Some portion of
the septa 9/10, 10/11, 11/12 having been left, the nerve-cord disappears for a
time in Segment x to reappear in Segment x1 (as defined by the septa, the
great development of the lumen of Segment x1 throws this region back to
Segment xiv, as defined by the intersegmental groove). The curved arrow
shows where the nerve-cord passes into Segment x11, and then under Seg-
ment x1. As a matter of fact, if Segment x1 was filled with ova, &c., not
only the nerve-cord but the ventral vessel and the cesophagus would not be
seen in this region.
Nephridia.—The position of the external apertures of the nephridia of
Segments 111 to x11 are shown on the right-hand side ; of Segments 1x to x11
on the left-hand side, and marked by a small circle with the numbers 38, 4, 5,
&e., by the side. In the case of the nephridia of Segments x—x11I a small
portion of the excretory duct is drawn to show how the ducts of the nephridia
of Segments x and x1 carry the septum forward to its true position. On the
left-hand side the marks x. 3, 2. 4, 7. 5, &c., enclosed by a line, indicate the
nephridia supplied by the blood-vessels which touch the line in question.
Vascular System.—The exact arrangement of the vessels must be fol-
lowed from the text. Note:—The ventral vessel, V. V.; the branches,
Br. V. T., of the ventro-tegumentary vessels of Segments 1 and 11; the
ventro-tegumentary vessels, V. T., of Segments 111, Iv, and v, the latter
supplying, besides the body-wall, &c., the salivary glands (vessels marked
Sal.) and the nephridia of those segments; the ventro-intestinal vessels
V. 1, the most anterior one in Segment v1; the “ventro”-tegumentary
380 ALFRED GIBBS BOURNE.
vessels, V.T.’, of Segments VI, vit, and vir1, which arise not from the ventral
vessel, but from the hearts of those segments; the ventro-tegumentary
vessel, V. T.”, which arises from the heart of Segment vit and supplies
Segments, Ix, x, part of x1, the nephridia of those segments, the sperm-sacs
(Sem.), the prostate (pr.), and the ovaries (ov.); the ventro-tegumentary
vessel of Segment x (V. T.’”) supplying only the ovary, which thus receives a
double supply ; the ventro-tegumentary of Segment x11, the arrangement of
which may serve as a type for those vessels in the rest of the body. Note
also the hearts (H. vi1.—H. 1x.), shown cut short with some of the vasa-
vasorum upon their walls. Note also the latero-longitudinal and subneural
system, coloured blue. The three big trunks are seen to join together just
behind septum 9/10 where the main trunk (shown cut short) passes upwards
to open into the hearts of Segment rx (ep. fig. 33). The anterior and outer
posterior trunks are seen to collect blood carried outwards by all the branches
of the ventral vessel and hearts (except the hearts themselves) of Segments
1 to rx; the inner posterior trunk is connected with the subneural vessel
(not shown) by vessels marked a and 4; with it is also connected the “ veins ”
of Segments x1 and xu. Except for the connection just mentioned, the
“vein” of Segment x1, really a branch of the subneural, serves as a type of
the arrangement of these vessels in the rest of the body; it is joined at z by
the dorso-tegumentary vessel of the segment (D. T.), and passes on from e¢ to
the subneural.
All the vessels are drawn in as accurately as possible as far as they go.
I have been unable to make quite sure about the ventro-intestinal of Seg-
ments x, x1, and xu. I think there are no such vessels in Segments x and
XI and a pair as usual in Segment xu, but these are not drawn. ‘The
nephridiopores of Segments Ix—xIII are specially inserted on the left-hand
side to show the relation of the blood-vessels to them.
Generative Organs.—These are only indicated to show their relation
to the various segments and septa. 4 and 9. Male and female pores.
Sp. sac. Sperm-sac. pr. Prostate. Ov. Ovary. ovd. Oviduct. ovs. Ovisac.
Sp. Spermatheca. cop. Copulatory pouch dissected out.
Fic. 28.—M. ophidioides. Plan of the blood-vessels in a segment in
the region of the gizzard, giz. D. v. Dorsal vessel. D. Zeg. Dorso-tegu-
mentary vessel; this joins the branch of the subneural at z. D. Jn¢. Dorso-
intestinal vessel ; its branches are seen to emerge from the gizzard wall and
then join together. V. Zeg. Ventro-tegumentary vessel; this passes round
towards the dorsal region, giving off branches to the nephridium, Nep/., and
body-wall; all the blood distributed by it is carried back to the subneural by
a branch of that vessel, Br. S. NV.
Fie. 29.—M. ophidioides. View of a portion of the latero-longitudinal
and subneural system. S. V. Subneural vessel; the most anterior branches
of this to receive dorso-tegumentary vessels are those of Segment xv; the
MONILIGASTER GRANDIS, A. G. B. 381
dorso-tegumentary vessel joins the subneural branch at z. The latero-longi-
tudinals a. 4., a. 6. have come down at a.a. from the head region; all the
blood brought forward by the subneural passes into them; more blood always
passes on the right side than on the left ; all this blood is carried by c.,c., ¢., ¢.
into the hearts of Segments viir and 1x. The subneural itself is seen to
come to an end shortly after giving branches to the latero-longitudinal
vessels. What little blood gets into this anterior prolongation must be
carried backwards to pass into the latero-longitudinal.
Fic. 30.—M. pellucida. View of the supra-intestinal vessel, S. I., and
its branches. The dorsal vessel is not drawn; it has no connection in this
species with the hearts of Segment 1x (H. 1x). The supra-intestinal arises
on the gizzard at 7, and receives vessels from the walls of that organ and
from those of the cesophagus in front, d., d. It also receives all the blood
from the latero-longitudinal and subneural vessels by e.,e. It then bifurcates
and empties its blood into the hearts of Segment 1x. a. Latero-longitudinal
vessel from the head region. 4. Portion of the latero-longitudinal which is
connected with the subneural. c. Vessels from the sperm-sacs. sp. Pros-
tates, &.
PLATE 26.
Fic. 31.—M. grandis. Diagram of the vessels in any Segment from x11
onwards. B.w. Body-wall. Inv. Intestine. On the right hand is shown
the nephridium. 2. The nephridiopore. The sets are shown in their right
position with regard to the gaps. The dorsal blood-vessel, D. V., is shown
with its dorso-tegumentary branch D. T., and its dorso-intestinal branch, D. I.
The former joins the branch of S. N., the subneural, at z. The ventral
vessel, V. V., is shown with its ventro-tegumentary branch, V. T., and its
ventro-intestinal branch, V. I. n., n., n. Nephridial arteries and veins.
Fic. 32.—M. grandis. Diagram of the vessels in Segment vi. The
heart is shown on the left side with the muscular bulb interposed between it
and the ventral vessel, V. V. The ventro-intestinal branch is shown going
to the cesophageal wall. The ventro-tegumentary vessel is shown arising
from the heart just beyond the muscular bulb, giving off a branch to the
heart-wall (this passes on to the wall of the dorsal vessel also) and branches
to the body-wall. The main trunk of the latero-longitudinal and subneural
system, viz. the anterior latero-longitudinal vessel, L. L., is shown in its
proper position, and is seen to receive veins from the cesophageal wall and
the wall of the heart and the body-wall.
Fic. 33.—M., grandis. Diagram of some of the vessels of Segment 1x
to show the latero-longitudinal vessels (each now composed of its three ~
branches, anterior and two posterior, cp. fig. 27) passing downwards and
382 ALFRED GIBBS BOURNE.
opening into the hearts of this segment. Their connection, and that of the
hearts, with the dorsal vessel is also shown.
Fic. 34.—M. grandis. Diagram of the arrangement of the blood-vessels
on one side of a segment. The ventro-tegumentary vessel and the branch of
the subneural are seen playing opposite roles. The dorso-tegumentary is seen
to join the branch of the subneural at z. The meaning of the various arrows
is explained in the text, p. 335.
Fic. 35.—M. sapphirinaoides. Small arteries and veins in the body-
wall seen in a longitudinal section. Sept. Septa. Zp. Epidermis. Cm.
Circular muscular layer. 7. m. and /. m*. Inner and outer longitudinal
muscular layers.
Fic. 36.—M. sapphirinaoides. Portion of the epidermis from the
same section as in the preceding figure. The actual “ capillaries” are seen.
Fic. 37.—M. grandis. Blood-vessels on a portion of the nephridium.
Fic. 38.—M. grandis. Diagram of the arrangement of the blood-vessels
in the intestinal wall. Letters as before. The arrows show the direction
of the blood-flow.
Fie. 39.—M. grandis. Portion of the dorsal vessel cut open in the
dorsal median line to show the apertures of the dorso-tegumentary and dorso-
intestinal vessels. «. Valve preventing backward flow in the dorsal vessel.
y. Valve at the mouth of the dorso-intestinal vessel preventing outward
flow.
Fie. 40.—M. grandis. View of the hinder surface of septum 7/8.
The relative positions at which the cesophagus, Oe., the dorsal vessel, D. Y.
the ventral vessel, V. V., the latero-longitudinal vessels, Z. Z., and the nerve
cord, z., pass through the septum are shown. The spermatheca, Sp., and its
duct are shown. The spermathecal duct penetrates the septum at 2. and
runs in the thickness of the septum, as shown by the dotted lines, to the
copulatory pouch, which is embedded in the thickness of the body-wall.
The sete are inserted, although, of course, they really lie further forward.
Fic. 41.—M. grandis.—Diagram of a section across the body in Segment
vit to show the ‘ mesenteries” marked by thick black lines (each really
double) which join the longitudinally running vessels to the intestinal and
body-walls. Letters as before.
PLATE 27.
Moniligaster grandis.
Fie. 42.—Plan of an entire nephridium of the left-hand series, drawn as
though the worm were opened along the dorsal median line and flattened
out. V. V. Position of the ventral median line, i.s., ¢.s.; 0.8.,0.s Inner
MONILIGASTER GRANDIS, A. G. B. 383
and outer seta lines. The position of the septum bounding anteriorly the
segment which contains this nephridium, at the place where the nephrostomial
duct perforates it, is shown at sepé., so that the vesicle and adjoining portions
of the nephridium are supposed to have been turned forwards. ‘The greater
part of the gland is shown in optical section; but at certain places, left un-
shaded, a view nearer the surface is taken. This will be intelligible on a
comparison with the text. s¢. Nephrostome. sf. a. Nephrostomial duct.
This duct joins the rest of the gland at a. The duct which passes from a to
4 is shown in three places in the surface views, and marked “ad.” The duct
which passes from 4 to c is shown in three places, and marked “dc.” The
duct passing from ¢ to dis marked “cd.” The duct passing from d to e is
marked “de.” The duct passing from e to fis marked “ef” The duct
passing from / to g is marked “fy.” The vesicle duct passing from g to 4, the
point where it opens into the vesicle is marked “gi.” ves. Vesicle; the
muscles shown in a portion of its walls only. y. Sphincter muscle. ez. apt.
Place where the excretory duct penetrates to the body-wall to pass to the
nephridiopore.
Fig. 43.—A nephrostome, fixed and stained, showing forty-five marginal
cells, two nuclei of central cells, and at the margin the flattened nuclei of
celomic epithelium cells.
Fic. 44.—Section through a nephrostome, with a portion of the duct,
showing a nucleus, c, of a central cell and a nucleus, m, of a marginal cell.
Fie. 45.—Portion of a nephridial ductule with undulating bundles.
Fic. 46.—Surface view of a portion of the nephridium in the region “ ad,”
drawn from a freshly mounted preparation.
Fies. 47 and 48.—Transverse sections of the ventral nerve-cord. Fig. 47,
from an interganglionic region; Fig. 48, passing through a ganglion and
through the origin of one of the nerve branches on one side. The two figures
drawn with the camera lucida to the same scale. ep. Celomic epithelium.
ce. t. Connective-tissue sheath. s. Space left owing to the contraction of 4.,
the hyaline sheath, in which are embedded the muscles, musc. g.f. Giant
fibre. 2. 7 Nerve fibres. g/. Neuroglia. gang. Ganglion cell. v. Blood
capillaries.
PLATE 28.
Moniligaster grandis,
Fic. 49.—View of a dissection to show the bands of muscle, ., which are
connected with the body-wall in the neighbourhood of the male pore (? copu-
latory muscles). JV.c. Nerve-cord, with the three nerves, 1, 2, 3, of Segment x
and the first nerve of Segment x1. Septum 9/10 is turned forward, and
Septum 10/11, to which is also attached Septum 11/12 (not shown) is turned
384. ALFRED GIBBS BOURNE.
backward. The excretory ducts of the nephridia of Segments x and x1,
neph. 10, neph. 11, are shown. pr. Prostate. v. d. Sperm duct. 0. s., i. 8.
Outer and inner seta lines.
Fic. 50. Diagrammatic drawing from a thick longitudinal section to show
the relations of the oviducal wall, ovd. Other letters as before.
Fic. 51.—The ovisacs, ovs., and the portion of Segment x1 which contains
the ovaries filled with ova, drawn the natural size.
Fic. 52.—“ Ciliated rosette,’ as seen on opening the sperm-sac. The
portion of the sperm-duct drawn is seen through the wall of the sac.
Fic. 53.—Sperm-duct in transverse section. A loop bound together by
the connective tissue of the ‘‘ mesentery,” the duct cut across twice.
Fic. 54.—Portion of the sperm-duct in longitudinal optical section. c. ¢.
Connective tissue. ep. Epithelium.
Fic. 55.—Longitudinal section passing through the papilla protruded at
the male pore. aé¢. Atrium. pv. Prostate. musc. Muscle covering the pros-
tate. v. ad. Sperm-duct about to open into the atrium.
Fie. 56.—Small portion of the same section more highly magnified, show-
ing the gland cells of the prostate and their ducts.
Fic. 57.—From a section across the ovary, ov., including a portion of the
oviducal wall. This shows the epithelium of the oviduct becoming continuous
at a. with the ccelomic epithelium.
Fics. 58 and 59.—Longitudinal sections from a young individual 140 mm.
long. 7. Testis. f Ciliated rosette. Other letters as before. The copulatory
pouch is not yet developed.
REVIEW OF SPENGEL’S MONOGRAPH ON BALANOGLOSSUS. 385
A Review of Professor Spengel’s Monograph on
Balanoglossus.!
By
E. W. MacBride, B.A.,
Fellow of St. John’s College; Demonstrator in Animal Morphology to the
University of Cambridge.
With Plates 29 and 30.
Tue eighteenth monograph of the Naples series lies at
length before us. As a monument of patient industry and a
mine of anatomical facts it stands in the foremost rank of
zoological treatises. It contains minute descriptions of all the
known species of Balanoglossus, as well as complete dis-
cussions, not only as to the relationship of this interesting
form to various phyla of the animal kingdom, but also as
to the special morphology of every organ in the body of
Balanoglossus.
As the conclusions at which Professor Spengel has arrived
are at variance with the views as to the relationship of the
Enteropneusta which have been widely held in England for
the last ten years, it seemed to me that a short account of the
principal additions to our knowledge contained in the mono-
graph might be useful to zoological students who lack the time
necessary to peruse so huge a tome as the one before us.
I shall append to the account of the new facts brought to
light a short discussion on Professor Spengel’s views as to the
phylogenetic position of the Enteropneusta, as it seems to me
that he has adduced nothing which militates in the slightest
1 «Die Enteropneusten des Golfes von Neapel und der angrenzenden
Meeresabschnitte,” J. W. Spengel, ‘Series of Naples Monographs,’ No. 18.
386 E. W. MACBRIDE.
against the view now so widely accepted as to the Vertebrate
affinities of the group.
Before entering on the new points contained in the mono-
graph it will be necessary to give a rapid sketch of the history
and the extent of our present knowledge of the anatomy and
development of Balanoglossus. Kowalevsky! was the first
who gave a tolerably correct account of the anatomy. Our
knowledge on this point was almost completed by Professor
Spengel’s? preliminary account. in 1884. Between the years
1884 and 1886 Bateson’ contributed a series of papers on the
development and morphology of Balanoglossus, which laid
a solid foundation for the interpretation of the anatomical
facts as indicative of Vertebrate affinities. Quite recently
Morgan‘ has published a paper on the growth and develop-
ment of Tornaria, which entirely confirms Bateson’s account,
making allowance for the difference between direct and free
larval development ; and Marion,’ Schimkewitsch,° and Kohler?
have given during the last few years descriptions of different
species.
Fig. 1, Pl. 29, is a diagram representing a horizontal section
through an adult. We notice that there are three regions to be
distinguished in the body, viz. the przoral lobe or proboscis,
the thick “collar” region, and the trunk. The proboscis
region is largely filled with muscle, but it also contains a small
coelomic space opening on the left by a ciliated pore. The
collar and trunk both contain a pair of such spaces, and the
1 «Anatomie des Balanoglossus delle Chiaje,” par A. Kowalevsky, in
‘Mém. Acad. Imp. Sc. St. Pétersbourg,’ 1866.
2 J. W. Spengel, “Zur Anatomie des Balanoglossus,” ‘Mitt d. Zool.
Station z. Neapel,’ Bd. v, 1884.
3 Bateson, ‘Development of Balanoglossus,” various papers, ‘ Quart.
Journ. Micr. Sci.,’ 1884—1886.
4T. H. Morgan, “Growth and Development of Tornaria,” ‘Journal of
Morph.,’ vol. v, 1891.
5 A. F. Marion, “ Sur deux espéces de Balanoglossus,” ‘ Arch. Zool. Exp.’
(2), Tome iv.
6 Schimkewitsch, ‘“‘ Ueber Balanoglossus Mereschkovski,” ‘ Zool. Anz.,’ 1888.
7 R. Kohler, “Contributions 4 l’Etude des Entéropneustes,” ‘ Inter.
Monatsschrift Anat. Hist.,’ Bd. iii, Heft 4.
REVIEW OF SPENGEL’S MONOGRAPH ON BALANOGLOSSUS. 387
collar ccelom on each side opens by a short ciliated canal into
the outer part of the first gill-slit. Projecting into the pro-
boscis celom from behind we observe a spongy mass of-tissue
with blood-vessels, Bateson’s proboscis gland, which is without
doubt an excretory organ. Posterior to this again, is a vesicle
(Pc.), the sac of the proboscis gland, immediately behind which
we see the heart. This last communicates by a narrow slit with
the great dorsal vessel. The alimentary canal commences just
posterior to the proboscis with a ventral mouth ; its first por-
tion in the collar region constitutes a simple buccal cavity ; the
next part in the anterior part of the trunk bears a series of pairs
of gill-sacs opening to the exterior by pores situated dorso-
laterally. Immediately ventral to these openings we see the
openings of the gonads, the sexual glands in both sexes being
represented by a series of simple or branched sacs opening
directly to the exterior.
Fig. 2 on the same plate represents a diagrammatic sagittal
section of the animal. We notice that the alimentary canal in
the trunk region is differentiated into two tubes—a lower one
called the cesophagus (sensu stricto), and an upper one, the
branchial tube, into which the gill-sacs open. We notice also
that the inner openings of these are elongated vertically, and
almost completely divided by a projection from their dorsal
edge, the tongue-bar ; the outer gill-pores (Brew.), on the con-
trary, are simple circular openings. In the collar region, dorsal
to the blood-vessel, we see a longitudinal band of tissue con-
nected before and behind with an invagination of the ectoderm.
This is the central nervous system, and it is connected with
the dorsal ectoderm by three cellular strands, the dorsal roots.
In its interior are a series of small isolated cavities lined by
columnar cells. In the front part of the collar region a dorsal
diverticulum of the pharynx is seen reaching far forward into the
base of the proboscis, and underlying the heart and sac of the
proboscis gland. Thisis the notochord; it has a cuticular
sheath, which is largely thickened, ventrally giving rise to a
chitinous skeletal rod. A portion of this rod is secreted by the
inner ends of the ectoderm cells on the base of the proboscis ; be-
VOL. 36, PART 3.—NEW SER. DD
388 E. W. MACBRIDE.
tween the ectodermic and notochordal portions a wedge of tissue
projects—really the modified tip of the ventral pocket of the
proboscis celom (Pr. cw. v.) ; Kohler, however, mistook it for
a blood-vessel, and described the skeletal rod as completely
divided into two by vascular tissue. In fig. 2 the great ventral
vessel is seen as wellas thedorsal. Fig. 3 is a transverse section
through the base of the proboscis, showing the mutual relations
of heart, notochord, proboscis gland, and proboscis celom.
Fig. 4is a transverse section through the collar region, the
main point to notice being that the dorsal vessel has on each
side of it a space half filled with muscular tissue—the peri-
hemal cavity (PH.). This, as can be seen in Fig. 1, is a for-
ward diverticulum of the trunk ceelom.
Fig. 5 is a transverse section through the branchial region
of the trunk. The figure-of-8 form of the alimentary canal is
well seen, the upper part being the branchial tube, the lower
the cesophagus. Where the one passes into the other there is
on each side a ridge of thickened epithelium, and in life these
two ridges are closely apposed, entirely separating the two
tubes. The tongue-bar is seen on each side of the branchial
tube in longitudinal section, and is seen to contain a portion
of the trunk celom. On each side of the dorsal surface is seen
a slight projection—the genital wing—in the proximal part of
which the gonad opens on each side. Sections of the dorsal
and ventral vessels are seen, and also of the dorsal and ventral
nerve-cords. The whole ectoderm is ciliated, and has a well-
developed plexus of nerve-fibrils amongst the bases of its cells.
The dorsal and ventral nerve-cords are mere local thickenings
of this plexus: the latter is confined to the trunk region; the
former, however, passes into the posterior end of the central
nervous system through the posterior ectodermic pit. Just
behind the collar a ring of nervous matter puts the dorsal and
ventral nerve-cords in communication with each other. The
nerve-plexus is well developed in the proboscis region, and is
especially thick round the narrow neck by means of which
the proboscis joins the collar, constituting a kind of anterior
nerve-ring.
REVIEW OF SPENGEL’S MONOGRAPH ON BALANOGLOSSUS. 389
The development has been worked out by Bateson as men-
tioned above, the main points observed being the following :—
The egg undergoes total and regular segmentation, and forms
a one-layered blastophere. Regular embolic invagination fol-
lows, giving rise to a gastrula, and the blastopore is situated
posteriorly and slightly dorsally. It closes in the position
about where the future anusis formed. The archenteron gives
rise to an anterior and to two lateral pairs of pouches, as
shown in fig. 6; the former becomes proboscis ceelom, the
latter the collar and trunk ccelomic cavities. The noto-
chord arises partly as an anterior diverticulum of the gut; its
posterior part, however, is constricted off as a dorsal groove of
the buccal cavity. The central nervous system is separated
from the ectoderm by delamination, but is added to anteriorly
and posteriorly by fore and aft invaginations from the anterior
and posterior surfaces of the collar. The species on which
Bateson worked develops up to a comparatively advanced stage
within the egg-membrane, and hence its external larval cha-
racteristics have been very much modified. The ordinary
mode of development seems to be through a most interesting
pelagic larva, Tornaria, the early history of which is unknown.
This larva shows a very close resemblance to the Bipinnaria
larva in the disposition of its cilia and in the possession of a
preoral coelomic vesicle opening by a pore on the left side,
whereas in the position of a well-marked apical plate, con-
nected by a double muscular band with the sides of the ceso-
phagus, it recalls certain features of the trochosphere.
Morgan has made the interesting observation that the central
nervous system in this larva is formed by the invagination of
a strip of ectoderm exactly as in Amphioxus.
The monograph of Professor Spengel commences with a
detailed description of one of the Neapolitan species, which is
chosen as a type. The other species are then described in the
order of their affinity to this type. Next follows a chapter on
the ontogeny. Then the special morphology of each organ is
discussed in detail. To this succeeds a chapter on the general
morphology of the animal, and finally there is a discussion of
390 E. W. MACBRIDE.
the various theories as to the relationships of Balanoglossus
which have been put forward, together with an exposition of
Professor Spengel’s own views on that subject.
Confining ourselves in the first instance to the purely
anatomical portion of the treatise, the main interest which
attaches to Professor Spengel’s work is the proof he has given
as to the extent to which homologous organs may vary within
a small sharply circumscribed group; in other words, it is
especially valuable from a systematic point of view. There
are, however, several points of general interest which he has
brought out which must be first dealt with. These concern
the vascular, muscular, and skeletal systems.
The vascular system is, according to Professor Spengel,
nothing else than a system of communicating spaces and clefts
devoid of any proper wall, the remnants of the embryonic seg-
mentation cavity. Their muscular coat, when they have one, is
derived entirely from the neighbouring organs. Thus the trunk
ceelom supplies longitudinal and circular muscles to the dorsal
and ventral vessels. The main part of the vascular system is
in the form of a very close dermal plexus, and of a similar one
between the splanchnic muscles and gut epithelium. These
plexuses are connected with each other through the dorsal and
ventral trunks by vessels running in the dorsal and ventral
mesenteries. In the trunk the ventral vessel takes up the whole
extent of the mesentery, and is in open connection with the
enteric plexus. The ventral vessel is joined to the dorsal just
behind the collar by a specially wide ring-shaped sinus. The
musculature of the heart is supplied by the ventral wall of the
vesicle, which Bateson called the sac of the proboscis gland.
Spengel has, however, shown that it is an entirely closed vesicle,
having no connection with the proboscis gland whatever. Itis
the same as the so-called “heart,” or pulsating vesicle of the
Tornaria. Spengel calls it ““ Herzblase’’—a term which I have
ventured to translate “ pericardium.” The true heart! is a space
between this vesicle and the notochord ; it communicates by a
1 Spengel calls this space the central blood-space, and refuses the term heart
to it on the ground that it is a mere split, and that its muscles are derived
REVIEW OF SPENGEL’S MONOGRAPH ON BALANOGLOSSUS. 391
very narrow channel with the dorsal vessel, and by several
chinks with the proboscis gland. The structure of the latter is
explained in a beautifully lucid manner ; it consists of nothing
but the posterior epithelial wall of the proboscis cavity, which is
pushed forward into the cavity by the forward growth of the
notochord. It is thrown into folds, to which correspond wide
blood-vessels connected with each other in a net-like manner.
The epithelium covering them consists of large pale cells with
often concretions of yellow granules, probably excretory in
nature. Blood is apparently carried from it by two lateral
vessels, which run along the notochord to the buccal cavity.
Here they run first obliquely and then vertically downward,
and unite underneath the gut to form the ventral plexus
of the collar region, which opens behind into the ventral
vessel.
The blood-supply of the gill-sacs is only part of the enteric
plexus; it is specially richly developed around the inner open-
ing of the gill-sacs, which appear to be their functionally
active part. Special vessels arise from the dorsal vessel (see
fig. 2) to join this network.
The blood is a coagulable fluid, and contains abundance of
those wandering cells or amcebocytes which are found in all
Metazoa, from the medusa to man inclusive. In Ptychodera,
the most differentiated genus, they appear to have given rise
to a continuous endothelium in the larger vessels, but no trace
of this is seen in the more primitive species.
The skeletal system next claims our attention. There is
no cuticle in the Enteropneusta; the entire surface of the skin
(with the exception of discharging gland-cells) is ciliated, and
the same remark applies to the gut-cells. The cells lining the
U-shaped inner openings of the gill-sacs have powerful cilia.
The Enteropneusta, however, unlike all Annelids and Arthro-
pods, possess to a great extent the power of secreting cuticular
or “ ground”’ substance from the inner or basal ends of the cells
composing the various epithelia. Spengel calls this substance
from the pericardium; but, as appears from his own description, similar
remarks apply to the whole of the vascular system,
392 E. W. MACBRIDE,
“ Grenzmembran,” implying that it is only found where two
tissues meet ; but it seems to me that he unnecessarily obscures
its nature by this mode of expression: since he allows that it
is in this “ border membrane” that the blood-vessels are per-
forated, it follows that the excretions of two epithelia are
sometimes separated by a remnant of the blastocele. This
cuticular substance is specially thickened (being deposited in
the form of concentric layers) to form the skeletal rod under-
lying the notochord, and the skeleton supporting the U-shaped
gill-slits. The latter, as being the simpler, will be explained
first. Adjacent gill-sacs are apposed along their inner borders.
This region is called by Spengel a gill septum. The apposed
epithelia give rise to a septal bar; this shows, however, its
double nature by a longitudinal furrow, and by diverging below
into two. The “tongue,” as we have seen, contains a portion
of the ceelom, and hence it gives rise to two bars, which owe:
their existence mainly to the epithelium of the gill-sac, though
the ccelomic mesoblast contributes also to a certain extent.
Each septal bar is connected by adorsal arch with the posterior
bar of the tongue in front, and the anterior one of the tongue
behind, so that the entire gill-skeleton consists of a number of
three-pronged bars, as shown in PI. 29, fig. 13a.
To understand the relations of the skeletal rod underlying
the notochord we must glance for a moment at the structure
of the notochord itself as shown in fig. 2. We can distinguish
in it a neck and a head portion. The former opens into the
buccal cavity behind, and has a comparatively wide lumen
throughout. The latter is almost solid (Bateson stated that
it was quite solid, but Spengel maintains the existence of a
narrow lumen), and runs out under the heart, pushing the
epithelium of the proboscis ccelom before it, thus giving rise to
two ventral pockets of this cavity (Pr. v. m.) separated by a
septum. Behind, where the head joins the neck, it gives rise to
a ventral sac with a widish lumen. Now the main portion of
the skeletal rod is a cuticular excretion of the cells of the
ventral side of the neck and the posterior aspect of the ventral
diverticulum of the head ; the ectoderm of the ventral side of
REVIEW OF SPENGEL’S MONOGRAPH ON BALANOGLOSSUS. 393
the proboscis stalk, however, also gives rise to skeletal matter,
which behind fuses with the notochordal excretion, but in front
is separated from it by a wedge of cells, the solid tips of the
ventral pockets of the proboscis cavity. This ectodermal rod is
called by Spengel “the tooth of the skeleton” (PI. 30, fig. 9 7.).
The notochordal skeleton is thinnest in front, where it supports
the ventral wall of the ventral sac of the notochord, and thickest
under the neck portion. Where the neck passes into the
buccal cavity the skeletal rod splits into two rods, which pass
to the right and left of the buccal cavity embedded in slight
lateral grooves in the wall of the same, to whose cells they owe
their existence. Now Bateson maintained that the hinder
part of the notochord was pinched off from the gut-wall by
lateral grooves, and Spengel has been able to supply a most
interesting confirmation of thisidea. If we cut sections of the
hinder part of the undivided skeletal rod, we find that whereas
the younger layers are arranged concentrically round a single
centre, the older layers form two systems of concentric curves
round two separate centres to the right and left respectively
of the new centre. Hence where now the skeleton is a single
rod secreted by the ventral wall of the notochord, it used to
be in the form of two rods lying in grooves of the lateral walls
of the buccal cavity, but these grooves have coalesced and
separated off the notochord. Pl. 30, figs. 8—13, show the
form of the collar skeleton in several of the species.
In addition to this primary skeleton there is developed in
all species, though much more in some than others, a supple-
mental skeleton, the chondroid tissue. It is the special
merit of Marion to have first called attention to this tissue.
It consists of a structureless ground substance with nume-
rous groups and strings of cells embedded in it, presenting in
transverse section a strong resemblance to cartilage. In lon-
gitudinal section, however, it is seen that these cellular masses
are outgrowths from the walls of the collar coelom, and in some
cases also from the wall of the dorsal pockets of the proboscis
cavity (i.e. the portion of the proboscis cavity which ends in
the pore and the corresponding blind portion on the right).
394 E. W. MACBRIDE.
In nearly all cases there is a small portion of this tissue also
derived from the ventral pocket; this projects, as we have
seen, into the niche left between the tooth and body of
the primary skeleton.
Professor Spengel gives a detailed description of the mus-
cular system. It gives characters which are of great use from
a systematic point of view ; but here we need only notice a point
or two of general interest. There is no dermis of con-
nective tissue; immediately beneath the epidermis is a
basal membrane of cuticular substance, and this serves for
origin and insertion of the muscles.
Fig. 8 (Pl. 29) shows the arrangement of the longitudinal
muscles in the proboscis. As there is an outer circular layer,
it follows that the longitudinal fibres must insinuate themselves
between the fibres of this to be inserted in the basal membrane
of the ectoderm. Besides these fibres there is a median plate
of dorso-ventral muscles, which make up the main mass of the
median septum separating the two ventral pockets of the pro-
boscis cavity. These do not cease, however, at the dorsal
edge of this septum, but are continued right to the mid-dorsal
line. The interesting point, however, is that the antero-
posterior extent of this plate is correlated to the length of the
head of the notochord ; hence the notochord in front of the
skeletal rod appears to have a supporting function, and many
of these fibres are, in fact, inserted in its basal membrane.
Fig. 9 shows the relation of the muscles of the anterior face
of the collar region to the skeletal rod. Spengel has made
some very interesting observations on the habits of Balano-
glossus. From these it appears that the animal progresses by
burying itself in the sand and passing the sand through its
alimentary canal, and hence that locomotion and nutrition are
effected by one and the same process. It further appears that
it is the proboscis and collar which are the active agents in
locomotion, the rest of the body being dragged passively
behind. Hence we can understand why the muscular, nervous,
and skeletal systems should be so highly developed in the an-
terior part of the body, and also why the animal should be
REVIEW OF SPENGEL’S MONOGRAPH ON BALANOGLOSSUS. 395
able to support with so little inconvenience serious truncations
ofits posteriorend. A further interesting point is that neither
proboscis nor collar is able to function unless it is turgid,
and that it is impossible for it to attain this turgidity even
when placed on a moist substratum if it is out of water.
Hence it is probable that water is taken in through the pro-
boscis and collar pores to set up this turgidity. This strik-
ingly recalls a point in Echinoderm physiology. The only
function which it is possible now to ascribe to the stone canal
and madreporic pores, in view of the fact that the current in
them is inwards,! is the maintenance of the turgidity of the
water-vascular system.
The fact that the currents through the proboscis pore and
the collar pores probably set inwards is not, as Professor
Spengel points out, irreconcilable with the fact observed by
Bateson, that carmine introduced into the proboscis cavity is
ejected by the pore, since the contraction of the muscular wall
on the fluid contents must necessarily cause ejection of the
latter, together with any particles of an excretory nature which
they may contain.
Fig. 9 shows us how the muscles of the anterior face of collar
radiate from the skeletal rod, the more ventral fibres forming
loops. The somatic wall of the collar cavity has an outer layer
of circular muscles only on its anterior half; its muscles are
mainly longitudinal. The splanchnic wall has an inner layer
of longitudinal muscles. In the more primitive species these
are gathered in front into a single bundle and inserted in the
divergent crura of the skeletal rod; in other species only the
more dorsal have this insertion. The circular muscles of the
splanchnic wall are derived from a forward diverti-
culum of the trunk celom, which Spengel calls the peri-
pharyngeal cavity. There is one on each side of the buccal
cavity, and their relations to each other and to the perihemal
cavities are shown in figs. 4 and 10 (P. PA).
The chief point to notice in the trunk muscular system is
1 Ludwig, H., ‘Uber die Function der Madreporenplatte der Echino-
dermen,” ‘ Zool. Anzeiger,’ No. 339, 1890.
396 E. W. MACBRIDE.
the interruption of the somatic muscular system along the
mid-dorsal and ventral lines and two dorso-lateral lines, the
** submedian” lines in which the gonads open, and in three
of the four genera also the gill-sacs; in Ptychodera, however,
the outer gill-pores are situated dorsally to-this line.
Both trunk and collar ceeloms are traversed by radiating mus-
cular fibres, and in the trunk both dorsal and ventral mesen-
teries persist, though in the collar they are partially absorbed.
As has been mentioned above, a specialised dermis is absent,
but there is a framework of connective-tissue cells ramifying
amongst the proboscis muscles, and clothing the radiating
muscles of the collar and trunk.
CLASSIFICATION OF THE SPECIES.
Pl. 30, figs. 2—7 are intended to give some idea of the range
of variation in external form found in the species of Entero-
pneusta. The main points to be noticed are the genital ridges
of Ptychodera, small and inconspicuous in P. minuta, large
and hiding the whole branchial region in P. erythrza, the
liver saccules of Ptychodera and Schizocardium, the excessively
short collar region of Glandiceps, and the immensely long
proboscis of Balanoglossus Kowalevskii.
Professor Spengel has described in all nineteen species, nine
of which are new. As, however, eight of these are founded on
the examination of single, more or less mutilated specimens,
their specific value must remain for the present somewhat
doubtful. The Enteropneusta have an almost cosmopolitan
distribution. Specimens have been found in the Mediterra-
nean, Atlantic shores of France, Channel Islands, Orkney
Islands, White Sea, Japanese coast, eastern coast of United
States, coast of Brazil, and coast of Peru.
Spengel divides them into four genera, viz. Ptychodera,
Schizocardium, Glandiceps, and Balanoglossus, all of which
appear to me to be well founded.
Ptychodera includes the oldest known Neapolitan species.
REVIEW OF SPENGEL’S MONOGRAPH ON BALANOGLOSSUS. 397
It is the most differentiated genus, and has the following
characters : .
The proboscis has the typical form, being about one and a
half times as long as broad; the collar is as long asitis
broad, and the crura of the skeleton only reach
through its anterior half.
The tongue-bars of the inner gill openings are
bound to the sides of the sacs by synapticula, as shown
in Pl. 29, fig. 13a. We see that whereas in Amphioxus the
synapticula run straight across the gill-slit, and are, according
to Professor Spengel, secondarily fused with the tongue-bar, in
Ptychodera they are at different levels on opposite sides of the
tongue-bar.
The body in the branchio-genital region is produced into
lateral flaps (Pl. 30, figs. 2 and 3), the genital wings,
which contain a large portion of the gonads.
The digestive region of the gutis produced into
papilla-like diverticula, which give rise to protrusions of
the dorsal surface of the animal. These liver saccules are
distinguished by the green matter which their cells contain ;
they open by narrow slits into the gut, and never contain food
material.
In the branchio-genital region behind the gill-sacs we find
the trunk celom divided into dorsal and ventral
chambers by oblique septa. Each chamber contains a fork
of the bifid gonad (see Pl. 29, fig. lla). These bifid gonads
open along the line of insertion of this lateral septum. If we
trace the septum backwards, it ends in a free edge at the com-
mencement of the liver region; traced forwards towards the gill
region, it becomes inserted on the skin at both ends; in other
words, the dorsal chamber becomes displaced from the middle
line by a median upgrowth of the ventral chamber, as shown
in fig. 114. In the gill region the dorsal chamber becomes a
small space containing no part of the gonad, and half choked
with muscular and connective tissue (Fig. 5). From the fact
that the dorsal chamber contains no radiating muscles, and that
the lateral septum ends in free edge behind, but fuses com-
398 _-E. W. MACBRIDE.
pletely with the skin in front, Professor Spengel concludes
that the dorsal chamber is a forward diverticulum of the trunk
ceelom behind the branchio-genital region.
The csophagus is constricted into two tubes—a
brauchial tube above and a true esophagus below—
which are separated from each other by the apposition
of two thickened ridges (fig. 14 d). The intestinal region
of gut (behind the liver region) shows a lateral paired
or unpaired groove with strong cilia. There is an
external circular layer of muscles throughout the
whole trunk, but the circular muscles of the proboscis are
very thin.
The peripharyngeal cavities are of great extent,
meeting each other inthe mid-dorsal and mid-ventral
lines; and since it is in the anterior walls of these cavities
that the efferent vessels of the glomerulus run, it follows that
in this genus they have a vertical course, and unite with each
other to form aventral collar plexus before opening
into the ventral vessel of the trunk.
-The central nervous system is connected with the dorsal
epidermis by several dorsal roots. These latter have a
sheath of nerve-fibrils, and hence there is a mantle of white
matter on the dorsal as well as on the ventral and lateral
surfaces of the cord. This genus contains nine species; the
peculiarities of the more minutely investigated are mentioned
below.
Pt. minuta (Plate 30, fig. 2) has very feeble genital ridges,
into which the gonads send only rudimentary projections ;
8—9 cm. long. Its collar skeleton is shown in fig. 8, and its
central nervous system has the structure described in the
introduction.
Pt. sarniensis, very similar, but much larger, about 50 cm.
long. Central nervous system traversed by a single canal
throughout, which, however, does not open into the posterior
ectodermic sac.
Pt. aperta has great genital wings starting from imme-
diately behind the collar, but not approaching each other; the
REVIEW OF SPENGEL’S MONOGRAPH ON BALANOGLOSSUS. 399
gonads send distinct branches into them. Central nervous
system as in Pt, sarniensis.
Pt. clavigera, very large genital wings converging in front,
so as to cover front part of gill region ; the gonad branch which
enters them is so large as to dwarf the ventral fork of the bifid
gland from which it arises; central nervous system as in
Pt. minuta,
Pt. aurantiaca, very like foregoing, but the still stronger
genital wings contain, besides the great branch from the
ventral fork of primary gonad, several longitudinal rows of
accessory gonads with independent openings. Fig. 10 shows
the form of its collar skeleton.
Pt. erythrza has immense genital wings, which completely
hide the branchio-genital region, and extend in lesser develop-
ment into front part of liver region. They, however, are
not exactly equivalent to the genital wings of other
species, since the lateral septa are inserted near their free
edges. Hence they correspond chiefly to a comparatively
narrow strip at the proximal end of the wings in other species.
The outer gill-pores are elongated into vertical slits,
and the ventral pocket of the proboscis cavity com-
pletely hollows out the “tooth” ofthe skeleton, con-
verting it into a thin-walled sac.
Pt. bahamensis very closely resembles the foregoing
species, but is very much smaller.
Spengel is inclined to divide this genus into three sub-
genera, viz. Ptychodera (sensu stricto), including Pt.
minuta and sarmensis; Tauroglossus, including Pt. aperta,
clavigera, aurantiaca; and Chlamydothorax, including
Pt. erythea and bahamensis.
The genus Schizocardium (fig. 7) was erected by Spengel
for the reception of a common species uf the Gulf of Rio de
Janeiro, of which he obtained numerous specimens and made
a very complete study. A second species of the same genus is
founded on a fragment dredged up off the coast of Peru. It
agrees with Ptychodera in possessing synapticula
inthe gill-skeleton, and liver saccules, but differs in
400 E. W. MACBRIDE.
the entire absence of the lateral septum in the
genital region,in the absence of genital wings, and
especially in the structure of the central organs of
the proboscis. '
The notochord is prolonged as a narrow solid cord
of cells almost to the anterior end of the proboscis; to this
corresponds a prolongation of the plate of dorso-ventral
muscles. The pericardium and heart are each pro-
longed into two great horns, which run alongside the
notochord. It is from this circumstance that the name of
the genus is derived. The mutual relation of these organs is
shown by the series of sections represented in Pl. 29, figs.
12 a—e. The pericardium has lateral as well as ventral
muscles, and the circular muscles of the proboscis are very
strong.
The collar region has the skeletal rods reaching through
two thirds of its length. ‘There is a large mass of chondroid
tissue easily distinguishable from the primary skeleton, and for
this reason the collar skeleton (Pl. 30, figs. 11 and 13) is of
clumsy build.
The central nervous system has no dorsal roots.
The peripharyngeal cavities are confined to the
lateral walls of the buccal cavity; hence their muscula-
ture is only dorso-ventral instead of circular: it is completed
above by transverse muscles belonging to the ventral wall of
the perihemal cavities, and below by muscles derived from
the collar celom; hence also there is noventral collar
blood-plexus, since the peripharyngeal cavities do not meet
ventrally.
The inner gill-pores are so elongated as almost to meet in
the mid-ventral line, leaving between them only (Pl. 29,
fig. 14c) a strip of hypobranchial epithelium. Though there
are no genital wings like those of Ptychodera, yet the body is
much broadened in the branchio-genital region, and these
lateral portions contain several rows of gonads.
In the genital region there are a varying number of
round pores opening from the gut to the exterior
REVIEW OF SPENGEL’S MONOGRAPH ON BALANOGLOSSUS. 401
somewhat dorsal to the submedian line—the line in
which the gill-pores and gonads open in this genus. Of these,
some just behind the gills are asymmetrically disposed ; whilst
separated from them by an interval, and placed just in front of
the liver region are a pair.
There is no circular muscular layer in the trunk;
it is replaced by a system of oblique dorso-ventral muscles
arising from the dorsal mesentery, and ventro-dorsal arising
from the ventral mesentery.
The genus Glandiceps is closely allied to Schizocardium.
It includes two species—one from the deeper zone of the
Mediterranean, another from the Bay of Tokio, in Japan ; of
the latter alone the whole body is known, and it will be taken
as type.
Glandiceps differs from Schizocardium in the absence of
synapticula in the gill-skeleton, and of liver saccules ;
also the pericardium has only two rudimentary ante-
rior horns.
The proboscis resembles that of Glandiceps in the
notochord and muscles. The collar region is extra-
ordinarily short (PI. 30, fig. 4), and the skeletal rods
reach right to the end. The peripharyngeal cavities are
entirely absent. The central nervous system has no roots.
The inner gill-pores are not so long as in Schizocardium
(Pl. 29, fig. 14.6). The genital region resembles that of Schizo-
cardium ; there are, however, additional gonads dorsal
to gill-sacs, opening in the submedian line.
In the body musculature and gut-pores Glandiceps resembles
Schizocardium. The chondroid tissue of the skeleton
reaches its maximum of development in this genus;
though there are no liver saccules, a liver region of the gut
can be recognised by its histological characters, and in this
region there is a dorsal siphon of non-glandular cells
opening into the intestine before and behind.
The genus Balanoglossus includes several species differ-
ing widely from each other, but agreeing in the follow-
ing characters :—The central organs of the proboscis are like
402 EK. W. MACBRIDE.
those of Ptychodera. The collar region is moderately
long, and the skeletal rods reach right through it.
They are consequently very long (Pl. 30, figs. 9 and 12). The
inner gill-pores only reach halfway down the side of the
cesophagus (fig. 14a), but the lower nutritive portion is not
constricted from the upper branchial. There are no synap-
ticula in the gill-skeleton, and no liver saccules.
There is no circular muscular layer in the trunk, nor any
system of muscles replacing it. The chondroid tissue
is very feebly developed. The central nervous system has
no dorsal roots.
B. Kupfferi (fig. 7) is the North Sea species. Its proboscis
has the normal length. It has two proboscis pores
opening into the right and left dorsal pockets of
the proboscis celom. (In Ptychodera minuta this
condition occurs as a variation; we also find one median pore
opening into both pockets.) The head of the notochord is
short; the neck has two lateral wings, giving rise to corre-
sponding projections of the skeleton (Pl. 30, fig.9 W). In full-
grown animals the hinder part of the neck is broken
up by invading portions of the skeleton. There are
no peripharyngeal spaces, but there is a layer of circular
muscles suspended in the midst of the collar cavity
by the radiating muscles. There are, as in Glandiceps, additional
dorsal gonads, and the eggs are large and yolky. There isa
continuous blood-sinus round the whole gut behind the collar.
B. canadensis has a short round proboscis, in which the
cavity is almost obsolete. The longitudinal muscles are
represented by a septum of radiating muscles with interposed
sheets of connective tissue. The head of the notochord has a
wide lumen and a strong cuticular membrane, into
which these muscles are inserted. The neck and the
corresponding portion of the skeleton as well asits
diverging cruraareabsent. There are two proboscis
pores, but they are very narrow,and do not appear to
reach the cavity. A sense-organ in the shape of a pit in
the mid-ventral line of the proboscis is present.
REVIEW OF SPENGEL’S MONOGRAPH ON BALANOGLOSSUS. 403
There are no peripharyngeal spaces. There are four
genital wings—a dorsal and a ventral pair—each with
several rows of gonads,and they do not extend behind
the gill region. The outer gill-pores are elongated to form
slits extending between the dorsal and ventral wings.
B. Kowalevskii, the species the development of which has
been worked out by Bateson, adheres very closely to the
normal type. Its chief peculiarity is the enormous length of
the proboscis (see fig. 6). The longitudinal muscles in this
part are arranged in regular concentric layers. The notochord
has a rather long head. Peripharyngeal spaces are
present, and are obviously diverticula of the trunk
celom, their walls containing not only circular but also
longitudinal muscles. The lateral walls of collar are
prolonged so as to cover three or four gill-slits (the
so-called atrial folds). The central nervous system
has no dorsal roots, but is connected with the dorsal
epidermis by a longitudinal ridge of cells, such as is
found in the young Ptychodera minuta. There are
no dorsal gonads, but there are gut-pores, though only the
more posterior paired pores are represented.
B. Mereschkovski is, as far as I can see, quite identical
with B. Kowalevskii; B. sulcata has a similar proboscis,
but its anatomy has not been investigated.
Balanoglossus Kowalevskii is in many respects the
central form of the whole group; from it to Glandiceps the
transition is not difficult, to which Schizocardium is also closely
related. Unless we suppose synapticula to have been twice
developed we must assume that the central organs of the
proboscis of Ptychodera have undergone reduction, since the
synapticula have evidently arisen between Glandiceps and
Schizocardium.
Ontogeny.
The chapter on Ontogeny is exceedingly disappointing.
Professor Spengel’s observations have been confined to
Tornariz caught with the surface-net, and his methods appear
to have been of a very crude nature. Though he has observed
VoL. 36, PART 3.—NEW SER. EE
4.04. E. W. MACBRIDE.
the metamorphosis of the larva into the young worm, he has
not seen the development of the notochord or central nervous
system. His youngest larve had a complete gut with mouth
and anus, and the proboscis celom opening already by its pore.
He has shown that the muscle connecting the apical plate or
brain with the proboscis cavity is really double, and that its
sheath is continuous with the peritoneum of the proboscis
cavity. Morgan,! in his last paper, which has appeared since
the publication of the monograph, confirms this, and goes
further ; he finds that the anterior unpaired part of the muscle
is an anterior prolongation of the proboscis cavity, and that
the paired portions are formed from its lateral walls, and
hence these muscles are strictly comparable to the
muscles of the preoral lobe of Asterid larve. Thus
one of the few resemblances to the trochosphere which
Tornaria was supposed to exhibit is cancelled. The only im-
portant organ, the origin of which Spengel has satisfactorily
determined, is the trunk ceelom, and here his results confirm
those of Bateson. In reading his discussion of the morphology
of Tornaria he gives one the impression that he has seen the
origin of the collar celom. This is not so, as can be seen
by a reference to the chapter on Ontogeny. He assumes
that it is segmented off from the front end of the trunk ccelom,
a supposition which is in contradiction to the results of both
Bateson and Morgan.
He believes that he has seen the collar pore arising as an
outgrowth of the first gill-sac. This view of its origin, which
is maintained in opposition to that of Bateson, who regarded
it as arising as a thickening of the skin on the posterior free
edge of the collar, is founded on the examination of one
specimen mounted whole. In his last paper Morgan
figures sections showing that Bateson’s view is perfectly right.
He tells us that there is an ectodermic invagination giving rise
to the collar pore at its upper end, and that both the first
and the second gill-sacs open into it in the species he
1 T. H. Morgan, ‘‘ The Development of Balanoglossus,” ‘Journal of Mor-
phology,’ vol. ix.
REVIEW OF SPENGEL’S MONOGRAPH ON BALANOGLOSSUS. 405
examined. Spengel states also that he has proved that the
pericardium originates from the dorsal ectoderm. Here again,
however, his main proof is one thick section, and again his
results have been contradicted by Morgan. It throws con-
siderable light on the prejudice with which Spengel has
approached the discussion of the results obtained by other
workers, that on this meagre evidence, and the assumption that
the intestine is a proctodeum, he should maintain as probable,
not only the ectodermic origin of the pericardium, but also
of all the other celomic cavities in the animal. His theory
of their method of formation is illustrated in Pl. 29, fig. 7. He
assumes because they are ciliated and separated by constric-
tions from the stomach that both cesophagus and intestine are
of ectodermal origin. The trunk ccelom arises from the latter,
and he supposes the collar celom to be segmented from it—a
process which he compares to the segmentation of the meso-
dermic band of an Annelid. He supposes the proboscis cavity
to arise as an ectodermal ingrowth, and that the pericardium is
its fellow, in spite of the fact that the probocis cavity is (apart
from the pore) a bilaterally symmetrical structure, and that
the pericardium is a median one.
Bateson, without definitely committing himself, suggested
as probable that the gonads were derived from the ectoderm.
Spengel, by the examination of young adults, has shown that
their connection with the ectoderm is secondary, and supposes
that they are derived from mesenchyme cells. Morgan has
proved them to originate, as in all other Ceelomata, from the
peritoneum.
Affinities of Enteropneusta with Chordata.
We have now reached the section of the monograph dealing
with Professor Spengel’s views on the relationships of the En-
teropneusta and the morphology oftheir organs. These views
may be briefly summarised thus: he denies utterly their rela-
tionship to the Chordata, and suggests, instead, distant Annelid
affinities.
I shall briefly recapitulate the arguments for the Chordate
406 E. W. MACBRIDE.
affinities of Balanoglossus as put forward by Bateson, and then
examine how Professor Spengel meets them. The Chordata
are absolutely distinguished from all other Metazoa by three
characteristics: (1) they have a dorsal central nervous system,
separated from the skin by invagination and containing a
central canal; (2) the anterior portion of the alimentary canal
has paired gill-slits; (3) they have an endodermic rod, which
has acquired a supporting function, and which is
separated from the mid-dorsal gut-wall. All these struc-
tures are found in Balanoglossus;! they have been
sought elsewhere throughout the whole animal kingdom in
vain. In addition to this, certain structures found in
Amphioxus alone are compared to some of the peculiar
structures of Balanoglossus. These are (1) the paired ante-
rior enteric diverticula of Amphioxus, the left of which opens
by a pore to the exterior, compared to the proboscis and its
pore in Balanoglossus; (2) the covering of the gill-slits by a
pair of atrial folds, from the walls of which a pair of excretory
pores (the atrio-co:lomic funnels of Lankester) are developed,
compared to the “atria” of B. Kowalevskii and the collar
pores; (3) the secondary gill-bars, compared to the tongue-
bars of Balanoglossus ; (4) the mouth adapted for digging in
both instances. Whether any or all of this latter series
of comparisons are valid is, however, of minor im-
portance; the theory of the Chordate affinities of the
Enteropneusta stands or falls with the validity of the
comparison between nerve-cord, notochord, and gill-
slits of Amphioxus, and the structures bearing the
same name in Balanoglossus.
With regard to the central nervous system, Professor
Spengel maintains that it is only the part of the dorsal nerve-
cord which belongs to the collar region; that it is found in the
Tornaria as merely a part of this, the ventral nerve-cord being
developed at the same time; and that the process of invagina-
1 Throughout this discussion I use the name Balanoglossus as a general
term, applicable to all the species of Enteropneusta, and do not in any way
restrict its application to Spengel’s genus so named.
REVIEW OF SPENGEL’S MONOGRAPH ON BALANOGLOSSUS. 407
tion has nothing to do with its formation, but only with its
removal from the surface, since nervous matter is only found
on its ventral and lateral surfaces. In adducing such con-
‘siderations as in any way telling against its similarity to the
nerve-cord of the Chordata, Professor Spengel betrays his hazy
conceptions of Vertebrate embryology. The nerve-cord of
Amphioxus, like that of Balanoglossus, appears as a
median differentiated strip of ectoderm before any
invagination commences; and so marked features do the
ventral position of the white matter and the presence of
numbers of non-nervous cells in the cord form in the develop-
ing lower Vertebrata, that one investigator has put forward
the startling suggestion that the nerve-cord of Vertebrates
represents the alimentary canal of Arthropoda with which the
ventral nerve-cords have fused. That the collar-cord is really
the central nervous system appears from Professor Spengel’s
account of its anatomy ; it is no argument against this to say,
as he does, that though ganglion-cells are more abundant in it,
yet they occur sparsely in the trunk and dorsal and ventral
cords. A similar mode of reasoning would prove that the
nerve-ring and radial cords of Echinoderms are not central
organs.
Professor Spengel’s objections to regarding the organ which
Bateson calls the notochord as homologous to the notochord
of Vertebrata are of two kinds, histological and topographical.
We shall deal with the former first.
Bateson stated that the head of the notochord was solid,
and that its tissue was remarkably like that of the notochord
of a young Elasmobranch. Spengel maintains that there is a
narrow lumen throughout, though sometimes interrupted ;
and though there is a superficial resemblance in transverse
section to notochordal tissue, yet it is only superficial, for the
cells are epithelial cells, modified by the development of huge
vacuoles, whereas he says true notochordal tissue consists of
spherical cells more or less flattened against each
other with vacuoles, and in the chinks between the spheres a
gelatinous substance, This kind of tissue he thinks it im-
408 E. W. MACBRIDE.
possible to conceive of as originating from epithelium. Here
again we are astonished by Professor Spengel’s strange ideas
about Vertebrate ontogeny. In Amphioxus the notochord
is a mid-dorsal strip of hypoblastic epithelium con-
stricted off by the meeting of lateral grooves, exactly
in the same manner as the hinder part of the noto-
chord is formed in Balanoglossus. Further, accord-
ing to Balfour,' the first changes which occur in it are
the production of vacuoles in the cells and the forma-
tion ofacuticular sheath, exactly comparable to the
thickened “border membrane” (Grenzmembran) in
Balanoglossus. Spengel says that there is no evidence of
the formation of skeletal products from this membrane in
Vertebrata. The answer to this is that the greater part of the
thick notochordal sheath in Amia is formed from the thickening
of this membrane, and that, indeed, it is only in virtue of
possessing such a firm membrane that a mass of vacuolated
cells can function as a supporting structure at all. The later
changes which occur in the notochord of Amphioxus, viz. the
interposition of discs of cuticular substance between the cells,
are not without parallel in some species of Enteropneusta. In
Balanoglossus Kupfferi the lumen of the neck is occluded
and its tissue broken into pieces by invading masses of cuticular
substance. We can, I think, go even further, and maintain
that in the chondroid tissue we have the first trace of
amesoblastic sheath. Spengel maintains that this tissue is
fundamentally different from cartilage, since in the latter the
cells are derived from a solid blastema, whereas in the former
we have a number of solid cellular outgrowths from the
celomic wall. This appears to me to be a difference of detail
rather than of principle, more especially as Hatschek! has
shown that the solid sclerotome of the Elasmobranch embryo
is represented by a hollow diverticulum in the larva of Am-
phioxus. Figs. la, 6, and ec in Pl. 30 are designed to show
that the arrangement found in Balanoglossus is to some extent
1 «Text-book of Comparative Embryology,’ vol. ii.
? Hatschek, “ Ueber den Schichtenbau von Amphioxus,” ‘ Anat, Anz.,’ 1888.
REVIEW OF SPENGEL’S MONOGRAPH ON BALANOGLOSSUS. 409
intermediate between those observed in Amphioxus and the
Elasmobranch embryo. Ch!. denotes the primary cuticular
sheath of the notochord, Ch®. the mesoblast concerned in the
formation of chondroid tissue, ‘membrana reuniens,’”’ and
cartilage, in Balanoglossus, Amphioxus, and Scyllium respec-
tively. The actual skeletal substance is indicated by dotting.
We see that whereas in Amphioxus the sclerotome is a single
hollow outgrowth on each side, in Balanoglossus it is repre-
sented by a number of more or less solid outgrowths, and in
Scyllium by a single solid one.
Spengel’s topographical arguments are (1) that the so-
called notochord appears late in development ; (2) that it shows
no relation to the blastopore or its line of closure, such as has
been described in Elasmobranchs; and (8) that it is on the
ventral side of the dorsal vessel. The circumstance that the
notochord and other chordate features appear comparatively
late in the development of the Enteropneusta is one of the
most interesting features of the group. They are, as Professor
Lankester has pointed out, the only members of the chordate
phylum which take us into ‘ prechordate”’ times, and give us
some reliable indication of the direction in which to look for
the ancestry of the Chordata. The reason for the precocious
appearance of notochord and nerve-cord in Vertebrate develop-
ment is the hurrying over of the earliest stages whilst the
embryo is still within the egg-membrane. In the larva of
Balanoglossus Kowalevskii the notochord, nerve-cord,
and the first pair of gill-slits are present at or immediately
after the time of hatching. A precisely similar instance of the
precocious appearance of an organ in embryonic as opposed
to larval development is afforded by the Vertebrate oviduct.
This appears in the chick on the fifth day, but in the frog only
after terrestrial life has commenced and the tail is being
absorbed.
The appearance, in the development_of Elasmobranch and
other heavily yolked eggs, of the formation of the notochord
and nerve-cord by the coalescence of the two lips of a slit,
giving the impression that in ancestral forms hoth were paired
410 E. W. MACBRIDE.
structures placed at the sides of a long slit-like blastopore,
seems to me to be only one of those numerous distortions of
development which yolk produces. Starting from a blasto-
sphere with a differentiated endodermic pole, let us suppose
that those cells which are destined in the gastrula to form the
ventral wall of the gut become swollen up with yolky inclu-
sions. Then the process of invagination will be asymmetrical ;
it will be modified to a growing of the cells which are to
form the dorsal gut-wall over the yolky cells like a lip. A
moment’s thought will convince anyone that this cannot take
place on a spherical surface without a simultaneous coales-
cence of the cells destined to form the right side with those
destined to form the left.
The position of the notochord with reference to the dorsal
vessel presents at first sight a serious obstacle to homologising
it with the notochord of Vertebrates. A little reflection will
convince us, however, that (1) when the Vertebrate notochord
was not as yet fully separated from the gut, it could not
possibly have been, as it is now, dorsal to the aorta; and that
(2) since, as Professor Spengel points out, blood-vessels are
merely blastoccelic spaces, there was very probably at that
period of its evolution a blood-vessel between the notochord
and dorsal ectoderm exactly where we find one in Balanoglossus.
As the notochord became separated from the gut it would press
on and obliterate the vessel above it. This has actually happened
at one point in Balanoglossus, the heart and dorsal vessel being
joined in the proboscis neck by only an exceedingly narrow chink.
As the dorsal vessel disappeared the ventral vessel would gain
in importance, and so constitute the central contractile organ
or heart. The chief point which Spengel adduces in favour of
his theory of Annelid relationship is the structure of the blood
system, and especially the direction of circulation. This latter,
according to an old observation of Kowalevsky’s, which, as far
as I can gather, Professor Spengel has not confirmed, is for-
wards in the dorsal vessel, and backwards in the ventral, thus
agreeing with the Annelid arrangement as against that found
in Vertebrates. ‘This fact Professor Spengel considers as
REVIEW OF SPENGEL’S MONOGRAPH ON BALANOGLOSSUS. 411
finally decisive against an assumption of chordate affinities,
How little weight is really to be attached to it in
the case of the Protochordata can be seen by con-
sidering the case of Tunicata, where the direction of
the blood-flow is periodically reversed. This fact seems to
have escaped Professor Spengel’s memory.
The attempt to prove that the gill-slits of Balanoglossus and
Amphioxus, in spite of their extraordinary similarity, are
morphologically different structures, is exceedingly weak. The
facts adduced against their homology in the two animals, viz.
the more dorsal position of the gill-pores in the Enteropneusta—
the presence of ccelom in the tongue-bar and its absence in the
primary branchial septum in the one case, and the opposite
arrangement in the other—the slightly different arrangements
of the gill-skeleton, are all points of detail.
As a matter of fact, the gillsof Balanoglossus are more
typically Vertebrate in structure than those of
Amphioxus, inasmuch as in the former case, as in
all the higher Vertebrata, we have gill-pouches,
whereas in Amphioxus we have merely slits. These
latter obviously correspond not to the small ex-
ternal gill-pores, but to the internal gill-slits of
Balanoglossus, the outer portions of the gill-sacs
having atrophied in Amphioxus in consequence
of the development of the atrial fold, just as in
Teleostei the development of the operculum has
hadasimilar effect. It is true that in many of the species
of Enteropneusta these inner slits do not as nearly reach the
middle line as they do in Amphioxus, but we trust no serious
morphologist will ask us to consider this an important
difference, more especially as in Schizocardium they nearly
meet in the middle line.
The fundamental plan of the gill-skeleton is the same in
Enteropneusta as in Amphioxus, as it may be considered to
consist in both cases of a series of U-shaped rods. These, in
the latter case, have all united to form a continuous lattice-
work ; the persistence of coelom in the tongue-bars of
412 E. W. MACBRIDE,
Balanoglossus renders this impossible, and so we have a series
of isolated tridents.
- Morgan has, in his last paper, made the exceedingly in-
teresting observation that in the young adult a celomic
space is found in the gill septum in the same
position as it occurs in Amphioxus, viz. on the
outer side of the skeleton ; and that, further, the single skeletal
rod found later here is represented, at this stage, by two
distinct rods. That the synapticula have been independently
acquired in each case I am quite prepared to believe, especially
as they are absent in the more primitive species of Entero-
pneusta; but one of Professor Spengel’s arguments on this
point seems to me to be absurd, viz. that in the one case the
synapticula carry blood-vessels, and in the other they do not.
A similar argument would prove that the interfilamentar
concrescences of the gills of Mytilus had nothing to do with
those in the gills of Anodon.
Professor Spengel’s statement that the synapticula in
Amphioxus are segmented off from the endostyle must be
received with great caution. He gives no adequate proof of
it in his paper! to which he refers, and it involves such
alarming conclusions with regard to the formation of the gill-
slits, that it is safe to maintain a completely sceptical attitude
towards it till further proof is adduced.
The blood-supply of the gills of the Enteropneusta is doubt-
less, as Professor Spengel points out, exceedingly different from
the branchial vessels of a Vertebrate; but its differences
depend entirely on its more primitive and undif-
ferentiated character. It is, in fact, nothing more than a
portion of the enteric plexus; indeed, the whole blood-system
of Balanoglossus is the most undifferentiated one which could
well be imagined, and it is perfectly easy to see how the
Vertebrate or, indeed, any other arrangement could have been
evolved from it.
As every one knows, the alternative theory of Vertebrate
1 J. W. Spengel, “ Beitrig zur Kenntniss der Kiemen des Amphioxus,”
‘ Zool. Jahrbiicher Abt. fiir Anat. und Ont.,’ 1891.
REVIEW OF SPENGEL’S MONOGRAPH ON BALANOGLOSSUS. 418
descent is Dohrn’s famous Annelid theory. The most striking
and, indeed, almost the only point of similarity between
Annelids and Vertebrates is the metameric repetition of many of
their organs which both exhibit. Bateson, in criticising this,
pointed out that whereas in Annelids this metamerism depends
on the repetition of mesoblastic segments, in Vertebrates the
repetitions of certain sets of organs have taken place indepen-
dently of each other ; the series of gill-sacs, for instance, bears
no relation to the series of myomeres. Spengel has made the
daring attempt to explain away this well-known fact. He says
that the gill-slits of Amphioxus are at first in strict correspond-
ence with the myomeres. Considering the fact that this remark
can only by any possibility be applied to the first dozen slits,
and that they even arise long after the complete ventral fusion
of the myomeres, it is difficult to see why any weight should
be attached to a chance correspondence between the first few
slits and the myomeres adjacent to them. Spengel’s only
evidence for this correspondence, such as it is, consists in
some figures of Amphioxus larve in Willey’s paper,! and if
Professor Spengel had taken the trouble to count the slits
and myomeres instead of resting satisfied with a superficial
examination of the figures, he would have seen that the corre-
spondence on which he relies does not exist.”
Having thus seen that no valid objections to the homology
of the nerve-chord, notochord, and gill-slits of Amphioxus with
the similarly named structures in Balanoglossus have been
brought forward, we are the less concerned to defend the other
homologies put forward by Bateson. No one, we suppose, will
deny that the “ tongue-bars” are similar structures in both
animals, The ‘‘ digging” mouth is probably in each case an
? Arthur Willey, ‘“‘ The Later Larval Development of Amphioxus,” ‘ Quart.
Journ, Micr. Sci.,’ 1891.
2 Professor Lankester has kindly looked carefully into this matter for me,
and he writes me that at the time when three gill-slits are present in the
larva if one counts the club-shaped gland as a fourth and makes allowance
for the obliquity of the myotomes, an apparent correspondence exists between
the two sets of organs. Such are the dimensions to which Spengel’s “ strict
metamerism ” of the gill-slits reduce on examination.
414 E. W. MACBRIDE.
independent adaptation to burrowing life, since it seems likely
that the main stem of the chordate phylum retained a pelagic
life; the highly developed sense-organs point in this direction,
As to the homology of the atrial fold in Amphioxus with the
posterior edge of the collar in Enteropneusta, so long
as we are ignorant of the representative of the collar
ceelom in Chordata, it must remain merely a tentative suggestion.
Morgan’s researches: have, however, destroyed one of Professor
Spengel’s main objections to this comparison, viz. his assertion
that the collar pore opens into the first gill-slit. With regard
to the homology of the anterior enteric diverticulum of
Amphioxus with the proboscis cavity, it is difficult to believe
that Professor Spengel is serious when he says that the
proboscis cavity is not asymmetrical, but only the proboscis
pore, and on that ground objects to its comparison with the
preoral ceelom of Amphioxus, more especially as, when speaking
of the Tornaria, he suggests that the proboscis cavity is really a
left-sided structure, the fellow of which is the pericardium.
Affinities of Enteropneusta with Annelida.
As mentioned above, Professor Spengel believes that Balano-
glossus is distantly related to the Annnelida ; and in support of
this view he endeavours to show that the Tornaria is a modified
trochosphere. Now Tornaria shows very great resemblance to
an Echinoderm larva, so that it was for a long time mistaken
for one; the only points in which it agrees with the tropho-
sphere, and differs from all Echinoderm larve, are the possession
of eye-spots on its apical plate, and the strong muscles running
from the plate to the esophagus. These latter, however, since
they are merely differentiations of the przoral celomic wall,
may be compared to the longitudinal muscles in the przoral
lobe of Bipinnaria. Hence in order to torture the Tornaria
into a trochosphere, he finds it necessary to make a series
of violent assumptions, which are not only not supported by ob-
servations, but in direct contradiction to such observations as
we have. In this connection, also, it is impossible to avoid
expressing indignation at the way in which the results of the
REVIEW OF SPENGEL’S MONOGRAPH ON BALANOGLOSsUS. 415
careful investigations of Bateson are set aside when they do
not happen to tally with Professor Spengel’s theories. Bate-
son’s researches established the following points of resemblance
between the development of the larva he investigated and that
of Echinoderms. (1) The blastopore closes in the position of
the futureanus. (2) The entire alimentary canal from mouth
to anus is of endodermic origin. (8) The mesoderm originates
as archenteric diverticula. In the trochosphere, on the other
hand, the blastopore when it persists becomes the mouth, there
is a stomodzeum and a proctodzeum, and the mesoblast originates
by budding from two large cells in the neighbourhood of the
blastopore.
Now Spengel says with regard to the first of these differences,
“T cannot put it otherwise, in spite of Bateson’s researches,
than that we are still ignorant of the earlier stages of
development ;”’ and “ Bateson’s statements as to the blastopore
do not appear to me to be more trustworthy than those of
others who have observed the persistence of the blastopore as
anus in other animals.” Bateson’s figures, however, give
demonstrative proof that his statement is correct. The Tornaria
larva has, as we know, a longitudinal ciliated post-oral band,
and behind this a perianal one. Now in Bateson’s larva this
perianal band is the only one present, and it appears very early.
In one of his figures we see it appear on an almost spherical
gastrula and in the centre of it we see the disappearing blasto-
pore. What further proof Professor Spengel would desire is
not quite clear tome. With regard to the absence of stomo-
deeum and proctodzum, he says, “‘In the absence of figures it
is impossible to form any opinion as to the value of Bateson’s
statements.” Now the mouth and anus do not appear till after
gill slits and notochord have been formed, and to demand
figures before accepting results about which it is difficult to
imagine that a worker who used modern methods, like
Bateson, could be mistaken, only exhibits the amount of
prejudice which has clouded Professor Spengel’s mental] vision.
With regard to the origin of the coelomic cavities as
endodermic pouches, Spengel complains that the transverse
416 E. W. MACBRIDE.
sections showing this are all taken from a single series ; that in
the case of the proboscis cavity the transverse section only
shows a constriction of the gut; and that in the single longi-
tudinal section figured which shows the anterior ceelom opening
into the archenteron, there is an abrupt break in the character
of the cells ; the gut-cells do not, he says, gradually pass into
celomic cells, thus suggesting that the opening figured is an
artefact. Now so little is the insinuation justified that Bateson
founded his results on a single series, that he does not
express complete certainty as to the original connection of the
collar ceelom with the gut, because he found the openings of
communication in ‘ very few” of the larve. Does Professor
Spengel mean to demand that a zoologist should figure all
the sections he has obtained which show a certain point? I
imagine that few editors of scientific journals would relish the
prospect. His objection to the longitudinal section can be
met with a direct denial; the section appears to any one
accustomed to the appearances presented by the developing
celom of Echinoderms to be perfectly normal, a transition
from gut-cell to coelomic cells can be made out. It is hardly
needful to add that the constriction of an anterior cceelom from
the gut, and a “constriction of the gut,’ are one and the
same. To what miserably small dimensions the supposed
“ trochophoral” peculiarities of the Tornaria reduce on
examination we have shown above; that the principal ciliated
ring of the trochosphere, the prototroch, is not represented in
Tornaria, Spengel himself has admitted. The absence of those
most characteristic organs of the trochosphere, the head
kidneys, he regards as of little importance, as also the presence
of a preoral ccelom totally unrepresented in the trochosphere.
His calm assumption, in defiance of evidence to the contrary,
that the esophagus and intestine are of ectodermal origin, I
have dealt with above. In this connection also it should be
remarked that he mistakenly attributes to Professor Huxley
the terms stomodzeum and proctodeum, which we owe to
Professor Lankester.
In justice, however, to Professor Spengel, we ought to add
REVIEW OF SPENGEL’S MONOGRAPH ON BALANOGLOSSUS. 417
that he does not appear to have convinced himself of the truth
of his own theory, because in discussing possible affinities of
the Enteropneusta with the Echinoderms, he admits that there
is a strong resemblance between the larvee, and proceeds even
to suggest homologies. He believes that the collar pores may
correspond to the two madreporic pores which have been
observed in the larva of Asterias by Field.!| Here again, how-
ever, Professor Spengel’s imperfect understanding of what has
been done in other groups has misled him. Neither Bury’ nor
Field has, as he imagines, described a transitory right hydro-
cele. Bury distinctly states that he regards the hydrocele as
a structure which has been from the first unpaired ; he does
describe a right anterior ccelom as well as a left, but the left
anterior ccelom is not the hydrocele. The oldest larva
described by Field did not show a trace of the hydroceele, though
some of the younger possessed two madreporic pores; but it
is now a dozen years since Ludwig; proved that the madreporic
pore primarily opens into the ccelom, and that its connection
with the hydrocele is secondary. The most interesting
feature of Enteropneustan development is the strong resem-
blance to the Echinoderm larva which Tornaria presents, which
renders at least plausible the suggestion that the Protochor-
data and the Echinoderms diverged from a common bilaterally
symmetrical pelagic ancestor. On what other grounds besides
the illusory resemblances of their larvee does Spengel ask us to
base our belief on a distant affinity between Enteropneusta and
Annelids? On the assertion that the collar celom is segmented
from the trunk ccelom, like a mesoblastic somite from the ger-
minal band of Annelids (this, as I have already said, is an
assumption contradicted by both Bateson and Morgan), and on
the direction of circulation in the dorsal and ventral vessels.
It is surely not going too far to say that a zoologist who
1 Field, ‘“‘ The Larva of Asterias,” ‘Quart. Journ. Micr. Sci.,’ 1892.
2 H. Bury, “Studies in the Embryology of Echinoderms,” ‘ Quart. Journ.
Mier. Sci.,’ 1889.
3 Ludwig, “ Entwickelungsgeschichte der Asterina gibbosa,” ‘Zeit. fiir
wiss Zoologie,’ Bd. xxxvii.
418 KE. W. MACBRIDE.
explains away the striking Chordate features presented by gill-
sacs and nerve-cord, and founds a new theory on such a basis
as I have just mentioned, can hardly expect to have much
weight attached to his judgment.
In conclusion I may be allowed to express my regret that
Professor Spengel has not condensed his work more. The
amount of apparently needless repetition is very great, which
is all the more to be regretted as the amount of literature
which every zoologist must read is immense and daily growing.
Those who believe in the Chordate affinities of Balano-
glossus will, however, derive consolation from the fact that in
the huge volume we have been considering, every possible
argument has been urged against this theory, and that, not-
withstanding the great industry and ability which Professor
Spengel has displayed in attacking it, all his attempts to shake
it have, in my mind, signally failed.
ZooLoGIcAL LABORATORY, CAMBRIDGE ;
May 3rd, 1894.
EXPLANATION OF PLATES 29 & 30,
Illustrating Mr. E. W. Macbride’s Review of Professor
Spengel’s Monograph on Balanoglossus.
List oF ABBREVIATIONS USED.
Al. Alimentary canal. 4/. Br. Dorsal branchial region of the alimentary
canal. A/. @. Ventral esophageal region of the alimentary canal. Br. Branchial
sac. Br. in., Br. ex. Internal and external openings of the samee Br. sk. 1.
Skeletal rod in primary gill-bar or gill-septum. Br. sk. 2. Skeletal rod in
secondary gill-bar or “ tongue-bar.” Bwcee. Buccal cavity or pharynx. C.
Collar. ©. cw. Collar celom. Ch. Notochord. Ch}. Primary cuticular
sheath of the same. Ch?. Secondary mesoblastic sheath of the same. @. NV.
Central nervous system. C.p. Collar-pore. D. NV. Dorsal nerve-cord of
trunk. D. R. Dorsal roots of central nervous system. D. V. Dorsal blood- -
vessel. Zp.Br. Epibranchial epithelial band. G/. Glomerulus. Gox. Gonad.
REVIEW OF SPENGEL’S MONOGRAPH ON BALANOGLOSSUS. 419
Gen. R. Genital ridge. H¢. Heart. Hé. aur. Auricular prolongations of
heart of Schizocardium. Hy. Br. Hypobranchial epithelial band of Schizo-
cardium. Zi. Liver saccules. JZ. S. Lateral septum of trunk celom.. M.
Mouth. Muse. Muscular fibres. Myce. Myocele. P. C. Pericardium. P. C.
Aur. Auricular prolongations of pericardium of Schizocardium. P. #. Peri-
hemal space. Pph. Peripharyngeal space. Pr. Proboscis. Pr. Ce. Pro-
boscis celom. Pr. Cw. V. Ventral pocket of proboscis celom. Pr. P. Pro-
boscis pore. Pr. vm. Ventral mesentery of proboscis celom. Proct.
Proctodeum. Sc/. Sclerotome. Sp. Cw. Splanchnocele. Stom. Stomodeum.
Sy. Synapticulum. 7. B. Tongue-bar. Zr. Trunk. Tr. ce. d. Dorsal
division of trunk celom. 77. ce. v. Ventral division of the same. 77. d. m.
Dorsal mesentery of trunk celom. 77. v. m. Ventral mesentery of trunk
celom. V. NV. Ventral nerve-cord of trunk. V.V. Ventral blood-vessel.
PLATE 29.
Figs. 3, 4, and 5, and Figs. 8—14 inclusive, are taken from the monograph,
though in some cases simplified in detail.
Figs. 2—13 inclusive are copied directly from the monograph.
(All figures diagrammatic.)
Fic. 1.—Horizontal longitudinal section of Balanoglossus.
Fic. 2.—Sagittal longitudinal section of Balanoglossus (Ptychodera).
Fic. 3.—Transverse section of proboscis region.
Fic. 4.—Transverse section of collar region.
Fig. 5.—Transverse section of branchiogenital region.
Fie. 6.—Horizontal longitudinal section of embryo of Balanoglossus
Kowalevskii, according to Mr. Bateson.
Fie. 7.—Ideal sagittal section of developing Tornaria, illustrating Professor
Spengel’s theory of the formation of the mesoderm.
Fic. 8.—Illustrates arrangement of fibres in the longitudinal muscular
layer of the proboscis.
Fig. 9.—Illustrates arrangement of fibres in the circular muscular layer of
the anterior wall of the collar cavity.
Fic. 10.—Sagittal section of collar region, showing mutual relations of
collar ceelom, trunk ccelom, perihemal and peripharyngeal spaces.
Fics. 1la, 114.—Transverse sections of trunk, showing insertion of lateral
septum. Fig. 11a is behind the gill region ; Fig. 114 is in the hinder part of
the gill region.
Fries. 12a, 4, c, d, e.—A series of five transverse sections through the
anterior part of the notochord and adjacent pericardium and heart in Schizo-
cardium.
VOL. 36, PART 3.—NEW SER. FF
420 BE. W. MACBRIDE.
Fic. 134.—Diagram of two “ inner gill-pores ” of Balanoglossus (Ptycho-
dera), showing their skeletal structures and the relation of the latter to the
slits. Br. sep. Gill septum.
Fic. 134.—Similar diagram of three complete gill-slits of Amphioxus.
Lr. sl. Gill-slits.
Fic. 14a.—Diagram showing the extent to which the gill-sacs are developed
in a vertical direction in Balanoglossus (sensu stricto).
Fig. 144.—Similar diagram of gills of glandiceps.
Fic. 14c.—Ditto, ditto, Schizocardium.
Fic. 14d.—Ditto, ditto, Ptychodera.
PLATE 30.
Fic. 1¢.—Diagrammatic transverse section of Amphioxus larva, showing
the division of ccelom into splanchnocele and myoccele, the hollow sclerotome
and its relation to the notochordal sheath.
Fie. 14.—Similar section of Balanoglossus. Ch?. The mesoblastic sheath
of ‘‘chondroid tissue.”
Fic. 1e.—Similar section of the Scyllium embryo. Sc/. The sclerotome,
here a solid outgrowth from the wall of the myocele.
Fic. 2.—Ptychodera minuta. Dorsal aspect. G. p. Outer gill-pores.
Magnification 2.
Fig. 3.—Ptychodera erythrea. Dorsal aspect. Natural size.
Fie. 4.—Glandiceps Lachsi. Dorsal aspect. Natural size.
Fic. 5.—Schizocardium brasiliense. From the side.
Fic. 6.—Balanoglossus Kowalevskii. Magnification +.
Fic. 7—Balanoglossus Kupfferi. Natural size.
Fic. 8.—Collar skeleton of Ptychodera minuta.
Fic. 9.—Collar skeleton of Balanoglossus Kupfferi. #. Head. W.
Wings. B. Body. JZ. Crura.
Fic. 10.—Collar skeleton of Ptychodera aurantiaca.
Fic. 11.—Collar skeleton of Schizocardium brasiliense.
Fig. 12.—Collar skeleton of Balanoglossus Kowalevskii.
Fie. 13.—Collar skeleton of Schizocardium peruvianum.
NOTES ON A GREGARINE OF THE EARTHWORM. 421
Notes on a Gregarine of the Earthworm
(Lumbricus herculeus).
By
Wm. Cecil Bosanquet, M.A.,
Fellow of New College, Oxford.
With Plate 31.
Wuite looking through some worms which had been obtained
for dissection in Professor Ray Lankester’s laboratory at
Oxford, Dr. Benham, Senior Assistant to Professor Lankester,
noticed one specimen the hinder end of which appeared to be
filled with small white bodies easily visible through the body-
wall of the worm. These were preserved by Mr. E. A. Minchin,
who identified them as a species of Gregarina, and by whose
kindness I was allowed to examine them. As they presented
some points of interest, perhaps these few notes may not be
out of place.
Previous Observations.—On referring to the literature
of the subject, I found that this species of Gregarine had been
before observed and described, as a very careful account of it
appears in the well-known paper of Lieberkiihn (1), who men-
tions Meckel as a still earlier observer of these creatures.
Lieberkiihn describes this Gregarine as occurring in great
numbers in the body-cavity of the worm, especially towards
the tail; as being round in shape and white in colour; and he
gives some account of its conjugation, spores, &c., but does
not name the species. Around Gregarine is also described by
Schmidt (2) as occurring in Lumbricus olidus, but this is
figured as possessing a shaggy cuticle, and therefore appears to
422 WM. CECIL BOSANQUET.
be a different species from the one occurring in L. herculeus,
of which the following is a brief account.
Description, Habitat, &c—The animal is of a pure
white colour, and quite opaque. It is rounded in form,
generally spherical, but some individuals are of a blunt oval
shape (fig. 1); and others, again, apparently young specimens,
appear as flattened oval discs. They vary in size from minute
specks to bodies with a diameter of over 1 mm. An average
specimen measured under the microscope gave the proportions
1:2 by'9 mm. A few individuals occurred in the hinder seg-
ments of nearly all the worms (L. herculeus) which I opened.
In one or two instances almost the whole body-cavity was
filled with Gregarines. (In a number of specimens of Allolo-
bophora fctida (?) which I examined I did not find any of
these animals.) They lie seemingly loose in the ccelom, but I
have found young specimens embedded in a growth of cells
along the intestine of the worm, and cysts very constantly in
the same position (see p. 429).
Influences affecting Development.—Gregarines in
different stages of development may apparently exist side by
side in the same host, but in most cases there appeared
a great preponderance of these in one condition, either as
cysts, or conjugating pairs, or mature individuals, which
fact would seem to point either to some external influence
acting on the parasites, e.g. some alteration in the nature of
their environment, leading them to a simultaneous change of
state, or, as seems perhaps more probable, to a definite cycle
in their lives in accordance with the seasons of the year—
since I found mature Gregarines plentiful in the autumn and
winter months, many instances of conjugation about the spring,
while during the summer I could not find any specimens other
than cysts and spores.
Evidence seems wanting as to whether the well-being of the
host is affected by the presence of Gregarines. On the one
hand, worms of the most healthy appearance contained a
considerable number of the parasites, while, on the other, a
very sickly and apparently moribund worm was crowded with
NOTES ON A GREGARINE OF THE EARTHWORM. 423
gregarines, all conjugating. Here it might be suggested that
if the gregarines were the cause of the illness of their host, the
state of health of the latter had in turn affected them, and that
this preparation for the formation of spores on their part was
in view of the imminent death of the worm, the destruction of
their home and their own dispersion over the soil,—to be taken
in perhaps as falciform bodies in the earth swallowed by other
worms, and so to find another habitation.
I have not been able to observe any movement in these
gregarines, nor does their circular form seem adapted to pro-
gression.
Microscopical Details.—Cuticle.—Coming to micro-
scopical details, little need be said about the cuticle. It is, as
described in other gregarines, very elastic, of a granular
appearance, presenting in sections a distinct double contour.
It dissolves rather slowly in strong acids, more rapidly in
potash, but in each case becomes so transparent before dis-
solving that its actual disappearance is difficult to follow.
Paraglycogen Granules.—In a teased specimen the bulk
of the animal appeared to consist of round granules of various
sizes, colourless or of a faint greenish colour by transmitted
light. They appeared quite circular, and the largest measured
as much as 15 w in diameter. In those Gregarines which I first
examined and which had been preserved in alcohol the granules
presented a very curious appearance, many of them containing
what looked like a crystal in the form of a three- or four-
pointed star (fig. 3, 6. c.). The same was seen in stained
sections, the apparent crystal being even more marked and
seeming in double-stained preparations to take a different
colour from the rest of the granule. The granules from a
fresh specimen were quite homogeneous, so that it was evident
that the crystal, if such it really were, was produced by the
reagents used in hardening. Experiments tried on the fresh
granules gave the following results :—The addition of absolute
alcohol and then of benzol produced at first no apparent
result ; on washing out the benzol, however, with alcohol the
star-like appearance was produced, and on again adding benzol
424 WM. CECIL BOSANQUET.
a more marked variety—an apparently more massive crystal—
was seen. Washing out the benzol once more and then adding
water caused the crystalline appearance to vanish almost
entirely, a faint semblance of crossing lines in the substance
of the granule being all that was left visible (fig. 3, 6. c. d.),
while this substance itself first became opaque and then cleared
again. The re-addition of the former. reagents caused the
reappearance of the crystals. Oil of cloves similarly used
seemed even more effectual in their production than benzol.
What the nature of the phenomenon may be is difficult to say.
It may be that true crystals are formed, modified by their
organic origin and surroundings, being precipitated from
alcohol and benzol and redissolved in water. On the other
hand, it may be suggested that such an appearance could be
produced by a splitting of the substance of the granule in
consequence of a shrinkage caused by the reagents. This
explanation seems perhaps to harmonise better with the
phenomenon of the faint lines left on the disappearance of the
crystal in water, i.e. a line of division remained to mark the
original cleft ; also in many cases where the granules appeared
split into two halves the line of cleavage corresponded very
closely with the dark lines forming the limbs of the stars
(fig. 3 g, A), which lines seemed in some instances to extend from
edge to edge of the granule. Further, in one case at least I
thought I could detect a forking of the end of one of the radiat-
ing lines which would not be possible in a crystal, and in one
instance when the objective was accidentally screwed down so
as to touch the cover-glass and crush the section beneath, the
flattening of the granules was accompanied by entire dis-
appearance of the “ crystals.”” This explanation (viz. as hollows
in the granules) is given by Wolters (8) of the rather different
crystalline appearance seen by him in another gregarine, but
the apparent staining of the “ crystals’ in some preparations
seems a difficulty in the way of this explanation.! The same
1 Actual crystals were found by Frenzel (4) in other Gregarines, but these
appear to occur not in, but among the granules. Might not these crystals be
comparable to the crystals found in plants, which apparently are excretory
products ?
NOTES ON A -GREGARINE OF THE EARTHWORM. 425
observer also suggests that the granules are a supply of food-
substance for the gregarine, used up when necessary, in refer-
ence to which view I may mention that in a teased specimen
containing chiefly sporoblasts, among which were a few
granules, the latter appeared oval or irregular in shape, as if
undergoing a process of corrosion and absorption.
The granules gave the usual paraglycogen reaction with
iodine and sulphuric acid, swelling up in the strong acid in a
very curious manner, witha distinctly double-contoured appear-
ance, and finally dissolving (fig. 3 7). They dissolved rapidly and
entirely in potash, even when diluted to 1 per cent. In many
preparations the granules appeared to form the whole substance
of the animal (fig. 2), but in others a certain amount of form-
less protoplasm was visible among them, and no doubt this is
really present in all cases. It was well shown in a preparation
stained with gentian violet and orange, the protoplasm taking
the latter colour, the granules the former. In many cases
sections showed a number of deeply-staining spots among the
granules, as if the protoplasm had collected together into
nodes; and in one series a distinct network was visible (fig. 4),
due no doubt, as Wolters holds, to the action of reagents.
Capsule.—The question of the amount of protoplasm
existing in the gregarine beside the paraglycogen granules
seems intimately connected with a curious appearance seen in
the greater part of the series of sections of mature gregarines
which I cut; this was a ring of staining substance which went
through as many as sixteen sections of 5 u thickness, gradually
contracting and disappearing. It was therefore a hollow
sphere, and it contained within itself matter in no respect
differing from the rest of the animal’s body (figs. 2a, 26). It
did not present any sharply-defined outline, such as would
suggest a containing-membrane, but appeared rather to be
a continuous “capsule ” of protoplasm within the gregarine,
which was in some cases centrally, in others eccentrically placed.
It seems possible that the appearance may be caused by the
reagents employed driving the fluid protoplasm before them
from the sides inwards as they penetrate, thus forming a ring,
426 WM. CECIL BOSANQUET.
or, as it rather appeared in some cases, a lump of protoplasm
towards the centre, with the granules which were originally
there appearing in the midst of it. This explanation was
suggested by Professor Lankester in the case of a somewhat
similar appearance found by Miss L. J. Gould (5) in Pelo-
myxa. However, in one instance there appeared to be two
such formations in the same animal (fig. 5) ; in another, the
whole central portion of the gregarine formed a distinct round
mass, staining more darkly than the rest of the substance and
divided off from it by a dark line, as to the nature of which—
whether it were a definite membrane or not—I was exceedingly
doubtful (fig. 6). Mr. Minchin suggests that the phenomenon
is rather to be regarded as a preliminary to the stage of spore-
formation, and it was he who applied to it the name “ capsule,”
which I have used. But its nature must for the present remain
obscure.
Nucleus.—The nucleus ina teased fresh specimen appeared
perfectly spherical, about 70 m in diameter, with a circular
nucleolus dimly visible. It was quite colourless and trans-
parent (fig. 7). In the process of hardening it invariably
shrank, and in sections was always much crumpled. It was
bounded by a membrane, which appeared double-contoured
under a high power. The ground-substance was finely granular
in appearance, and it contained as a rule two nucleoli, but
sometimes more—in one instance as many as five (figs. 8, 9),
The nucleoli were highly vacuolated, the vacuoles in some
instances seeming to have some contents capable of being
stained (fig. 10).
Life History.—The following is a brief account of those
stages in the development of this gregarine which I have been
able to observe :
Conjugation.—Conjugation first takes place, two round
individuals becoming pressed and flattened one against the
other, the cuticles of the two coalescing at the portions where
they are applied one to the other. I have not been able to
observe any mitosis in the nuclei, but in those cases in which
the conjugating gregarines each contained a single undivided
NOTES ON A GREGARINE OF THE EARTHWORM. 427
nucleus the structure of this was peculiar, almost all the
contents of the nuclear membrane being gathered into a single
large mass, situated at one side (fig. 11). This is no doubt
preparatory to division, which probably takes place with mitosis
as described by Wolters in other Gregarines. The next
stage which I have myself seen is one in which the nucleus
has broken up into minute portions, of about the same size as
the larger paraglycogen granules ; these fragments are scattered
pretty evenly throughout the body of the animal, and are very
inconspicuous. They were pointed out to me by Mr. Minchin,
who kindly examined some of my preparations, and when
once demonstrated were fairly easily seen, as in a preparation
stained with hematoxylin and eosin they were dyed a distinctly
bluer colour than the surrounding granules (fig. 12). I did
not see any breach of continuity in the septum formed by the
coalesced cuticles which separates the conjugating Gregarines,
through which any interchange of substance might occur ; nor
have I seen the nuclei or nuclear fragments apparently attracted
towards one another in this species. In a specimen of Mono-
cystis agilis, from one of the vesicule seminales of the
worm, the nuclei of the two conjugating individuals were
drawn together to the opposite sides of the septum (fig. 21),
and probably the latter is ultimately entirely absorbed, as it
does not appear in cysts containing spores.
Sporoblastomeres and Sporoblasts.—The next stage
is the formation of sporoblastomeres (Minchin) by the se-
paration of the substance of the animal into masses, each
surrounding a fragment of the nucleus. These appeared in
fresh preparations to be surrounded by a transparent coat, but
this is either not universally the case, or disappears in per-
manent preparations (figs. 13 a, 6). In some the original
granules still appear embedded; at a later stage these dis-
appear, and the masses become homogeneous; they are of
different sizes, and contain one or more nuclei (fig. 130). I
have found specimens containing one, two, and four nuclei, sug-
gesting a tetraschistic division into the ultimate sporoblasts.
These latter become elongated and boat-shaped before secret-
428 WM. CECIL BOSANQUET. —
ing their spore-coat (fig. 14). The nucleus is placed almost
invariably to one side, on the equator of the sporoblast.
Spores and Falciform Bodies.—The formation of spores
from the sporoblasts appears to begin at the outside of the
cyst, and in one preparation the spores seemed to be so ar-
ranged as to point outwards, with their long axis on the radii of
the cyst (fig. 15), the centre being occupied by sporoblasto-
meres. In many cases a number of granules remain unab-
sorbed till the end. It may be noted that the characteristic
knobs at the end of the spores seem to disappear entirely in
permanent preparations, either dissolving in the reagents or
becoming invisible owing to their possessing an index of re-
fraction equal to that of Canada balsam. In glycerine these
processes were still visible. In two cases out of a large number
of cysts examined I have seen aspecimen of the curious tri-
angular spores figured by Lieberkuhn, which are probably
monstrosities. The development of the spores proceeds as
follows :—The nucleus divides into two, then each fragment
divides again, and the four resulting fragments divide once
more. There are thus formed eight small round nuclei, lying
in the protoplasm of the spore (fig. 16). The exact process of
division, with or without mitosis, I have not been able to
follow. The protoplasm surrounding these nuclear fragments
divides, and a portion surrounds each, these portions elongating
aud forming the falciform bodies. The nucleus in these lies
almost at the extreme end (fig. 17). The arrangement of the falci-
form bodies in the spore does not appear to be constant, the most
favourite being perhaps that in which the nuclei or heads lie
towards the centre, the tails pointing four to either end of the
spore (fig. 17 6). The falciform bodies are coloured very prettily
by the Ehrlich-Biondi stain; the nuclei appearing green, the
body pink.! Their shape seems to vary considerably, some being
longer, others thicker ; but it is possible that these differences
represent stages in their elongation and development.
1 In the mature Gregarines the nucleoli were stained a bright pink by this
reagent, the ground-substance of the nucleus taking the same colour, but
very faintly,
NOTES ON A GREGARINE OF THE EARTHWORM. 429
Position of Gregarines in the Worm.—A few more
points remain to be noted. With regard to the position in the
worm of the Gregarines at various stages, it may be mentioned
that the adult gregarines and the cysts which I first found,
also the conjugating pairs, lay in the ccelom of the worm, as
before stated ; later on, in the summer, I had great difficulty
in finding any in this position, but discovered that the cysts
were to be found embedded in certain oval or sausage-shaped
masses of tissue, which appeared to lie loose, or but slightly
attached, beneath the gut in the hinder segments of the worm.
These masses seemed to consist of altered and degenerated
nephridia, and were in some cases composed entirely of cells, in
others of an apparently structureless matrix, with a few cells
here and there (fig. 18). The Gregarines occurred in them along
with a number of encysted nematodes, so that it was impossible
to say whether the growth in which they lay was due to the latter
or to the gregarines; but Dr. Benham assured me that the
formation was certainly abnormal in the worm.' If the Grega-
rines were the cause of this growth, the fact might have a
pathological interest, in view of the suggestion that some
cancerous tumours are caused by a psorosperm, as is tlie case in
the disease peculiar torabbits caused by Coccidium oviforme.
The sausage-shaped growths in the worm were in some cases
honeycombed with cysts of small size closely packed together
(fig. 18) ; in other cases only a few larger cysts occurred in
each. In one or two specimens of mature Gregarines, which I
found apparently loose in the celom, there were a certain
number of tissue-cells attached to the cuticle, showing either
that it had once been embedded in a similar growth, or rather
perhaps that it was about to embed itself, but interrupted
before the process had proceeded far. The suggestion that the
falciform bodies make their way out of the worm to find another
1 Metchnikoff (6) records that the gregarine-cysts in the vesicule semi-
nales become surrounded by a mass of phagocytes which attack and some-
times succeed in killing them. The masses of cells above described seem
also to be of an inflammatory nature, but it seems doubtful whether they
injure the enclosed cyst, in which I have not noticed any marked signs of
degeneration.
430 WM. CECIL BOSANQUET.
host seems all the more likely from the fact that the cysts
occur in the same masses of tissue as the nematodes, which
are known to act in this manner.
Spores of two Sizes.—It is also of interest to note that
there appear to be two different kinds of spores occurring in
similar positions in cysts which are also similar, one being a
large variety measuring approximately 30 pw in length by 12 u
in breadth, while the other is almost exactly half this size,
viz. about 15 w by 5 mw (fig. 19). The description of spore-
development given above applies in the main to the larger
variety which are more easily observed, but, as far as I have
been able to make out, in the case of the smaller kind very
similar changes occur. The size of individual cysts differs
considerably in both kinds, but the cysts containing the smaller
spores are not themselves necessarily smaller than those in
which the large variety is found. I have found in one or two
instances a cyst containing large spores surrounded by tissue,
in which were embedded, apparently loose, a number of the
smaller kind. In cutting some sections through portions of
the vesiculz seminales of a worm, I also found cysts containing
spores of two different sizes, corresponding very closely with
the two varieties found in the hinder end of the worm. It
would be natural to suggest that these two varieties in the
vesiculz are respectively the spores of Monocystis magna
and agilis, and that the spores in the posterior segments may
also belong to these gregarines, the large round gregarines
described being a form of M. magna. Against this theory,
however, stands not only the distance from the known haunts of
M. magna and agilis, but also the large size of the cysts of
small spores, which could hardly be formed by M. agilis,
while the cysts in the vesicule containing large spores were
themselves no larger than those containing small ones ; and it
seems necessary to suppose, in default of further evidence,
either that two sorts of spores are formed by each gregarine, or
that still another species of gregarine remains to be described.
It therefore appears best for the present to consider the round
gregarine of the tail segments as distinct from the other
NOTES ON A GREGARINE OF THE EARTHWORM. 431
gregarines of the earthworm, and I would propose for it the ,
name Monocystis herculea, on account of its large size and
its occurring in Lumbricus herculeus.
As to the exact stage of development represented in fig. 20
I am still in doubt. Small round or oval bodies with one or
more nuclei such as are there depicted occurred in a small
number of cysts which were embedded, generally filling the
whole cyst, though in the case figured they formed a sort of
chain lying across it, apparently contained in a special mem-
brane. It seems most probable that they are the sporoblasto-
meres of the smaller variety of spores, as from their containing
more than one nucleus they can hardly be the final sporo-
blasts, and they are considerably smaller than the sporoblasts
shown in fig. 14. On the other hand, they rather closely
resemble the contents of the spores (fig. 16) where the nucleus
is dividing, but they have no trace of any coat.
I may perhaps note in conclusion that whereas in the round
gregarine described in this paper at the time of conjugation
two spherical individuals become pressed and flattened against
one another, in the specimens which I have seen of Mono-
cystis magna the two gregarines lie side by side at length
(fig. 21), and do not become circular as I understand Wolters
to state.
It remains only for me to offer my most sincere thanks to
Mr. Minchin for his continual kindness and assistance through-
out the time during which I have been working at this subject.
To him are entirely due any merits which this paper may
possess. I mustalso thank Professor Ray Lankester for kindly
allowing me the use of his laboratory, and Dr. Benham for
many hints and suggestions from time to time.
432 WM. CECIL BOSANQUET.
WoRKS REFERRED TO IN THE ABOVE PAPER.
=
. LIEBERKUHN, N.—‘‘ Evolution des Gregarines,” ‘Mém. Cour. et Mém.
des Savants Etrangers, Acad. Royale de Belgique,’ vol. xxvi, 1853,
pp. 3—40.
Scumipt, Apotr.—‘ Beitrag zur Kenntniss der Gregarinen und deren
Entwickelung,” ‘ Abhand. herausg. von der Senckenbergischen Natur-
forsch. Gesellschaft,’ Bd. i, 1854-5, p. 174, et seq.
3. WottEeRs, Max.—“ Die Conjugation und Sporenbildung bei Gregarinen,”
‘Archiv fiir mikr. Anat.,’ xxxvili, 1891, pp. 107 and 109.
4, FrenzeLt, Jou.— Ueber einige Argentinische Gregarinen,” ‘ Jenaische
Zeitschrift f. Naturwissenschaft,’ Bd. xxvii, 1892, p. 314, et seq.
. Goutp, Miss L. J.—‘ Quart. Journ. Micr. Sci.,’ vol. xxxvi (1894).
. Metcunixorr, E.—‘ Lectures on the Comparative Pathology of Inflam-
mation,’ translated by F. A. Starling and E. H. Starling, M.D., 1893,
pp. 69, 70.
ad
Q ow
EXPLANATION OF PLATE 31,
Illustrating Mr. Wm. Cecil Bosanquet’s paper, “‘ Notes on a
Gregarine of the Earthworm (Lumbricus herculeus).”
Fic. 1.—Mature gregarines and conjugating pairs, enlarged about four
times.
Fic. 2 a.—Section of a mature gregarine, showing nucleus and capsule.
x about 90.
Fic. 2 6,—Portion of the same section, showing capsule. x 280.
Fic. 3.—Various granules, showing the action of reagents. a. Granule
seen fresh in normal saline. 4%. c. After addition of alcohol and benzol.
d. On subsequent addition of water. e¢./ Same granule in different focus,
showing central spot dark and light respectively. yg. 4. Split granules.
i. Granule swollen by sulphuric acid. x about 480.
Fic. 4.—Portion of substance of gregarine, showing apparent network of
protoplasm. x about 300.
Fic. 5.— Portion of gregarine, showing two capsules. x 110.
Fic. 6.—Portion of gregarine, showing central mass divided off from the
rest. x 110.
NOTES ON A GREGARINE OF THE EARTHWORM. 433
Fic. 7.—Nucleus in fresh teased preparation. x 300.
Fic. 8.—Nucleus (crumpled), showing two vacuolated nucleoli. x 600.
Fic. 9.—Nucleus with five nucleoli. x 700.
Fic. 10.—Nucleus and nucleolus, showing staining substance in the
vacuole. x 600.
Fic. 11.—Nucleus at time of conjugation. x 450.
Fic. 12.—Nuclear fragments among granules, previous to formation of
sporoblastomeres. Zeiss, homog. immers., 2 mm., oc. 4.
Fic. 13 a.—Sporoblastomeres, fresh, showing coat. x 450.
Fie. 13 4.—Sporoblastomeres, showing division of nucleus and granules
embedded. From a permanent preparation in which the coats are not seen.
x 450.
Fic. 14.—Sporoblasts. x 720.
Fic. 15.—Spores in cyst, showing arrangement (semi-diagrammatic).
x about 100.
Fic. 16.—Spores in section, showing development of falciform bodies by
division of nucleus. x 400.
Fies. 17 a, 6.—Falciform bodies in spore. Sections of spores, showing
only half the number. Crouch, homog. immers. ;4,, 0c.2. x 800.
Fic. 18.—Portion of growth of cells surrounding the cysts. x 100.
Fic. 19.—Spores of two sizes. x 1000.
Fic. 20.—Doubtful stage, ? sporoblastomeres of small spores. x 700.
Fie. 21.—Conjugation of Monocystis agilis, showing mutual attraction
of nuclei. x 420.
Fic. 22.—Conjugation of Monocystis magna. x 16.
CE UMOTAITAAS TOO. IATA A oe Bee
4
4 WE ees ite vated ba w » lao). ni espl payee “
A100 x buat batifnasay ow) aniwails. dh Mauss earl Waoe i bh
a: HOT % sfonde Oy tile, gral
sid af Oodaladiiy ualiiola. gulwode aaloalsua bags aol lies Olea
OR pis “
O58. > mor taguiitoa ty /iiht i aualiw ere
ly “adivmianll ab! anaderny ealdahye? singe ATTA Tihs Ane ;
ve 00 ity eat en dé aN BINDS |
= yee dod wal ewls phate SUT Na ieoeg 1)
waluanrg, bit! dgaluiy 1S ahiatvibe qhiwnile e rit paldeun} i= te .
ove Jing O18 wlava Wh) oni Te nonate ay: Ionaeeigg a pine :
, i, is ab
et » stent) Saree 4
ADK sii asthe Fase) nioniatia” Geruals ety al a es |
_ OUE s
4 said solid 10 -LrumeniQval Sarwouds noloos Ni eon ene m
DUB x Abalone 1a) aia
sitwats all ld sont - .areere Ri oiticel Willa Aa Ww an
7 UX 8 605 ein somed widow) sadn out |
, Ob xs aver uid D ibinoye elie ye Arvin in noibny Toe, a )
‘ P - ;
-- U0E ” coxa wip elngu— OL ae ’
t
a he
—s
QOY 4 esroqs Leute Jo enokiedeatdotgs-? yore DT Ota
wile ila Wilt wih Wis Al [iba eita Lybee wie noieq pias, a
; . ja UES xe tole
F OFX nigedt ehiuysbaclt Ip nolisagiio den
—
ae
SOME ABNORMAL ANNELIDS, 435
Some Abnormal Annelids.,
By
E. A. Andrews.
With Plates 32—34.
Amonest the Annelida, as amongst other groups of non-
Vertebrates, various abnormalities and monstrosities may be
found upon examining large numbers of individuals. Thus in
some groups, as in the Eunicidz, double and supernumerary
antenne and parapodia are not so very unusual, though they
have received little or no attention as yet. Again, more fun-
damental abnormalities affecting the integrity of the body
somites occur in widely separated groups of Annelids, as has
been recently demonstrated by C. J. Cori, T. H. Morgan, and
F. Buchanan. Even repetitions of considerable regions of the
body, bifurcations of the main axis that produce a double
monster have been observed again and again. As naturally to
be expected from the accessibility to observation and familiar
occurrence of the terrestrial Annelids, the earthworms, it is
chiefly amongst this group that cases of doubleness have been
observed and recorded, though such also exist amongst the
marine forms.
From an examination of the bibliography of this subject as
recently given by H. Friend,! it appears that some twenty
cases are upon record, cases of double head or double tail,
chiefly in the earthworms.
In the hope that the recording and description of such cases
may ultimately contribute to a rational comparative teratology,
1 * Nature,’ February, 1893.
VOL. 86, PART 4,—NEW SER. aa
436 E. A. ANDREWS.
and also prove of value in the understanding of some fun-
damental problems of regeneration, two additional cases of
doubling of the posterior end in earthworms are here pub-
lished, together with eight cases in a marine Annelid.
ALLOLOBOPHORA FETIDA.
The two abnormal specimens of the “ Brandling” here de-
scribed were taken at Bryn Mawr, Penn.,in April, 1892, at the
period when the conjugation process seems to be at its maxi-
mum. Some conception of the probable frequency of such dupli-
cations in this animal may be got from the facts that the first
abnormal individual, which we will call A, was found amongst
480 preserved specimens collected at one time by Dr. T. H.
Morgan (to whom I am indebted for much of the following),
while the second individual, B, was found amongst 560 collected
at different times : a third collection of 220 specimens carefully
examined when alive contained no bifurcated individuals.
The first specimen, A, as seen in fig. 1, is perfectly normal
down to and including the 73rd somite, where a marked bi-
furcation occurs. An examination of the anterior region proved
that the sexual organs were normal and mature, and that the
sperm receptacle contained ripe sperm, indicating that con-
jugation had taken place.
Posterior to the 73rd somite there are two large abneenal
rings marked, like the normal somites, by dark brown bands on
the dorsal surface, as seen even in this alcoholic specimen.
From the second of these spring the two tails or terminals ;
one having 18 somites and an anal piece; the other, on the
right, only 7 somites and an anal piece. There are thus
73 +2 + 18 = 93 somites in what seems the main axis of the
animal and seven in what appears as a lateral outgrowth from
the second abnormal ring. This lateral terminal springs from
the right side and lies in the same plane as the main trunk, as
is evident from the enlarged views, figs. 2 and 3.
In the dorsal view, fig. 2, we see that the dorsal pores are
everywhere present in the normal positions, except that there
are none visible at the base of the right terminal, between its
SOME ABNORMAL ANNELIDS. 437
first somite and the second abnormal ring of the trunk, which
we have marked 75. The dorsal pore between the two ab-
normal rings 74 and 75 is removed to the right, so that it
lies rather in a curved series from the trunk into the right
terminal than in the straight series of the trunk and main
terminal. When, however, the dorsal body-wall is removed and
viewed from the ventral aspect, having been made translucent
by glycerine, we see, as in fig. 4, that the dorsal pores of the
right terminal B are not really in the series of the trunk H;
while the latter are in the same series as those of the main
terminal T. The slender muscles or groups of longitudinal
muscle-fibres that pass from one dorsal pore to the next are
seen to bend, from the pore 74—75 over towards the animal’s
right to end at the pore 75—76, crossing the darkly pigmented
band of 75 and half the light bands 74—75 and 75—76. The
muscle that runs up from the lateral terminal becomes lost in
the general dorsal muscle and is not connected with any other
pore than the one 1—2 whence it starts.
The ventral view, fig. 3, together with fig. 2, shows the true
shape of the abnormal rings 74 and 75. While 74 appears but
slightly notched on its left side, it is subdivided ventrally by a
groove coming down from the right side; it is a ring much
wider on the left than ontheright. So the ring 75 is, inversely,
much wider on the right than upon the left; it is marked ven-
trally by grooves that but imperfectly separate it from the
somite | of the right terminal. The two abnormal rings 74 and
75 together make a cylindrical mass of less diameter than the
trunk or the main terminal. It is from the expanded right
side of this second ring, 75, that the right terminal arises, while
the main terminal is continued from its distal face.
As is seen in the ventral view, the parallel lines that mark
those thin areas of the body-wall muscle along which the
ventral setz are inserted, continue from both trunk and main
terminal out into the lateral terminal, or in other words, the
right line of ventral sete is continued as the right or anterior
line of ventral setz of the lateral terminal, and the left or pos-
terior line of ventral sete of the lateral terminal is continuous
438 E. A. ANDREWS.
with the right ventral line of the main terminal. This con-
tinuity of structure from trunk to terminals is seen again in
the body-walls and in all longitudinally elongated organs.
Thus in fig. 4 we see the muscle fibres of the longitudinal
system of the dorsal body-wall passing from the trunk into both
terminals.
Dissections and longitudinal sections of the terminals and their
union with the trunk show that each is perfect in all the organs
of a normal posterior end of an earthworm aud that all the lon-
gitudinal organs of each terminal become continuous with
those of the trunk. Thus, as seemed apparent from the
surface (fig. 3), there is an anal opening for each terminal.
The short right one has, like the larger one in each of its
somites, the normal character and arrangement of body-wall,
septa, digestive tract, blood-vessels, nerve cord, nephridia, and
sete. The nephridia terminate in nephrostomes and pass into
the body-wall, presumably to open to the exterior, though this
was not seen in the sections.
The only departures from normal structure are found in the
peculiar rings 74 and 75. The digestive tract, as shown some-
what diagrammatically in fig. 5, sends out a large branch into the
right terminal ; the dorsal blood-vessel does the same, and so
does the ventral or subintestinal vessel, as was demonstrated in
serial sections. The ventral nerve-trunk bifurcates in such a
way that fibres pass from each of the three diverging trunks
into each of the other two. The “ giant fibres ’’ also pass from
one to the other; fibres from the trunk or body pass into each
terminal and fibres pass from each terminal across into the
other terminal. All this takes place in a nerve mass but little
smaller than and not unlike the normal, except where the actual
separation of the fibres takes place.
The ventral nerve thus formed as a lateral outgrowth from, or,
more accurately, as one fork of the nerve-cord of the body, has
in each terminal the normal ganglionic swellings for each
somite.
The septa that subdivide the body-cavity are normal in posi-
tion and structure in the body and in each terminal except in
SOME ABNORMAL ANNELIDS. 439
these abnormal rings 74 and 75. As far as could be made out
by dissection and by section the arrangement of the septa was
as indicated in the diagram, fig. 5. A large septum appears to
separate 74 from 75 internally and to correspond to the external
groove between those two peculiar rings. The body-cavity of
75 is bifurcated like its wall and its contents, supplying the base
of each terminal, and is separated from the somites 76 and 1 of
the two terminals by two distinct, complete septa. While this
ring is not subdivided internally, the anterior one, 74, presents
on the left four small,incomplete septa extending from the body-
wall to the digestive tract, as seen in fig. 5. Each compart-
ment of the body-cavity so formed has its own cluster of small
sete, five pairs in place of one on this left side. On the right
side no setz could be found
The nephridia that are normally present in both terminals, as
indicated in fig. 5, seem to be absent in 75 on its right, where
we would expect one either posterior to or anterior to the
branch of the intestine that runs out into the right terminal ;
it may have been displaced and not recognised in the serial
sections. The nephridia of 74 are multiplied on the left to
correspond to the number of subdivisions of the body-cavity,
five nephridia taking the place of one.
One other abnormality shown in the diagram should be
mentioned, as it makes the exception to the rule that all the
abnormalities are within the two rings 74 and 75; the ring 1
of the right terminal has no nephridium upon the left side, as
far as could be ascertained.
The ring 75 thus bears both the terminals and contains all
the bifurcated organs ; it appears to be a single somite.
The ring 74 is single upon the right but subdivided upon the
left. From the small size and apparent newness of these small
septa and sete we may judge them to have been interpolated
here to make what were else a normal ring so much longer
upon the left than upon the right. If this elongation of the
left side of 74 were non-existent the two terminals would appear,
obviously, as two sister structures or as a bifurcation of the
main axis, not as a main terminal and a side outgrowth.
440 E. A. ANDREWS.
The abnormality of this specimen, A, thus expresses itself
not only in the duplication of the posterior somites but in an
elongation of the trunk upon one side with partial formation of
somites there.
The second specimen, B, is much less conspicuously bifid,
having only a small bud-like outgrowth upon its left side near
the posterior end, as shown in fig. 6. There is here no doubt
that the right terminal is the normal continuation of the body
while the left terminal is of much subordinate importance.
This individual was sexually mature like the other; dis-
section of the anterior region showed that the reproductive
organs were perfect and mature and that the dorsal receptacle
contained sperm; there were also two “ spermatophores ”’ at-
tached to the exterior upon the ventral side.!
A very important peculiarity of this specimen is seen in fig.
6, that is that all the somites posterior to the 59th are short
and crowded with numerous double or abnormal rings, and that
all this posterior region is noticeably narrower than the anterior,
so that we at once infer, from the specimen, that all the region
posterior to the 59th somite is a new formation due to re-
generation of the original terminal. This seems to be a pretty
safe conclusion, though there is still the chance that the nar-
rowness of this region may have been due to incomplete de-
velopment, to some deficiency in the early embryo or larva.
The entire number of somites as counted on the right side,
where several split or double rings increase the number, is
103, and an anal piece so bent ventrally as to be invisible in the
dorsal view, fig. 6. Of these about nine are posterior to the
bifurcation, counting the anal piece as one; in the left lateral
process there are also nine divisions, including the final or anal
piece, which, however, shows no sign of an anal opening.
The dorsal and ventral views, figs. 7 and 8, show that there
are considerable irregularities in the somites anterior to as
1 T hope to show elsewhere that in A. foetida the dorsally placed sperm-
receptacles are filled during conjugation, and that the so-called ‘ spermato-
phores” are almost accidental structures, which fall off soon after the act of
conjugation.
SOME ABNORMAL ANNELIDS. 441
well as posterior to this left lateral outgrowth, while this itself
seems normal in the character of its rings. Thus the somite
99 is a single ring while 89 is split on the left side, or forms as
it were a ring the ends of which somewhat overlap one another.
91 and 92 as counted on the right are continuous by a spiral
of one turn. Thus, as may be easily seen in a model con-
structed from the above figures, the anterior part of 91 isa
complete ring that is continued from its middle dorsal part as
a spiral passing to the left, down, and across the ventral side, to
come up again into the dorsal side on the right ; here it forms
the anterior part of 92 and is continuous with the complete
ring that the posterior part of 92 forms.
The dorsal pore seems to be lacking at the beginning of this
spiral, where 910 leaves 91 a, but it is present at the other end
where 92a joins 92. The spiral may be designated 915 + 92a.
The next rings, 93 and 94, are complete and single.
The lateral process springs from the left of a complex spiral
beginning in 95. Here again a clay model or a glass tube or
bottle with the intermetameric groove marked upon it will make
clear the actual state of things represented in figs. 7 and 8.
The anterior part of 95, 95a, isa complete ring whence a spiral,
950, runs over toward the right and so around across the ventral
side, in a direction opposite to that of the spiral 91—92. This
spiral 955 comes up on the left as 96 and so across and down
to end on the ventral side in a complete ring, 97, fig. 8. This
spiral thus makes one and a half complete turns from the ring
95a to the ring 97,
Certain complications must next be considered. The ring
95a is subdivided ventrally by a faint groove, seen only in
fig. 8, into two half rings 95a! and 95a? corresponding to
the single ring 95a as seen on the dorsal side. Again, the
region on the left whence springs the lateral terminal is an
elongated elliptical expansion of the union of 95a and 950, as seen
from the left side. This enlarged lateral region ends ventrally,
fig. 8,in a sharp angle between the divaricated 95a and 95d ;
the enlarged region crowds the halves of 95a (95a'and 95a?) in
front of it and crowds behind it the ring 95 0, where continued
4.42 E. A. ANDREWS.
as 96, on the left side of the animal. The dividing line be-
tween 95 a and 95d, bifurcates to send one limb anterior to
the enlarged region, the other posterior to it ; the former limb
stops abruptly at the union of lateral and dorsal surfaces, and
SO appears as a mere notch in the left edge of 95 @ as seen in
fig. 7; the posterior limb continues as the line between 956
and 96, fig. 7. An additional complication appears in fig. 8,
where a small furrow, X, is seen posterior to the lateral out-
growth; it extends only a short distance upon the ventral and
lateral faces of the enlarged region.
The lateral process may thus be regarded as springing from
the enlarged left side of the spiral as it starts from 95 to run
over to the right, the enlarged area dropping down to the left
enough to be visible upon the ventral surface.
The existence of this spiral makes less obvious the similarity
of this case to the first, A; but for the spiral the lateral pro-
cess would arise from the left side of 95 4 much as does the
process from the right side of the ring 75 in the specimen A.
Posterior to the ending of the spiral in the ring 97 there are six
normal rings and a normal anal piece and anal opening (fig. 8).
Turning near to the left termimal we observe that it has
eight rings, the first somewhat subdivided by a short groove,
and an anal piece that is not perforated by any anal opening.
The colour bands and rete are evident externally. The
absence of anal opening is found upon examination to be
accompanied by the entire absence of any digestive tract.
The entodermal parts of this lateral outgrowth of eight
somites were entirely wanting.
In a cross-section of the middle part of the process we find
the body-wall made up of the normal epidermis, longitudinal
and circular muscles and peritoneal lining of the body-cavity.
The histology of these is, however, much as in a young imma-
ture animal or as in a regenerated posterior part. There are
dorsal and ventral setz, a ventral nerve-cord, and large nephridia
of the type found in this species. The saccular terminal parts
of these nephridia are, however, very much distended, so that
the pair of nephridial sacs take up most of the space usually
SOME ABNORMAL ANNELIDS. 443
occupied by a digestive tract. There is no dorsal blood-
vessel, but the subintestinal blood-vessel is represented by a
trunk that pierces all the septa and is traceable to the sub-
intestinal blood-vessels of the trunk, of which vessel it isa side
branch.
The only organ that seems to indicate a digestive tract is a
peculiar strand of muscles forming a slender cord running the
whole length of the body-cavity dorsal to the nerve-cord. This
muscle pierces the septa, and closely accompanies the above-
mentioned ventral blood-vessel, lying just dorsal to it. It may,
perhaps, be regarded as a representative of the musculature of
the digestive tract.
The body-cavity contains numerous amceboid corpuscles.
The ventral nerve-cord presents a ganglionic enlargement in
each somite, and is accompanied by subneural and by lateral
blood-vessels. Its sheath is but little developed, and its histo-
logy presents indications of immaturity, though showing the
normal ganglion cells, giant fibres, and fibrillar substance.
In such a section the only departures from the normal are
in the absence of digestive tract and dorsal vessel, in the in-
complete histological state, and in the presence of a free supra-
neural muscle-cord.
A transverse section near the posterior tip, fig. 9, shows
most of the structures as seen at the middle of the process,
but the tissues are still less perfected and the ventral blood-
vessel and its accompanying supra-neural muscle are absent.
There is in addition at this part of the process a peculiar
septum dorsal to the nerve-cord, which here is just separating
from the body-wall. This partition is only the last septum
much inclined along with the ventral bending of the tip of the
process, so that it cuts off a small terminal part of the body-
cavity on the ventral side around the nerve-cord from a larger,
really more anterior, part that lies dorsally and contains the
large nephridia. As the nerve-cord sinks down completely
into the body-wall near the extreme tip of the process, there
is left above it a small pocket of the body cavity, shut off
from the rest by the above oblique partition yet open anteriorly
444, E. A. ANDREWS.
into the cavity of the penultimate region 8. Owing to the
absence of the digestive tract, the appearance of the tip of the
process is quite peculiar in section, and this backward exten-
sion of the penultimate body-cavity space makes it the actual
termination of all the sections. Thus the last appearance of
the body cavity is as two small pits in the body-wall, on either
side of a ridge that passes along where the nerve-cord is about
to emerge from the ventral body-wall; these are ventral and
belong to the region 8. Dorsally, the terminal region 9 ends
bluntly, with no trace of any anal invagination or indications
of a digestive tract.
In other words, the septum between 8 and 9 sends back a
nearly horizontal arch over the nerve-cord, and then cuts off a
small ventral coelomic space that ends at the extreme tip of
the process as a pair of pits, one right and one left of a slight
median ridge.
Where this peculiar left process joins the main trunk of the
animal we find all its longitudinal organs continuous with
those of the trunk; the nerve-cord joins the main nerve-cord,
the ventral blood-vessel the main sub-intestinal vessel, and
the longitudinal and circular muscles are directly continuous
with those of the main trunks.
The digestive tract of the trunk is in no way affected by the
presence of the lateral process.
As might be anticipated from the external views (figs. 7
and 8) the arrangement of the septa in this region of union is
by no means simple nor readily made clear. As will be seen
in the horizontal section (fig. 10), the septa here are much
enlarged and distorted, so that they lie beyond the planes of
demarcation between external rings.
With some considerable success we may attempt to trace a
septum for each external groove, but errors easily creep in.
In the lateral process itself the septa are but elongated
and forced inwards towards the trunk, so that from 3—2 and
2—1 we pass to the much distorted septum 1—95 that reaches
in almost to the digestive tract of the trunk. Each body-
cavity has its appropriate nephridia.
SOME ABNORMAL ANNELIDS. 445
In the trunk many irregularities occur. A remarkable long
septum runs nearly parallel to the digestive tract anteriorly
separating 94 from 95 a, and posteriorly 95 a from 95 6. It
is connected with the walls of the digestive tract by two short
septa that may he regarded as the continuations of septa 94—95 a
and 95 a—95 4, though this continuity was not demonstrated.
An examination of the series of sections shows that this
longitudinal septum is not continuous, but of slight extent.
It may be regarded as a veil stretched between 94—95 a and
95 a—95 b, which seem to be complete transverse septa. Some
additional complexities affect this region.
As indicated in this figure, nephridia are found in most of
the chambers of the body-cavity, right and left. In 93, on the
right, there is one of the peculiar masses of old setz and
leucocytes so often found near the posterior end of earth-
worms. In addition, this part of the body-cavity contained
a normal nephridium. As far as could be ascertained the
nephridia are absent in the left of 94 and 96, but they may
have been destroyed in a dissection previous to sectioning.
PODARKE OBSCURA.
This small polychztous Annelid is exceedingly abundant in
the “eel: pond” at Wood’s Holl, Mass., where it may be seen
crawling upon the eel-grass or upon the muddy bottom, or
occasionally swimming free.
My attention was called to the occurrence of bifid monsters
in this species by Mr. A. L. Treadwell, of Miami University,
Oxford, Ohio, who observed as many as fifteen cases of bifur-
cations of the posterior end, none of the anterior end, amongst
1500 or more individuals examined in the summers of 1891
and 1892. Although less than 1 per cent. were abnormal, yet
a collection of about 100 taken at one time showed five or six
bifid monsters.
Regarding these he writes :—‘ The tail sometimes branches
equally, so that there is no difference between the sides, and
sometimes the second portion is much smaller than the other.
44.6 E. A. ANDREWS.
It may in some cases branch out one third of the way towards
the head from the tail, and is then usually at right angles to
the body proper. This branch may be so small as to be a
mere bud, or it may attain a considerable length. I think
that both sides are supplied with intestine and nerve-cord.”
To his courtesy I am indebted for three of these specimens,
as well as for five others obtained the past summer, 1893, at
the same place.
For convenience these eight cases of bifurcation of the body
posterior to the head region may be referred to as the specimens
A, Bs Cy VG Weeks YG aa:
The first, A, was in alcohol, having been carefully hardened
for examination by means of sections. The other two, B and
C, were stained and mounted in Canada balsam.
The general flat, broad shape and colour of this Annelid are
indicated in fig. 16, which, however, is not as dark as the
animal usually appears.
In the specimen A the body is bifurcated in such a manner
that there are two equal terminals, one dorsal and one ventral,
as seen in the side view (fig. 11). At the same time each
terminal may be regarded as a right or a left branch, since
their tips diverge widely from side to side as well as up and
down, and since, moreover, the terminals and trunk come
together obliquely, that is, the dorsal terminal is somewhat
twisted towards the right, the ventral one slightly towards the
left, or as we may express it, the median planes of the three
parts diverge and are twisted so that they have no line in
common, only a point.
The parapodia continue along the ventral or right terminal
with little interruption in their line of sequence, but the para-
podia of the dorsal terminal are separated from those of the
trunk by a wide interval, and do not have the same direction as
those of the trunk. In fact, the dorsal terminal looks like a
dorsal interpolation grafted upon the trunk where it bends
down, and somewhat to the right, as the ventral terminal.
In this specimen the number of setigerous somites in the
trunk or region anterior to the bifurcation is 27, while in each
SOME ABNORMAL ANNELIDS. 44.7
terminal there are just 8, so that there are 27 + 8 = 35 seti-
gerous somites in the main line, with a duplication of the
posterior 8 arising between the 27th and 28th somites.
Before speaking of the internal anatomy of this individual
we may note that the other two, B and C, have almost the
same character, the same oblique insertion of the terminals,
but different numerical proportions. The specimen B has 23
somites in the trunk, 12 in the right and 11 in the left ter-
minal. The specimen C has only 18 somites in the trunk, 10
in the left and 8 in the right terminal, which latter is broken.
In all cases the terminals are all complete, as far as can be
judged from an external examination of the general form, in
proportions, parapodia, anal termination, and ventral nerve
trunk.
On studying the internal structure of A, by means of serial
sections, we find at once that the two terminals almost exactly
repeat the normal anatomy of the trunk, and are thus like
two normal posterior ends of two normal Annelids. Each ter-
minal has the normal digestive tract, nerve-cord, blood-vessels,
and muscles, as well as sete and parapodia and even repro-
ductive organs—the mother-cells of sperms.
At the point of divarication the body-cavity and the digestive
tract branch, as shown in fig. 12, which represents a median
ventral section of the specimen A.
Each terminal has its anal opening and its ventral nerve-
cord extending forward from the anus throughout all its
somites; but while the nerve-end of the ventral terminal is
continued on as the ventral cord of the trunk, the nerve-cord
of the dorsal terminal stops suddenly in the first somite, as
shown in the figure. It thus appeared that the dorsal terminal
was imperfect in having its nerve-cord imperfect in the first
somite and without connection with the nerve-cord of the main
trunk of the animal. A careful examination of the entire series
of sections shows that this is probably the case. The nerve-
trunk in the first somite is free from the epidermis, in which
it normally remains in Podarke, and gives off large nerves
along certain muscles which radiate upward from the nerve-
448 E. A. ANDREWS.
cord to the intestine in each somite. Here, in the intestine,
the nerves were lost, and could not be followed into any con-
nection with the main trunk of the animal. The nerve-cord
ends bluntly as an upturned mass, suggesting a former con-
nection with, and rupture from, the main ventral] nerve-cord of
the animal.
The lateral nerves that run out in the epidermis to the para-
podia are found in the first somite as in the others, but there
is no indication of any lateral nerve establishing a connection
between the ventral nerve-cord of this dorsal terminal and the
nerve-cord in the main trunk.
To determine if this peculiar behaviour of the nerve-cord was
found in the other two specimens, B and C, they were prepared
and cut into serial sections, crosswise. It was then easy to see
that the same conditions obtained here as in specimen A. In
both B and C the body-cavity bifurcates, and the digestive tract
bifurcates and runs out into each terminal as if to an anal
opening (which, however, was not demonstrated). Yet the
nerve-cord does not bifurcate, but passes directly into only one
of the terminals, the left one in each case.
The other terminal, the right, has its normally formed nerve-
trunk that stops near the base, with no indication of any direct
connection with the main system.
In other respects the two terminals in B and C are normal
and alike. The dorsal and ventral blood-vessels and the
muscles are all perfectly normal; and the body-cavity, as in
the more anterior region of the body, contains large masses of
young reproductive cells, sperm mother-cells, in each terminal.
In these three specimens we find the terminals normal repe-
titions of the posterior end of the body, except that the more
dorsal terminal, the one more to the left in A, but the one
more to the right in B and C, has no apparent connection with
the nervous system of the anterior part of the body, and is thus
isolated in a peculiar manner. This isolation was entirely un-
expected, since in all other cases of bifid Annelids the nerve-
cord, as far as there are any statements made about it, is said
to bifurcate also and to send a branch into each terminal.
SOME ABNORMAL ANNELIDS. 449
The five other specimens taken by Mr. Treadwell in the
summer of 1893 may now be described in detail.
The individual V, taken August 18th, was not normally
active, but rather torpid; it lived but three days in captivity.
The bifurcation, though visible to the naked eye when once
noticed, was not conspicuous, but, on the contrary, easily over-
looked.
The supernumerary terminal has the form of a small bud
from the right side, as seen in fig. 13, which appears to the
naked eye as a new parapodial-like mass interpolated between
parapodia 33 and 384 on the right side, but is in reality a com-
plete posterior end with five normal somites and normal anal
tip and cirri.
This small lateral terminal is so far imperfect in that the
basal or first somite has an imperfect parapodium upon its left
or posterior side, as seen in the figure. This parapodium is
small and not complete ventrally, so that a ventral view would
show only four parapodia upon the left side of the small ter-
minal and five upon the right. This lateral process or terminal
stands out nearly at right angles to the main trunk, but points
somewhat downward or ventrally. Its ventral surface is con-
tinuous with that of the trunk, but its dorsal surface does not
extend up to the level of the dorsal surface of the trunk, but
ends between the parapodia as indicated in the figure. :
Judging from the size and appearance of the last or 5th
somite, the lateral terminal is young and actively growing. It
is well pigmented, and otherwise like a normal terminal in
appearance. The somites 2, 3, 4, show the same dark masses
upon each side of the intestinal tract that normally occur upon
the intestine in the rectal region in Podarke.
Posterior to this minute terminal there are 13 somites in the
main line of the trunk, which makes a total of 46, in addition
to the five small ones of the lateral outgrowth. The body ends
abruptly in an anal piece that has evidently recently replaced
one lost by accident ; as yet no anal cirri have developed, but
the long dorsal cirri of the last somite are directed backward,
as is common when the anal end is lost.
450 E. A. ANDREWS.
Serial sections show that the minute lateral terminal has the
normal internal structure: the main digestive tract and the
ventral nerve-cord send each a branch into the terminal and
so through the length of it. This side intestine ends in an
anal opening at the tip of the terminal.
The animal is evidently animmature female with developing
ova in the ovaries of the trunk and very young ovaries in the
basal somites of the terminal also.
This individual is also remarkable for the abnormal de-
velopment of many of the parapodia upon the main trunk. As
represented in fig. 14, the tip, in many of the parapodia, is not
a single cone but bears from its posterior face a conspicuous
outgrowth, x , which causes the parapodium to appear forked
or bifid from a dorsal view.
The specimen W, taken August 17th, has a large, swollen
dorsal and a small ventral terminal as seen in the side view,
fig. 15.
In life the swollen dorsal terminal is lighter in colour than
the dark trunk and appears as if filled by a milky mass as of
sperm. It is wider than the trunk at this region and has nine
somites that are short, as if contracted. The tail end is trun-
cated, the anal piece together with some few somites having
been lost. It stands up abruptly from the posterior edge of the
25th setigerous somite.
The small ventral terminal has but little colour, an appear-
ance of being a new growth and an anal end that seems to be
growing. It is made up of six setigerous somites and a seventh
with slight parapodial outgrowths as well as an anal piece with
one anal cirrus present.
It stands nearly horizontal but yet pointing decidedly down-
ward, at times making an angle of nearly 45°. Having but
little pigment in its dorsal surface, the blood in the dorsal
vessel may be seen passing forward towards the trunk.
The animal crawls and swims with the dorsal terminal
raised stiffly. When stimulated by touches from a needle, the
dorsal terminal seems less sensitive than the ventral one, partly,
perhaps, because it has no anal cirri. While a gentle touch to
SOME ABNORMAL ANNELIDS. 451
the ventral terminal will often produce active crawling or even
swimming, such a touch often causes the dorsal terminal merely
to shrink without setting up any locomotion in the entire
animal.
Sections show that there is a complete bifurcation of the
digestive tract, so that the small ventral terminal has an anal
opening that communicates with the main intestine, while the
dorsal terminal is also provided with a branch of the intestine
proceeding towards the wanting anal piece.
The large dorsal terminal contains much sperm or masses of
nearly ripe sperms, and the trunk as well as the small terminal
contains male cells. The nerve-cord continues from the main
trunk directly down into the ventral terminal, and so out to its
anal tip. A normally developed nerve-cord is present in the
dorsal terminal, but there it ends abruptly without any trace-
able connection with the nerve-cord of the trunk.
We meet here again the same peculiar condition found in A,
B, and C, and have an explanation for the differences observed
in stimulating one or the other of the terminals: the dorsal
one has an interruption in its nervous connection with the
anterior part of the body, and hence when stimulated does not
readily set up movement in that part of the body.
Sections of this region show the same anatomical relations
as are indicated in fig. 12.
The individual X was a large ripe female, found among
several hundred adults taken at different times in July. Being full
of ripe eggs the body readily ruptured, and was not studied alive.
An examination by section showed that the eggs and ovaries
fill not only the main trunk, but also both terminals.
Of the two terminals the smaller one arises as a lateral out-
growth from the right, much as in fig. 13. It is, however,
much larger than in that specimen, though shorter and
narrower than the main terminal.
It springs from the right side between the parapodia of the
29th and 380th somites, which it forces widely apart. At first
turned outward, it then bends backward, so as to lie more
nearly parallel with and dorsal to the main terminal. It is
VOL. 36, PART 4,—NEW SER. HH
452 E. A. ANDREWS.
inserted obliquely, so that its median plane is not vertical, but
inclined so that the right side is turned downward and the left
side upward. It is so attached as to occupy all the space
between parapodia 29 and 30, and then extends upward some-
what upon the dorsal aspect of the trunk, so that its left or
more posterior face is continued out from the dorsal surface of
the somite 30 where it joins 29. Ventrally the attachment
extends no further than the ventral part of the parapodia. The
parapodia upon the left of the bud, near its base, overhang the
back of the main terminal.
The length of the lateral terminal is 2} mm., that of the
entire main axis 12 mm., of which 9 mm. is anterior to the
point of bifurcation and 3 mm. posterior to it.
The main axis has 42 somites, of which the last three are
quite small and young; 29 are anterior to the bifurcation, and
13 posterior to it. The lateral terminal also has 13 normal
somites, and an anal piece on which the anal cirri are lacking,
from accident apparently.
Owing to the rupture of the body in dying, no idea of the
internal anatomy of the point of divarication of the terminals
was obtained. Both terminals have the normal internal organs,
even ovaries, as above stated.
The specimen Y was taken amongst two or three hundred,
August 3rd. It has the general appearance of W, as shown
in Fig. 15.
Of medium size, it swam about actively or crawled upon the
bottom of the aquarium.
The main trunk has 27 setigerous segments, the ventral
terminal 6, and the dorsal 7, but it is truncated by the loss of
its normal tip and several somites.
The dorsal terminal is much stouter and longer than the
ventralone. The latter does not extend across the entire width
of the trunk, but arises as a sort of outgrowth from the left half
of the ventral face.
The somites 25, 26, 27, are contracted and narrow from side
to side ; 25 and 26, judging from the appearance of the para-
podia, have been injured upon the right side.
SOME ABNORMAL ANNELIDS. 453
When the animal was alive the dorsal terminal stood nearly
vertical and the ventral nearly horizontal while swimming, but
in crawling the ends were inclined at about 45° to the horizon
and so at right angles to each other. After death the dorsal
terminal projects horizontally nearly in the line of the main
trunk, while the ventral one now stands nearly at right angles
to the rest. The ventral terminal is of some use in crawling,
since its parapodia perform regular, rhythmic crawling move-
ments. The dorsal terminal is of no use in crawling, as it is
raised above the surface, yet its parapodia at times move, but
in an irregular manner. The dorsal terminal may move back
and forth to vary its angle with the trunk, and at times it
makes undulating movements from side to side, as does the
trunk, in swimming.
The colour of the animal is normal ; the dorsal side of each
terminal is marked by the normal transverse bands of dark.
As the animal swims, the anterior face of the dorsal terminal
and the posterior face of the ventral terminal, being the dorsal
aspects, are dark brown, while the posterior face of the first
and the anterior face of the second, being the ventral aspect,
are light coloured.
As the colour bands of the dorsal terminal are somewhat
lighter than those of the ventral, it has a new or grey appearance
which does not harmonise with the anatomical conditions found
upon sectioning.
When stimulated by the touch of a needle to the head,
antenne, or cirri, the anterior region shrinks back. When the
end of the ventral terminal is so stimulated there results a
forward movement of the whole body, there may be even a
rapid swimming away. Stimuli applied to the dorsal terminal
cause local movement, but no change in the anterior region
unless the stimulus is strong, when the animal may crawl, but
not swim. It thus appears that the dorsal terminal is not as
perfect a medium for transmitting stimuli to the entire animal
as is the ventral.
Examination of the internal anatomy shows the same con-
dition of affairs found in A, as seen in fig. 12, While the
454, BE, A. ANDREWS.
digestive tract bifurcates to furnish an arm to each terminal
with an anus at the tip of the ventral and, presumably, one at
the tip of the dorsal when the tip was present, the nerve-cord
does not bifurcate.
The nerve-cord of the trunk runs down to the anus as what
seems to be, judging from its histological appearance, a new
nerve-cord along the median ventral line of the ventral terminal.
The nerve-cord of the dorsal terminal ends abruptly where the
ventral surface is continuous with the dorsal surface of the
other terminal. It has no connection, so far as discovered,
with the main nerve-cord.
The sections also show that Y was a female with ovaries
both in the trunk and in each terminal ; the ovaries are as yet
only small masses of cells with large nuclei. ,
The last specimen, Z, was found amidst several hundred,
August 6th. It differs from all the others in having the
bifurcation far forward as a small lateral terminal upon the
left, as seen in fig. 16.
At rest upon the bottom of the aquarium the animal is
straight or variously bent, generally so that the lateral process
projects at acute or obtuse angles or at right angles with the
straight or convex side of the body, rarely is the animal bent so
that the process projects from a concave line. The curvature
of the body may be uniform anterior to and posterior to the
lateral process, or there may be a separate curvature on each
side the region of the process; posterior to the process double
curves, right and left, were observed.
In swimming, the body is nearly straight and the process
projects nearly at right angles.
From the general coloration of the body, as indicated in
fig. 16, the process makes an exception as regards the left or
anterior side of its dorsal face. On this half the transverse
colour bands are very imperfect, existing only upon the side
between the parapodia and not extending up on to the dorsal
surface. The posterior or right half has, however, the normal
arrangement of bands. Along the lighter coloured left side
there extends a clear streak, like a scar, reaching from near the
SOME ABNORMAL ANNELIDS. 455
third left parapodium back nearly to the anal end in a some-
what oblique course.
The dorsal blood-vessel is plainly seen in this process and
left terminal as well as in the main trunk and in the main
terminal. The vessels contract in each terminal with about
the same rhythm, so that blood flows simultaneously from each
forward into the anterior region of the body.
In other respects the lateral process has the normal appear-
ance of a normal terminal, excepting that the anal piece and
the anal cirri are smaller than in the main terminal, appearing
to be young.
The ventral surface shows a ventral blood-vessel that bifur-
cates to enter each terminal, and there is also an indication, in
the live specimen, that the ventral nerve-cord also bifurcates.
When the lateral terminal was touched it contracted, and
sometimes there was also a movement of the anterior region, but
rarely was there any movement of the posterior region or
locomotion of the entire animal. If the anal region of the
main terminal was touched a locomotion of the entire animal
was pretty sure to follow.
When not stimulated the animal exhibited but little
motion.
As seen in fig. 16, the number of setigerous somites is 16
anterior to the bifurcation, 14 in the lateral terminal and 25
posterior to it, making a total of 45, of which 31 are in the
main axis and 18 in the lateral duplication.
The entire length, in life, was 103 mm.; the length of the
process 2 mm. This process springs from the 17th seti-
gerous somite at a point 4 mm. posterior to the head end
and 63 mm. from the anal end.
An examination of the internal anatomy shows that the
process has the normal organs; its digestive tract is a branch
of that in the main trunk; the nerve-end is a branch of the
main cord ; the body-cavity is filled with a number of sperm
mother-cells. There is a complete bifurcation or branching of
all longitudinal organs in such fashion that each terminal gets
one branch; those to the smaller terminal appear as side
456 E. A. ANDREWS.
branches from the main system found in the trunk and
directly continued into the main terminal. The side terminal
repeats all the organs of the main terminal, having digestive
tract, nerve, blood-vessels, muscles, reproductive cells (male),
and anal opening, just as in the main terminal. It is an
exact repetition, and differs only in small size, fewer somites,
and abnormal coloration upon its left dorsal half.
The following table will serve as a summary of the general
character of the above eight abnormal specimens of Podarke.
The first column gives the number of somites anterior to the
bifurcation, the second the number in the right terminal, the
third those in the left, the fourth the entire number in the
chief or main axis, while the entire number of all the somites
is given in the last column.
Aig By dashing sear, Lanse Eee
B; Dine iy SQ eee AK edb eG
CG, J 1S. ett 98) oe wllOe. 2acSb ene
Ne 2 DO nth aE OO be ope fOr ones tak O pee
W, 3 Bb ee Ose 8 ee ee
x, OO ee Sf et LS. oe werden) aD
yy, 9 OF sw Ee Oe 8a ee 40-5
(Ane LGriy.) (eaWBeas WAAR eae: eee
CoNncLUSION.
From the above description we find that these ten cases of
bifid monsters agree with those previously described amongst
the Annelids in that we have to deal with a complete duplica-
tion of part of the body, the main axis and all its appurtenances
being exactly duplicated for a certain distance. The only
exception, and a very marked one, is the earthworm B, which
has a side duplication in which the entodermal part is absent.
This absence of entoderm in what is otherwise a normally
developed and metameric region of the Annelid, is of consider-
able interest with reference to the ccelom theory and its appli-
cations to the explanation of metamerism. When so perfect a
metameric part of an Annelid may be formed, even abnormally,
without the presence of an entodermal core, we have an addi-
tional reason for doubting that the ultimate explanation of
SOME ABNORMAL ANNELIDS. 457
metamerism can be found in the idea that “ gut-pouches” are
the first steps.
The questions as to the time of formation and the causation
of such monsters remain still open. Whether they arise in the
adult or in the embryo is as yet undetermined. Yet the balance
of evidence seems to point to the conclusion that such bifurca-
tions may be produced in the adult.
While such a very unusual, in fact unique, specimen as the
earthworm B might have arisen from such an abnormal larval
form as figured and described by Mr. Wm. E. Ritter! for
Polygordius, we might just as well imagine it formed as the
result of abnormal regeneration after traumatic interference. All
the other cases may be easily explained upon the assumption
that removal of the normal terminal, or injury to it, has resulted
in regeneration of such abnormal sort that more than the last
part has been re-formed, either two complete terminals to
replace one, or else one complete terminal instead of the
injured part of the normal one.
Some cases lend themselves to this explanation more easily
than to any other ; in fact, some cannot be explained upon any
other assumption than that the duplication has arisen in the
adult. Such a case is that of the doubled-headed Annelid
observed by Langerhans, which, as previously noticed,” could
scarce have attained such dimensions without the former
presence of a larger head that was then replaced by two small
ones in abnormal regeneration.
Again, in the earthworm B there are indications that the
entire region posterior to the 59th somite has been regenerated,
which would mean that the bifurcation was formed in a late
stage. Then, in the eight cases studied in Podarke, anatomical
and physiological facts point to the conclusion that in five of
them, namely A, B, C, W, Y, one of the terminals has no, or
no direct, nervous connection with the head-end ; while its well
developed and large nerve-cord ends abruptly, as seen in fig.
12, in such a way as to force upon us the idea that there was
1 ‘American Naturalist,’ December, 1892, pp. 1047—50,
2 *Thid., September, 1892, p. 730,
458 E. A. ANDREWS.
formerly a connection of this nerve-stump with the main
nerve-cord, an idea that is strengthened by the often immature,
young appearance of the other terminal. We naturally suppose
here that a wound has healed over in such a fashion that the
old nerve-cord does not reunite, but that the distal end remains
as an isolated stump while the proximal end grows down into
the new terminal that is formed, abnormally, in place of the
injured tissue, or at the place of injury to the tissue of the
normal terminal.
As there thus seemed some probability that operative inter-
ference would result in the regeneration of the terminals in
some cases, many experiments were performed on earthworms
and upon Podarke. The results were negative. After the most
various injuries,—removal of the posterior end, lateral, dorsal,
ventral injuries, hemisection, cutting or cauterising,—there
resulted a normal healing of the wound or else the regenera-
tion of a normal terminal, never a duplication. This, however,
is not astonishing, considering the rare occurrence of such
duplication even in nature, where, no doubt, similar experiments
are constantly being made. In fact, if the hypothesis be true
that such bifurcations arise in process of regeneration, we
would not expect it to occur often, as in nature both Allolo-
bophora fetida and Podarke obscura very frequently
lose and regenerate the posterior end, but very rarely produce
a bifid termination. Thus, though such negative experiments
tend rather to impress one with the remarkable tenacity with
which the organism adheres to its normal form, and suggest to
one’s mind the presence of some controlling law like that
denominating the formation and regeneration of a crystal, they
furnish no evidence against our assumption.
Granting provisionally that bifurcation in Annelids in all
cases may be, and in some cases must be, due to abnormal pro-
cesses of regeneration, we may advance the working hypothesis
that these processes consist in the interference with the
normal position or conditions of the cells concerned in rege-
neration. That is, we need not invoke the activity of any
unusual amount or character of idioplasm, no special manipu-
SOME ABNORMAL ANNELIDS. 4.59
lation of “‘ supplementary determinants” seems necessary if we
seek to apply here the facts worked out for the cleaving ovum
by Driesch and by Wilson. As there the partial separation of
cells or their affectation by various external agents may de-
termine their action in groups to produce more or less separate
individuals in place of one single individual, so here we may
suppose that injuries and other external agents affect the re-
generative tissue so that the same cells that else had formed
one normal terminal now form two more or less separate
ones.
Jouns Hopkins UNIVERSITY,
BattimmorE; December 7th, 1893.
EXPLANATION OF PLATES 32—34,
Illustrating Mr. E. A. Andrews’ paper, “Some Abnormal
Annelids.””’
All the figures are drawn with a camera lucida excepting Fig. 1.
Figs. 1—10 are of Allolobophora fetida. Figs. 11—16 are of
Podarke obscura.
Fie. 1.—Dorsal view of specimen A taken soon after death in weak alcohol.
Fic. 2.—Dorsal view of the posterior end of A, much magnified.
Fic. 3.—Ventral view of the same region.
Fic. 4.—View of the longitudinal muscles of the body-wall at the region of
union of the bud B, the tail T, and the anterior region H of this same speci-
men; seen from the ventral aspect. Three dorsal pores are shown in the non-
pigmented intersegmental regions.
Fic. 5.—Somewhat diagrammatic longitudinal view obtained from sections
of the region of union of the two terminals and the trunk, specimen A.
Fic. 6.—Dorsal view of specimen B,
Fic. 7.—Dorsal view of the posterior part of specimen B, much enlarged.
460 E, A. ANDREWS.
Fic, 8.—Ventral view of the posterior part of specimen B, much enlarged.
Fic. 9.—Transverse section of the left terminal, or lateral process of speci-
men B near its posterior tip, showing absence of digestive tract, the character
of the body-wall, nerve-cord, ventral sete, and nephridia.
Fic. 10.—Longitudinal horizontal section of the region of union of the
two terminals and the trunk of specimen B, showing arrangement of body-
wall, body-cavity, digestive tract, septa, and nephridia.
Fic. 11.—Right side of the posterior end of Podarke A, from an alcoholic
specimen. The dorsal terminal should appear as if somewhat rotated towards
the. right.
Fie, 12.—Vertical median section of Fig. 11, showing the bifurcation of
the digestive tract and the interrupted nerve-cord of the dorsal terminal, as
well as the connected nerve-cord of the ventral terminal.
Fie. 13.—View of part of the right side of Podarke V to show the small
right terminal from its dorsal aspect.
Fic, 14.—Dorsal view of one of the abnormal parapodia of the main trunk
of V, showing the posterior process x.
Fic. 15.—Side view of Podarke W from an alcoholic specimen with pro-
boscis partly extended.
Fie. 16.—Dorsal view, much enlarged, taken from the specimen Z in life.
STUDIES ON THE NERVOUS SYSTEM OF ORUSTACEA. 461
Studies on the Nervous System of Crustacea.
By
Edgar J. Allen, B.Sc.Lond.
With Plates 35 and 36.
I.—Some Nerve-elements of the Embryonic Lobster.
Tue observations to be recorded in the following paper were
made at the Laboratory of the Marine Biological Association
in Plymouth, with the assistance of a grant made me by the
Government Grant Committee of the Royal Society. My
thanks are due to the officials of the laboratory for their con-
stant support, to the committee of the British Association,
Robert Bayly, Esq., of Plymouth, and Professor E. B. Poulton,
for supplying me with nominations to the laboratory, and to
Professor Weldon for advice and help.
The remarkable results obtained by recent investigators of
the nervous system of Vertebrates, by making use of embryos
and very young animals, led me to try a similar plan in an
attempt to obtain an insight into the structure of the Crustacean
nervous system. The two methods used have been Ehrlich’s
methylene blue method, as modified by Biedermann! and
Apathy,” and the rapid method of Golgi, as modified by Ramon
y Cajal, Kolliker, and Nansen. Of these, up to the present,
1 Biedermann, “Uber den Ursprung und die Endigungsweise der
Nerven in den Ganglien wirbelloser Thiere,” ‘Jena. Zeitschr.,’ Bd. xxv,
1891.
2 Apathy, ‘ Erfahrung in der Behandlung des Nervensystems fiir histolo-
gische Zwecke. I. Methylen blau,” ‘ Zeitschr. wiss, mikr.,’ Bd. ix, 1892.
462 EDGAR J. ALLEN,
the methylene blue has yielded by far the most satisfactory
results, and the observations recorded in the present communi-
cation have been made by its means.
After experiments upon a number of embryos and larve of
smaller Decapods (Palemon, Palemonetes, Crangon,
Virbius), which failed largely on account of the rapidity with
which the fine fibres assumed the beaded form, and subse-
quently broke up under the influence of the reagent, an
excellent object was found in embryos of the lobster (Homa-
rus vulgaris) The fibres are here much coarser, and
appear to be able to retain their normal shape for a longer
time. Another great advantage of lobster embryos is their
size and the comparative ease with which the thoracic ganglia
can be exposed. All that is necessary for this purpose is, with
a pair of needles, to burst the yolk and remove it, together
with the investing membranes of the egg. It is well also to
remove the connective-tissue bands (endosternite) which lie
across the thoracic ganglia, especially over the portion imme-
diately behind the esophagus. With a little practice this may
be readily done in healthy embryos without injury to the
underlying ganglia. These operations may be performed in
sea water, or in dilute methylene blue solution.
With regard to the strength of the methylene blue, it has
been my practice to keep a standard solution of ,, per cent.
in normal salt solution, and this has been diluted with fifteen
or twenty volumes of a mixture of three parts of sea water to
one of fresh immediately before use. The staining is most
satisfactory when the temperature is from 20° to 25° C.
The embryos, after being prepared as described, are placed
on a slide, with the thoracic ganglia uppermost, and covered
with methylene blue solution. The process of staining may
then be watched under the microscope from its commencement.
For this purpose no cover-glass should be used, and the object
should be illuminated with an Abbé condenser, the diaphragm
of which is left open.
The stain enters the ganglia in two ways: (1) through the
lateral nerves, and (2) through any wounds which may exist in
¢
7
STUDIES ON THE NERVOUS SYSTEM OF ORUSTACEA. 463
the ganglia themselves. By taking advantage of the latter
fact, many of the results to be recorded have been obtained.
The ganglionic cord has been injured or cut across in various
places, and some elements can be got to stain with considerable
certainty by wounding in particular places. No very definite
directions, however, can be formulated on this point, but a
little practice will give the desired result. Accidental surface
wounds, made when preparing the embryo, are often of great
value in this way.
The preparations may be fixed in a solution of ammonium
picrate containing an excess of ammonium carbonate, and
mounted in glycerine diluted with an equal volume of the fix-
ing solution. A large number of preparations were fixed in
this way. They do not appear to be quite permanent, but
preparations which I have kept for five months, although they
have lost some of their original brilliancy, are still of great
value. On the whole, however, it was found more practical
not to fix at all, but to examine a very large number of prepara-
tions in the fresh state. A number of diagrams of the ganglia
similar to the ground plan of fig. 1) were traced, and into one
of these the results of a preparation were drawn. In this way
from fifteen to twenty preparations could be made, and their
results recorded, in a day. Not only is the number of pre-
parations made greater, but the results obtained are more
numerous than if the preparations were fixed, for in order to
obtain good permanent preparations it is essential to stop the
staining process when it is at its best. If, however, the embryo
be allowed to remain in the blue solution, although the colour
in most of the elements fades, individual fibres, and especially
fresh cells, may continue to stain, and the results thus obtained
are often the most valuable.
In dealing with the results obtained by methylene-blue
staining, it must be always borne in mind that one can never
be quite sure that the whole of the element has taken up the
blue. With this reservation, the results to be recorded, which
have been obtained after examining many hundreds of pre-
parations, are those only which I regard as practically certain,
464 EDGAR J. ALLEN.
More doubtful points will be reserved for further investigation
and a future communication.
Embryos at various stages of development have been used,
from the time when the eye-pigment first becomes visible
until the time of hatching. An element which has once
stained at any particular stage is found in practically the
same condition at all later stages, excepting that the finer
branches become more numerous, and, although the results
as given all apply to late embryos, they have been obtained
from embryos at different stages. For instance, in tracing
certain fibres, which travel the whole length of the cord,
although they can be readily recognised in late stages, they
have only been actually traced throughout their entire length
in much earlier ones. As a general rule, a fibre is most easy
to trace through its whole length at the earliest stage at which
it will stain.
In the lobster embryo, as in the embryos of other Decapod
Crustacea, the thoracic ganglia are fused together into one
mass (figs. 1—4), which is united to the brain by a very short
commissure (fig. 1, com.). At this stage the whole of the
commissure which exists is the portion which in the adult lies
between the commissural ganglia, from which the stomato-
gastric nervous system arises, and the first thoracic ganglion,
the stomatogastric commissural ganglia (fig. 1, st. com. gang.)
being now immediately connected with the ganglia from which
Antenne II are innervated (fig. 1, Ant. II), and forming part
of the brain. The posterior portion of the brain of one side is
connected with that of the opposite side by a transverse bridge
(tr. br.), which lies immediately behind the cesophagus, and
which in the adult lies behind the stomatogastric commissural
ganglia. Eleven ganglia may be recognised in the fused mass
in the thorax. Of these the first six go to form the anterior
thoracic ganglion of the adult. In the present paper, for the
sake of clearness, the ganglia will be distinguished as Th. I—
XI. The ganglion cells of each of these eleven ganglia are
arranged in four masses (blue in fig. 1), two lateral and two
median, a median dorsal and a median ventral mass, one of the
STUDIES ON THE NERVOUS SYSTEM OF ORUSTACEA. 465
latter only being shown in the figure. The spaces between the
ganglion cells (left white in fig. 1) are occupied by masses
of nerve-fibre, the “ punkt-substance” of Leydig, or to adopt
the more appropriate name suggested by His,' the neuro-
pile.
When, shortly after the hatching of the larva, the thoracic
ganglia separate, the median masses of ganglion cells of the
embryo divide transversely, the anterior portion going to one
ganglion, the posterior to the ganglion immediately behind it.
In fig. 1 the dotted line between Th. VI and Th. VII represents
the line of this division.
In the abdomen the six ganglia are separate in the embryo,
as in the adult.
In the following description a ganglion cell, together with
the fibre springing from it, will be termed a nerve element.
These elements may be divided into three main groups:
I. Elements which lie entirely in the ganglionic chain, and
which must be supposed to serve the purpose of co-ordinating
the action of its various parts.
II. Elements which consist of a ganglion cell in the cord,
and a fibre which runs out at a lateral nerve-root. Some at
least of these elements, possibly all, are connected with muscles,
and are motor elements.
III. Sensory elements, consisting of cells outside the
central ganglionic chain and fibres running from them to a
ganglion.
I. Co-orpinatinc Exements. (Figs. 2—4.)
These elements may be divided into four classes:
A. Elements (fig. 2) made up of a cell in the brain or one of
the ganglia, and a fibre which runs posteriorly to the end of
the cord, giving off in most cases collateral branches to the
neuropile in each ganglion through which it passes.
B. Elements (fig. 3) made up of a cell in one of the ganglia
1 His, ‘‘ Histogenese und Zusammenhang der Nervenelemente,” ‘ Archiv
Anat, u. Entwickelungsgesch. Anat. Abt.,’? Supp. Bd., 1890.
466 EDGAR J. ALLEN,
of the cord, giving off a fibre which runs anteriorly to the
brain.
C. Elements (fig. 4 red) made up of a cell in one ganglion,
and a fibre which runs posteriorly and ends in the next gan-
glion of the cord.
D. Elements (fig. 4 blue) made up of a cell in one of the
ganglia of the cord and a fibre which runs anteriorly, giving
off collateral branches in the next ganglion, and ending in the
next ganglion but one.
\ A.
\\ These elements are represented in fig. 2. In this figure
only cells of the left side have been drawn, but exactly similar
elements exist starting from cells on the right side. The
figures 1—4 are necessarily somewhat diagrammatic, but in
almost all cases the individual elements are drawn from actual
preparations, and I have endeavoured to make the diagrams
represent as nearly as possible the actual appearance of pre-
parations.
The elements of this class, for reasons to be explained, must
be regarded as only placed provisionally in Group I. It may
be necessary subsequently to place them in a group by them-
selves. Two kinds of A elements may be distinguished—(a)
those whose fibres decussate (blue in fig. 2), and (4) those
whose fibres pass down on the same side of the cord (red in
fig. 2).
Staining of the following elements of this class has been
obtained :—
A (a) Br.—A pair of elements each consisting of a large
cell on the ventral surface of the brain, from which a moderately
thick fibre at first runs forwards and upwards to the dorsal
surface. After turning outwards the fibre runs backwards to
a point immediately in front of the cesophagus, where it passes:
across to the other side and runs down the ganglionic cord.
On entering the first thoracic ganglion the fibre becomes very
broad, having a diameter many times greater than that of any
other fibre in the body, and, maintaining this exceptional size,
STUDIES ON THE NERVOUS SYSTEM OF CRUSTACEA. 467
runs down the cord to the sixth abdominal ganglion. Here
the fibre divides into several branches (fig. 5, A (a) Br), but
_I have never obtained complete staining of these.
According to Retzius,! in the adult Astacus the branches of
these fibres pass directly into the nerves which leave the
posterior end of the ganglion. If this be the case, the elements
must be placed in a class by themselves, as they serve the
purpose of putting some organ at the posterior end of the
abdomen into direct communication with the brain. This is
rendered more probable by the fact that in their course down
the ganglionic cord the fibres give off no collateral branches,
or at any rate none have ever stained, and I believe none to
exist. In the brain, however, a few branches go to the neuro-
pile (fig. 2), the most important being at the angles of decus-
sation. These fibres, both in their narrow portion in the brain
and in the broad portion in the thorax and abdomen, often,
under the influence of the reagent or from some other cause,
assume the beaded form so frequently observed in nerve-fibres.
This pair of elements evidently corresponds to the pair of
so-called giant fibres in the adult. The two fibres stain very
readily in embryos at all stages, but the cells are a little more
difficult to demonstrate. JI have, however, traced the fibres
into them with absolute certainty on so many occasions that
there can be no doubt as to their position and significance.
They are often best seen by examining the brain, after the
staining is complete, from the ventral surface. Fig. 6 repre-
sents such a view.
A (a) Ant II.—This element commences with a cell in the ©
ganglion of the second antenna. After leaving the cell the
fibre gives off two stout branches to the neuropile, and then
passes through the transverse bridge behind the cesophagus
to the opposite side. Here it turns upwards, gives off a branch
to the neuropile of the brain, and then turning again down-
wards, travels along the ganglionic cord.
This fibre has been traced with certainty as far as Abd. 5,
1 Retzius, “Zur Kenntniss des nervensystems der Crustaceen,” ‘ Biol.
Untersuch.,’ neue Folge i, 1890.
VOL. 36, PART 4,—NEW SER. Ha
468 EDGAR J. ALLEN.
but as it showed no sign of breaking up or ending there, it
probably goes to Abd. 6. This is rendered almost certain from
the fact that the element agrees in other respects with the
members of the series now being described, which have been
actually traced all the way. On its way down the cord, the
fibre gives off collateral branches to the neuropile of the dif-
ferent ganglia.
A (a) Th. I1—The cell of this element lies anteriorly
in the lateral ganglionic mass of Th. I. The fibre gives
rise to a figure of very characteristic shape in Th. I and II,
sending off a peculiar branch, which passes downwards on the
same side as that on which the cell is situated as far as Th. II,
where. it turns inwards towards the centre of that ganglion.
The main fibre after passing to the opposite side turns down
the cord, and has been traced to Abd. 6. The actual ending
in Abd. 6 has never stained. The fibre gives off collateral
branches in the ganglia through which it passes.
A (a) Th. I1.—The cell lies at about the centre of the lateral
mass of ganglion cells. The fibre gives off an upward and
downward branch to the neuropile of the same side, then
passes across to the opposite side, and runs down the cord.
It gives off collateral branches, and has been traced to Abd. 6.
A (a’) Th. II.—This is an element starting from a cell in the
ventral median mass of ganglion cells of Th. II. The fibre
passes upwards and outwards to Th. I, where it decussates and
then runs down the cord on the opposite side. It has been
traced to the end of the thorax, but does not appear to stop
there. It has collaterals, at least during the upper part of its
course.
A (a) Th, III. | —These fibres are similar in their relations
A (a) Th. V. to A (a) Th. II, but they have only been
actually traced to the end of the thorax.
It may be explained here that the great difficulty in tracing
fibres from the thorax to the abdomen of lobster embryos is
due to the abdominal flexure. It is impossible to straighten
out the abdomen without such injury as interferes with the
staining, or causes it to disappear. Hence it is only when
STUDIES ON THE NERVOUS SYSTEM OF CRUSTACEA. 469
one or at most two fibres are stained upon the same side of
the cord that it is possible to trace them with certainty through
their whole course. This, however, does not apply to the giant
fibres (A (a) Br), which can always be easily recognised.
A (b) Br.—This element starts with a cell of moderate size
on the ventral surface of the brain immediately anterior to the
large cell of the giant fibre (A (a) Br). The fibre passes first
forwards and upwards, giving off numerous branches to the
auterior lobes of the brain on both sides, and then backwards
through the brain and down the ganglionic cord of the same side
to the sixth abdominal ganglion. It give off collateral branches
to the neuropile of the ganglia through which it passes.
A (b) Th. I1I1.—The cell of this element lies on the inner
ventral border of the lateral mass of ganglion cells of Th. III.
The fibre runs forwards and inwards, and then forwards and
outwards to Th. II. After giving off a branch which enters Th. I
it turns downwards on the same side of the cord, sending off a
stout branch in Th. III, and has been traced to the end of the
thorax. It gives off collateral branches during its whole course.
A (b) Th. I1V.—The cell lies in the median ventral mass of
ganglion cells, but it is difficult to decide whether it should be
considered as belonging to Th. III or Th. IV. The fibre passes
forwards and outwards toa point in Th. III, where it gives off
a stout transverse arm, which runs to the neuropile of the
opposite side and there bifurcates. The main fibre turns
downwards, gives off a branch which runs forwards to Th. II,
and then runs down the cord on the same side as that on which
the cell is situated. It has been traced to the end of the
thorax, and gives off numerous collateral branches.
B.
These elements are represented in fig. 3. In all cases
similar elements exist upon both sides of the cord, although,
for the sake of clearness, only one is sometimes inserted.
The elements consist of fibres starting from a cell ina thoracic
ganglion and running forwards to the brain. Of this group
the following have stained :
4.70 EDGAR J. ALLEN.
B Th. I1.—The cell lies in the ventral median mass of ganglion
cells. The fibre runs forwards and outwards, and then decus-
sates. After reaching the opposite side it goes forwards to the
brain. It has been traced as far as the cross in the figure.
No collaterals have been observed after the fibre leaves Th. I.
A similar element occurs on the opposite side.
B Th. II.—Cell lies in lateral mass. Fibre gives off a large
arborescent branch in Th. II, and one also which passes down-
wards to Th. III. It then runs across to the opposite side,
turns forwards, and goes to the brain. No collaterals have
been observed, with the exception of those figured.
This element generally forms a conspicuous object when
stained, on account of the comparative stoutness of the fibre at
the part where the branches occur, and the richness of the
arborescence.
B Th. Vi —These elements will be considered together,
Bi Th. Wall, as they appear to be very intimately associ-
ated, ending in the same part of the brain, and being generally
found stained in the same preparations. The cell of each
element lies in the anterior portion of the lateral ganglionic
mass. Soon after entering the neuropile the fibre gives off
two branches, one running forwards and breaking up in the
hinder portion of the ganglion immediately anterior, whilst
the other runs backwards and breaks up in that immediately
posterior. The main fibre turns inwards, crosses its fellow of
the opposite side, and then bends forwards, running close to
the median ganglionic cells, until it enters the brain. In the
brain the fibre continues to run forwards, giving off many
branches, and ends at about the level of the nauplius eye.
These fibres have not been observed to give off collaterals
during their course through the thorax.
The two pairs of elements just described very frequently stain
with methylene blue, and the characteristic figures produced by
their decussation are very useful landmarks in preparations.
B Th. VI.—This element resembles B Th. I more nearly than
B Th. II, V,or VIII. The cell liesin the ventral median mass
of ganglion cells. ‘The fibre curves outwards and backwards,
STUDIES ON THE NERVOUS SYSTEM OF ORUSTACEA. 471
and then forwards and inwards, gives off a branch to the
neuropile of the same side, and then passes across to the other
side. It runs forwards to the brain along the outer border of
the neuropile, and breaks up at a point nearly as far forwards
as the terminations of Th. V and VIII, but somewhat lateral
to them.
No collateral branches have been observed in ganglia other
than that in which the element originates.
C.
These elements are best observed in the posterior ganglia of
the thorax of very late embryos or young larve. Their exist-
ence in practically the same condition can, however, be made
out very much earlier, in fact soon after the eye-pigment begins
to deposit. They are represented in fig. 4 in red, whilst fig. 7 is
a drawing of one of these elements made from a fresh prepara-
tion. Several of them are also shown in fig. 8, which is drawn
directly from a preparation fixed in ammonium picrate.
These elements have been clearly demonstrated in each seg-
ment between Th. VI and Th. XI, and there is evidence of their
existence in the more anterior segments. Starting from a cell
in the anterior portion of the lateral mass of ganglion cells, the
fibre passes inwards and backwards for a short distance, and
then turns till it runs almost directly backwards, giving off to
the neuropile numerous branches, which are themselves
arborescent. After a short backward course it again turns in-
wards, and finally runs directly backwards between the fibre of
series D (blue in fig. 4) and the median ganglionic mass.
Each element ends in a tuft of fibres (figs. 4 and 7), which lies
at the posterior end of the ganglion next to that in which the
element originates, and immediately opposite the terminal
tuft of one of the D elements.
D:
These elements are also best observed in very late embryos.
They appear to be intimately associated with those of group C,
and generally stain in the same preparations. They are repre-
4,72 EDGAR J. ALLEN.
sented in blue in fig. 4, and are also seen in fig. 8, which is
drawn from a fixed preparation. This latter figure illustrates
the extent to which the systems C and D will stain in a favour-
able preparation. The elements can be recognised in six seg-
ments, although in some cases the cell has not stained. The
cells of elements D lie in the posterior portion of the lateral
ganglionic mass. The fibre passes forwards and inwards through
the neuropile, giving off numerous arborescent branches to the
latter. On entering the next ganglion in front, the fibre has
reached the outer border of the median ganglionic mass, and
after giving off a little tuft of branches immediately in front
of the tuft in which one of the elements C ends, it pursues a
direct anterior course, ending in a tuft of branches in the
ganglion next but one to that in which it started. This
terminal tuft lies opposite the terminal tuft of one of the C
(ried) elements, and immediately behind the lateral tuft of the
D (blue) element of the next ganglion. These relations will,
I think, be more readily understood from a study of fig. 4 than
by any further description. It should, however, be stated that
the three tufts standing opposite each other are at exactly the
same level in the cord, being all three in the focus of the micro-
scope at the same time. The fine fibres of which they are
composed are often obscured owing to the rapidity with which
they assume the beaded condition.
Il. Motor ELements.
Under this head will be described those elements which
consist of a cell in the ganglionic cord, giving off a fibre
which, after sending arborescent branches to the neuropile,
passes out from the cord by one of the nerve-roots. Many of
these elements (not improbably all of them) are motor, and in
some cases I have been able to trace the fibre from the cell,
through its whole course, until it breaks up on a muscle.
Such a case is represented in fig. 9, which was drawn from a
preparation preserved in ammonium picrate.!
1 The slight break in the fibre, which was not present when the prepara-
tion was fresh, is, even on an examination of the preserved preparation, due to
STUDIES ON THE NERVOUS SYSTEM OF CRUSTACEA. 473
Fig. 1 contains types of most of the elements of this kind
which have stained. In this figure the general course of the
nerves springing from the ganglionic cord is indicated. In the
more posterior ganglia (Th. [V—VIII) it will be observed that
two nerve-roots exist, one anterior (ant. n. 7.) and one posterior
(post. n. r.). Fibres passing out at the anterior root go for the
most part to the limbs, whilst those passing out at the posterior
root supply the wall of the body, and their general distribution
is indicated. The anterior root is in reality a double root, the
two halves separating at a later stage. In Th. II and III this
separation has already taken place. In fig. 1, for the sake of
clearness, those elements which originate in cells of the central
ganglionic mass are coloured red, whilst those having their
origin in the lateral mass are blue. Each of the ganglia Th. VI
to Th. IX contains exactly similar elements. One element of
each kind which has stained has been drawn upon both sides
in Th. VIII, whilst to render them clearer only one element
of each colour has been inserted in Th. VI, VII, and IX.
The lettering will be continuous with that of the elements
already described.
E. Type: the Blue Element of fig. 1 in Th. VI
and VII.
This element starts from a cell in the anterior portion of the
lateral mass of cells. The fibre runs first inwards, and then
curves downwards and outwards, leaving the cord by the ante-
rior nerve-root. The fibre gives off three main branches to the
neuropile, which gives a very characteristic trident shape to
this portion of the element. In late stages the three main
branches give off numerous smaller ones, but these have been
omitted in fig. 1 for the sake of clearness.
Two or even three elements of this type are sometimes
stained on the same side of a ganglion at one time.
The element has stained in Th. III, Th. V—XI. In fig. 1,
E Th. I1I, Th. V—VIII are inserted.
an accident in putting it up. The drawing has, however, been made from this
preparation, in preference to others, on account of the whole being more
clear and satisfactory.
474, EDGAR J. ALLEN.
The element E, Ant. 2 resembles in some points the members
of this group.
F, Type: the blue element in Th. IX, fig. 1.
This element starts from a cell in the lateral mass, which
lies very near the ventral surface. The fibre passes backwards
and inwards, and then suddenly turns and runs forwards and
outwards until it reaches the level of the anterior nerve-root,
when it turns backwards and outwards and so leaves the cord.
At the points where the fibre turns, branches are given off to
the neuropile. Only one element of this kind has been ob-
served to stain in the same ganglion at one time.
The element has stained in Th. V—IX. In fig. 1, F Th. V,
Th. VIII, Th. IX are inserted.
The element F Th. I may also belong to this group.
G. Type: the red element of Th. VI in fig. 1.
This element starts from a cell in the median mass of gan-
glion cells.' From the cell the fibre passes outwards, and after
a somewhat curved course leaves the cord by the anterior root.
It gives off two main branches to the neuropile, one of which
gives off a third, the whole producing a figure somewhat
resembling the trident figure of element E. Two or three
elements of this type may stain on the same side of one
ganglion.
The element has stained in Th. II, III, VII—IX. In fig. 1
it is inserted in Th. II, III, VI, VIII, IX.
G Th. III is an element which stains in almost every pre-
paration of young stages, and forms a useful landmark in the
cord.
H. Type: the red element in Th. X.
Originates from a cell in the median mass, from which the
1 As has already been stated (p. 5), the median masses of cells in the
embryo divide transversely soon after the larva is hatched, the anterior portion
going to one ganglion, the posterior portion to the ganglion behind it. The
division takes place between the two groups of red cells in fig. 1, Th. VI
and VII, at the level of the dotted line.
STUDIES ON THE NERVOUS SYSTEM OF ORUSTACEA. 475
fibre takes a slightly curved course outwards and downwards,
to pass from the cord by the anterior root. It gives off three
or four small branches to the neuropile. The element stained
in Th. V—X. Inserted in fig. 1 in Th. V, VII, VIII, and X.
J. Type: the red element passing through the
posterior root in Th. VII.
The fibre of this element has a somewhat curious course. It
starts from a very small cell in the median mass, runs for some
distance outwards, and then takes a circular course through
the neuropile, forming a complete loop, after which it passes
outwards to the posterior root. From the upper margin of the
loop three branches are given off, whilst from its lateral side a
straight arm passes inwards and meets a similar arm from the
opposite side. <A direct fusion of the two arms has, however,
never been observed. This element has stained in Th. VII
-—IX, and is inserted in each case in the figure.
K,
This element has been satisfactorily demonstrated up to the
present only in Th. III. The cell lies in the median ventral
mass, and from it the fibre passes outwards for a considerable
distance, then suddenly turns and runs inward until it crosses
its fellow of the opposite side, from which point it continues
its former direction till it passes out at the posterior root of
the ganglion, on the opposite side to that upon which the cell
lies. This element differs, therefore, from the elements pre-
viously described in the fact that fibres of opposite sides de-
cussate,
L.
This is an element in Th. I, and is coloured red in the figure.
The cell lies near the median mass. The fibre passes back-
wards and slightly outwards, then turns upwards and passes
out at the anterior root. A branch is given off at the point
where the fibre turns, which soon bifurcates, one branch run-
ning backwards, whilst the second runs transversely to the
opposite side of the ganglion.
476 EDGAR J. ALLEN.
III. Sensory ExLements.
These are nerve elements in which the cell lies outside the
ganglionic cord.
M.
Up to the present I have only been able to obtain satis-
factory staining of such elements in lobster embryos in the
abdomen. If the ganglionic cord in the thorax be cut across
at any point behind the csophagus, numerous longitudinal
fibres take up the methylene blue, and many of these continue
to stain until, on turning the embryo over, they may be seen
to pass into the abdomen. In some preparations some such
fibres are seen to pass out at the lateral nerves of the abdo-
minal ganglia, and may be traced backwards and upwards to
the dorsal surface of the abdomen, where they end in cells
similar to those described by Lenhossék! and Retzius? in the
earthworm, and by Retzius* in Polychetes and Molluscs.
These cells vary somewhat in shape. Two of them are repre-
sented in fig. 10. The cell itself is generally spindle-shaped,
the distal end either running out and ending in a sharp fibre
(fig. 10, 6), or being flattened out as in fig. 10 (a). In other
cases the distal end of the cell appears rounded, as in the lower
cell of fig. 11, but this may be due to incomplete staining.
The fibre arising from the proximal end of the cell, which
lies on the dorsal surface of the abdomen, is moderately fine,
and as previously stated passes forwards and downwards to the
nerve cord. The cell generally lies vertically over the gan-
glion behind that into which the fibre enters (fig. 11). Within
the ganglion the fibre bifurcates, forming a Y-shaped figure,
one arm of which passes forwards and the other backwards.
As to the endings of these fibres, I regret that I can at present
* Lenhossék, “ Ursprung, Verlauf, und Endigung der sensibeln nerven-
fasern bei Lumbricus,” ‘ Arch. mikr. Anat.,’ 39, 1892.
? Retzius, “Das nervensystem der Lumbricinen,” ‘ Biol. Untersuch.,’
neue Folge iii, 1892.
5 Retzius, “Das nervensystem der Lumbricinen,” ‘ Biol, Untersuch.,’
neue Folge, iv, 1892.
STUDIES ON THE NERVOUS SYSTEM OF CRUSTACEA. 4.77
make no definite statement. When the fibres are stained in
two or three successive ganglia, as is generally the case, the
appearance presented is that given in fig. 11. The fibres from
different ganglia run very close together, and it is impossible
to differentiate them individually. In a few cases, however,
the fibre of only one ganglion has stained. The most interest-
ing of these is represented in fig. 12, A. The abdomen was first
examined at a somewhat early stage of the staining, when in
the ganglion of Abd. II the Y-shaped bifurcation was coloured.
The anterior arm of the Y was actually traced as far as the
posterior end of Th. XI (last thoracic ganglion), but could not
be followed further on account of the flexure. The posterior
arm was traced through Abd. III and through Abd. IV, but the
staining had ceased before Abd. V was reached. After the
staining had proceeded for some time longer, the appearance
represented by Abd. II was that shown in fig. 12, B, and the
corresponding fibres in the other abdominal ganglia were com-
mencing to stain.
As these elements often stain when the ganglionic cord is
cut across at the level of the cesophagus, and when there is no
visible wound behind that point, it seems possible that they
may run forwards as far as the brain, although this is at
variance with the conclusions arrived at by Lenhossék and
Retzius for similar fibres in other Invertebrates.
I may here mention that similar cells to those just described
I have found in large numbers on the wall of the cesophagus of
Astacus.
N.
In Th. I and Th. II there are two nerve-endings in connec-
tion with which no cell has ever stained in the cord, although
the fibres stain with very great frequency. It is not unlikely
that these are sensory nerves, whose cells lie outside the cord.
That of Th. I, which enters at the posterior root, is figured in
fig. 18, whilst that of Th. II is inserted in fig. 1 at N. The
latter is the more striking of the two on account of the fact
that it not only branches in Th. II, but also sends stout
478 EDGAR J. ALLEN.
branches which arboresce in Th. I and in Th. III. The fibre
therefore influences three adjacent ganglia.
I have made numerous observations of certain other nerve-
endings in the thoracic ganglia, which are very possibly the
endings of sensory nerve-fibres, but the account which I could
give of these is at present not sufficiently complete to make it
worth while describing them here, and I prefer to reserve them
for future investigation.
THEORETICAL.
Without entering into a general discussion, which will be
more suitable after the investigations have been continued to
a further stage, I shall now endeavour to draw attention to
some points of interest in the results described above.
In the first place, the elements C and D (fig. 4), which form
what may be termed the short internal connections of the cord,
will be considered. I am of course aware that there is a pos-
sibility that this arrangement of fibres is purely embryonic,
and has not yet reached the active condition. This view, how-
ever, appears to me to be improbable, firstly on account of the
fact that there is practically no change in the arrangement
from quite early embryos, in which the eye-pigment has just
begun to deposit, to the oldest larvze (about one week) which
I have been able to examine; and secondly, because the
elements take up the methylene blue in a way which, according
to present experience, only active nervous tissue does.
At the outset, the question arises as to what is the signifi-
cance of the three tufts of fibres which stand opposite to each
other at the point where each element ends, and of the lateral
branches, which both elements give off to the neuropile, where
they cross. My observations agree entirely with those of
Retzius, Kélliker, and the majority of recent investigators, in
the fact that I have never, here or elsewhere, been able to
observe anastomosis of the fibres of different elements. There
can, however, be little doubt that the view now commonly
held that it is by means of the finer branches that the
nervous energy passes from one element to another is the
STUDIES ON THE NERVOUS SYSTEM OF CRUSTACEA. 479
true one, although in what way this takes place remains
unknown. The constant relative position of these three tufts
of fibres would seem to be inexplicable upon any other assump-
tion, and it appears at least worth while to make the sugges-
tion that the nervous energy resembles a static electrical charge,
in the fact that the discharge takes place most readily through
points. Wherever the methods of Ehrlich and Golgi have been
applied to demonstrate the endings of nerve-elements, the
breaking up of the fibre into finer and finer branches which
end freely has been shown to occur, and the suggestion is
that it is by means of these fine branches that the nervous
discharge takes place.
On the view suggested, each nerve-element resembles an
electrical condenser capable of charging itself, and being
suddenly discharged by an appropriate stimulus. The tufts of
fibres at the ends of the elements C and D (fig. 4) would be
comparable to the “ brushes” of an electrical machine. In
the same way we should suppose that the nervous energy
passes from one element to some other element by means of
the numerous branches given off to the neuropile.
Now it must be borne in mind that the elements repre-
sented in fig. 1, some of which at any rate we know to be
motor elements (cf. p. and fig. 9), also send off numerous
branches to the neuropile, which interlace with the branches
of the elements C and D. One such element has been inserted
in black in Th. VIII (fig. 4). Although the speculation has
perhaps little value, it is at least interesting to consider what
would happen upon the suggested theory, supposing one of the
elements to be in any way stimulated. Imagine, for instance,
the element D Th. VIII to be caused to discharge, either by an
impulse received from a sensory nerve-ending, or by an impulse
coming along a fibre from the brain. The main discharge
would, we must suppose, pass into such an element as E, and
so along the motor nerve to the muscle. A portion, however,
would pass along the element D to the lateral tuft, from which
it might pass to the terminal tuft of the element D of Th. IX
and, we may suppose, cause this element to discharge. Another
4.80 EDGAR J. ALLEN.
portion would pass on to the terminal tuft of Th. VIII D, and
would discharge Th. VII D. In this way the discharge would
pass right along one side of the thorax, and the particular
motor fibre concerned would be stimulated in each segment,
the stimulus being conveyed in its turn to the muscles. If
we suppose the fibres of the element C to discharge into those
of D at the point where the fibres cross, the use of this series
of elements is difficult to understand, as the elements D are by
themselves capable of discharging each segment. There is,
however, an alternative supposition. It may be that the
branches of C influence not either D or E, but some other
motor element, say F. Then by means of the terminal tufts
of the D (blue) elements the C (red) elements would all be
discharged, and they in their turn would stimulate all the F
elements, which would cause the particular muscle innervated
by F to discharge. Thus by the stimulation of the element
D Th. VIII the two muscles innervated by elements E and F
in each segment, upon one side of the thorax, would be made
to contract.
We now pass to the elements A and B, by means of which
the brain and anterior ganglia are connected with the remain-
ing ganglia of the cord (elements A, fig. 2), and a particular
ganglion is put into direct communication with the brain
(elements B, fig. 3). The fact that the fibres of elements A
send out collateral branches into the neuropile of each ganglion
indicates that they control some element which also has
branches in that particular ganglion, and the suggestion would
be that by their means a series of elements are stimulated all
along the body by an impulse from the brain. On the other
hand, by means of elements B a particular ganglion would be
placed in more direct communication with the brain. This
communication will be independent or correlated with a
stimulus to (or from) all the ganglia through which the fibre
passes, according to whether the absence of collaterals is the
true condition of the element, or is due merely to imperfect
staining.
With regard to the giant fibres (A (a) Br) a good deal
STUDIES ON THE NERVOUS SYSTEM OF CRUSTACEA. 481
also depends upon whether or not the absence of collaterals
represents the true condition. This is almost certainly the
case on account of the ease and frequency with which these
elements stain, and the readiness with which their branches in
the brain are demonstrated. It will be remembered, too, that,
according to Retzius, the branches into which the fibres divide
in the last abdominal ganglion pass out through the nerves of
that ganglion, and it therefore seems probable that they serve
the purpose of putting some organ into direct communication
with the brain. It may be that these branches go directly to
the muscles which control the tail fin, so that the steering
apparatus of the animal is under the direct control of the
brain. This, at any rate, is the most obvious suggestion.
It is of course possible that the other elements of the class
A serve a similar purpose, but the fact that they have collateral
branches in each ganglion seems to me to be opposed to this
view.
The above remarks are intended merely as an indication of
the nature of the problems suggested by the observations
recorded in the previous part of the paper, problems which can
only be finally solved by means of physiological research, for
which research, however, an accurate knowledge of anatomical
details is an essential preliminary.
EXPLANATION OF PLATES 35 and 36,
Illustrating Mr. Edgar J. Allen’s paper on ‘Some Nerve
Elements of the Embryonic Lobster.”
Fic. 1.—Brain and thoracic ganglia of Homarus embryo. Motor ele-
ments. @s. Gisophagus. st. com. gang. Commissural ganglion from which
stomatogastric nerves arise. ¢r. dr. Transverse bridge behind csophagus.
com. (sophageal commissure. ant. x. r, Anterior nerve-root. post. n. r.
482 EDGAR J. ALLEN.
Posterior nerve-root. Azé. 17. Ganglion of Antenna II. 7h. J—X. Thoracic
ganglia. J—VJ. Form anterior ganglion of adult. H#—. Elements, chiefly
motor. See pp. 495—499. Somewhat diagrammatic.
Fig. 2.—Brain and thoracic ganglia of Homarus embryo. A. Elements.
Fibres run from cell in anterior ganglion, and pass down ganglionic cord .
For individual elements see pp. 488—491.
Fic. 3.—Ditto, B elements. Fibres run from cell in ganglion forwards to
brain (see pp. 491—493).
Fic. 4.—Ditto, C and D elements. OC. Fibres run from cell backwards
and end in next ganglion. JD. Fibres run from cell forwards and end in next
ganglion but one. E. Motor element. (See pp. 493, 494).
Fie. 5.—Branching of giant fibres in sixth abdominal ganglion. A (a) Br.
Giant fibre.
Fic. 6.—View of brain of Homarus embryo from the ventral surface.
A (a) Br. Cell in which giant fibre originates.
Fic. 7.—Element C. Camera drawing from fresh preparation. Late em-
bryo. xX 250.
Fic. 8.—Posterior thoracic ganglia (Th. VI—Th. XI) of Homarus
embryo. Elements C and D stained. Camera drawing from preparation
fixed in ammonium picrate. x 100.
Fic. 9.—Posterior thoracic ganglion (Th. XI), showing nerve-cell (~. c.) and
fibre going to muscle (musc.). Camera drawing from preparation fixed in
ammonium picrate. 250.
Fic. 10.—Sensory cells in ectoderm of dorsal surface of abdomen of
Homarus embryo.
Fic. 11.—Three abdominal ganglia (Abd. II—IV) of Homarus embryo,
showing sensory elements M.
Fic. 12.—A. Second abdominal ganglion (Abd. II) of Homarus embryo
showing bifurcation of fibre of sensory element M. JB. Ditto, at a later
stage of staining.
Fic. 13.—Nerve-ending in Th. I.
STUDIES ON THE NERVOUS SYSTEM OF CRUSTACEA. 483
Studies on the Nervous System of Crustacea.
By
Edgar J. Allen, B.Sc.Lond.
With Plates 37 and 38.
II.—The Stomatogastric System of Astacus and Homarus.
III,—On the Beading of Nerve-fibres and on End Swellings,
II.
Wirn a view to making myself acquainted with the use of
Ehrlich’s methylene blue method, in order to study the struc-
ture of the nervous system of the larve and embryos of Decapod
Crustacea, I made preliminary trials upon the stomatogastric
nervous system of Astacus. As several points of interest
appeared, I carried the observations further than I had intended,
and I now propose to give some account of them, especially as
I am able to supplement them in one or two particulars by
observations which I have made during my later work on the
embryos of Homarus.
The greater portion of the research recorded here was carried
out in the zoological laboratory at University College, and my
thanks are due to Professor Weldon for the help and advice
which he has given.
Mertnuop.—My attention was chiefly directed to the study of
that portion of the stomatogastric system which puts it into
communication with the central nervous system of the animal.
The following mode of procedure was adopted :—The animal,
after being killed with chloroform, was opened from the dorsal
VOL. 386, PART 4,—NEW SER. KK
484, EDGAR J. ALLEN.
surface, the anterior gastric muscles were carefully cut through,
as much of the green gland as possible removed with forceps,
and the brain freed in front from the carapace. The whole of
the head in front of the mouth was then cut off with a pair of
stout scissors, and the anterior thoracic ganglion exposed from
the ventral surface. When this has been done the whole
stomach can be removed with the cesophageal nerve-ring
attached. In some cases the thoracic ganglion was not
dissected out, the commissures being simply cut across as far
back as possible. The stomach thus removed from the animal
was then placed in normal salt solution, and its anterior face,
with the cesophagus and the anterior gastric muscles attached,
cut off. After the esophagus had been slit longitudinally along
its posterior wall, the whole piece thus obtained was spread out
on a slide with the nerve-ring lying uppermost. The prepara-
tion was then covered with a dilute solution of methylene blue
in normal salt solution. In order to keep the tissue properly
covered, it is an advantage to have a glass or metal cell attached
to the slide. The strength of methylene blue found most use-
ful was from 1: 50,000 to 1: 10,000. It is often good to start
with the more dilute solution, and gradually add the stronger
as the staining proceeds. In cold weather the slide should be
kept warm; indeed, the staining always takes place more
rapidly at a temperature of 25°—30° C.
One great advantage of this method, due largely to Bieder-
mann! and Apathy,” over the original method of Ehrlich, lies
in the fact that the object can be examined with the micro-
scope from time to time, and the staining process watched.
When the staining is at its best it may be fixed by placing the
tissue in a solution of ammonium picrate containing excess of
ammonium carbonate (Apéthy), and, after a few hours, mount-
ing in glycerine mixed with an equal volume of the fixing
solution To obtain good preparations a great deal depends
1 Bieaermann, “ Uber den Ursprung und die Eudigungsweise der Nerven
in den Ganglien wirbelloser Thiere,” ‘Jena. Zeitschr.,’ Bd. xxv, 1891.
2 Apathy, “Erfahrung in der Behandlung des Nervensystems fiir histo-
logische Zwecke,” ‘ Zeitsch, wiss. Mikr.,’ Bd. ix, 1892,
STUDIES ON THE NERVOUS SYSTEM OF CRUSTACEA. 485
upon judging properly the moment to stop the staining process.
If the latter be allowed to continue for too long a time, the
nervous elements lose their colour, and a diffuse staining of
other tissues takes place.
It is necessary to make a large number of preparations to
obtain satisfactory results, and care must be taken not to pull
the tissue about very much in getting it ready. It must also
be so arranged that all the nerves lie freely exposed to the
methylene blue solution, and not covered by other tissues. If
these precautions are attended to, very beautiful results can be
obtained.
The diagram, fig. 1, represents much the appearance of a good
methylene blue preparation, and the individual elements there
recorded are taken from actual preparations. The figure repre-
sents the nerve-ring turned through nearly a right angle until it
comes to lie upon the anterior surface of the cesophagus, as it
will do in preparations made as described above. From the
brain (Br.) the two cesophageal commissures run down upon
either side, and contain numerous stained fibres which often
are beaded in various ways. This beading, which will be after-
wards discussed in more detail, is not the natural condition of
the fibre, and can give rise to a variety of appearances, some
of the more typical being recorded in the commissures drawn
in the figure.
About one third of the way down the commissure is the
ganglionic swelling (com. gang.) from which the nerves of the
stomatogastric system spring. Behind this swelling lies the
transverse bridge (¢7. dr.), which runs from one commissure to
the other, and in the natural position of the parts is situated
immediately behind the esophagus. It is worth noting that
all the fibres observed to pass through this transverse cord
come down the commissures, a few possibly from the com-
missural ganglia, the majority from the brain.
From each of the commissural ganglia two nerves spring,
which unite with their fellows of the opposite side in what is
practically a somewhat diffuse ganglion, which lies in the
middle of the anterior surface of the esophagus, and will be
486 EDGAR J. ALLEN.
termed the cesophageal ganglion (@s. gang.). From this
ganglion the azygos nerve (azy.) takes its origin, and runs up
the anterior face of the cesophagus and stomach to the ganglion
which lies between the anterior gastric muscles. This ganglion
will be termed the gastric ganglion (gast. gang., fig. 2). There
are thus four ganglia connected with the stomatogastric
system—the pair of commissural ganglia, the cesophageal
ganglion, the gastric ganglion.
Returning now to the two pairs of nerves which spring from
the commissural ganglia, and bearing in mind that in the
figure (fig. 1) the brain is supposed to be turned back upon
the cesophagus, it will be seen that one of these pairs lies more
anterior and ventral (vent. n.) (antero-lateral nerve of Huxley’s
description), the other more posterior and dorsal (dors. n.)
(medio-lateral nerve, Huxley).’
The ventral nerve, after giving off close to its point of
origin with the commissural ganglion a small nerve (a) which
breaks up on the wall of the csophagus, runs downwards
towards the mouth for some distance, and then turns upwards
to the csophageal ganglion. At the point where the nerve
turns upwards a downward branch (6) is given off, which
breaks up and, together with its fellow of the opposite side and
a median nerve (c) coming from the csophageal ganglion,
forms a large plexus of fibres (plex.) at the border of the
mouth. An examination of the point where this branch (0)
takes its origin from the ventral nerve shows that the plexus
receives two sets of fibres, some coming from the direction of
the oesophageal ganglion, others from that of the commissural
ganglion. It is also somewhat curious that most of the fibres
which have stained in the median nerve (c) running from the
cesophageal ganglion to this plexus appear to come, not from
the ganglion itself, but through the ventral nerve (vent. n.)
from the direction of the commissural ganglion. What the
meaning of this fact is I do not know.
The ventral nerve (vent. n.) itself obtains fibres from various
sources. Many of its fibres are seen to enter the commissural
' Huxley, ‘Anat. Inv.,’ p. 286.
STUDIES ON THE NERVOUS SYSTEM OF CRUSTACEA. 487
ganglia, and there arboresce, as in fig. 1 on the right side. In
Astacus, however, I never succeeded in demonstrating the con-
nection of any of these fibres with ganglion cells. Many cells
were stained, as in the ganglion on the left side of fig. 1, but
their connections could not be made out. In the lobster
embryo I was fortunately able to fill up this gap in my previous
work. Fig. 2 represents the stomatogastric system of such an
embryo, and the element A is seen to start from a cell in the
commissural ganglion (which in the embryo forms practically
a part of the brain) and enter the ventral nerve.
Other fibres of the ventral nerve come from the cesophageal
ganglion. Figs. 3 and 4 represent enlarged views of this ganglion
drawn directly from single preparations. Two kinds of cells
are found in it, both of which send fibres to the ventral nerve,
the one kind being unipolar, the other bipolar. The bipclar
cells (figs. 1 and 4, C,, C,) are spindle-shaped cells sending off
a fibre from each end. One fibre runs upwards to the azygos
nerve (azy.), the other downwards through one of the ventral
nerves. Some of these cells lie in the portion of the ventral
nerve which adjoins the ganglion (figs. 1, 4, C,) rather than in
the ganglion itself. The unipolar cells (figs. 1, 3, 4, D, E, M)
give off a single branch which, however, generally soon
divides, one of the branches bifurcating again almost imme-
diately, so that there are really three fibres having their origin
in the one unipolar cell (figs. 3, 4, D,, D,, Ds). Of these three
fibres, one runs upwards to the azygos nerve (D,), whilst the
other two run downwards through the ventral nerves to the
commissural ganglia (D,,D,). It is probable that by means of
these elements the gastric ganglion and the two commissural
ganglia are put into direct communication. In fig. 4 a uni-
polar cell (£) is drawn, which gives off one branch which does
not divide but passes directly into the right anterior nerve.
It is not improbable, however, that this is due simply to in-
complete staining, and that the element is in reality of the
same type as that already described (D).
In fig. 3 is a cell (F) which appears to be closely applied to
a single fibre. This probably represents an intermediate stage
488 EDGAR J. ALLEN.
between the unipolar and bipolar type of cell. It may be
regarded as a bipolar cell in which the two poles have come to
lie closely together, or as an unipolar cell the single process of
which immediately bifurcates. It is interesting to note that
in the lobster embryo all the cells of the cesophageal ganglion
(fig. 2, r) which have stained, are of this type.
His! has shown that the cells of the spinal ganglia of human
embryos are at first bipolar, giving off one nerve process to the
cord, the other to peripheral parts. In later stages the body
of the cell becomes excentric to the axis passing through the
two fibres, and by degrees a single process with a T-shaped
fibre is formed. V. Lenhossék* has found both bipolar cells
and unipolar cells with a T-shaped fibre in Pristiurus embryos,
and discusses the transition of one form into the other.
Ramon? also notes the fact that in embryonic birds the cells of
the spinal ganglia are all bipolar, whereas in adult mammals
they are unipolar with a bifurcating fibre.*
In addition to those already mentioned, fibres from several
other sources enter the anterior nerve. Some fibres pass down
from the azygos nerve (hence probably from the gastric
ganglion) directly into one or other of the anterior nerves.
Such fibres are represented (c) in fig. 3, and they are also in-
serted in fig. 1. A fibre, which passes from the dorsal nerve of
one side to that of the other (fig. 1, m), has also been observed
to give off a branch which passes down one of the ventral
nerves.
Lastly, the esophageal ganglion is connected directly with
the brain by an anterior median nerve (ant. med. n., figs. 1,
3, 4) which Huxley describes as “ the anterior continuation of
the azygos nerve.”* Fibres passing from the brain along this
1 His, “ Histogenese und Zusammenhang der Nervenelemente,” ‘ Arch.
Anat. u. Entw. gesch. Anat. Abt.,’ Supp. Bd., 1890.
2 'V. Lenhossék, “ Beobactungen an den Spinalganglien und den Riick-
enmark von Pristiurusembryonen,” ‘ Anat. Anz.,’ 7, 1892.
3 Ramon y Cajal, “Sur l’origine et les ramifications des fibres nerveuses
de la moelle embryonnaire,” ‘ Anat. Anz.,’ 5, 1890.
4 Compare also Retzius, ‘ Biol. Untersuch.,’ neue Folge iv, 1892.
§ I have only observed one such nerve (cf. Huxley, ‘ Anat. Inv.,’ p. 287).
STUDIES ON THE NERVOUS SYSTEM OF CRUSTACEA. 489
nerve bifurcate in the cesophageal ganglion, a branch then
passing into each of the ventral nerves (figs. 1 and 3, H). It
is probable that in this way the commissural ganglia are
placed in direct communication with the brain. As, however,
I have not followed one of these fibres individually to the com-
missural ganglion, there is of course the alternative that it
may pass out at the brauch (0) and enter the plexus (plez.)
at the border of the mouth.
Passing now to the dorsal nerves (dors. n.) arising from the
commissural ganglia, I was here again able to complete the
observations I had made on Astacus by those on the lobster
embryos. In the latter, fibres were often observed starting
from cells in the commissural ganglia (fig. 2, B), and after
giving off numerous arborescent collateral branches to the
neuropile, passing into the dorsal nerve and so to the
azygos.
After leaving the commissural ganglion, the dorsal nerve
gives off a branch (d) which runs upwards on the wall of the
cesophagus and soon breaks up into a large number of in-
dividual fibres, each one having in its course a moderately
large bipolar cell (3, fig. 1). These cells exactly resemble
sensory cells described by Retzius! in many Polychetes and
Molluscs.
As near as I could determine in preparations in which the
general tissue was not stained, they lie on the exterior surface
of the cesophagus, whilst the distal fibres proceeding from the
cells, sometimes after a very long course, end in a much deeper
layer, probably on the interior face of the esophagus. The
proximal fibres from these cells all enter the branch (d) of
the dorsal nerve, and the fibres of this branch enter the com-
missural ganglion.
A few fibres passing through the branch (d) break up on the
1 Retzius, ‘ Biol. Untersuch.,’ neue Folge iv, 1892: 1. “ Das Sensible
Nervensystem der Polychaten;” 2. ‘‘ Das Sensible Nervensystem der Mol-
luscen.” Compare also Lenhossék, ‘Arch. mikr. Anat.,’ 39, 1892, and
Retzius, ‘Biol. Untersuch.,’ neue Folge iii, 1892, for similar cells in Lum-
bricus.
490 EDGAR J. ALLEN.
wall of the csophagus as at K (fig. 1), that is to say, in the
typical manner of motor nerve-endings.
At a slightly higher point to that from which the branch
above described arises, the dorsal nerve gives off a second
branch (£), which runs forwards on the esophagus, and breaks
up into a mass of fibres, each of which has a bipolar ganglion-
cell in its course (LZ). These cells are confined to two limited
areas on the wall of the cesophagus, which may almost be re-
garded as definite sense-organs. The cells of the first branch
() are much more scattered than these; indeed, they may be
found upon all parts of the cesophageal wall. These cells of
the second branch are similar in shape to those of the first,
but are much smaller, and the fibres springing from their
distal ends are shorter. Their fibres can be readily traced
through the wall of the esophagus to its inner surface, where
they appear to end. From the position of these elements it is
not unlikely that they form organs of taste.
After giving off these two branches, the dorsal nerves unite
with each other and with the cesophageal ganglion, or rather
unite to help to form the esophageal ganglion, for it would
seem that the whole tract of nervous tissue where the ventral
and dorsal nerves unite should be included under that name.
The ventral portion alone contains ganglion-cells, but in the
dorsal portion the fibres give off numerous collateral branches
(fig. 1, mew.), forming a small neuropile, so that this part must
also be regarded as belonging to the ganglion.
Several kinds of fibres enter the dorsal nerves from the ceso-
phageal ganglion. Some fibres spring directly from cells in
the ganglion, as at fig. 8 m, whilst others take their origin from
fibres which come down the azygos nerve, bifurcate in the
upper part of the cesophageal ganglion, and send one branch
down each dorsal nerve (figs. 1—3, n). One such fibre in the
lobster embryo (fig. 2, N) has been traced to a bipolar ganglion-
cell which hes at the anterior end of the gastric ganglion.
Some fibres pass directly from the azygos nerve into one of
the dorsal nerves (fig. 3, 0).
The azygos nerve itself, after leaving the cesophageal gan-
STUDIES ON THE NERVOUS SYSTEM OF CRUSTACEA. 491
glion, passes up the anterior face of the stomach to the gastric
ganglion. This ganglion is spindle-shaped (fig. 2, gast. gang.),
and from the opposite end of the spindle to that at which the
azygos enters a stout nerve is given off, which bifurcates
on the dorsal surface of the stomach (fig. 2). The gastric
ganglion of Astacus is enclosed in a connective-tissue sheath,
which can with care be removed. When this is done, the large
unipolar cells can be seen protruding on either side, each in-
vested in a nucleated sheath. One such cell with its sheath
is represented in fig. 5. The central portion of the ganglion
is occupied by a mass of neuropile made up of the arborescent
branches springing from the processes of the ganglion cells,
and the arborescent ends of fibres which enter the ganglion by
the azygos nerve. As far as I have observed, only one fibre
passes straight through this ganglion, namely, the fibre (NV)
which springs from the bipolar ganglion-cell already described
(fig. 2, N) in the lobster embryo, and which is also frequently
stained in Astacus. All other fibres appear either to break up
in the neuropile of the ganglion, or to be derived from the
ganglion cells.
Three pairs of lateral nerves spring from the gastric ganglion.
These are represented in fig.2. The same figure shows several
fibres starting from ganglion-cells and entering these nerves.
In the adult Astacus the processes of the cells of the gastric
ganglion are remarkable for the extraordinary richness of
the arborescence of the branches which they give off to the
neuropile. Figs. 6 and 7 represent two such cells with their
processes. The cell of fig. 6 gives rise to one main fibre, which
passes backwards to the bifurcation on the dorsal surface of
the stomach. The figure, which is a camera drawing ren-
dered as accurately as possible in all its details, illustrates
not only the remarkable richness of the arborescence, but
also the characteristic forms in which the finer branches
terminate.
Fig. 7 represents a cell from which two main fibres originate.
Looking at the stomatogastric system as a whole, it is pro-
bable that there are three main centres from which motor
492 EDGAR J. ALLEN.
fibres start and in which sensory fibres end, namely, the two
commissural ganglia and the gastric ganglion. The two com-
missural ganglia are in communication with the central
nervous system by means of fibres which enter it from the
commissures, and have also a special direct communication
with the brain by means of fibres which pass through the an-
terior median nerve, and bifurcate in the cesophageal ganglion
(figs. 1, 8, oH). The gastric ganglion is placed in common
communication with both commissural ganglia by means of
elements originating in cells in the cesophageal ganglion, the
processes of which divide into three main branches (D). It is
also placed in independent communication with each commis-
sural ganglion by means of elements originating in cells of the
latter, and running through the esophageal ganglion.
III.
I propose to add here a few remarks upon the varicose or
beaded appearance so often presented by nerve-fibres in pre-
parations made by the methylene blue method, or by the
method of Golgi, and on the swellings which occur on the
finer nerve-endings. These contain the conclusions arrived at
from observations on the adult Astacus, on embryonic lobsters,
and on the larve of Palemon and Palemonetes.
In the two cesophageal commissures of Astacus in fig. 1, I
have inserted numerous typical beaded fibres, which illustrate
the great variety of appearances presented. In some cases the
beads are almost spherical and are joined together by long,
thin threads; in others they are elongated spindles, joined by
shorter threads, whilst in yet other cases they form long
cylinders with pointed ends, the individual ones hanging to-
gether by the points.
Figs. 8, 9, 10, and 11 are drawn on a larger scale from pre-
parations stained with methylene blue and fixed with ammo-
nium picrate.
Fig. 8 represents a fibre in which many of the beads are
STUDIES ON THE NERVOUS SYSTEM OF CRUSTACEA. 493°
almost spherical, and are connected by moderately long
threads ; whilst in fig. 9 they have the form of elongated
spindles, and the threads are shorter. In both cases it will
be noticed that the beads vary greatly in size and shape, and
that the length of the connecting threads also varies.
Fig. 11 represents a nerve-fibre ending in the commissural
ganglion, in which the beading has been carried to a somewhat
remarkable extent, out of all proportion to the size of the
fibre.
Fig. 10, a, B, c, represents three portions of one fibre, which
was stained in an cesophageal commissure of Astacus. Near
the brain the fibre is practically cylindrical (first portion of A).
The undulations become more strongly marked, producing the
appearance of elongated spindles lying close together (B) as
we pass further back; whilst near the commissural ganglion
the spindles are joined together by fairly long threads, and the
result is a typical beaded fibre (C). Thus we have the various
stages all represented on a single fibre. Such cases are not
uncommon, and they alone make it practically certain that the
cylindrical form is the normal condition of the fibre, the beaded
condition being due to some external cause. It was observed
that when the object was unduly stretched or pulled about in
the process of preparation, the number of beaded fibres very
largely increased.1
My observations on embryonic and larval forms have been
1 The larger fibres, when stained with methylene blue and preserved in
ammonium picrate, generally have the surface covered with deeply stained
granules, which are either arranged irregularly, as in figs. 8—11, or may lie in
longitudinal rows, as in fig. 7. It is not possible to say, in my opinion,
whether these are entirely artificial, or whether they may be taken as an
indication of the structure of the protoplasm composing the fibre. The latter
view has been held by Dogiel (‘ Arch. mikr. Anat.,’ 41, 1893), who appears to
consider the fibrillar structure as demonstrated. I should like to draw the
attention of those interested in this question to the large black chromatophore
masses of Mysis flexuosa. When these are expanded, the pigmented pro-
toplasm in the larger processes is arranged in parallel strands. After soaking
for some time in sea water, these strands break up into short pieces and give
rise to an appearance which resembles very closely that represented on the
nerve-fibre in fig. 7.
494 EDGAR J. ALLEN.
even more conclusive in showing that the cylindrical form is the
normal condition of the fibre. In the lobster embryos, where
I have been able to identify a good many different elements, it
has been observed that the same fibre, although generally
cylindrical, may at times appear beaded, and that a particular
fibre may be cylindrical when it first stains, but afterwards
become beaded. This latter fact, which can be readily ob-
served, places the question beyond doubt. That the result,
however, can be produced by mechanical means, and is not
always entirely due to the reagent, is shown by the following
observation. In the lobster embryo there is a single pair of
fibres running along the whole length of the cord, the so-
called giant fibres, which have a diameter very much greater
than any other fibres of the cord. These can be seen quite
clearly in fresh embryos, before staining has taken place, and
are generally cylindrical. In an embryo, however, which has
been roughly handled in the preparation, I have observed that
these fibres, whilst still unstained, are distinctly beaded.
The rapidity with which a fibre beads appears to depend toa
large extent upon its absolute diameter, and this point is of
great importance in selecting suitable objects upon which to
study the histology of the nervous system. The fine branches
are always the first to assume the varicose condition, and in
them the process of beading is carried to such an extent that
the fibre breaks up entirely into drops, which do not even hold
together. For this reason it is often difficult to obtain stain-
ing of the whole of an element in the lobster embryo at the
same time, the finer branches being completely broken up into
drops before the cell has commenced to stain. The finer longi-
tudinal fibres also become beaded much more quickly and more
frequently than the coarser ones.
In the larve of the smaller Decapods, as for example Pale-
mon, Palemonetes, &c., where the fibres are all very fine, it
becomes impossible to obtain useful preparations on account of
the rapidity with which the beading takes place. The whole
of the neuropile rapidly becomes converted into a mass of small
blue drops, which appear to have lost all continuity. Hence
STUDIES ON THE NERVOUS SYSTEM OF CRUSTACEA. 495
in selecting suitable objects for study, those animals in which
the nerve-fibres are coarsest will be most likely to give good
results.
Ramon! has also noticed, in studying the retina of Verte-
brates, that the more delicate the structure, the more difficult
it becomes to obtain satisfactory impregnation by Golgi’s
method. This tendency on the part of fine nerve-fibres to
break up into drops appears to me to explain, to a large extent,
the impossibility of satisfactorily making out any structure in
the neuropile (punkt-substance) when preserved by ordinary
methods.
Another phenomenon, which at first sight closely resembles
the beading of nerve-fibres, is the formation of small swellings
on the terminal branches of a nerve-ending. These have been
often described, and for a number of typical instances I may
refer to fig. 6, and the finer branches of fig. 11. It will be
seen that these swellings usually occur at the angles, where
fine twigs are given off, and it is not unlikely that this is
always the case, though in some instances the branch has not
stained.
From the facts recorded above concerning the beading of
fibres, especially the direct observation that the finer branches
do assume the varicose appearance much more rapidly than
any others, I should have been led to cousider it probable that
these end-swellings were post-mortem products, and were
found in that particular position so constantly on account of
the fineness of the fibres. In examining living lobster em-
bryos, however, I was struck with the appearance presented by
the terminal branches of the bright red pigment cells or chro-
matophores, which occur in large numbers. Fig. 12 (a) repre-
sents such a chromatophore, whilst fig. 12 (b), (c), (d), (e), and
(f) shows individual terminal branches of other cells. Such
appearances can be observed in favorable cases even before
the embryo is removed from the egg membranes, and when
there cannot be the slightest doubt that the animal is alive
1 Ramon y Cajal, ‘ La Cellule,’ ix, 1893 (quoted in ‘ Journ. Micr, Soc.,’
pt.5, 1893).
496 EDGAR J. ALLEN.
and in its normal condition. It will be seen that the terminal
swellings are of exactly the same nature as those found on the
nerve-endings. (Compare figs. 6 and 12.) Hence this is
clearly a form which fine branches of living protoplasm are
capable of assuming. Under these circumstances, in the
absence of direct observation to the contrary, it seems more
natural to suppose that the terminal swellings on the nerve-
endings also represent the normal condition.
In this connection the recent observations of Ballowitz!
upon the chromatophores of fishes are of interest, confirming
the fact that the retractile pigmented protoplasm represents
only a portion of the protoplasm of the cell process, the process
itself remaining always fully expanded. Should this view be
correct, the question naturally arises, does the portion of a
branching nerve-fibre which takes up methylene blue represent
the whole protoplasm, or only that portion which corresponds
to the pigmented protoplasm of the chromatophore ?
Both the phenomena of beading and the formation of end-
swellings appear to be due to a simple physical cause, namely
the difference of surface tension between two fluids. A fluid
cylinder surrounded by some other fluid of different surface
tension is in a condition of unstable equilibrium, and tends to
break up into spherical drops. A stream of water issuing
through a circular orifice and allowed to fall for some distance,
goes through much the same series of changes as those which
have been described for a beaded nerve-fibre, until it finally
breaks up into spherical drops. I have been able to produce
fibres which have almost exactly the same size and assume the
same variety of shapes as beaded nerve-fibres in a very simple
way. A thick syrup of gum and sugar is made, and a drop of
this is placed in the centre of a glass slide, which has been
covered with a layer of paraffin oil. If fine threads are drawn
from the syrup across the oiled surface of the glass, with the
point of a knife or a rough needle, on examining under the
1 Ballowitz, ‘Die Nervenendigungen der Pigmentzellen,’”’ ‘ Zeitschr.
wiss. Zool.,’ 56, 1893; “ Ueber die Bewegungserscheinungen der Pigment-
zellen,” ‘ Biol. Centr, bl.,’ xiii, 1893, p. 625,
STUDIES ON THE NERVOUS SYSTEM OF CRUSTACEA. 497
microscope these threads will be seen to be beaded, and gene-
rally show all the various forms assumed by nerve-fibres.
EXPLANATION OF PLATES 37 and 38,
Illustrating Mr. Edgar J. Allen’s paper on “Studies on the
Nervous System of Crustacea.”
Fig. 1.—Diagrammatic view of the connections of the stomatogastric
system of Astacus with the central nervous system. Br. Brain. com. (so-
phageal commissure. com. gang. Commissural ganglion. ¢r. dr. Transverse
bridge behind cesophagus. «@s. gang. (isophageal ganglion. azy. Azygos
nerve. ant. med.n. Anterior median nerve. dors. 2. Dorsal nerve. vevt. n.
Ventral nerve. aand 4. Branches of ventral nerve. c. Median nerve from
cesophageal ganglion. dande. Branches of ventral nerve. mew. Neuropile.
C,, C,. Bipolar ganglion cells. D. Unipolar ditto. H. Fibre connecting
brain with commissural ganglia. . Motor nerve-ending. J and Z. Sensory
nerve-cells. WV. Fibre connecting gastric ganglion with commissural ganglia.
O. Fibre connecting commissural ganglia.
Fie. 2.—Ditto, of lobster embryo. Lettering same as Fig. 1. 4 and B.
Cells of commissural ganglia. /. Cell of cesophageal ganglion. gast. gang.
Gastric ganglion. m.z. Nerve from commissural ganglion to region round
mouth.
Fie. 3.—(sophageal ganglion of Astacus. Drawn from one preparation.
Lettering as in Fig. 1.
Fic. 4.—Lower portion of cesophageal ganglion of Astacus, enlarged.
Drawn from one preparation. As in Fig. 1.
Fie. 5.—Ganglion cell from gastric ganglion of adult Astacus, showing
connective-tissue sheath with its nuclei. JV. Nucleus of cell. 2. Nuclei of
sheath. Cell process stained with methylene blue; cell itself only slightly
stained. xX 420.
Fic. 6.—Ganglion cell of gastric ganglion of adult Astacus, showing
branches to neuropile. Methylene blue, fixed with ammonium picrate.
Camera drawing. X 250.
Fic. 7.—Ditto, showing two main fibres (I and II). Methylene blue, fixed
with ammonium picrate. Camera drawing. X 250.
498 EDGAR J. ALLEN.
Figs. 8 and 9.—Beaded nerve-fibres of Astacus. Methylene blue and
ammonium picrate. x 420.
Fic. 10,—Three portions of one fibre from commissure of Astacus. Same
mode of preparation. x 420.
Fre. 11.—Nerve-ending in commissural ganglion of adult Astacus. Methy-
lene blue and ammonium picrate. x 420.
Fie. 12.—(a). Chromatophore of lobster embryo. (b), (¢), (d), (e), (f).
Endings of branches of chromatophores from lobster embryo, showing end-
swellings.
THE SENSORY CANAL SYSTEM OF FISHES. 499
The Sensory Canal System of Fishes.
Part I.—Ganoidei.
By
Walter Edward Collinge,
Demonstrator of Zoology and Comparative Anatomy, Mason College,
Birmingham.
With Plates 39 and 40.
ContTENTS.
PAGE | PAGE
i. Introductory . : 7) A99 x. Acipenser sturio . 521
ii. Historical . : . 500 xi. Summary and Conclusion 524
iii, Nomenclature : . 502 xii. Comparison of Polydon
iv. Function : : . 504 with Acipenser . 525
v. The Cranial and other xiii, Comparison of the Sela-
Bones ; é . 506 choid Ganoids with the
vi. Polyodén folium . 507 Elasmobranchii . . 526
vii, Innervation of the Canals, xiv. Comparison and review of
&e. . 5 5 . 514 the Septum in the Ga-
viii. Psepherus gladius . 519 noidei sling The . 527
ix. The Cranial and other xv. Classification . : . 529
Bones ; : . 520 | xvi. Bibliography . : . 530
i, INTRODUCTORY.
I HAVE been led to take up these investigations upon the
sensory canal system of fishes at the suggestion of Professor
Bridge, to whom I am deeply indebted for the very generous
manner in which he has placed at my disposal specimens of
very many valuable Ganoids and Teleosts, and for his valuable
criticism, assistance, and advice.
voL. 86, PART 4,—NEW SER. L L
500 WALTER: EDWARD COLLINGE.
The importance of the sensory canal system in fishes has not
yet been sufliciently estimated, partly, I believe, from the absence
of any continued and systematic investigations concerning the
same, and the need of more detailed and reliable information
upon the development and early stages. The present series of
papers will endeavour in some measure to meet the first require-
ment, and may in a few instances add to our knowledge of the
ontogeny.
It seems strange, considering the important modifications
that the skull and cranial nerves have undergone, due very
largely to the presence of a sensory canal system, and also the
bearing it has upon the origin of various sensory organs, that
it has not*hitherto been subjected to a more prolonged and
thorough examination. So far as I can gather from the
voluminous literature upon the subject, previous writers, with
one or two exceptions, have been content to simply describe
the course of the canals and their branches, omitting any
account of the innervation or histology, whilst comparisons
with other species, orders, &c., are almost unknown.
In pursuing these inquiries I have placed myself under
many obligations to Professor T. W. Bridge, M.A., of Mason
College, Birmingham, for his very generous and continued
assistance and advice; to Dr. Giinther, F.R.S., of the British
Museum, London, for the facilities which he has at all times
given me for examining the collections under his care. I take
this opportunity of expressing my thanks to the Council of the
Birmingham Philosophical Society for a grant from their
‘Research Fund’ in aid of these investigations, and to Pro-
fessor G. B. Howes, Professor J. Cosar Ewart, F.R.S., and Mr.
Samuel Garman for assistance or advice they have from time
to time so willingly rendered.
ii. HisToRIcAL,
The earliest reference to the system of organs known as
‘‘muciparous canals,” “sense-organs of the lateral line,”
“branchial sense-organs,” “lateral canals,’ &c., is the de-
scription by Stenonis (78) in 1664 of the mucous canals in a
THE SENSORY CANAL SYSTEM OF FISHES, 501
species of skate, and in 1669 of a similar system in one of the
sharks. Lorenzini (50) in 1678 confirmed the observations of
Stenonis and separated the canals into two systems, viz. sensory
and ampullary canals. Monro secundus (56) in 1785 investi-
gated and figured the canals in the head of a cod and a skate ;
he further traced the innervation of the ampullary canals.
Seeing that Monro regarded the system as one for the secretion
of mucus, it seems probable that he was not acquainted with
Lorenzini’s work. Geoffrey St. Hilaire (34), 1802, regarded
the mucous ducts as the analogues of the electric organs of the
torpedo. Jacobson (88), 1813, was the first to put forward a
theory that this system of canals were sensory organs for the
transmission of vibrations of the water to the nerves. His
theory was more or less supported by Treviranus, Knox, and
others. Mayer, 1843, Jobert de Lamballe (40), 1858, and
M‘Donnell (51), 1861, all regarded them as electric organs ;
while Blainville, 1822, Savi, 1840, and Robin (64) held that
they were mucous canals.
Wagner, 1847, H. Miiller, 1851, Kolliker, 1856-8, Max
Schultz, 1862, Boll (10), 1868, and many other anatomists
gave considerable attention to the subject, but little advance
was made until 1868, when Leydig (48) published his com-
prehensive paper. Commencing in 1850, his series of papers
form by far the most complete and important contribution to
the subject that has yet been published.
More recent contributions on the development, innervation,
and histology have been advanced by Gotte, Semper, Balfour
(8 and 4), Solger (72, &c.), Hisig (28), Dercum (21), van
Wijhe (82 and 83), Hoffmann (35), Wright (84), Fritsch (28
and 29), Sappey (67), Beard (7), and others.
De Sede (22) in 1884 published an interesting paper detail-
ing his results upon the function. He was the first to compare
the system with a view to ascertaining its value in classification,
and in the Selachia accords to them greater importance for
such purposes than in the Teleostei. Garman (82) in 1888, in
an exceedingly valuable paper, described and figured the
course of the canals in a large series of Selachia and Holo-
502 WALTER EDWARD COLLINGE.
cephali, with a view to ascertaining, like de Sede, their value
in classification. For this system of canals, &c., Garman pro-
posed the term Tremognosters, but it has not met with general
acceptance, the term sensory canal being much more appro-
priate.
Allis (1) in 1889 dealt in an able manner with the topo-
graphy and development of the system in Amia calva.
Pollard (62) in 1892, following Garman, compared the canals
in a number of Siluroids and discussed their value for purposes
of classification.
In 1892 Ewart (25 and 26), following up his researches
upon the cranial nerves of Elasmobranchs, published a valu-
able account of the system in Lemargus and Raia batis,
the first that has attempted with any completeness to deal with
the general anatomy of the system in the Elasmobranchii. He
emphasised the necessity for studying this system of sensory
canals and their innervation together. He has shown that the
canals in the Elasmobranchii, instead of being innervated by
the branches of the trigeminal and facial nerves—as was
generally supposed—are supplied by the latter only and the
vagus. He further pointed out that certain branches of par-
ticular cranial nerves are developed solely for the innervation
of, and in connection with, the sensory canals, and disappear
almost, if not entirely, in the higher Vertebrates.
In 1893 the writer published short accounts of the system
in Polypterus, Calamoichthys (18), and Lepidosteus
(19), and in certain fossil fishes (20).
For many of the earlier historical details I am indebted to
Mr. 8. Garman, C.M.Z.S., of the Museum of Comparative
Zoology, Cambridge, U.S.A., to whose memoir the reader is
referred for more detailed accounts of the opinions of the
various writers. 7
iii, NOMENCLATURE.
It will be as well at the outset of these investigations to
revise the nomenclature of the subject, and so avoid an endless
confusion and series of explanations which otherwise must
THE SENSORY CANAL SYSTEM OF FISHES. 503
necessarily ensue. And here let me remark in passing that
the system adopted by some authors—particularly Garman
(82) and Allis (1)—of naming every separate branch a canal is
one which I cannot assent to, for even were there a separate
source of innervation to each and all of these branches, it is
very questionable whether any purpose is served by adding a
burdensome nomenclature; and further, when we come to
review the sensory canal system in fishes generally, we find
that the conclusions arrived at do not support such a method.
In the most generalised forms the canals are simple open
grooves, or grooves covered by scales, e.g. Heptanchus.
There are a few exceptions to this in the deep-sea Teleosts
such as Cottus bathybius, Liparis micropus, Lycodes
murena, &c. The gradual specialisation to dermal canals,
then canals partly dermal and partly borne by short drainpipe-
like canal bones, then bony channels, and finally passing
through the cranial and other bones, is readily traced in the
Ganoids and Physostomous Teleosts.
From what little we know of the development of these canals,
the gradual evolution of the system is further borne out, for
we find that they make their appearance first as isolated
grooves, which coalesce and form canals, and these again join
with others to form a network more or less distributed over the
head.
The whole question of the origin and evolution of this
system I hope to discuss in some detail in a later paper. For
the present, therefore, I think it advisable to treat the whole
series of sensory organs, canals, pits, pores, &c., as one system,
and to divide what have been termed canals into a series of
branches. We must not lose sight of the fact that, whatever
their origin or function, this system is of great importance to
the fish, and is consequently subject to endless modifications,
and therefore, in the present state of our knowledge, it would,
I think, be unwise to lay down any hard-and-fast rule re-
specting this subject of nomenclature. The scheme I have
tabulated below, I think most morphologists will agree, is one
based on broad and general principles, and will greatly facili-
504. WALTER EDWARD COLLINGE.
tate future investigations, obviating complications and mis-
understandings that are sure to ensue from a disregard of some
such system.
1. The system of canals, branches, sense-organs, &c., I shall
term the sensory canal system.
2. The canal passing from the tail along the lateral portions
of the trunk will be spoken of as the lateral canal—the
term main canal is used in referring to the anterior portion
of the lateral canal traversing the head. The various branches
from the main canal which traverse the region of the head will
be referred to respectively as the supra-orbital, the sub-
orbital, and the opercula-mandibular or hyomandi-
bular branches ; and the supra-occipital, the pre-orbital,
and the ethmoidal commissures. Where a dorsal canal is
present on the body it is spoken of as such.
8. A system of fine dermal canals running from the main
canal or a branch of the same, and opening by a series of fine
branches to the surface by isolated pores, will be termed
cluster pores (= peripheral organs of Allis).
4, The fine pore-like openings spoken of as “ pin-hole”
pores by many authors I shall term primitive pores, as
illustrating the most generalised form, e. g. certain Elasmo-
branchs and Polyodon, Psepherus, and Acipenser.
5. Those canals which are unbranched and radiate from a
given number of centres in the region of the head, having an
expanded proximal end or ampulla, and opening to the surface
by their distal end, and often spoken of as the canals of
Lorenzini, I shall—following Ewart—refer to as ampullary
canals.
6. The series of organs known as smell-buds, sense-organs,
pit-organs, Merkel’s buds, branchial sense-organs, &c., will be
termed sensory organs.
iv. FUNCTION.
Of the function of this system but little is known. Lorenzini
(50), 1678, was probably the first to regard it as a sensory one.
Jacobson (88), 1813, stated its function more definitely; but,
THE SENSORY CANAL SYSTEM OF FISHES. 505
as has been previously pointed out, until the publication of
Leydig’s papers most authors either regarded its function as
an electric one or as a system of canals simply for the secretion
of mucus.
Leydig drew attention to the relations of the system with
the auditory organ, and compared the canals, &c., with the
ampulle; later, however, he suggested that the system was
one of some unknown function, which he termed a sixth sense.
The most probable theory is that advanced by Schulze (69)
in 1861, viz. that this system of canals, &c., is one for the
perception of wave vibrations and oscillations ; to which it were
as well, perhaps, to add Krause’s theory, 1875, that they also
serve to give notice of chemical or physical changes in the
water.
Dercum (21), 1879 ; Emory (24), 1880; and Bodenstein (9),
1882, have each pointed out relations to the auditory organ.
Beard (7), 1886, in a brilliant paper, showed that the organs
of this system have some physiological relationship with the
gill-clefts, and that the auditory and olfactory organs are but
specialised sense-organs of the sensory canal system.
Ayers (2), 1892, in a voluminous and highly speculative
treatise, has also pointed out the relations of the auditory
organ to the sense-organs of this canal system; and although
this author sums up many such weighty questions in Vertebrate
morphology upon very little evidence and in a somewhat hasty
manner, he has nevertheless brought forward a number of
points which greatly tend to strengthen Beard’s theory,—one
of the most important, perhaps, being his discovery that “in
Elasmobranchs the structural connection between the ear organs
and the surface canal organs is for a long time maintained after
the ear has migrated to its internal home, and in some forms
may be said with truth to persist during the life of the
individual” (p. 315).
A number of authors—LEisig (23), Whitman, Leydig—have
pointed out the relations of the segmental sense-organs of
Invertebrates to those of Vertebrates. Some of these I shall
have occasion to discuss in a later paper.
506 WALTER EDWARD COLLINGE.
v. THE CRANIAL AND OTHER BONES.
A brief description of the cranial and other bones of the head
will facilitate reference to the course and distribution of the
sensory canal system.
The nomenclature is mainly that used by Bridge (16).
Post-temporals (Pl. 39, fig. 1, pt.) are two somewhat
Y-shaped lateral splints, the innermost area being much the
shorter. Upon the surface of these elements are a series of
canal bones, which conduct the posterior portion of the main
canal.
Dermo-sphenotics (Pl. 39, fig. 1, d. sph.).—The dermo-
sphenotics are almost hidden by the development upon their
surface of a series of much expanded canal bones. Posteriorly
the dermo-sphenotics are bounded by the post-temporals, ante-
riorly by the dermo-ect-ethmoids. On their inner side they
are produced into a lateral process, the sutural margin of
which is closely interwoven with the parietal and frontal
sutures,
Over the dermo-sphenotics the main canal divides into the
supra- and sub-orbital and hyomandibular branches.
The dermo-ect-ethmoids (Pl. 39, fig. 1, d. ec. eth.) are
two lateral parostoses lying over the olfactory region and form-
ing the lateral boundaries of the frontals.
The parietals (Pl. 39, fig. 1, pa.) are two elongated splints
on either side of the median line. They do not oppose each
other mesially, excepting in the most posterior portion, at least
not in the specimens I have seen. Bridge (16), however, says
they do, and figures them so, from which it would appear that
there is a considerable amount of variation. Anteriorly they
dovetail with the frontals, posteriorly terminating in a number
of fibrous-like rays. The parietals extend from the occipital
region to just behind the orbit. They are the largest and the
thickest of the cranial splints.
The frontals (Pl. 39, fig. 1, fr.) lie on either side of the
dermo-ethmoid, extending forwards for some distance. They
THE SENSORY CANAL SYSTEM OF FISHES. 507
are two irregular and unequal splints, and, like the parietals,
do not meet in the median line.
The dermo-ethmoid (Pl. 39, fig. 1, d. eth.) lies between
the anterior portion of the frontals. Anteriorly it terminates
in a long pointed process. y
_ The circumorbital series are the canal bones which
conduct the suborbital branch of the main canai of the head—
after passing over the dermo-sphenotic—along the posterior
and inferior borders of the orbit. They area series of tube-like
bones and pass beneath the dermo-ect-ethmoid.
The parasphenoid extends from the occipital region to
a point slightly anterior to the nasal capsules. Its anterior
portion is invested by two median splints—the vomers.
The hyomandibular is a slender shaft-like bone. Its
proximal end fits in a groove on the lateral border of the otic
capsule. Its distal end is attached to the symplectic cartilage.
The axis of the hyomandibular “is inclined backwards at an
angle of less than 30° with the cranio-spinal cartilage.”
The maxilla is a thin splint-like bone closely adherent in
the anterior portion to the scale-like mesopterygoid and the
surrounding cartilage ; the middle portion is separated from the
pterygoid process by the levator mandibularis muscle, whilst
the posterior portion adheres to the obitar process of the ptery-
goid bar.
The branch of the sensory canal passing along the maxilla
is a dermal one, and does not enter into the substance of the
bone at all.
The mandible consists of a long dentary splint closely
applied to the Meckelian cartilage. There is no angular element
or os articulare.
The mandibular branch of the sensory canal, like the
maxillary, is a dermal one, and passes over the surface of the
splint.
vi. POLYODON FOLIUM.
So far as I am aware there is no reference in any of the
accounts of the anatomy of Polyodon to the sensory canal
system, excepting a passing notice by van Wijhe (82).
508 WALTER EDWARD COLLINGE.
In P. folium there are present all the canals and sensory
organs previously referred to, excepting the ampullary canals,
which have not as yet been observed in any order of fishes but
the Elasmobranchii.
The system extends from the dorsally deflected terminal
portion of the trunk, which I shall speak of as the dorsal
caudal fin, to the tip of the rostrum. There is a certain
polarity about the system, it having its greatest development
on the head and rostrum, and the posterior end of the trunk
of the body and the dorsal caudal fin. Short branches of the
lateral canal cover each side of the body, terminating in series
of cluster pores.
The general distribution of the canals and branches is as
follows :
The lateral canal commences about 180 mm. from the tip of
the dorsal caudal fin and passes along the side of the body,
giving off in its course the above-mentioned branches, which
are mostly ventral ones. The cluster pores on these branches
exhibit slight modifications from those upon the head. The
canal at the anterior end of the body makes an upward curve
and enters upon the skull in the post-temporal region as a
dermal canal. Traversing the region for a short distance, it
enters a canal bone and gives off a small branch on the inner
side of the canal in the occipital region. The canal is con-
tinued forwards in a series of canal bones lying on the dorsal
surface of the dermo-sphenotic of Bridge, giving off a short
lateral branch at the most posterior end of the bone. In the
anterior portion the canal bone spreads out in a wing-like
manner and divides into two dorsal branches, viz. a supra- and
sub-orbital, and a ventral branch, the hyomandibular (PI. 39,
fig. 1). The first-mentioned branch passes upwards and for-
wards, skirting the lateral border of the frontal; it then takes
a downward course along the dermo-ect-ethmoid and _ passes
between the nasal openings, below which it unites with the
sub-orbital branch. The sub-orbital branch is given off in the
anterior portion of the dermo-sphenotic, and runs as a bony
canal posterior and inferior to the orbit, meeting anteriorly
THE SENSORY CANAL SYSTEM OF FISHES. 509
with the supra-orbital branch,—thus bringing about a con-
dition not at present known to occur in any other investigated
fish.
Following the coalescence of these two branches on the ven-
tral surface of the rostrum, the course now lies forwards and
inwards, the canal being conducted by a series of canal bones,
to which reference will be made later. It continues forwards,
diverging laterally in the anterior portion, and meets with its
fellow of the opposite side.
Traversing the lateral and posterior portion of the parietals
are two small branches passing from the main canal towards
the median line, but which, however, do not form a commis-
sure. The hyomandibular branch, as previously pointed out,
passes off from the posterior lateral border of the dermo-
sphenotic. It is continued as a dermal canal to the angle of
the mouth, where it divides into two finer branches, the dorsal
one passing along part of the maxilla, and the ventral one along
the mandible. A series of small branches pass off from the
hyomandibular portion and are distributed over the opercular-
flap, and terminate in cluster and primitive pores (Pl. 39, fig.
3, ¢c. p. and p. p.).
It will be seen from the preceding account that the system,
in so far as the canals and branches are concerned, is one of a
very simple character.
We will now examine the various branches, pores, &c., in
detail.
1. Tue Lateran Canat.—Commencing about 180 mm.
from the tip of the dorsal caudal fin as a fine dermal
canal giving off numerous branches, the lateral canal passes
forwards and downwards, and then dorsally again after this
slight ventral flexure. It continues along the dorso-lateral
border of the body, becoming dorsal previous to entering upon
the head. During its course from the caudal to the cranial
region there is a gradual, but distinctly appreciable, enlarge-
ment in its diameter.
There are from thirty-two to thirty-five branches given off
during the whole length of the canal, of which twelve are
510 WALTER EDWARD COLLINGE.
dorsal and twenty-two ventral, fourteen of the twenty-two
being situated on the dorsal caudal fin. The number and
position may vary on either side of the body. None of the
branches are of any great length, ranging from 12 to 17 mm.
Kach branch terminates in from four to seven still smaller
branches opening to the surface by a series of pores (Pl. 39,
fig. 2,2. c. b.). There are seldom less than two and rarely
more than twelve in number. These are the cluster pores.
Along the whole length of the canal and its branches are two
series of sensory organs, about which, from the condition my
specimens were in for histological purposes, I can say but little.
The first are a series, sixteen to twenty in number, of fairly
large circular markings—3}$ to 4 mm. across—slightly raised
above the epidermis. They are distributed at irregular in-
tervals along the canal and open to the surface by a transverse
aperture or slit. Sections cut by a freezing microtome showed
them to consist of a layer of epithelial cells, with numerous
goblet cells forming a follicular-shaped organ. At the base of
the follicle a sensory organ was present, in which no difference
could be seen trom those found in the cluster and primitive
pores, excepting in size. From their structure and position I
regard them as modified cluster pores and synonymous with
the sensory follicles which Fritsch (29) speaks of as “ spalt-
papillen.” Externally they are very like the “ seitenorgane ”
figured by Leydig (49, Taf. viii, fig. 13, 0.) in Petro-
myzon. .
The second series are much more numerous. They are dis-
tributed over the canal and its branches, and also in the imme-
diate neighbourhood. ‘They are present in the greatest number
at the posterior end of the body. ‘There are from 800 to 1000
on each side of the body. They are very small, the largest not
measuring more than 8 mm. across. They appear as a small
ring enclosing a depression in the centre of which arises a
-small, hard, conical body, opening at the apex by a minute
slit-like pore. In one or two cases they were observed in
groups (Pl. 39, fig. 6, 6 and c); these were not on the canal,
but in close proximity. Those upon either the canal or its
THE SENSORY CANAL SYSTEM OF FISHES. 511
branches were always in the form of single papille. In the
dead Polyodon which had been in alcohol for some years,
the apex of the papilla was hard and glossy as if for pro-
tection.
2. THE Main Canat or THE HEAv.—The main canal is a
direct dermal continuation of the lateral canal. On the lateral
border of the post-temporal region it passes on to the head as
a dermal canal, being continued in a canal bone. It passes
forwards along the dermo-sphenotic, on which bone it gives off
a short lateral branch which terminates in a cluster pore.
Wing-like processes extend over the anterior portion of the
dermo-sphenotic, by which the canal is divided into three
branches—the supra- and sub-orbital, both dorsal, and a pos-
terior ventral one, the hyomandibular.
The Supra-orbital Branch.—Leaving the main canal on
the inner border of the anterior portion of the dermo-sphenotic,
it passes forwards over the posterior lateral border of the
frontal, giving off in its course three small branches (PI. 39,
fig. 1) which continue as dermal canals for a short distance,
and then each divides into still smaller branches, all of which
open to the surface by a series of cluster pores. From the
frontal bone the branch then takes a downward course, across
the dermo-ect-ethmoid suture, between the nasal openings, and
joins the sub-orbital branch.
The Sub-orbital Branch.—The sub-orbital branch passes
off from the main canal on the outer border of the anterior por-
tion of the dermo-sphenotic, and is conducted around the pos-
terior and inferior borders of the orbit in a series of small canal
bones. On the ventral surface beneath the dermo-ect-ethmoid
it joins with the supra-orbital branch, and the two are con-
‘ducted as one canal in a series of canal bones which pass
forwards and inwards towards the parasphenoid and continue
along its ventral and lateral border. In their course forwards
two lateral branches pass off, and break up in a series of very
minute dendriform branches.
Approaching the anterior portion of the rostrum the dia-
meter of the canal becomes less, diverges laterally, and passes
512 WALTER EDWARD COLLINGE.
around the anterior border, joining with its fellow half of the
opposite side in the median line (Pl. 39, fig. 11).
The Hyomandibular Branch.—Leaving the main canal
slightly posterior to the division resulting in the supra- and
sub-orbital branches, the hyomandibular branch passes back-
wards as a dermal branch over the region of the hyomandibular
bone to the angle of the mouth, where it divides into two smaller
branches of much less diameter, viz.a mandibular and a maxil-
lary branch; the former is the larger, and passes along the
mandible, the latter traversing the upper jaw for a short
distance. Neither meet with the companion branches of the
opposite side. Inthecourse of the hyomandibular branch from
the dermo-sphenotic to the angle of the mouth there are twelve
branches given off. They are similar in all appearances to
those previously described on the lateral canal, only here there
are dense patches of primitive pores also, in close connection
with the terminations of the branches (Pl. 39, fig. 3). The
branch @ shows both cluster and primitive pores at the termina-
tions of the smaller branches. Judging from the two examples
I have investigated of Polyodon, I should say that it is a
feature of rare occurrence for a branch to terminate in primitive
pores only, as in Pl. 39, fig. 3, d.
3. THE CommissurES.—There are no true commissures
present in the sensory canal system of Polyodon. In the
numerous fishes in which this system has been investigated, the
Polyodontidz are the only family in which as yet the com-
missures are known to be absent.
On either side of the occipital series of canal bones branches
from the main canal pass inwards, as if to meet in the median
line, but, leaving their bony channel, they divide into three
series of cluster pores, those of the one side having no con-
nection with their companions of the opposite side (Pl. 39,
fig. 1).
A similar condition is found in the region of the pre-orbital
commissure. From the supra-orbital branch three smaller
branches pass off as short bony channels, and are continued for
a short distance as dermal canals, each terminating in a series
THE SENSORY CANAL SYSTEM OF FISHES. 518
of cluster pores. There is no connection between the two sides
(P1. 39, fig. 1).
4, Tur Ciuster PorEs.—Distributed over the head and
along the sides of the body are a series of organs which I have
termed cluster pores. They are subject to great variation in
both number, location, and form—a fact which seems to have
been lost sight of by many observers.
On a specimen of Polyodon, 1660 mm. in length, there
were counted 132 series or patches of these pores, distributed
as follows :
On each side of the head : . 12 = 94
On the dorsal surface of the head . 387 7=)" 38
On each side of the body ‘ woont—— WAU
132
Each series of pores arises as a small single branch from the
lateral or main canal, or one of its chief branches ; it traverses
the surface for a short distance and then divides into from
three to twelve still smaller branches, each of which opens to the
surface by a small oval or circular pore, the peripheral border
of which is pigmented. In a few cases division of the terminal
branches was noted, the pores assuming the form of the figure 8.
Fig. 6d (Pl. 39) illustrates one taken from the lateral canal.
In longitudinal and transverse sections the following struc-
ture was noted. They agree almost in every detail with the
“spalt-papillen” of Fritsch. The neck is surrounded by a
layer of epithelial cells, at the base of which a series of
columnar supporting cells—whose nuclei stain deeply—sur-
round a series of pyriform sense-cells containing large oval
nuclei at their inner and lower ends and hair-like processes at
the opposite ends. The supporting cells are surrounded by the
epidermis. At the base of the layers of columnar epithelium
nerve branches pass, which break up into a series of fine
terminal fibres and pass into the sense-cells (Pl. 39, fig. 4).
5. THE PrimitivE Pores.—These are strictly confined to
the region of the head and rostrum. In Polyodon they are
developed to an enormous extent, covering the whole of the
514 WALTER EDWARD COLLINGE.
head, rostrum, and gill-flaps in the greatest profusion (Pl. 39,
fig. 5).
On the specimen in which the cluster-pores were counted
there were about 3500 groups of primitive pores, the number
of pores in each group varying from seven to seventeen.
The diameter of a pore was almost equal to the distance
from the surface to the sensory organ lying at its base.
Histologically they appeared to be miniature cluster-pores.
It has been very generally supposed that the sensory organs of
the lateral canal differed from those in the canals of the head.
Excepting in size I have been unable to distinguish any dif-
ferences worthy of note in those in Polyodon and Acipenser.
Ewart (26) states that this difference does not exist, and in
Raia finds that “ parts of the cranial canals exactly agree in
structure with the canals of the trunk” (p. 98).
So far as I have been able to ascertain, these primitive pores
are found only in the Selacheoid Ganoids in the form I have
figured on Pl. 39, figs. 2 and 3, p.p., viz. in series aggregated
into distinct pigmented patches.
vil. INNERVATION OF THE CANALS.
The cranial nervous system of Polyodon exhibits a number
of interesting features at present not known to occur in any
other Ganoid.
From the large number of sensory organs I expected to find
great branching of the facial nerve and possibly the trigeminal
also, and generally a condition not unlike that figured by Ewart
in Lemargus (25). Indeed, I have endeavoured to inter-
pret the cranial nerves in the light of this author’s recent
researches, which have placed the study of the cranial
neurology of fishes and Vertebrates generally in quite a new
aspect.
In Polyodon there are three features which have greatly
modified the number and distribution of the cranial nerves, viz.:
(i) The unusual number of sensory organs ; .
(ii) The backward position of the suspensorium (hyomandi-
bular) ; and
THE SENSORY CANAL SYSTEM OF FISHES. 515
(iii) The largely developed rostrum.
The principal nerve groups innervating the sensory canal
system are the trigeminal, the facial, and the vagus. The
fact of the trigeminal actually innervating a part of the sen-
sory canal system is of special interest, seeing that Ewart (25)
found that in Lemargus and Raia the innervation proceeded
from the facial and vagus only.}
The cranial nerves of Polyodon have been briefly de-
scribed by van Wijhe in his memoir on the cranial nerves of
Ganoids? (82). For purposes of comparison I have reproduced
his figure, and it will at once be seen that there are a number
of important differences between his description and figure and
those here given (Pl. 40, figs. 9 and 10).
Van Wijhe (82, p. 240) states that the specimen he worked
at was a young one and only small. It is probably owing to
the fineness of the nerves in such a specimen, of which he
frequently makes mention, that our accounts differ so widely
from one another.
The Trigeminal group may be divided into five branches,
viZ.—
1. The Ramus opthalmicus superficialis.
2. The Ramus opthalmicus profundus.
3. The Ramus maxillaris.
4. The Ramus mandibularis.
5. The Ramus oticus.
1. The ramus opthalmicus superficialis is the most anterior
branch of the group, and passes forwards dorsal to the orbit
and on the inner side of the olfactory capsule.
2. The ramus opthalmicus profundus is given off from the
main branch of the trigeminal, and lies ventral to the ramus
opthalmicus superficialis and much deeper. In front of the
1 Ewart treats of the ramus oticus as distinct from the trigeminal nerve.
2 « Als ich Spatularia untersucht, kannte ich die Wichtigkeit des Ver-
laufes der Schleimcanale noch nicht, sodass ich nur sagen kann, dass der
mandibulare Zweig vor dem Spritzloche das Cranium verlasst, langs dem ~
Vorderrande des Hyomandibulare und dann zwischen Unterkiefer und Hyoid
verlauft. Hr liegt unbedeckt in der weissen Haut und fallt sogleich in die
Augen; seine Wande sind verknéchert”’ (p. 247, § 4).
voL. 36, PART 4,—NEW. SER. MM
516 WALTER EDWARD COLLINGE.
olfactory capsule they run almost parallel to each other. Ac-
cording to van Wijhe it divides into two branches, a dorsal
one which passes over the olfactory capsule, and another to the
fore part of the orbit.
3. The ramus maxillaris is a large nerve passing beneath the
orbit and dorsal to the ramus palatinus of the facial. It passes
along the whole length of the rostrum.
4, The ramus mandibularis passes ventral to the hyomandi-
bular bone across the cheek in an oblique direction. Anterior
to the symplectic cartilage it divides into two branches, the
anterior one sending branches to the muscles and primitive
pores in the region of the maxilla, whilst the posterior branch
innervates the muscles in the mandibular region.
5. The ramus oticus is the most posterior branch of the
trigeminal group. It is a dorsal branch which passes back-
wards. Van Wijhe describes it as passing dorsally through a
canal, and terminating just in front of the foramina of the
vagus. In front of the vagus, however, it turns forwards again
and innervates the cluster pores between the occipital region
and the spiracle (Pl. 40, figs. 11 and 12, 1. of.).
The Facial group consists of the following five branches—
1. The Ramus palatinus?
2. The Ramus buccalis.
3. The Ramus mandibularis.
4. The Ramus hyoideus.
5. The Ramus opercularis.
1. The ramus palatinus?—Presumably this is the ramus
palatinus of van Wijhe. It is the most anterior branch of the
facial, and passes beneath the orbit and olfactory capsule,
having an almost parallel course to the ramus buccalis (Pl.
AO wig, dd; 7.5p-),
2. The ramus buccalis is the chief branch innervating the
canals and sensory organs. In Polyodon it is developed to an
unusual extent. Its course lies ventral to the orbit, imme-
diately in front of which it divides into two branches, these
again dividing more anteriorly. These are termed respectively
a, b, c, d, and e. |
THE SENSORY CANAL SYSTEM OF FISHES. Silk?
a. The branch a@ is the most dorsal one (PI. 40, fig. 11, a.),
and gives off innumerable fine branches from either side, which
pass to the combined supra- and sub-orbital branch of the main
canal. In the space of 26 mm. fifty-three branches were
counted and traced to the canal. In Lemargus, Ewart (20)
states that there are over 1500 twigs given off from the facial
nerve to the sensory and ampullary canals. In Polyodon
there must be 2000 to 3000. In the anterior portion the
branch further divides into smaller branches, which terminate
either in primitive pores or the anterior portion of the canal.
6. This is only a small branch passing along the rostrum for
a little more than half its length. It supplies the primitive
pores on the dorsal region of the rostrum.
ce and d. These two branches supply the primitive pores on
the lateral and ventral portion of the rostrum; most of the
branches from d pass to the two lateral branches given off from
the combined supra- and sub-orbital branch of the sensory
canal.
e. There is a small branch passing off from the ramus buc-
calis posterior to the orbit, labelled e (PI. 40, fig. 12). It
innervates the circum-orbital series of primitive pores.
Van Wijhe figures the ramus buccalis as a single branch
passing directly forwards beneath the olfactory capsule (Pl. 40,
fig. 9, r.m.8.).
3. The ramus mandibularis—Van Wijhe’s description and
figure do not at all correspond to the condition I have found in
Polyodon. He figures a ramus mandibularis passing across
the centre of the hyomandibular bone, a condition common to
the Teleostei, but not at all correct for this fish. In his figure
(Pl. 40, fig. 9) it is at once evident that he has the hyomandi-
bular at a wrong angle. Supposing his figure were otherwise
correct, which it is not, it would appear in the position I have
shown it in in fig. 10,4. m. Matters are further complicated by
his figuring the ramus hyoideus as a branch of the ramus mandi-
bularis (fig. 9,7..). This latter is shown as passing beneath
the suspensorio-quadratum ligament, and is described as divid-
_ ing into an internal and external branch. He speaks of the
518 WALTER EDWARD COLLINGE.
former as innervating the ventral portion of the mandibular
branch of the sensory canal.
I find that the ramus mandibularis—which is quite distinct
from the ramus hyoideus—passes beneath the proximal head of
the hyomandibular to the angle of the jaw, where it bifurcates
and further divides into smaller branches (P1. 40, fig. 12, 7”.2’.).
It innervates some of the primitive pores of the sides of the
head and mouth, but not the mandibular branch of the sensory
canal, as described by van Wijhe, which is supplied by the
ramus opercularis superficialis No. 4 (Pl. 40, fig. 12, 7.0. 4).
4, The ramus hyoideus is a large branch passing beneath
the hyomandibular around the angle of the jaw and along the
cerato- and epi-hyal. It does not branch from the ramus
mandibularis, but is the ventral division of the ramus opercu-
laris superficialis (P]. 40, fig. 12, 7..). It passes from the
brain to the angle of the jaw beneath the hyomandibular
branch of the sensory canal, and dorsal to the ramus man-
dibularis.
I regard van Wijhe’s ramus hyoideus (PI. 40, fig. 9, 7. A.)
as homologous with the ramus opercularis superficialis No. 4,
(Pl. 40, fig. 10), while I think he has mistaken the true
hyoidean branch for what he describes and figures as the
ramus mandibularis ext. (Pl. 40, fig. 9, 7. m”.).
5. The ramus opercularis superficialis arises with the ramus
hyoideus. Dorsal to the hyomandibular branch of the main
sensory canal it divides into four branches. Nos. 1 and 2
pass backwards and break up into a series of finer branches,
The branch, 7. 0. 3, fig. 10, traverses the region between Nos. 1
and 2 and the branch 4, This latter passes dorsally along the
whole length of the hyomandibular branch of the main sensory
canal, which it innervates, and its cluster of primitive pores.
At the angle of the jaw, in the region of the symplectic car-
tilage, it divides into two branches, which innervate the canals
of the upper and lower jaws and the majority of the primitive
pores.
The ramus opercularis superficialis would seem to be a
special branch of the facial for the innervation of the primi-
THE SENSORY CANAL SYSTEM OF FISHES. 519
tive pores, &c., on the sides of the head and gill-flap in
Polyodon.
The Glossopharyngeal.—Ewart has suggested that
possibly the most anterior portion of the lateral canal may be
innervated by nerve-fibres from the glossopharyngeal nerve
previous to its leaving the cranial cavity, but so far he has
failed in Lemargus to find any branches which pass to either
the sensory or ampullary canals from this nerve. In Polyodon
I have met with similar results, the anterior portion of the
lateral canal being innervated by the branch /’. (Pl. 40, fig. 11)
of the vagus. This branch also innervates the cluster pores in
the occipital region.
The Vagus.—The lateral division of the vagus supplies the
lateral canal and its associated organs. It gives off a number
of branches all along its course, which innervate the branches,
cluster pores, and other sensory organs situated in the region
of the canal.
Vill. PSEPHERUS GLADIUS.
The sensory canal system of Psepherus approaches very
closely to that described in Polyodon, to which it holds the
same relation as Calamoichthys does to Polypterus.
From an external examination of specimens in the British
Museum little or no difference could be observed in the form
and distribution of the canal and its branches.
I have described in Polyodon a series of sensory organs in
the region of the lateral canal which I am inclined to regard
as modified cluster pores, and synonymous with the ‘ spalt-
papillen” of Fritsch. In Psepherus these are rather larger
and more numerous.
A partial dissection of the head of a small specimen was
made, in which the dermal portions of the main canal and its
branches were all found to agree with the condition present
in Polyodon.
In the smallest specimens examined there were no cluster or
primitive pores visible on the lateral canal, and no trace of
branching. Ina specimen about 320 mm. long the branches
seemed to be in a very early stage of formation.
520 WALTER EDWARD COLLINGE.
ix. THE CRANIAL AND OTHER BONES.
But a very brief reference is necessary to the cranial and
other bones of Acipenser. The nomenclature used is mainly
that of Parker (58).
The post-temporals (Pl. 40, fig. 13, p. ¢.) are two large
dermal scutes lying at either side of the head, posterior and
lateral to the dermo-occipital. They are bounded in front by
the epiotics and conduct the main canal.
Epiotics (Pl. 40, fig. 13, ep.).—T wo irregular scutes having
a lateral position to the dermo-occipitals. They form the lateral
posterior border of the parietals and the posterior border of
the squamosals. In their anterior portion the occipital com-
missure passes off from the main canal.
Squamosals (Pl. 40, fig. 13, sg.).—T wo large scutes form-
ing the lateral borders of the parietals. In their anterior
portion the main canal divides into the supra- and sub-orbital.
The dermo-ect-ethmoids (= the prefrontals of many
authors; Pl. 40, fig. 12, d. ec. eth.) lie in front of the squamosals.
Ventrally they are bounded by the first of the circum-orbital
series ; anteriorly they form the dorsal boundary of the orbit,
extending as far as the nasal capsules. The chain of canal
bones conducting the sub-orbital branch passes over the poste-
rior portion of either side.
Dermo-occipital (Pl. 40, fig. 18, d. oc.),—A median some-
what dagger-shaped scute forming the posterior border of the
parietals; the most anterior portion passes between the parietals
for some distance and is overlapped by them. It is traversed
by the occipital commissure.
The parietals (Pl. 40, fig. 12, pa.) are two large median
scutes bounded posteriorly by the dermo-occipital and epiotics,
laterally by the squamosals, and anteriorly by the frontals and
dermo-ethmoid. In very large specimens which I have
examined the occipital commissure passes through the poste-
rior portion of the parietals as well as the epiotics and the
dermo-occipital scute.
The frontals (Pl. 40, fig. 12, fr.) lie at either side of the
THE SENSORY CANAL SYSTEM OF FISHES. 521
dermo-ethmoid. They conduct the supra-orbital branch of the
main sensory canal, and are bounded laterally by the dermo-ect-
ethmoids.
The Sub-orbital Series (Pl. 40, fig. 13, s. or.).—There
are only two of the orbital scutes visible externally. The upper
forms the posterior border of the orbit, and the lower both
posterior and inferior borders. The sub-orbital branch of the
sensory canal passes through both and then into the two
internal ones lying anterior to the orbit.
There is no pre-opercular scute in Acipenser, or any
branch of the main sensory canal corresponding to the operculo-
mandibular branch in Lepidosteus, &. What Parker (58)
has termed pre-operculum is in all probability the jugal.
There are no mandibular or maxillary branches of the
sensory canal in Acipenser, the primitive pores in the imme-
diate region probably functioning in their place.
x. ACIPENSER STURIO.
References to the sensory canal system of Acipenser are
found in the writings of M‘Donald (52), van Wijhe (82), and
others, but most of these are only in relation to the form or
position of the scutes on the lateral portion of the body, con-
ducting the lateral canal.
Van Wijhe seems to have found some difficulty in tracing
the canals, &c. His account is as follows :—
“Was das System der Schleimcanale anbelangt, so sind die
Rohre beim Stor ausserordentlich fein, nur mit Mihe kann
man sie an einigen Stellen durchschimmern sehen. Um
ihren Lauf wahrzunehmen wurden sie mit einer feinen Borste
sondirt, und dann der Knochen, der sie umschloss, nach der
Aussenseite aufgeschlitzt.
“Der Hauptstamm liegt im Supraclaviculare, lateralen
Supratemporale (Occip. externum, Gegenb.; epioticum,
Huxley) Squamosum, Frontale, asale Nund in der Scheide-
wand beider Nasenlocher derselben Seite.
“ Die supratemporale Quercommissur liegt hier, eben so wie
bei den andern Fischen in den Supratemporalia, deren sich
522 WALTER EDWARD COLLINGE.
beim Stor drei vorfinden, ein paariges, oben genanntes, und ein
unpaares medianes (Supraoccipitale).
“Der suborbitale Zweig verlasst den Haupstamm in der
Mitte des Squamosums, tritt dann in das (dermo-) Post-
frontale und liuft weiter nach vorn durch die beiden Infra-
orbitalia.”’
In a foot-note he says :—
“Die Lage der Schleimcanale ist von hohen Interesse bei
der Bestimmung der Deckknochen des Schadels, weil sie in
Hinsicht auf diese ziemlich constante Beziehungen zeigt.”
Commencing some 13} mm. from the tip of the dorsal caudal
fin the lateral canal traverses a number of isolated scutes, an-
teriorly becoming more dorsal it enters the skull by traversing
the dermal ossifications which connect the pectoral arch.
Short branches are here given off which ramify the bone. In
the post-temporal an occipital commissure passes off from the
main canal, which is continued forwards through the epiotic
and squamosal. In the squamosal the canal divides into a
supra- and sub-orbital branch, the former continuing through
the frontal, and passes outwards and downwards behind and
posterior to the anterior nasal opening and then along the
rostrum, becoming connected with its fellow half of the oppo-
site side, and also ventrally with the sub-orbital branch. This
latter branch passes from the division of the main canal in the
squamosal into the pre-frontal, and then through a circum-
orbital series which surround the posterior and ventral borders
of the orbit, some of which are not visible externally. There
are four bones in this series. The canal issues from the most
anterior as a dermal canal, and passes into another series of
canal bones, which are not unlike in appearance a series of
miniature drain-pipes ; these conduct the canal to the tip of the
rostrum, where it now passes into a tooth-shaped bone, in
which it branches in various directions, the main branch join-
ing the supra-orbital branch above.
1. The Lateral Canal.—In Acipenser the lateral canal
commences about 134 mm. from the tip of the dorsal caudal
fin, in a series of canal bones. Previous to the commencement
THE SENSORY CANAL SYSTEM OF FISHES. 523
of the canal there are a series of these bones. The canal is
conducted along the sides of the body by alternating series of
these small bones and of isolated scutes, into which latter the
canal enters posteriorly and on the ventral side, passing out
anteriorly and on the dorsal side. The number of these small
drain-pipe-like bones between each scute varies from four to
ten. The greater portion of the canal is exceedingly small in
diameter, increasing slightly in the anterior portion.
There are no branches given off from the lateral canal,
neither in or on the scutes or from the intervening portions.
There is also a complete absence of cluster or primitive pores.
2. The Main Canal of the Head.—The main canal
enters the skull by traversing the lateral border of the post-
temporal. It is continued forwards through the epiotic, in the
anterior portion of which it gives off the occipital commissure.
Entering the squamosal the canal passes forwards to a central
position and divides into two branches, the supra- and sub-
orbital.
The Supra-orbital Branch.—Leaving the main canal in
the squamosal, the supra-orbital branch takes an inward and
forward direction; passing through the frontal bone, in
which it gives off two lateral branches, it takes an out-
ward course between the nasal apertures and is continued
on the dorso-lateral border of the rostrum, again passing
into a series of drain-pipe-like bones. In the most anterior
portion it makes a ventral turn and joins with the sub-orbital
branch by passing through a small tooth-shaped bone (PI. 39,
fig. 7, a and 0).
The Sub-orbital Branch.—After separating from the
main canal, the sub-orbital branch passes in a lateral direction
into the pre-frontal and then through a series of circum-orbital
bones, of which there are four, the second forming the
posterior ventral angle, while the third and fourth are not
visible externally. Leaving these bones as a dermal canal, it
enters a series of drain-pipe-like bones by which it is conducted
along the ventral and lateral borders of the rostrum. In the
anterior portion it makes a slight lateral divergence and passes
524 WALTER EDWARD COLLINGE.
around the anterior border of the rostrum, joining with its
fellow of the opposite side.
These bones which I have spoken of as drain-pipe-like canal
bones are exceedingly interesting in that they probably repre-
sent the earliest trace we have of an ossified investment of the
sensory canal system.
In Polyodon and Psepherus they are long thin-walled
bony channels in the greater portion of the head, none occur-
ring on the lateral canal. They never assume the disc-like
form found in Acipenser (Pl. 39, fig. 7, c). They average
from 3 to 9 mm. in length, some few on the dorsal surface of
the head being much larger.
In Acipenser they are present both on the head and in
connection with the lateral canal. Those on the head vary in
shape and size, some being pipe-like in form, others irregular.
The simplest form in which they were found is shown in
fig. 7, c (Pl. 39). These were taken from the anterior portion
of the sub-orbital branch. They are thin disc-like pieces of
bone measuring from 43 to 7 mm. across and about + mm. in
thickness. Similar forms, only thicker, were found in the same
species conducting the lateral canal.
8. The Pores.—In Acipenser there are no structures
that can be regarded as cluster-pores. On the ventral surface
of the rostrum there are numerous patches or groups of primi-
tive pores which are similar in all respects to those previously
described in Polyodon. In a sturgeon measuring 10 feet
3 inches in length upwards of eight hundred of these groups
were counted.
xi. SUMMARY AND CONCLUSION.
In summarising the more important features resulting from
this investigation, it will already have been observed that
perhaps the most important evidence obtained is that bearing
upon the Elasmobranch character of Polyodon and of the
Selachoid Ganoids generally.
The Selachoid Ganoids are the first group of fishes as yet
THE SENSORY CANAL SYSTEM OF FISHES. 525
known which show the gradual specialisation of a dermal
canal to one enclosed in a bony tube and then in bony plates.
The system of pores, sensory organs, &c.,is one showing the
modifications and evolution of a sense-organ.
The form, number, and branching of the cranial nerves has
been very largely modified by the presence of an innumerable
series of sensory organs (cluster and primitive pores, &c.) dis-
tributed over the cephalic region.
The Ganoids as a group have very rightly been divided by
Bridge (16) into Selachoidei and Teleosteoidei. For the reten-
tion of the group there seems to be sufficient morphological
evidence. Nota few zoologists would join the Ganoids with the
Teleosts. Possibly there are forms included in the Ganoids
whose affinities are undoubtedly with the Teleosts, but the
division termed by Bridge Selachoidei is much more closely
allied to the Elasmobranchs. Indeed the position, origin,
number, and course of the cranial nerves in Polyodon suggests
a much closer relationship to this latter group than even Bridge
supposed, who was the first, I believe, to institute a comparison
between them.
The conclusions of Ewart (26) are very largely confirmed
and augmented by this investigation.
From the variability of the sensory canal system in the
Ganoids but little importance can be attached to it for pur-
poses of classification, other than of the most generalised
nature.
xii. COMPARISON OF PoLYODON WITH ACIPENSER.
As we should naturally expect, there is a great difference in
the sensory canal system of the two forms. Polyodon, on the
one hand, may be regarded as a form on the very border-line
between the Elasmobranchii and the Selacheoid group of
Ganoids, whilst Acipenser, on the other hand, may be
placed on the border-line between the latter group and the
Teleosteoid Ganoids.
Although widely separated from one another by very many
important differences, the two species show at the same time a
526 WALTER EDWARD COLLINGE.
number of features common to both. These may be summarised
as follows:
Polyodon differs from Acipenser in—
(a) The canals being largely dermal ones, e.g. the lateral
canal part of the main canal of the head and its branches—none
passing through any of the cranial elements.
(6) The presence of branches on the lateral canal and on the
hyomandibular branch of the main canal.
(c) The absence of commissures.
(d) The peculiar course and coalescence of the supra- and
sub-orbital branches, by which a single canal is formed which
does not pass along the lateral borders of the rostrum, but on
either side of the parasphenoid.
(e) The presence of small mandibular and maxillary branches,
a feature of great interest considering the many Elasmobranch
affinities of the fish.
(f) The presence of cluster pores and other sensory organs
on the lateral canal.
The sensory canal system of Polyodon agrees with that of
Acipenser in—
(a) The presence of a lateral and main canal with supra- and
sub-orbital branches.
(b) The presence of a series of small bony elements—drain-
pipe-like canal bones—conducting either the canals or their
branches.
(c) The presence of primitive pores upon the head.
(d) The absence of either a pre-orbital or an ethmoidal com-
missure.
xiii. CoMPARISON OF THE SELACHOID GANOIDS WITH THE
ELASMOBRANCHII.
A comparison of the sensory canal system of the Selachoid
Ganoids with that of the Elasmobranchii shows that there
are many important features common to the two.
The distribution and number of the canals and branches
in all the Selachoid Ganoids are undoubtedly of an Elasmo-
branch character. Firstly, there is the branching of the
THE SENSORY CANAL SYSTEM OF FISHES. 527
lateral canal in Polyodon and Psepherus, developed to
a greater extent in such forms as Tenura lymna, Dasy-
batus dipterurus, Alopias vulpes, Rhinobates plani-
ceps, &c.! |
The persistence of dermal canals in Polyodon and
Psepherus, and the very slight development in these two
forms, and complete absence in Acipenser, of any true man-
dibular or maxillary branches, are prominent characteristics of
the Selachians and Batoidea. In the distribution and form of
the sensory organs (cluster pores, &c.) there are many points
of agreement.
The total absence or only slight development in the Elasmo-
branchs of commissures agrees with the condition found in
Polyodon and Psepherus, whilst in Acipenser the pre-
sence of only one—the occipital—and the arrangement of
the canals and branches in the commissural regions of all
three genera, exhibits many close affinities to such forms as
Heptabranchias maculatus, &c.
The number, form, and distribution of the cranial nerves of
Poly odon resemble, in the branching of the buccalis division
of the facial nerve, and in the distribution of the otic, &c.,
the condition described and figured by Ewart (25 and 26) in
the Elasmobranchii.
Indeed, after the many anatomical relations that Polyodon
exhibits with the Elasmobranchii—described by Bridge some
years ago (16)—I should have been greatly surprised had
there not been a similar agreement in the sensory canal
system.
xiv. CoMPARISON AND REVIEW OF THE SYSTEM IN THE
GANOIDEI.
In comparing the sensory canal systerm in the existing
Ganoids, it must be borne in mind that they are an order of
fishes the families of which are widely separated from one
1 The sensory canal system of all these forms has been carefully figured by
Mr, 8. Garman (82).
528 WALTER EDWARD COLLINGE.
another, the intervening gaps being but very imperfectly
bridged over by a series of still less perfect extinct forms whose
fossil remains show, in most cases, few if any indications of the
course of this system of canals (20).
Of the eight known genera of recent Ganoids I have exa-
mined specimens of all excepting Scaphirhynchus, which
in all probability is very similar to Acipenser.
The PoLtyoponTIp# are undoubtedly the most generalised
family, and most closely related to the Elasmobranchii.
In the ACIPENSERIIDZ we have a great advance upon any-
thing seen in either of the genera in the preceding family, and
yet the absence of mandibular and maxillary branches and the
persistence of large numbers of primitive pores are features
which are truly Selachian. Acipenser is the first species in
the class Pisces in which the canal or its branches enters into
the cranial elements.
Of the Teleosteoid Ganoids we have three families widely
separated from one another.
The most interesting feature in the LEPIDosTEIDZz is un-
doubtedly the greatly developed system of dendritic branches
passing off from the main canal and entering into the various
cranial elements. These fine branches anastomose and form a
dense network. In this feature the system resembles in many
ways that present in the Selachians. There are no branches
on the lateral canal. In the pre-orbital region there are a
number of fine branches which anastomose and form a com-
missure connecting the two supra-orbital branches, a condition
peculiar to Lepidosteus osseus, so far as is at present
known (19).
The Potyrrerip& show a number of features not previously
met with. In Polypterus (18) we note the absence of dendri-
form branching and primitive pores, and the small number of
pores generally, the presence of a branch from the operculo-
mandibular branch, which passes across the cheek-plate, and a
fine canal passing through the series of canal bones (= inter-
calary ossicles of Traquair) which connects the main canal of the
head with the operculo-mandibular branch. There are also
THE SENSORY CANAL SYSTEM OF FISHES. 529
traces of a rudimentary or degenerate canal in the pre-oper-
culum.
The system in Calamoichthys is practically the same as
that described in Polypterus.
The family Am11p# has been the most thoroughly studied of
any of the Teleosteoid Ganoids (1). The branching of the
lateral canal still persists in the posterior border of the scale,
the canals of the head are almost wholly contained in the bones,
dermal portions being exceedingly small. The cluster and
primitive pores are much more specialised in character and
fewer in number. There is little or no branching of the
system upon the head, excepting in the pre-operculum, an
operculo-mandibular branch being present. There is no con-
nection between the two mandibular portions in the median
symphysis of the lower jaw. Both occipital and ethmoidal
commissures are present, and a similar form of branching in
the pre-orbital region as was noticed in Polyodon.
Not a few writers have examined the sensory canal system
with a view to ascertaining its value in classification. I am
not aware that it has been used for such purposes, and from
the variability of its nature I should much doubt its value if
applied in any than a very general manner.
xv. CLASSIFICATION.
GANOIDEI.
Group 1. Srnacuorpet, Bridge.
Sub-order 1. CnHonproste1, Miiller.
Fam. i. PoLyoponTip.
Polyodon, Lacép.
— folium, L., p. 507.
Psepherus.
— gladius, Martens, p. 519.
Fam. 11. ACIPENSERIDA,
Acipenser, Artedi.
— sturio, L., p. 521.
Scaphirhynchus, Heckel, p. 528.
— catephractus, Gray.
530 WALTER EDWARD COLLINGE.
Group 2. TExxostzoipet, Bridge.
Sub-order 2. Hotostet1, Miiller.
Fam. i. LepiposTE1pz&.
Lepidosteus, Lacép.
— osseus, L., p. 528.
— platystomus (Raf.).
Fam. ii. Potyprerips.
Polypterus, Geoff.
— bichir, G., p. 528.
Calamoichthys, J. A. Smith.
— calabaricus, J. A.S., p. 529.
Fam. ii. AMIIDA.
Amia, L.
— calva, L., p. 529.
xvl. BIBLIOGRAPHY.
1. Attis, E. P.—‘ The Anatomy and Development of the Lateral Line
System in Amia calva,” ‘Journ. Morphology,’ 1889, vol. ii, pp.
463—569, pls. xxx—ulii.
2. Ayers, Howarp.—‘‘ A Contribution to the Morphology of the Verte-
brate Ear, with a Reconsideration of its Functions,” ‘Journ. Mor-
phology,’ 1892, vol. vi, pp. 1—360, pls. i—xii.
8. Batrour, F. M.—‘ A Monograph on the Development of Elasmobranch
Fishes,’ London, 1878.
4, Batrour, F. M—‘ A Treatise on Comparative Embryology,’ 2 vols.,
London, 1880-1.
5. Bearp, J.—On the Segmental Sense-organs of the Lateral Line and
the Morphology of the Vertebrate Auditory Organ,” ‘Zool. Anz.,’
1884, vol. vii, pp. 161-162.
6. Bearp, J.—‘‘ On the Cranial Ganglia and Segmental Sense-orgaus of
Fishes,” ‘Zool. Anz.,’ 1885, vol. viii, pp. 220—223.
7. Bearp, J.—‘ The System of Branchial Sense-organs and their Asso-
ciated Ganglia in Ichthyopsida,” ‘ Studies from the Biol. Laboratory
of Owens College,’ 1886, vol. i, pp. 170—218, pls. vii—ix.
THE SENSORY CANAL SYSTEM OF FISHES. 531
8. Buave, Jutius.— Untersuchungen iiber den Bau der Nasenschleimhaut
bei Fischen und Amphibien, namentlich iiber Endknospen als Endap-
parate des Nervus olfactorius,” ‘Arch. f. Anat. und Phys.,’ 1884,
pp. 231—309, Taf. xii—xiv.
9. Boprnstxin, Emiu.—“ Der Seitenkanal von Cottus gobio,” ‘ Zeit. f.
wiss. Zool.,’ 1882, Bd. xxxvii, pp. 121—145, pl. x.
10, Bout, Franz.—‘ Die Lorenzinischen Ampullen der Selachier,” * Arch.
f. mikr. Anat.,’ 1868, vol. iv, pp. 375—3891, pl. xxiii.
11. Bout, Franz.—‘ Beitrage zur Physiologie von Torpedo,” ‘ Arch. f. Anat.
und Phys.,’ 1873, pp. 76—102.
12. Bout, Franz.— Le vesicole de Savi della Torpedine,” ‘Roy. Acad. de
Lincei,’ vol. ii, p. 385.
13. Bout, Franz.—* Die Savi’schen Blaschen von Torpedo,” ‘ Arch. f. Anat.
und Phys.,’ 1875, pp. 456—468, Taf. xi.
14, Bout, Franz.—‘ Ueber die Savi’schen Blaschen von Torpedo,” ‘ Monats-
berichte Berl. Akad.,’ 1875, p. 238.
15. Bripez, T. W.—“ The Cranial Osteology of Amia calva,” ‘Journ.
Anat. and Phys.,’ 1877, vol. ii, pp. 605—622, pl. xxiii.
16. Bripecz, T. W.—‘On the Osteology of Polyodon folium,” ‘ Phil.
Trans. Roy. Soc.,’ 1878, vol. clxix, pp. 683—733, pls. lv—lvii.
17. Carus, G.—‘ Lehrbuch der vergleichenden Zootomie,’ 1843, i.
18. CottincE, WALTER E.—* Note on the Lateral Canal System of Polyp-
terus,” ‘Proc. Birmingham Phil. Soc.,’ 1898, vol. viii, pt. il, pp. 255—
262, pls. i—iii.
19. Conuincr, Watrer E.—* The Lateral Canal System of Lepidosteus
osseus,” ibid., pp. 263—272, pls. iv—ix.
20. Cottincr, WattER E.—‘“‘ The Morphology of the Sensory Canal System
in some Fossil Fishes,” ibid., vol. ix, pp. 1—14, pls. i, 1.
20a. Cottincr, WatteR E.—“On the Sensory Canal System of Fishes,”
‘ British Assoc. Rept.’ for 1893, 1894, pp. 810—811.
21. Dercum, F.—“ The Lateral Sensory Apparatus of Fishes,” ‘ Proc. Acad.
Nat. Sci. Phil.,’ 1879, p. 152.
22. Dz Sipe pe Lrécrix, P.—‘ Recherches sur la ligne laterale des Pois-
sons Osseux,’ 1884.
23. Hisic, H.— Die Seitenorgane und beckerforungen Organen der Capi-
telliden,” ‘ Mittheil. a. d. Zool. Station zu Neapel.,’ 1879.
24. Emory, Carto.—‘ Le Specie del Genere Fierasfer nei Golfo de
Napoli e Regioni limitrofe,’ ‘Fauna und Flora des Golfes von
Neapel,’ 1880.
25. Ewart, J. C.—‘‘ The Lateral Sense-organs of Hlasmobranchs. I. The
Sensory Canals of Lemargus,” ‘Trans. Roy. Soc. Edinb.,’ 1893,
vol, xxxvii, pp. 59—86, pls. i, ii.
VOL. 86, PART 4.—NEW SER. NN
aR WALTER EDWARD COLLINGE.
26.
27.
28.
29.
30.
31.
32.
33.
34,
35.
36.
37.
38.
Ewart, J. C., and Mrtcnett, J. C.—“ II. The Sensory Canals of the
Common Skate (Raia batis),” ibid., pp. 87—106, pl. iii.
FRranque, Heyricus.— Amix calve Anatomia,” ‘Berolini,’ 1874,
pp. 1—14, pl. i.
Fritscu, G.— Ueber Bau und Bedentung der Kanalsysteme unter der
Haut der Selachier,” ‘ Sitzungsberichte d. Kénigl. Akad. d. Wissensch.
zu Berlin,’ 1884,
Fritscu, G.—‘ Die Electrischen Fische, die Torpedinen,’ Leipzig, 1890.
Froner, A.—“ Ueber Anlagen von Sinnesorganen am Fascialis, &c.,”
‘Arch. f, Anat. und Phys.,’ 1885, pp. 1—55, Taf. i, ii.
Garman, SamurL.—Chlamydoselachus anguineus,” ‘Bull. Mus.
Comp. Zodl.,’ Cambridge, U.S.A., 1885, vol. xii.
Garman, Samuzt.—‘ On the Lateral Canal System of the Selachia and
Holocephala,” ibid., 1888, vol. xvii, pp. 57—119, pls. i—liii.
Garman, SAMUEL.—“ The Vesicles of Savi,” ‘Science,’ 1892, vol. xix,
No. 474, p. 128.
Gzorrroy St. Hitatre.—“ Sur l’anatomie comparee des organes élec-
triques de la raie torpille, du gymnote engour dissant, et du silure
trembleur,” ‘Ann. du Mus.,’ 1801, vol. i, p. 392.
Horrmann, C. K.—“ Zur Autogonie der Knockenfische,” ‘Arch. f.
mikr. Anat.,’ 1884, Bd. xxiii, pp. 45—108, pls. iv—vi.
Houssay, F.—“ Analysis and Critique of P. Mitrophanow’s Paper on
the Metametric Significance of the Organs of the Lateral Line of
Vertebrates,” ‘Arch. de Zool. Exp. et. Gen.,’ 1891, vol. ix.
Hyrti, D.—“ Sur les sinus caudal et céphalique des Poissons, et sur les
systéme de vaisseaux latéraux avec lesquels ils sont en connexion,”
‘Ann. des Sci. Nat.,’ 1843, vol. xx, pp. 215—229, pls. vi, vii.
Jacosson, L.—“ Extrait d’un Mémoire sur un Organe particulier des
Sens dans tes Raies et les Squales,” ‘Nuov. Bull. d. Scis., par la Soc.
Philomatique de Paris,’ 1813, vol. vi, p. 332.
. Kyox, R.—“On the Theory of a Sixth Sense in Fishes,’ ‘ Edinb. Journ.
Sci.,’ 1825,
. LAMBALLE, JoBERT DE.—‘Des Appareils électriques des Poissons
électriques,’ 1858.
. Lampatie, Jopert pE.—‘ Etudes d’Anat. Comp. sur les Organes du
toucher chez diverses mammiferes, oiseaux, poissons, et insectes,’ 1872.
. Luypie, F.—“ Ueber die Schleimkanale der Knockenfische,” ‘ Arch. f.
Anat. und Phys.,’ 1850, pp. 170—181, Taf. iv.
. Luypie, F.—‘ Ueber die Haut einiger Siisswasserfische,” ‘ Zeit. f. wiss,
Zool.,’ 1851, Bd. iii, pp. 1—12, pl. i.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
THE SENSORY CANAL SYSTEM OF FISHES. 533
. Leypie, F.—“ Zur Anat. u. Hist. der Chimera monstrosa,” ‘Arch.
f. Anat. und Phys.,’ 1851, pp. 241—271, Taf. x.
. Lrypic, F.—‘ Beitrage zur mikroskopischen Anatomie und Entwicke-
lungsgeschichte der Rochen und Haie,’ Leipzig, 1852.
. Leypic, F.—‘ Anatomische-histologische Untersuchungen iiber Fische
und Reptilien,’ 1853.
. Leypie, F.—‘ Lehrbuch der Histologie des Menschen u. d. Thiere,’ 1857.
. Lrypie, F.—‘‘ Ueber Organe eines sechsten Sinnes,” ‘ Nova Acta Acad.
C. L. Nat. Curios,’ 1868, vol. xxxiv, pp. 1—108, pls. i—v.
Lerypie, F.—‘* Neue Beitrage zur anatomischen Kenntniss der Haut-
decke und Hautsinnesorgane der Fische,” ‘ Festschrift d. Naturf. Ges.
zu Halle,’ 1879, pp. 1—58, Taf. vii—x.
Lorenzini, 8.—‘ Ossewazioni intorno alle Torpedini,’ Fireuze, 1678 ;
London, 1705, Angl.
M‘Downett, R.—“ Electric Organs of the Skate,” ‘ Nat. Hist. Review,’
1861, p. 59.
M‘Downet1t, R.—“ On the System of the Lateral Line in Fishes,”
‘Trans. Roy. Irish Acad.,’ 1862, vol. xxiv, pp. 161—187, pls. iv—vi.
Matsranc, M.—“Von der Seitenlinie und ihren Sinnesorganen bei
Amphibien,” ‘ Zeit. f. wiss. Zool.,’ 1875, vol. xxvi, pp. 24—82,
pls. i—iv.
MerkeEtL, Fr.—‘ Ueber die Endigungen die Sensiblen Nerven in der
Haut der Wirbelthiere,’ Rostock, 1880.
Mirropuanow, P.—‘“‘ Ueber die Erste Anlage die Seitenorgane bei
Plagiostomen,” ‘ Warschauer Univ. Nachbrichten,’ 1889.
Monro, A.—‘ The Structure and Physiology of Fishes,’ 1785.
Miier, J.—‘ De glandularum secernentium structura,’ 1830.
Parker, W. Kitcoen.—‘‘On the Structure and Development of the
Skull in Sturgeons,” ‘ Phil. Trans.,’ 1882, pp. 139—185, pls. xii—xviii.
Parker, W. Kitcuen.—“On the Development of the Skull in Lepi-
dosteus,” ‘ Phil. Trans.,’? 1882, pp. 443—492, pls. xxx—xxxviii.
Parker, W. Kircuen, and Bertany.—‘ The Morphology of the Skull,’
London, 1877.
Pottarp, H. B.—“On the Anatomy and Phylogenetic Position of
Polypterus,” ‘ Zool. Jahrb.,’ 1892, pp. 387—428, pls. xxvii—xxx.
Pottarp, H. B.—‘ The Lateral Line System in Siluroids,” ‘ Zool. Jahrb.,’
1892, pp. 525—550, pls. xxxv—xxxvi.
Rirter, W. E.—“ On the Eyes, the Integumentary Sense Papille, and
the Integument of the San Diego Blind Fish (Typhlogobius cali-
534 WALTER EDWARD COLLINGE.
64
65.
66.
67.
68.
69.
70.
aa;
72.
73.
74.
75.
76.
77.
78.
79.
formensis, Steindnachner),” ‘Bull. Mus. Comp. Zodl.,’ Cambridge,
U.S.A., 1893, vol. xxiv, pp. 51—102, pls. i—iv.
Rosin, C.—‘* Recherches sur un appareil qui se trouve sur les poissons
du genre des Raies,” ‘Ann. Sci. Nat.,’? 1847 (8rd sér.), vol. vii,
pp. 193—302, pls. iii, iv.
SacemMenL, M.—‘ Beitrige zur Vergleichenden Anatomie der Fische.
I. Das Cranium von Amia calva,” ‘Morph. Jahrb.,’ 1884, Bd. ix,
pp. 177—228, pl. x.
Sacement, M.—“‘III. Das Cranium der Characiniden,” ‘ Morph.
Jahrb.,’ 1885, Bd. x, pp. 1—119, pl. cii.
Sappey, P. C.—‘ Etudes sur l’appareil mucipare et sur le systéme
lymphatique des Poissons,’ 1880.
Scuuretpt, R. W.—‘‘ The Osteology of Amia calva,” ‘Ann. Rep.
of the Comms. of Fish and Fisheries for 1883,’’ Washington, 1885.
Scuutze, F. E.—‘ Ueber die Nervenendigung in den sogenannten
Schleimkanalen der Fische und iiber entsprechende Organe der durch
Kiemen athmenden Amphibien,” ‘Arch. f. Anat. und Phys.,’ 1861,
pp. 759—769, Taf. xx.
Scuuizn, F, E.—“ Ueber die becherformigen Organe der Fische,” ‘ Zeit.
f, wiss. Zool.,’ 18638, vol. xii, pp. 218—222, pl. xxiii.
Scuuize, F. E.—“ Ueber die Sinnesorgane der Seitenlinie bei Fischen
und Amphibien,” ‘ Arch. f. mikr. Anat.,’ 1870, Bd. vi, p. 62.
SoueER, B.—*“‘ Ueber die Seitenorgane der Fische,” ‘ Kais. Leop. Akad.
der Naturf.;’ 1878, Heft. xiv, p. 74.
SotceR, B.—‘‘ Neue Untersuchungen zur Anatomie der Seitenorgane
der Fische. I. Der Seitenorgane von Chimera,” ‘Arch. f. mikr.
Anat.,’ 1879, Bd. xvii, pp. 95—1138, pl. viii.
SotcrR, B.—‘ II. Die Seitenorgane der Selachier,” ibid., 1880, Bd.
Xviii, p. 450.
Soterr, B.—‘‘ III. Die Seitenorgane der Knochenfische,” ibid., 1880,
Bd. xviii, pp. 364—390, pl. xvii.
Soteer, B.—‘ Ueber den feineren Bau der Seitenorgane der Fische,”
‘Sitzungsb. nat. Ges. Halle,’ 1880, p. 105.
SotereR, B.—‘‘ Bemerkung wber die Seitenorgankelten der Fische,”
‘Zool. Anz.,’ 1882, vol. v, p. 660.
Srenonis, N.—‘ De Muscalis et Glandulis observationum specimen cum
duabus epistolis quarum una ad. Guil. Pisonum de anatome Rajae,
&c.,’ Amst., 1664.
Strenonis, N.—‘ Elementorum Myologie specimen; accedit Canis Car-
charise caput dissection, et dissectus piscés e Canum genere,’ Amst.,
1669.
THE SENSORY CANAL SYSTEM OF FISHES, 535
80. Traquarr, R. H.—‘On the Cranial Osteology of Polypterus,”
‘Journ. Anat. and Phys.,’ 1870, vol. v, pp. 166—183, pl. vi.
81. Trrevrranus, G. R.—‘‘ Ueber die Nerven des fiinften Paars als Sinnes-
nerven,” * Vermischte Schriften anat. und phys.,’ 1820.
82. Wine, J. W. van.—“ Ueber das Visceralskelett und die Nerven des
Kopfes der Ganoiden und von Ceratodus,” ‘ Neiderland. Arch. f.
Zool.,’ 1882, Bd. v, pp. 207—320, Taf. xv, xvi.
88. Wise, J. W. van.— Ueber die Mesodermsegmente und die Entwicke-
lungsgeschichte der Nerven des Selachierkopfes,” ‘ Natuurk. Verhandh.
Akad. Amst.,’ 1882, xxii.
84, Wricut, R. Ramsey.—“‘On the Skin and Cutaneous Sense-organs of
Amiurus,” ‘Proc, Canadian Inst.,’ 1884, vol. ii, pp. 251—269, part
of pl. i.
DESCRIPTION OF PLATES 39 and 40,
Illustrating Mr. Walter Edward Collinge’s paper, “ The
Sensory Canal System of Fishes.”
Fig. 1.—Dorsal view of the skull of Polyodon folium, showing the
cranial elements and canal bones. The stellate-shaped bones on the rostrum
are shown on the left-hand side, and the epidermis also in the most anterior
portion. Lettering as below.
Fic. 2.—Lateral view of Polyodon, showing the course and branching of
the lateral canal, its position on the caudal fin, and the groups of primitive
pores on the head, &c. Lettering as below. x 4.
5
Fie. 3.— Diagrammatic view of the side of the head of Polyodon, showing
the course of the main canal of the head and its branches.
Lettering.—com. Commissure in occipital region. c. p. Cluster pores.
d. ec. eth. Dermo-ect-ethmoid. d. eth. Dermo-ethmoid. d.sph. Dermo-sphenotic.
Jr. Frontal. yom. 6. Hyomandibular branch of the main sensory canal.
Z.c. Lateral canal. 7. ¢. 4. Branches of the lateral canal. m. Mouth. WM. C.
Main canal of the head. Mz. 6. Mandibular branch of the main canal. Mz.d.
Maxillary branch of the main canal. zar. Nares. op. Operculum. or. Orbit.
pa. Parietal. pp. Primitive pores. p¢. Post-temporal. sp. Spiracle. 8. or.
Suborbital branch of the main canal. Sp. 0. Supra-orbital branch of the
main canal. «. Branches in the pre-orbital region.
Fic. 4.—Longitudinal section through a cluster pore from the occipital
region of the head of Polyodon. Lettering as below.
536 WALTER EDWARD COLLINGE.
Fie. 5.—Longitudinal section through a primitive pore from the opercular
region of the head of Pol yodon.
Lettering.—ep. Epithelial cells. x. 7. Nerve-fibres. p. Pit. sez. o. Sense-
organs.
Fic. 6.—Sensory organs from the lateral canal of Polyodon.
Lettering—a. Modified cluster pore. 4. Sensory organ, unknown. ¢,
Sensory organ, unknown, in group. d. Cluster pore undergoing division.
Fie. 7.—Canal bones.
Lettering.—a. Lateral view of the tooth-shaped bone in the anterior part
of the rostrum of Acipenser, connecting the supra- and _ sub-orbital
branches. x 8. &. The same, inner view. The apertures, a’, represent the
passage of various branches, the 4 indicates the course of the main branch.
x 8. ce. Dise-like bone from the lateral canal of Acipenser. x 5. @.
Dise-like bone from the lateral canal of Acipenser; natural size. d. Canal
bones from the head of Polyodon. x 2. e. Canal bones from the sub-
orbital branch of Acipenser. X 2.
Fic. 8.—Lateral view of the cranium of Polyodon, showing the exit of
the cranial nerves. The opisthotic and pterotic are added from Bridge
(16, pl. 55, fig. 3).
Lettering.—a. w. Anterior nares. d. sph. Dermo-sphenotic. fy. g. Articular
groove for the head of the hyomandibular. op. Opisthotic. ptr. Pterotic.
The foramina for the exit of the cranial nerves are indicated by their respective
numbers I to X.
Fic. 9.—Copy of van Wijhe’s figure (82, pl. xv, fig. 4) of Polyodon,
showing the relations of the cranial nerves to the hyomandibular, &c.
Lettering as below.
Fic. 10.—Corrected figure showing, in my opinion, the right position of
the hyomandibular bone and cranial nerves, &c.
Lettering.—c.h. Cerato-hyal. e¢. 4. Inter-hyal. 4.4. Hypo-hyal. 4. m. Hyo-
mandibular. 7. Mandibulo-hyoidean ligament. /’. Quadrato-suspensorium
ligament. Je. p. Ethmo-palatine ligament. J. Meckel’s cartilage. m. ad.
Mandibular adductor muscle. pm. Outgrowth of masseter muscle. p. g.
Palato-quadrate. g. Quadrate. 7. m. Ramus mandibularis, fascialis. 7. m’.
Ramus mandibularis, fascialis internus. 7. m'’. Ramus mandibularis, fascialis
externus. 7. m. 7% Ramus maxillaris inferior, trigemini. 7. m. s. Ramus
buccalis. 7.0.p. Ramus opthalmicus profundus. +7. 0. s. Ramus opthalmicus
superficialis. 7. o¢. Ramus oticus. sy. Symplectic.
Fies. 1] and 12.—Diagrams illustrating the sensory canals of the head and
part of the lateral canal of the trunk, and the nerves which innervate their
sense-organs. The canals and branches are coloured yellow, the trigeminal
nerve blue, and the facial red. The cranial nerves are numbered respectively
Ito X. On the left-hand side the connection between the supra- and sub-
THE SENSORY CANAL SYSTEM OF FISHES. 537
orbital branches is not shown, in order to show the branching of the olfactory
nerve. Z. C. The most anterior portion of the lateral canal, where it
joins with the main canal of the head, MW. C.; in its course forwards it gives
off the branches Oc. com., Hyom., S. or., and Sp. O. Hyom. The hyoman-
dibular branch, innervated by the ramus opercularis superficialis No. 4,
7.0.4. It divides into two smaller branches, viz. a mandibular, Mz., and a
maxillary, Mz. §. or. The sub-orbital branch passing posterior to the orbit
and meeting with the supra-orbital, Sp. o., in the region marked Sp. o%. In
the region Sp. o'. short branches are given off. .§.,8. O!. The combined supra-
and sub-orbital branches which give off in their anterior portion the branches
S. 8. O?. and S. §. 0%.; 8.8. O*. indicates where the combined branches make
an outward curve and join with their fellow of the opposite side.
Lettering.—Hyom. Uyo-mandibular branch. 7. Lateral division of the
vagus. J/'. Branch of lateralis vagi. JZ. C. Lateral canal of the trunk. M.
Mouth. M. C, Main canal of the head. MM/z. Mandibular branch of the hyo-
mandibular branch. Mz. Maxillary branch of the hyomandibular branch. J.
Nares. Oc. com. Branches from the main canal in the occipital region. Or.
Orbit. 7. 6. Ramus buccalis, with the branches a, 6, c, d, ande. 7. m.
Ramus mandibularis, trigeminal. 7. m'. Ramus mandibularis, facial. 7. me.
Ramus maxillaris. 7.0. Ramus opercularis superficialis, with the branches
1, 2, 3, and 4. 7. 0. p. Ramus opthalmicus profundus. +r. o. s. Ramus
opthalmicus superficialis. 7. o¢. Ramus oticus. 7. p. Ramus palatinus.
S. or. Sub-orbital branch of the main canal. Sp. Spiracle. Sp. o. Supra-
orbital branch of the main canal. §. 8. 0. The combined supra- and sub-
orbital branch. Vg. Vagus. V. gi. Ganglion of vagus.
Fic. 13.—Dorsal view of the head of Acipenser. ‘The sensory canal
system is marked in red. Lettering as below. x 3.
Fic. 14.—Lateral view of the same. x 2.
Lettering.—d. eth. Dermo-ethmoid. d. ec. eth, Dermo-ect-ethmoid. d. oe.
Dermo-occipital. ep. Epiotic. fr. Frontal. 2. C. Main canal of the head.
oc. com. Occipital commissure. op. Operculum. or. Orbit. pa. Parietal
pt. Post-temporal. §. 07. Sub-orbital branch. s. 0, Sub-orbital bones. Sp. O.
Supra-orbital branch. Sz. Squamosal.
r
’
.
4 « es
j
} j 7 ais tA Ay fe be
- i]
: i i} > ww a we
: i \ aud
F ‘
Ta ee A (i LP . i } i? t ‘EG OF
: sare ‘ a! R dae J , ‘
} \
: . -
; aily ' ; iG
J é or
/
4 ue i ‘ — ;
i 4 ‘on. 6
' m!
- ! 4
: ea be
. , f
7 mm A ’ i Lvs.
ait ' f +)
i , ;
‘
=
{ 4 ‘ f
af AY
a j i hy int ; }
‘
¢ re lah sree ’ Y Vier Mf
,
ae . ri rf
‘ i i
7 A i i
a #
inwe. ii’ a } ‘ ‘
f
.
"is ‘ i al j : ne rw
‘ }
‘ _ ’ ar Ste 2
y b> Al, a ; ( }
’ f ime A
. ag ‘ lal a, NS (Pea lim, cena
aan ‘ a
, 4 a oy 1 ] j ie nike ee ae
k i : Poth Sey oe ¢ 6 Gi Oy tie
mal F 1 ie ¥) pf i " 4
: ‘Y Awa “t : une
hiy { ' :
: 4 yi } uf ff 4 inate aL® j
a £ -,
° aft a ¥ aii? : ’ L iy
‘ : aif i wins
é _-< " J a ="
ee) ee NEP, ay UST hy Aes SY ined porkta
; ' é ¥¢ an ee 8, . IOI VS aA a a "
j tn an ‘ia » ae iti ME iy (alti J
oP re Boi it Lihat Nite by OMNI Tear His We A fa a La
‘iy wa An LL iT Mi
INDEX OV Ol. 3.6;
NEW SERIES.
Allen, Edgar J., studies on the ner-
vous system of the Crustacea, 461
and 483
Andrews on some abnormal Annelids,
435
Annelids, abnormal, by EH. A. An-
drews, 435
Bacteria, morphology of, by E. Klein,
al
Baer’s law, by Adam Sedgwick, 35
Balanoglossus, Spengel’s monograph
on, reviewed by McBride, 385
Beddard on the Oligocheta of tropical
Eastern Africa, 201
Bosanquet on a Gregarine from the
Earthworm, 421
Bourne, Professor A. G., on Moni-
ligaster grandis, 307
ES on certain points in the
anatomy and development
of some Karthworms, 11
Canal system of Ganoid fishes, 499
Collinge on the sensory canal system
of Fishes, 499
Crustacea, nervous system of, by
Allen, 461 and 483
Dendy on Lelapia australis, a
living representative of the fossil
Pharetrones, 127
VOL. 36, PART 4.—NEW SER,
Earthworms, certain points in the
anatomy and development of, by
Professor A. G. Bourne, 11
Earthworms, various, 201 and 307
Euphrosyne, by McIntosh, 53
Fishes, sensory canal system of, by
W. E. Collinge, 499
Ganoidei, sensory canal system of,
499
Gould, L. J., on the minute structure
‘of Pelomyxa palustris, 295
Gregarine, notes on one from the
Earthworm, by W. C. Bosanquet,
421
Giinther, R. T., on Limnocnida
tanganyice, 271
Hubrecht’s Spolia nemoris, 77
Klein, E., on the morphology of Bac-
teria, 1
Lankester, E. Ray, notice of his
editorship, with portrait, i—iil
Lelapia australis, a living repre-
sentative of the fossil Pharetrones,
by Dendy, 127
Limnocnida tanganyice, by R.
T. Ginther, 271
Lobster, nerve elements of, by Allen,
461
00
540
Magelona, by McIntosh, 53
McBride, review of Spengel’s mono-
graph on Balanoglossus, 385
McIntosh, a contribution to our
knowledge of the Annelida, 53
Moniligaster grandis and other
species, by Professor A. G. Bourne,
307
Nervous system of Crustacea, by
Allen, 461 and 483
Oligocheta of tropical Hastern Africa,
by Beddard, 201
Ornithorhynchus, the bill and hairs
of, by Professor Poulton, 143
INDEX.
Pelomyxa palustris, minute struc-
ture of, by Lilian J. Gould, 295
Poulton on the structure of the bill
and hairs of Ornithorhynchus, 143
Rudiments in embryonic development,
by Adam Sedgwick, 35
Sedgwick on the law of development,
commonly known as von Baer’s
law, 35
Spengel’s monograph on Belanoglos-
sus reviewed, 385
Spolia nemoris, by
Hubrecht, 77 _
Sponges (Lelapia), by
Dendy, 127
Professor
Arthur
PRINTED BY ADLARD AND SON,
BARTHOLOMEW CLOSE, E.C., AND 20, HANOVER SQUARE, W.
BS Hi hy
ee iy vi)
a ¥
' ‘ pl ey
)
hi
AY Ah 4 4
ACMI
3 2044 106 2
sy oa an antares ee —
sad a a a eee mone I a rt
asa eS ae Sau
hr g me
a .
aN reeamater mao ne
heron ee neces