Quarterly journal of microscopical science

SAE Ohms
olen

Te iy co eBE IO te ints rte

yee
ert arte
F enbebebetobran anton

ere Serena 9
a ee ee

Rebound 48

HARVARD UNIVERSITY

Wel
Ash

LIBRARY
OF THE
MUSEUM OF COMPARATIVE ZOOLOGY

1521

i! ny 2

\

ee i ye
| ida ae

QUARTERLY JOURNAL

OF

MICROSCOPICAL SCIENCE:

EDITED BY

E. RAY LANKESTER, M.A., F.R.S., F.LS.,

Fellow of Exeter College, Oxford, and Jodrell Professor of Zoology in University
College, London ;

WITH THE CO-OPERATION OF
W. T. THISELTON DYER, M.A., C.M.G., F.R.S., F.L.S.,
Assistant Director of the Royal Gardens, Kew ;

eg UGE IN Ds. Eko.
Joint-Lecturer on General Anatomy Gad Pirolaly in ‘i Medical School of
St. Bartholomew's Hospital, London ;

H. IN, MOSEDEY, McAy i it.o5) 2 .lS:.,
Linacre Professor of Human and Comparative Anatomy in the University of Oxford,
AND

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

VOLUME XXIII.—NeEw Srnizs.
With Aithographic Plates and Engrabings on Wood,

LONDON: :
J. & A. CHURCHILL, 11, NEW BURLINGTON STREET.
PSs:

=: Gea EET

Bor eee fe fe y
a ee | ME TY Ws

te

rid
‘

Fo o>
Weil Mit ie .

} id Ass pA i

7” ay

- >. }
ossygcsl Wt Aaa foe Sey
agi ny qytidagaa Rika a aly, Dp |

ingare ACTA Oey

f
}
{

CONTENTS.

CONTENTS OF No. LXXXIX, N.S., JANUARY, 1883.

MEMOIRS:

PAGE

On the Relation of Pathogenic to Septic Bacteria, as illustrated
by Anthrax Cultivations. By HE. Kuery, M.D., F.R.S., Joint
Lecturer on General Anatomy and Physiology in the Medical
School of St. Bartholomew’s Hospital, London . te i

Sigcestions as to

y Epwarp B. Pouxron, M.A.

The Tongue of Perameles nasuta, with some
the Origin of Taste Bulbs,
(With Plate I)

ae Nips and Livy, Matter. by Louis Etspere, M.D., of New
Yor

. : : : : i
‘Ihe Life History of the Liver-fluke (Fasciola hepatica). By A.
P. Tyomas, M.A., F.L.S., Balliol College, Professor of Natural
Science in University Coilege, Auckland, New Zealand, late De-
monstrator in the Anatomical Department, University Museum,

Oxford. (With Plates II and ITI) : : : aets09

Note on the Early Development of Lacerta muralis. By W.F.
R. Wrxpon, B.A., Schoiar of St. John’s College, Cambridge,
Assistant Demonstrator ia the Morphological Laboratory of the
University. (With Plates 1V, V aud V1) : : . 134

On a Crustacean Larva, at one time supposed to be the Larva of
Limulus. By the late R. V. WitLemors-Sunm, Ph.D., Naturalist
on board H.M.S. ‘Challenger.’ (With Plate VII). - 145

On Plasmolysis and its bearing upon the Relations between Cell-
wall and Protoplasm. By F. O. Bowrr, M.A., Lecturer on
Rotany at the Normal School of Science, South Kensington.
(From the Jodrell Laboratory, Royal mt Kew.) (With
Plate VIII) . . . . oe bt

69

“a eye
nerd ‘tie
24 ovaries em
’ ; la coment
atu ie
cate isp
. ™ | iia | vate
seit 42 1 fut peed eg
ieee
i if
— ys - sa\1 4 1 vif. e
all i + a ; ' "
: ) aseee?
avs
$0s
ia vegoem om tm
] tho rue hib¢ \e onto ol ui
oir ; . VA»A TAD Waee

fe (eek! Se: iy | nage) yromalig a i. /
5. if 27 Kal Téci | yf
Ooo ov eaecnd DA hee loot le ny

CONTENTS.

CONTENTS OF No. XCI, N.S., JULY, 1883.
MEMOIRS:

On the Ancestral Form of the Chordata. By A. A. W. Huprecnt,
Professor of Zoology at the ies of Utrecht. (With Plate
©. 099 ae ;

The Renal Organs (Nephridia) of Patella. By J. T. Cunninenam,
Fellow of University College, Oxford

A Rare Form of the Blastoderm of the Chick, and its Bearing on
the Question of the Formation of the Vertebrate Embryo. By
C. O. Wuirmay, Ph.D. (With Plates XXIV and XXV)

On the Development of the Pelvic Girdle and Skeleton of the Hind
Limb in the Chick. By Atice Jounson, Newnham se
Cambridge. (With Plates XXVI and XXVII) .

The Development of the Mole (Talpa Europea). The Formation
of the Germinal Layers, and Early Development of the Medullary
Groove and Notochord. By Waiter Heare, Demonstrator in
the Morphological Laboratory at Cambridge. (With Plates
XXVIII, XXIX, XXX and XXXI)

The Tongue of Ornithorhyneus paradoxus: the Origin of Taste
Bulbs and the parts upon which they occur. By Epwarp B.
Poutton, M.A., of Jesus and Keble Colleges, Oxford. (With
Plate XXXII)

Observations upon the Foetal Membranes of the oe and other
Marsupials. By Henry F. Oszory, ScD., Assistant Professor of
Natural Science, Princeton, U.S.A. (With Plate XX XIII)

CONTENTS OF No. XCII, N.S., OCTOBER, 1883.

MEMOIRS:

Observations on the Genus Pythium (Pringsh.). By H. Marsna.n
Warp, M.A., Fellow of Christ’s College, Ca mbridge, Assistant
Lecturer in Botany at the Owens College, Victoria University.
(With Plates XXXIV, XXXV and XXXVI)

On Budding in Polyzoa. By Aurrep C. Happon, M.A., Professor
of Zoology in the Royal College of Science, Dublin. (With

Plates XXXVII and XXXVIII). .

PAGE

399

412

453

473

485

516

vi

CONTENTS.

The Structure and Relations of Tubipora. By Sypney J. Hickson,
B.A., B.Se., Scholar of Downing College, Cambridge, and As-
sistant to the Linacre Professor, Oxford. (With Plates XXXIX
and XL). 2 ‘ ; :

On the Malleus of the Lacertilia, and the Malar and Quadrate Bones
of Mammalia. By M. L. Dotto, Assistant-Naturalist in the
Royal Museum of Natural History, Brussels. (With Plate XLI)

Notes on Echinoderm Morphology, No. VI. On the Anatomical
Relations of the Vascular System. By P. HerBeRT CaRPENTER,
M.A., Assistant Master at Eton College.

Recent Researches upon the Origin of the Sexual Cells in Hydroids.
A Review. By A.G. Bourys, B.Sc., Lond.

On the Osteology and Development of Syngnathus peckianus
(Storer). By J. Puayrarr McMurnricn, M.A., Professor in the
Ontario Agricultural College, Guelph; Canada. (With Plates
XLII and XLIII.)

TirLtzE, ConTENTs AND INDEX.

PAGE

556

579

623

On the Relation of Pathogenic to Septic Bac-
teria, aS illustrated by Anthrax Cultivations.

By

E. Klein, M.D., F.R.S.,

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

Tue research, of which in the following report I propose to
give the first instalment, had for its object, first, to investigate
whether and how far the Bacillus anthracis undergoes any
change, morphologically and physiologically, when cultivated
artificially ; and secondly, whether ordinary bacteria of putre-
faction and septic fermentations can by artificial cultivations
be so modified as when introduced into the body of an animal
to be productive of disease, that is to say, whether it is possible
for an innocuous saprophyte to assume the properties of an
obnoxious pathogenic organism.

It is well known that bacteria of ordinary putrefaction may
be introduced, either by ingestion with the food into the ali-
mentary canal, or by inoculation into the skin, the mucous
membranes, the subcutaneous or submucous tissue, or by direct
injection into the vascular system, without there being pro-
duced in the animal experimented upon any appreciable dis-

1 Reprinted from the ‘ Reports of the Medical Officer of the Local Govern-
ment Board for 1881.’

VOL. XXIII.—NEW-SER. A

2 DR. E. KLEIN.

order that could be directly brought into connection with the
bacteria, provided that these latter be introduced in small
quantities only. Large quantities, on the other hand, are pro-
ductive of putrid intoxication, not so much on account of the
presence of bacteria as on account of the now well known and
generally accepted chemical putrid poison of Panum, Bergmann,
and others.

It is equally well known that pathogenic bacteria prove their
efficacy when introduced in minimal quantities, for it is essen-
tial to their character to find in the animal organism a suitable
soil for multiplication and by their increase of numbers tu
produce directly or indirectly a definite disorder in the animal
economy.

In order, then, to investigate whether saprophytic bacteria
have assumed the properties of pathogenic organisms it is
necessary to bear in mind that they must be expected to show
these properties after their introduction in minimal quantities
into the animal organism ; that is to say, they must be capable
to resist and overcome the effects of the healthy tissues—effects
proving invariably deleterious to ordinary saprophytic bacteria—
and, having done so, of starting a definite disorder in the
tissue.

It will be admitted, I presume, that it is not necessary that
the disorder be of a general nature; in some well-established
instances, such as anthrax, febris recurrens, pneumo-enteritis
of the pig, the malignant cedema (Koch), a general disorder
ensues on the introduction of the pathogenic organism, but in
other instances, such as the actinomycosis in cattle and man,
known through the researches of Bollinger, Jahn, Israel, Pon-
fick, the effect of the invasion by the actinomyces is at first, at
any rate, of a purely local nature, being generally limited to
the lungs; similarly the introduction of tubercular virus into
the anterior chamber of the eye is followed by an eruption of
tubercles in the iris (Cohnheim and Salomonsen), and of the
same character is the pulmonary tuberculosis. occurring in dogs
after inhalation (Tappeiner) of the tubercular virus. The ma-
lignant ulceration in mice (Koch) is in the same way at first a

RELATION OF PATHOGENIC TO SEPTIC BACTERIA. 3

purely local effect of the introduction of a pathogenic organism.
And it will be necessary to postulate at least this of a would-
be pathogenic organism, viz. that its effect after its introduc-
tion into an animal should be local, but need not be so if
general.

As is well known, Professor v. Nageli is a firm advocate of
the “sporting” of harmless saprophytes, they becoming con-
verted under certain conditions into pathogenic organisms.
But, on the other hand, Professor v. Nageli also maintains that
pathogenic bacteria may become converted into harmless sapro-
phytes. His views are stated with great clearness in his work,
‘ Die niederen Pilze,’ &c., Munich, 1877. From many years’
studies, carried on with patience, and after a strictly experi-
mental method, he arrived at these important conclusions.

Professor v. Nageli makes the very widest allowance for
bacteria, Inasmuch as a harmless form, when brought under
suitable conditions, may become the origin of an infectious dis-
ease, may through generations retain this power, and when
again placed under different conditions may change into an
inactive form.

Dr. Hans Buchner, a pupil of v. Nageli, and working under
this latter’s directions, put these general statements to a special
test, and succeeded, or maintains to have succeeded, in con-
firming them. He claims to have succeeded in changing, by
successive artificial cultivations under constantly varying con-
ditions, the Bacillus anthracis, of previously deadly power,
into a perfectly inactive and harmless bacillus, which in mor-
phological respects appeared then identical with the motile
Bacillus subtilis (Cohn) of hay infusion. But he also
thinks he has succeeded, what is of even greater and more
fearful consequence, in transforming, through artificial cul-
tivations under ever varying conditions, the notoriously
harmless Bacillus subtilis of hay infusion into deadly
Bacillus anthracis.

Dr. Buchner’s paper, “‘ Ueber d. exper. Erzeugung d. Milz-
brandcontag. aus d. Heupilzen . . . .,” which gives the results
of a very large number of observations, is published in the

4. Dk. E. KLBIN.

‘Sitzungsb. d. Math.-Physikal. Classe d. k. b. Akad. id. Wiss.
zu Miinchen,’ 1880. Heftiii, p. 368 et passim. I shall have
opportunity to return below to Buchner’s assertions in detail,
as I shall have to criticise some of his facts and deductions, but
at present I wish to point out that one of Buchner’s funda-
mental propositions, viz. that the Bacillus anthracis and the
bacillus of hay are morphologically (with the exception of the
motility of the latter) identical, is altogether erroneous. The
two kinds of bacilli are not identical, and never become iden-
tical, however long they may be cultivated in artificial cultiva-
tions ; and on this point I must with Koch (Cohn’s ‘ Beitriige,’
ii, Bnd. iii, and ‘ Aetiol. d. Milzbr., p. 21), most decidedly
oppose Buchner. It is true that Buchner admits some very
essential differences between the two, but these differences
refer to chemical and functional relations. I shall point out
below in detail these differences.

Buchner cultivated, at a temperature of 35°—37° C., the
Bacillus anthracis, originally derived from the spleen of a
white mouse dead of anthrax, in 0°5 per cent. solution of
Liebig’s meat extract, with or without the addition of peptone
or sugar. As a first result of his observations on white mice
Buchner found (p. 383) “that the infectious activity of the
fungus becomes the more diminished the more generations it
had passed in the artificial cultivations.”” But on looking
carefully into his facts we notice that the above result does not
come out in so simple and regular a manner as is represented
in the above sentence; for in one series of cultivations of
Bacillus anthracis, carried on in a nourishing fluid of 10
parts of Liebig’s meat extract, 8 parts of peptone, and 1000
parts of water, Buchner found (p. 383 et passim) that the
inoculations with the first, second, third, and fourth remove or
generation produced always anthrax, whereas those with the
fifth, sixth, seventh, and eighth did not yield any positive
results ‘if the same quantity of infective material was used,”
but “if larger quantities are used positive results were ob-
tained.” In like manner in other series of cultivations he finds
great differences as regards the activity of the bacillus of the

RELATION OF PATHOGENIC TO SEPTIC BACTERIA. 5

various “generations.” Thus he found (p. 384) that minute
quantities of the first cultivation proved effective, whereas the
second, third, and fourth, in small quantities, proved ineffectual ;
but the fifth cultivation proved effective in large quantities ;
the sixth was ineffective. In a series of cultivations in meat
extract, peptone, and sugar, the second cultivation proved
active; the third and fourth inactive; the fifth active; the
seventh, eighteenth, and even the thirty-sixth generation proved
active. From all these and similar observations Buchner de-
duces that the Bacillus anthracis undergoes a gradual
change, by which sooner or later it is rendered altogether
harmless and inactive. This change is, however, not one of
morphological character. I shall, further below, when de-
scribing my own observations on these subjects, have te return
to Buchner’s observations, and I shall then show that all his
facts are easily explained by the aid of my own observa-
tions, but not by the theory of a gradual diminution of
activity of the Bacillus anthracis. There is such a thing
as a real diminution of activity of the Bacillus anthracis in
artificial cultivations; the inactivity on white mice of some
cultivations of the Bacillus anthracis and not of others is
due to a variety of circumstances, one of which, at any rate, is
this, the absence of spores. If Buchner had tested his cul-
tivations on guinea-pigs or rabbits he would have obtained
altogether a different result, always supposing that he worked
with pure cultivations of Bacillus anthracis.

Kech (‘Zur Aetiologie d. Milzbr.,’ p. 22 et passim) con-
siders it probable that Buchner’s uncertain and unequal results
are explained by the fact that his (Buchner’s) cultivations not
being absolutely guarded from contamination with other non-
pathogenic bacilli, he may have had, and probably did have,
in some of his cultivations the Bacillus anthracis, originally
sown, diluted, or altogether suppressed by the overgrowth of
the non-pathogenic bacillus. With Koch I fully think this
objection well justified, especially since Buchner does not admit
a difference between non-moving non-pathogenic bacilli and
the non-moving Bacillus anthracis.

6 DR. E. KLEIN.

[1 am acquainted with Dr. Greenfield’s paper on cultivations
of the Bacillus anthracis communicated to the Royal
Society on June 17th, 1880; but I am unable to find in it any-
thing to which I can attach importance, with the exception of
some observations that repeat earlier experiments of Dr.
Burdon-Sanderson and of Dr. Buchner. In my opinion some
of Dr. Greenfield’s observations contain internal evidence that
he was occasionally operating with some harmless bacillus and
not with anthrax bacillus at all; and I cannot admit his
claim to speak with authority on the etiology of splenic fever.
It may be noticed in this connection that Koch, in his last
elaborate paper on the subject, has not made a single mention
of Dr. Greenfield’s assertions. ]

Buchner noticed, as his cultivations rose in degree of genera-
tion, that the bacilli showed a change in their general mode of
growth, inasmuch as, unlike their previous behaviour so often
described by Pasteur, Koch, and others, they ceased to form
the beautiful cloudy and flaky felt-work (Pasteur’s “ en fila-
ments tout enchévetrés, cotonneux ”’) rising from the bottom of
the culture-vessel into the otherwise perfectly clear nourishing
fluid, but they graduaily assumed the tendency to adhere to the
walls of the vessel. After a duration of cultivation of ninety
days (Buchner calls this the 900th generation), this condition
became very pronounced, and in later generations the bacilli
formed the same pellicle on the surface as do the harmless hay
bacilli. On changing the mode of growth he ultimately suc-
ceeded in obtaining bacilli that in this respect did not differ
from the typical hay bacilli. With reference to this point I
am inclined to think that Koch (‘ Aetiologie d. Milzbr., p. 22)
is right in refusing to accept this as proven; he says that
Buchner had not sufficiently guarded himself against outside
contamination, and therefore it is quite possible that he intro-
duced into his cultivations at one or another step a common
bacillus which after several cultivations became so numerous
as to replace altogether the original Bacillus anthracis.

With reference to Buchner’s cultivations, by which (using
blood for his nourishing fluid) he gradually changed the

RELATION OF PATHOGENIC TO SEPTIC BACTERIA. 7

common hay bacillus into the Bacillus anthracis, Koch
(1. ¢., p. 26) maintains that what Buchner really had before him
in those cases in which he had animals die after the inocula-
tion with the hay bacillus cultivated in blood for many genera-
tions, was not anthrax bacillus, but the baccillus of malig-
nant edema(Koch). I think Koch’s criticism is very thorough
and has every probability for itself. I certainly should be
surprised to find that a cultivation in blood treated in
Buchner’s manner (1. c., p. 405), viz. without any precaution
against accidental contamination, and which naturally would
undergo putrid changes, did not yield the cedema bacillus ;
this, when cultivated for several successive generations, as was
the case in Buchner’s experiments, would at last yield a culti-
vation in which the cedema bacillus is the only organism
present.

I now come to the most important research of this cycle, viz.
that of M. Pasteur. A saccinct summary of his results he
himself has given us in a remarkable address given during the
last International Medical Congress in London in August,
1881. This address has been reprinted, with a translation, as
a Parliamentary paper, under the title “ Animal Inoculation.”
It refers to the micrococcus of fowl cholera and to the
Bacillus anthracis. It is only the latter that interests us
here. By numerous previous observations M. Pasteur has
found that cultivating the Bacillus anthracis in chicken
broth at a temperature of 42° and 43° C., the bacillus, al-
though vigorously growing in the shape of the characteristic
convolutions of threads, nevertheless does not form spores (p.
10).

‘<TIn a month or six weeks the culture dies ; that is to say, if
one impregnates with it fresh decoction, the latter is com-
pletely sterile. Up to that time life exists in the vessel
exposed to air and heat. If we examine the virulence of the
culture at the end of two days, four days, six days, eight days,
&e., it will be found that long before the death of the culture
the microbe has lost all virulence, although still cultivable.
Before this period it is found that the culture presents a series

8 DR. E. KLEIN.

of attenuated virulences; everything is similar to what happens
in respect to the microbe of chicken cholera. Further, each of
these conditions of attenuated virulence may be reproduced by
culture. Lastly, since splenic fever does not recur (ne réci-
dive pas), each of our attenuated anthracoid microbes consti-
tutes for the superior microbe a vaccine, that is to say, a virus
capable of producing a milder disease. Here, then, we have a
method of preparing the vaccine of splenic fever.”

[These are stated as general propositions, and M. Pasteur
makes them without mention of any particular kind of animal.
He cultivates the anthrax bacillus at 42 C°. in fowl-broth,
and treats as indifferent the class of animal into which he
inoculates the cultivation. ]

“T was asked to give a public demonstration at Pouilly-le-
Fort, near Melun, of the results already mentioned. This
experiment I may relate in a few words. Fifty sheep were
placed at my disposition, of which 25 were vaccinated, and the
remaining 25 underwent no treatment. A fortnight after-
wards the 50 sheep were inoculated with the most virulent
anthracoid microbe. The 25 vaccinated sheep resisted the
infection ; the 25 unvaccinated died of splenic fever within 50
hours. Since that time the capabilities of my laboratory have
been inadequate to meet the demands of farmers for supplies
of this vaccine.! In the space of fifteen days we have vacci-
nated in the departments surrounding Paris more than 20,000
sheep, and a large number of cattle and horses. This experi-
ment was repeated last month at the Ferme de Lambert, near
Chartres. It deserves special mention. The very virulent |
inoculation practised at Pouilly-le-Fort, in order to prove the
immunity produced by vaccination, had been effected by the

1 Tt is matter of regret that the exact methods of preparation used by M.
Pasteur in his laboratory are not made public; so that the present research
has had to be conducted in ignorance of his details. The fact of his success in
producing what he calls a “ vaccine”—a something which when inoculated
into sheep produces some modified splenic fever that protects the sheep against
the after-production of fatal splenic fever when the virulent material is inocu-
lated into the sheep—may be taken as established.

RELATION OF PATHOGENIC TO SEPTIC BACTERIA. 9

aid of anthracoid germs deposited in a culture which had been
preserved in my laboratory more than four years, that is to say,
from the 2lst March, 1877. ‘There was assuredly no doubt
about its virulence, since in 50 hours it killed 25 sheep out of
25. Nevertheless, a commission of doctors, surgeons, and
veterinary surgeons of Chartres, prejudiced with the idea that
virus, obtained from infectious blood, must have a virulence
capable of defying the action of what I call cultures of virus,
instituted a comparison of the effects upon vaccinated sheep
and upon unvaccinated sheep of inoculation with the blood of
an animal which had died of splenic fever. The result was
identical with that obtained at Pouilly-le-Fort; absolute resist-
ance of the vaccinated, and death of the unvaccinated.”

I have hitherto had no experience with the inoculation of
sheep with cultivated Bacillus anthracis (but hope soon to
be able to gain some) ; and I cannot therefore say anything
about it, nor do I for one moment question the absolute reli-
ability of M. Pasteur’s successful vaccination of sheep with
Bacillus anthracis, and the immunity thus conferred upon
them, altho gh no such uniform results were obtained by his
assistant when repeating M. Pasteur’s experiments in Buda
Pesth, but what I will take the liberty of questioning is the
general application of these results by M. Pasteur and his fol-
lowers to anthrax in animals other than sheep, or to the other
infectious maladies. For I am able to show, that not only
does no such mitigation of activity on rodent animals take
place in the Bacillus anthracis when artificially cultivated
_and precluded from forming spores, but that no immunity is
conferred on rodent animals, if not succumbing to the effect of
such cultivated Bacillus anthracis. There are a number of
statements by M. Pasteur, such as the oxygen of the air being
the cause of the attenuation of the virulence; further, the
inability of the cultivated Bacillus anthracis to form spores
at a temperature of 42° and 43° C.; then the assertion that the
attenuated virulence once obtained is transmitted to the next
cultivation, the accuracy of some of which my experience
obliges me to question, of others directly to contradict.

10 DR. E. KLEIN.

Koch (‘Zur Aetiologie des Milzbrandes, in Mittheil. de
Kais. Gesundheits-Amtes.,’ Bnd. i, Berlin, 1881) made some
interesting contributions to the etiology of anthrax. The most
important points in his publication are the criticisms of
Buchner’s and Pasteur’s work on the subject. Although a
great deal of what Koch says when speaking in a perfectly
objective manner of Buchner’s observations is justified, I do
not think that there exists the same justification for all he says
of Pasteur’s work. That in the present extended knowledge
of anthrax, both in its etiology and pathology, but especially
the former, we owe more to Koch’s brilliant researches than to
the researches of all other observers taken together, including
Pollender, Brauel, and Davaine, the discoverers of the Ba-
cillus anthracis, will, I think, be readily conceded by all
who have had the opportunity of repeating some of Koch’s
experiments and observations, and have read his several com-
munications on this subject, and it is not want of respect to
the other workers in this field to concede this much. It will
likewise be conceded, I think, by all who read Pasteur’s various
contributions on the subject of the etiology of anthrax, that
Pasteur would have profited by a more careful study of Koch’s
observations and writings, particularly that the error of Pas-
teur’s earth-worm theory ! might have been avoided if he had
felt the significance of Koch’s previous observations on the
incapability of the Bacillus anthracis to form spores within
the body of an animal owing to the want of sufficient amount
of oxygen, and especially of Koch’s valuable observations on
the inability of the Bacillus anthracis to form spores in the
depth of the soil. And, again, objection may perhaps be taken

' According to this theory spores having been formed in the bacilli within
the organs of a buried animal that has died of anthrax, such spores are taken
up by earth-worms, carried up to the surface, and then deposited with their
castings. From the surface of the soil they find easy access into the mouth or
nostrils of animals grazing on that soil.

I shall show in a future report that the bacillus threads, as such, do not
survive even the initial stages of decomposition of the buried body. And it
cannot be supposed that earth-worms can feed on buried bodies in the few
days that may elapse before decomposition has set in.

RELATION OF PATHOGENIC TO SEPTIC BACTERIA. 11

with reason to Pasteur’s dogmatic way of dealing with patho-
logical facts, such as his arbitrary definitions and descriptions
of what constitutes anthrax and what constitutes malignant
oedema.

But while allowing all this, and perhaps even more, I think
all must extremely regret the tone in which Koch’s criticism
is made; all must think that his criticism would have been
much more valuable had it kept within strict bounds of an
objective statement. Koch shows by means of most valuable
observations, some of earlier date, published in his previous
writings (Cohn’s ‘ Beitrage zur Biologie d. Pflanzen,’ 2 Heft),
some of recent date, that no spore formation is possible in
Bacillus anthracis at a temperature below 15° C. or 59°
Fahrenheit. At a depth of 1 meter the temperature of the
soil in middle Europe is so low that the formation of spores in
Bacillus anthracis is practically impossible in an animal
buried unopened at that depth. Koch also proves (l.c., p.
20) by very instructive and direct experiments with earth-
worms and spores of Bacillus anthracis mixed with earth,
that Pasteur’s earth-worm theory cannot be correct.

Various considerations of the distribution of anthrax in
Germany and of the manner of the outbreaks of its epidemics
lead Koch to the assumption (p. 29, et passim) that the
natural habitat of the Bacillus anthracis is really in the
soil, and that its casual introduction into the body of an
animal and the production of anthrax in it, is only ‘a casual
excursion of a micro-organism not generally limited to such
a parasitism.”

As mentioned just now, there is a great deal of evidence
for such an assumption, especially the way in which in Ger-
many the disease makes its appearance in animals grazing in
fields and meadows which occasionally become flooded. In
these instances it is assumed that the Bacillus anthracis
growing in vegetable infusions (that it does readily do so in
some of them is shown by Koch) of a distant locality or in
the depth, is carried by means of water to the surface, and is
left here when the water is receding or drying up, to find

12 DR. E. KLEIN.

ultimately entrance into the body of an animal grazing on that
field. Hay infusions and other vegetable matter, it is true,
owing to their acid reaction, do not, according to Koch, form
a suitable nourishing fluid for the Bacillus anthracis ; but
Koch draws attention to its being a known fact that the
dangerous anthrax localities have a chalky or loamy soil, and
it is therefore possible that even where the vegetable infusions
in the soil would be acid (hay, some kinds of straw, barley,
grass, &c.), the carbonate of lime of the chalky soil would
suffice to neutralise the infusions.

Koch (l. c., p. 25) failed to find any diminution in viru-
lence of the Bacillus anthracis cultivated artificially up to
the fiftieth successive cultivation, an experience which I am
able, as I shall presently show, to confirm.

Before I give a detailed description of my own observations
on the cultivation of Bacillus anthracis, I have to state
the method of cultivation which was employed in this inquiry.
As nourishing fluid I have employed broth prepared from fresh
pork. About a pound anda half to two pounds of pork are
boiled in water for an hour or so down to about two pints of
fluid. The fat scum is removed, and the fluid, provided the
pork employed has been lean, filters tolerably clear through
filter paper. To obtain it, however, perfectly limpid, the broth
is cleared 4 la cuisine with egg-albumen and then filtered.
The filtrate, which will be spoken of as “ the pork broth,” is
of aneutral or faintly acid reaction ; in this latter case a sufli-
cient amount of carbonate of sodium is added in order to make
it neutral; it is then received in long-necked flasks, which are
then plugged with cotton wool.

In all cases, and I wish to state this once for all, where a
cotton-wool plug is spoken of, whether in connection with a
flask or a test-tube, it will be understood that a cotton-wool
plug of about two inches length is meant, in some instauces
two plugs one above the other, each about an inch long, being
used. The cotton wool, the flasks, beakers, filters, filter paper,
test-tubes, and all vessels used, are invariably disinfected by

RELATION OF PATHOGENIC TO SEPTIC BACTERIA. 13

exposing them for several hours, generally repeated several
times, to a temperature varying between 140° to 150° C. To
use cotton wool disinfected by prolonged (for several days or
weeks) steeping in absolute alcohol, or concentrated carbolic
acid solution, is not absolutely reliable. Overheating the
cotton wool in an air chamber to the above temperature till
singed has proved invariably and absoiutely safe for all culti-
vations; I have had to my regret failures in my cultivations
which could be referred to cotton wool soaked in concentrated
carbolic acid solution even for months. The same is to be said
of the flasks and test-tubes used. No amount of cleaning,
even with strong acid, is to be relied on; nothing but over-
heating gives reliance. I at first always used to heat the
vessel well all round over the open flame of a Fletcher’s
burner almost till the glass becomes glowing, and soon after
when the glass is still hot, but not as to do more than just
singe the cotton wool, I place in its neck the cotton-wool plug,
this having previously been overheated in the air chamber.
But lately I have overheated the vessels in the air chamber
to about. 140°—150° C. for several hours, several times re-
peated, having previously cleaned them with distilled water,
and dried them as far as possible; and I have found this
perfectly sufficient to disinfect them thoroughly. It cannot be
too strongly insisted on with Koch that the flasks and test-
tubes, and especially the cotton wool used as plugs for the
vessels, should be thoroughly sterilised by overheating, for it is
as much and as often that cultivations become thereby con-
taminated as by the non-sterility of the nourishing fluids or
the accidental entrance of organisms from the air.

The filtered nourishing fluid (pork broth) having been re-
ceived in a clean flask plugged, as above stated, with long and
clean cotton wool, is boiled for about ten to fifteen minutes. I
never fill the flask to more than half its volume with the broth,
in order to avoid the fluid rising too high during boiling, and thus
wetting the cotton-wool plug. This, although not necessarily
fatal, owing to the sterillity of the cotton wool, nevertheless [
always avoid in this and other cases, for the sake of cleanli-

14 DR. E. KLEIN.

ness, and to avoid all possible contamination. Immediately
before turning off the flame of the burner, and while the fluid
is still boiling, I place over the mouth of the flask a cotton-
wool cap, and keep this pressed over the mouth and upper
part of the neck of the flask by an inverted beaker pushed
firmly over it. The flask is then placed into an incubator,
and kept here at a temperature of about 32°—35° C. After
two or three days the flask, plugged but without the cotton-wool
cap, 1s again placed over the gas-flame, and the broth subjected
to boiling for five to ten minutes. While still boiling the
cotton-wool cap and beaker are placed over the mouth and
neck, and the flame is turned off. Such a flask with broth
may now be considered absolutely sterile; it may be kept in
the incubator for weeks and months—it will always remain
absolutely limpid and free of any organisms. Such broth will
in the following be always spoken of as “ sterile pork broth.”

This broth [ use either as such, i. e. as pure broth, or in
combination with gelatine, as “ gelatine pork,” in order to have,
as recommended by Koch (1. c.), a nourishing material, not of
fluid but of solid consistency. I consider, with many others,
this method of Koch’s, viz. of using gelatine as an admixture
to a nourishing fluid, aud thus converting it into a solid state,
a very great advance indeed in the methods of cultivating
bacteria, especially in securing pure cultivations not contami-
nated accidently, for then the sowing of a particular species
of bacterium is possible in a particular spot or spots, the
growth and progress can be easily watched and controlled,
and accidental contaminations can be readily recognised ; but
I shall show below that it is quite possible also without the
gelatine admixture after the method I use to be almost abso-
lutely guarded from accidental contamination, i.e. to have
pure cultivations. Koch has very minutely described the ad-
vantages of the gelatine method and his modus procedendi,
and he has given numerous photographic illustrations of
various species of bacterium in pure cultivations effected by
his gelatine method.

Koch recommends, in order to solidify the nourishing

RELATION OF PATHOGENIC TO SEPTIC BACTERIA. tes

material (in his case it was a solution of Liebig’s meat extract),
to mix with it purified and well sterilised and neutralised gela-
tine solution in such a proportion that the gelatine would form
2—8 per cent. Such a nourishing mixture is solid at ordi-
nary temperatures, and represents an excellent soil for sowing
on or in it the desired species of bacterium in dots or lines ;
kept in flat glass dishes or slides the examination with the
microscope can be easily carried out from time to time, and it
can easily be ascertained how and whether the sown species is
making progress, and accidental contaminations can thus be
easily detected and removed, all growths, owing to the solid
state of the nourishing material, being naturally limited to the
spot or line on which the bacterium has been sown. It is
necessary to keep the dish or glass in a chamber (under a bell-
jar) saturated with meisture. This is, in short, the essence of
Koch’s method. He maintains that such a gelatine material
remains solid at a temperature of 20°—25° C., sufficiently
high for the growth of all species of bacteria.

All this sounds very excellent, but when one comes to work
with it practically one finds that everything is not as perfect
and excellent as one imagines at first.

As is well known from the researches of Brefeld, Grawitz,
Wernich, and others, nourishing material in a solid state, such
as gelatine, boiled potato, bread, paste, &c., has been used for
the sake of obtaining pure cultivations, and for the sake of
easily watching and keeping under control the progress and
growth of particular organisms, e. g. Penicillium, Aspergillus,
Micrococcus prodigiosus, &c.; but most of these observa-
tions were carried on at ordinary temperatures. Koch, how-
ever, recommends it, after many observations, in the above
form for pure cultivations, even in the incubator, at 20°—25°
C. for all species of bacteria (Micrococcus, Bacterium termo,
and various species of baccilli, &c.).

The first difficulty one has to overcome is to obtain a sterile
and neutral clear and limpid gelatine solution. I have tried
every obtainable kind of gelatine, in which I was much assisted
by Dr. George Maddox, to whom I am under great obligations,

16 DR. E. KLEIN.

such as ordinary French gelatine, best Swiss gelatine, much
recommended to photographers by Dr. Eder of Vienna, best
French gelatine, called gold-label gelatine, isinglass, a peculiar
lichen-gelatine used by Chinese cooks to get a very firm jelly, and
various other kinds of gelatines ; and after a great many experi-
ments, both time-consuming and patience-trying, to enumerate
which would be a very unnecessary infliction on the reader, I
have found best answering our purpose a gelatine solution
prepared in the following manner: one part of “ gold-label
gelatine” (the tablets in which it is sold being cut up into
small strips) is soaked overnight in six parts of cold water, it
is then dissolved on the water bath ; this solution has a slightly
acid reaction ; to it is added carbonate of sodium just sufficient
to give it a neutral reaction. While quite hot it is filtered
through filter paper once or twice. (It must be borne in mind
that the filter paper, the vessels receiving the solid gelatine or
the filtrate, and all other vessels subsequently used for its re-
ception, are perfectly disinfected by overheating them.) The
process of filtering is carried out by using hot filters and filter
paper, keeping up the warmth by placing at two opposite sides,
as close to the filter as practicable, Bunsen burners. The fil-
trate is tolerably clear, but can be obtained perfectly clear by
adding to it after neutralisation egg albumen, and then boiling
it for several minutes. In this latter case it may be filtered
through calico or flannel previously disinfected. To the fil-
tered gelatine are then added three parts (not three times its
volume) of the above pork broth, so that we have now alto-
gether one part of solid gelatine, six parts of water, and three
parts of pork broth, which would be equal to one part of solid
gelatine in nine parts of fluid, or 11, per centum, This mix-
ture is received in several sterilised flasks, closed well with
long and thoroughly sterilised cotton-wool plugs, and is sub-
jected to boiling for 5—10 minutes. While still boiling, and
just before removing from the flame, the mouth of the flask is
covered with a cotton-wool cap, and a beaker is inverted over
it. The flasks are then placed in the incubator and kept there
at 32—35° C. for twenty-four hours, after which they are again

RELATION OF PATHOGENIC TO SEPTIC BACTERIA. bia

subjected to boiling for about five minutes. This I have found
to be sufficient to keep them sterile for ever after. This mix-
ture, which I will speak of as “ sterile gelatine pork,” remains,
even in the smallest quantity, solid up to a temperature of
25° C., a temperature generally sufficiently high for the growth
of bacteria.

Koch (‘ Zur Unters. d. Pathog. Organ.,’ p. 24) recommends
a mixture of gelatine and nourishing fluid in such proportions
that the gelatine amounts to about 24 to 3 per cent., and he
states this to have served in a solid state for the cultivation of
bacteria, not only at the ordinary temperature of the room, but
at temperatures varying between 20° and 25° C.

Now, no kind of gelatine which I have been able to lay hold
of has kept solid at a temperature of 20°—25° C. in such per-
centage, nor as 4 per cent. mixture, not even as 7°5 per cent.
Ten per cent. mixture is the lowest that I have been able to
keep solid at such a temperature. It is true a good many bac-
terial organisms grow tolerably well in a temperature about
15°—18° C., at which the 24—3 per cent. gelatine mixture is
solid, but their growth is very slow. In some instances, e. g.
Bacillus anthracis, the growth progresses tolerably well,
but in others it is extremely slow. To make spores of hay
bacillus sprout at such a temperature is exceedingly difficult,
and so it is also with the spores of some other kind of bacilli.
I have seen bacilli which absolutely refuse to grow at such a
temperature. Of micrococci some grow well, others do not.

It is clear, then, that if, as is the case in laboratory experi-
ments, one requires growth of a particular organism to take
place within a reasonable time, not to mention those cases in
which organisms do not grow at all at so low a temperature,
the above temperature, viz. 15°—18° C., is not sufficiently
high, and it is necessary to use gelatine mixtures stronger than

$—3 per cent. The above 11 per cent. mixture of gelatine
and pork keeps well and solid at 25° C., and at this tempera-
ture all bacteria that I have tried grow well and abundantly.

Thus far I have been describing the manner in which I pre-
pared the nourishing material, viz. sterile pork broth and sterile

VOL, XX111.—NEW SER. B

18 DR. E. KLEIN.

gelatine pork, which is to serve as stock for the cultivations. I
have used also other nourishing material, such as beef broth,
rabbit broth, &c., for the cultivation of various organisms ; but
the subject of the present report is the observations made with
Bacillus anthracis, and for this I have used, hitherto with
satisfactory results, the pork broth and gelatine pork only, and
I shall not enter on the present occasion into a consideration
of the other nourishing materials.

I now come to the description of the method of using the
above stock of nourishing material for the special cultivations
of the Bacillus anthracis.

(A.) A number of disinfected test-tubes and small flasks are
used, the latter of the capacity of an ounce or so, plugged with
disinfected cotton wool, the plug lifted, and each charged as
rapidly as possible for a fraction of their volume with the
nourishing material from the stock flask, and then plugged with
cotton wool. In the case of the gelatine pork, this is of course
first liquefied over the flame. The stock flask, if not emptied
by this process of charging, is subjected to boiling from five to
ten minutes. When charged and plugged each test-tube and
small flask is subjected to boiling for a few minutes; the boil-
ing is effected over a small flame in order to prevent the
over-boiling ; this is not so much to be feared in the case of the
flasks as in that of the test tubes. Thorough boiling for once
is generally sufficient to destroy every organism that may have
accidentally entered during the process of charging. Kept for
an indefinite time in the incubator at 32°—35° C. the fluid in
them remains bright and clear.

(B.) Glass cells of exactly the same nature as those that
were uf so great use to me in my research on the pneumo-
enteritis of the pig (see these Reports for 1877, p. 210), in the
majority of instances without any addition, in some with the
addition of a thin glass tube cemented to the glass slide and
leading into the cell; the outer opening of this glass tube is
plugged with cotton wool. This tube was chiefly added with
the view of facilitating the formation of spores, but as a rule I
found, ceteris paribus, if the other conditions for the spore

RELATION OF PATHOGENIC TO SEPTIC BACTERIA. 19

formation are present, the amount of air present in the glass
cell sufficiently large to enable the spore formation.

As in my former work so also now, I use olive oil to fix the
cover glass over the glass ring forming the sides of the cells.
The cover glass before being used is well heated over the flame.
A small quantity of the nourishing fluid (pure pork broth or
liquefied gelatine pork) is withdrawn from the stock flask by a
freshly drawn-out small capillary pipette; this is effected in
this manner: the cotton-wool plug of the stock flask is drawn
up for about half its length, and the one end of the pipette
being drawn out into a long capillary tube is gradually pierced
through the remaining half length of the plug and pushed
down till it reaches the fluid; the pipette is filled and with-
drawn, and the plug is again pushed down into its previous
position. By this means absolutely no access is allowed to
particles from the air into the stock flask, and at the same time
the capillary tube, while being pushed through the cotton-wool
plug, is cleaned from accidentally adhering particles. It must
be borne in mind that for the above purpose the cotton wool
must have been well sterilised by heat, because if not so,
the nourishing material in the stock flask is sure to become
contaminated by impurities adhering to the cotton-wool fibres,
some of these being pushed down as well as carried down into
the fluid by the capillary tube. From this pipette a drop is
quickly deposited in the centre of the cover glass, and this is
inverted and fixed on the ring of the glass cell, a drop of dis-
tilled water having been previously placed at the bottom of the
cell at a peripheral place. The cell is now “charged” and
ready to receive the organism that is to be cultivated in the
drop of nourishing material attached to the centre of the lower
surface of the cover glass. The process of charging the test-
tubes and flasks, as well as the glass cells, being carried out
in the air, is of course subjected to the complication of « con-
tamination with air organisms. In the case of the test tubes
and flasks this is remedied by subsequent boiling of the
charged and plugged vessels ; but in the case of the glass cells
a sterilisation after charging is for obvious reasons impossible,

20 DE. E. KLEIN.

and it is therefore necessary to take one’s chance, so to speak,
of having a number of failures owing to accidental contamina-
tion. And it is this very point, viz. the chance of contamina-
tion with air organisms, which makes the Koch’s method, as
recommended by him, impracticable in the case of many culti-
vations, as I shall have to point out below in detail.

It depends very much on the place and season where and
when the charging is carried out, as regards the accidental
contamination with air organisms. I have made some com-
parative studies on these points, and 1 think it worth while to
enter here more fully into them.

At first, when working at the laboratory of St. Bartholo-
mew’s Hospital Medical School, I charged my test-tubes from
my stock flask under carbolic-acid spray, the carbolic acid being
of the strength of about 5—6 per cent. From my note-book I
gather that in one series I charged sixteen test-tubes carefully
under the carbolic-acid spray, and placed them into the incu-
bator at about 85° C. Of these test-tubes one went bad in
the course of twenty-four hours, at which time it became
turbid owing to the presence of actively moving bacilli. In
another series of fourteen test-tubes two went bad. In a
third series of twenty-two test-tubes every one went bad,
although the method of charging under the carbolic-acid spray
was the same as in the other cases; but the conditions of the
atmosphere were not the same. While I had tolerably good
results in July and August I had very bad results in October,
and my failures, both in preserving sterile my stock fluids and
my test-tubes charged with them, became during this month so
numerous and persistent that I had to give up work altogether
for this period. To have cultivations exposed to the air and
not afterwards sterilised, as is almost the general rule in Koch’s
method of gelatine cultures, and to keep them pure was alto-
gether out of the question. The cause of these universal and
unconditional failures was not far to seek. During October
we had a good deal of dry weather with strong winds, and the
laboratory in which I worked faces Smithfield hay market,

RELATION OF PATHOGENIC TO SEPTIC BACTERIA. 21

whence a good deal of dust was blown into the laboratory. The
dust contained an enormous number of spores, especially of
bacilli, as was proved not only by the direct observation, but
also by the fact that every kind of nourishing fluid, Cohn’s
nourishing fluid, hay infusion, beef broth, mutton broth, pork
broth, &c., previously sterile, when exposed to the air on such
windy days for a second became very difficult of sterilisation ;
boiling for ten minutes and sometimes fifteen minutes, or even
more, did not produce sterilisation. After forty-eight hours’
incubation the fluid was invariably swarming with bacilli.

During July, August, and September, when days were
tolerably still, especially on rainy days, and there were no high
winds, test-tubes containing sterile nourishing fluid could be
kept open, i.e. the cotton-wool plug could be altogether re-
moved under carbolic-acid spray for several seconds, and
without being subjected to boiling after this remained sterile at
a temperature of 35° C., only a relatively small percentage,
varying from five to seven, being lost by air organisms. ‘This
is not at all an unsatisfactory result, considering that the
laboratory faced the hay market, and considering’ how easily a
contamination could occur under these conditions. But ata
time with high winds the contamination was so serious that
even prolonged boiling after exposure did not sterilise. This
is not to be wondered at if we remember that fluids containing
hay bacillus spores and some other bacillus spores require for
absolute sterilisation boiling extending up to and even over
half an hour (see Cohn’s ‘ Beitrage,’ ii Bnd., ii Heft). The
results obtained subsequently, when resuming my work, not
in the previous locality, but in the laboratory of the Brown
Institution, near Vauxhall, situated in a less contaminated
atmosphere, were very much more satisfactory. Comparative
experiments which I here made showed that exposing to the
air for half a minute sterile nourishing fluids contained in test-
tubes during windy weather yielded about 50 per cent. failures,
while exposing them to air under the carbolic-acid spray yielded
no failures in one series, it yielded 5 per cent. failures in
another series.

22 DR. E. KLEIN.

As I mentioned above, I charge my test-tubes rapidly either
with pure pork broth or with liquefied gelatine pork without
spray, and then boil them well for a few minutes, and doing
this in the laboratory of the Brown Institution I find it sufficient
to thoroughly sterilise the fluids.

In connection with this subject I would draw the attention
of the reader to the very important investigation made by Mr.
Watson Cheyne (‘ Transact. of the Pathol. Society of London,’
1879, p. 577) on the value of the carbolic-acid spray in the
inoculation of artificial cultures.

The plan of Koch of spreading out a large drop or small
quantity of liquefied gelatine nourishing fluid on a glass slide,
or on a flat glass dish, and having inoculated this on its surface
with the desired organism to place it in a chamber closed by a
bell-jar or the like, which is kept moist by putting into it (the
chamber) moist filter paper, has not been found practicable,
owing to the fact that the gelatine nourishing fluid having
solidified again dries up too soon, before the sown organisms
have had time to make a start, the moisture all settling on the
inner surface of the bell-jar. This condition I have invariably
found to obtain, even when the chamber was closed air-tight,
the bell-jar with ground margin being fixed by lard or oil on
toa ground-glass plate. Not only in ordinary temperature, but
still more so in the incubator, was this drying up of the gela-
tine nourishing fluid found to happen, and I have therefore
modified Koch’s plan by using the arrangement mentioned
before, viz. the closed glass cells, and the test-tubes plugged
with cotton wool.

The next important step in the cultivation of bacteria in the
nourishing material hitherto described, as contained in the test-
tubes or glass cells, is the inoculation of these materials with
the organisms it is desired to grow, i. e. the process of sowing.
It is, of course, obvious that if it is desired to cultivate a single
species of organism, it is necessary to sow a single species, and
to prevent contamination with air organisms, the nourishing
fluid itself being sterile. With reference to the first, it is
necessary to be certain that the material containing the seed

RELATION OF PATHOGENIG TO SEPTIC BACTERIA. 23

and to be transferred into the nourishing material contains no
other but the desired species. This is, however, not always a
simple matter. It is simple enough in the following cases :—
If I transfer to my nourishing material a droplet of blood taken
from the heart or the spleen of an animal just dead or dying of
anthrax, I am certain to have no other organism in the blood
except the Bacillus anthracis; or if I have an artificial cul-
tivation of Bacillus anthracis which from certain definite
naked-eye appearances (see below), and still more from micro-
scopic examination of anilin-stained specimens, I can pro-
nounce with certainty to be a pure cultivation of Bacillus
anthracis, I shall be certain that I shall again, ceteris
paribus, obtain a pure cultivation, if sowing out from this
cultivation. Again, if I take an infusion of hay in which fer-
mentation produced by the hay bacillus has been completed—
that is to say, in which the bacillus has passed its whole cycle
and has yielded an abundant crop of spores forming a fine brown
precipitate at the bottom of the infusion—and if I boil this in-
fusion for several minutes, I shall be sure to destroy everything
living except the spores of the hay bacillus, and if I sow out
from this so-boiled infusion I shall have the satisfaction to find
that the new growth contains only hay bacillus.

The above modified use of Koch’s method, viz. charging the
covering glass of the glass cell with a drop of liquefied gelatine
nourishing fluid, and when this has become solidified again to
inoculate it in one or two straight lines with matter containing
the bacteria to be sown, i. e. to dip a needle previously heated,
or the end of a freshly drawn-out capillary tube into the fluid
containing the seeds, and then to draw this needle or the capil-
lary tube quickly across the surface of the drop of gelatine
nourishing material once or twice; this method, I say, is in-
valuable for the study of the gradual changes those bacteria
undergo when subjected to incubation, the manner in which
they multiply ; further, to ascertain whether the desired or-
ganism has been sown, and whether only one kind of organism
or several are growing in the nourishing material; for the
glass-cell specimen can be easily examined, even with high

24 DE. EB. RLEIN.

powers of the microscope from time to time, without in the least
disturbing the growth. Koch, in his paper above quoted, has
minutely described all these advantages, and therefore I need
not further enter into this part of the subject, as I have no
doubt it must be obvious to every one who has the slightest
acquaintance with artificial cultivations of bacteria.

if you have sown in this manner a particular organism
well known to you, it is of course easily ascertained on micro-
scopic examination immediately after, whether the same is
present in any part of the line you have drawn over the gela-
tine drop in the above glass-cell specimen with your needle or
capillary tube. Thus, inoculating the gelatine drop with the
Bacillus anthracis or with its spores, or the spores of hay
bacillus, with sarcina, with torule, with Micrococcus pro-
digiosus, &c., you can at once find these seeds in the streak
you have drawn on the gelatine drop; according to the number
of seeds present in the material to be sown there will be more
or less numerous seed in that streak. If in addition to this
you have sown only one species of those named any accidental
contamination will soon be detected under the microscope in
the gelatine drop, say after a day or two or longer.

But supposing you are sowing a material of which you do
not know whether it contains any organism, or, if so, what kind
of organism, the case is altogether different, and ‘the value of
this method is not obvious ; on the contrary, may lead to serious
errors; in this way: the inoculation of the solidified gelatine
nourishing material, whether in my glass-cell specimens or
after Koch’s plan, on glass slides or flat dishes, must take
place in the air, and there is no means to prevent contamina-
tion with air organisms. Under ordinary circumstances and
working quickly the chances of such contamination are not very
great, but are, nevertheless, objectionable. Now, supposing
that you inoculate your gelatine in several specimens with the
material to be tested for organisms, you may find after a day or
two or more of incubation that in one or more of the specimens
in’ the streak you have drawn there is no growth whatever of
any organism, but outside it at other points an organism or

RELATION OF PATHOGENIC TO SEPTIC BACTERIA. 25

several organisms begin to grow, you will justly say that all
these organisms are accidental contaminations, air organisms ;
but if you found in the streak one species or more growing
you cannot conclude from this that you have transferred this
or these species from your original material, because your moist
needle-point or capillary glass tube may have caught these seeds
while passing through the air ; and this has actually happened
to me, not once, but repeatedly. I have several instances in
which I have sown or meant to have sown in one streak over
the gelatine drop of my cell specimen a particular species of
bacillus, and to my great annoyance I found, after several days’
incubation, in that very streak growing three different species
of organisms, viz. one kind of micrococcus and two different
species of bacilli. In another instance I wished to test a fluid
for the presence of an organism or organisms, consequently I
sowed it out on the gelatine in several of my glass cells, and I
obtained in the streak drawn over the gelatine drop two species
of organisms, a micrococcus and a bacillus; but as I ascertained
with a more precise method, the fluid contained no organism
whatever. These facts, it must be conceded, prove that the
method of Koch, although of great value in certain cases, is
less to be recommended in others, and therefore does not deserve
that unqualified praise which its author accords to it (I. ¢.), say-
ing as much as that this is the only method after which cul-
tivations of micro-organisms are to be carried on. I shall
presently show that there is a more reliable method (provided
the question is one of transferring one definite organism from
one fluid into a vessel containing the nourishing material), a
method in which the chances of contamination are less and the
method also, for other reasons more practicable.

The method of inoculation of the nourishing material which
I at first used was under the protection of the carbolic-acid
spray: a freshly drawn out capillary pipette is dipped into
the material to be sown, the cotton-wool plug of the test-
tube or flask containing the nourishing material is lifted under
carbolic-acid spray on one side just sufficient to admit the end
of the capillary pipette ; this being done the plug is again closed

26 DR. E. KLEIN.

over the mouth, the capillary tube is pushed down into the
nourishing fluid and then quickly withdrawn, and the plug
completely replaced. In this manner I have been very
successful in inoculating, without contamination with air
organisms, nourishing fluids with the special organism desired
to be sown.

But this method is in so far unpleasant as the spray pre-
vents one from seeing easily the capillary tube while being
pushed down into the test-tube. Although I used this method
a good deal, I have nevertheless recently employed a much
simpler method, which yields as good if not better results.
In the carbolic-acid spray the chances of contamination with
air organisms are small, as I have above stated, and when a
contamination with them occurs it is probably through the
spray catching them and carrying them into the test-tubes ;
but it must be obvious that this is really only a remote chance,
considering that in my casesI only momentarily lift one side
of the plug sufficient to admit the end of a capillary tube.

The best and most practicable method which I am now in
the habit of using, and which is almost absolutely safe
against accidental contamination, is this: the cotton-wool plug
of the test-tube or flask containing the nourishing material is
pulled out for about half its length; a capillary pipette
having been charged with the fluid to be sown is then gradu-
ally and carefully pierced through the remaining part of the
cotton wool (thereby clearing itself of adhering particles),
introducing it between it and the sides of the vessel; it is then
pushed down into the nourishing material and a trace of
the seed fluid emptied into the former. The capillary pipette
is quickly withdrawn, and the cotton-wool plug pushed down
into its old position. If the nourishing material is gelatine
pork, it is of course easily possible at will to deposit the seed
from the capillary tube either on to the free surface or into
the depth. If the seed fluid is to be obtained from an arti-
ficial cultivation contained in a test-tube or a flask, it is with-
drawn with a freshly made capillary pipette in exactly the
same manner as the seed material is introduced into the new

RELATION OF PATHOGENIC TO SEPTIC BACTERIA. 27

cultivation just described. It must be borne in mind that for
the success of this method it is imperative, in a greater degree
than in the other previously mentioned methods, that the
cotton-wool plug is thoroughly sterilised. For it is obvious
that if this is not the case, by the piercing of the cotton-wool
plug with the capillary pipette wool fibres are always carried
down into the nourishing material, and if these are not
thoroughly sterilised a contamination of the latter must inevit-
ably follow.

I had charged twelve test-tubes with pork broth, and had
them well plugged with cotton wool, well boiled on two suc-
cessive days, and placed into the incubator at 32°—35° C.;
they were kept here for two weeks, and remained perfectly
limpid and sterile. I then inoculated six of them in the above
manner with Bacillus anthracis of an artificial cultivation,
viz. introducing the bacilli by piercing the capillary tube con-
taining them through the cotton-wool plug. After twenty-
four hours all showed signs of accidental contamination. I
remembered that I had kept the test-tubes for several hours,
and at two successive days, at 140°—150° C.; but the cotton
wool had been tightly compressed in a beaker, and exposed only
for about an hour to a temperature of about 120° C. From the
remaining six test-tubes I removed the plugs of this cotton
wool, and closed them with fresh plugs of thoroughly sterilised
cotton wool. They were well boiled and kept in the incubator
for several days; as they remained quite limpid they were
inoculated after the same maner and with the Bacillus
anthracis of the same cultivation as in the case of the first
six test-tubes; the result was completely satisfactory; no
accidental] contamination occurred. From this it is clear that
the test-tubes and the nourishing material were sterile in both
instances, and also the bacillus to be sown was the same, and
in a pure state in both cases, but in the first the cotton wool
was at fault, hence the accidental contamination introduced
into the nourishing material.

As a rule, in cases where the naked-eye appearances do not
and cannot give indications of the actual state of the cultiva-

28 DR. E. KLEIN.

tions, i.e. whether pure or not, as is the case in most culti-
vations of bacteria, except, perhaps, of Bacillus anthracis,
I have employed both methods, i.e. I cultivated it in the test-
tube or flask, and at the same time controlled it under the
microscope, by cultivating in the above glass cell a specimen
in a drop of solid gelatine nourishing material.

In the cultivations of Bacillus anthracis in the above-
named neutral pork broth in test-tubes or small or large flasks
with which I worked, after three or four or more days’ incuba-
tion, even at a temperature so low as 20° to 25° C., a beautiful
whitish crop of the bacilli is visible at the bottom of the
vessel in the shape of a fluffy, nebulous, more or less filamen-
tous mass as incubation proceeds, gradually extending into the
further layers of the fluid, this latter being tolerably bright
and limpid. ‘These appearances have been well described by
Pasteur, and have been also noticed by Buchner (l.c. p. 876).
The cultivations which I carried on in the pork broth, from
one transfer to the other, all showed these characteristic appear-
ances, except in those few instances in which, as mentioned
above, an accidental contamination occurred. ‘These appear-
ances are so striking and peculiar that it can with certainty,
from the naked-eye inspection alone, be recognised whether a
given cultivation is one of pure anthrax.' If the growth after
the first few days does not present the peculiar nebulous and
filamentous masses at the bottom of the fluid, if the super-
natant fluid remains clouded and turbid, and especially if a
scum appears on the surface, either only where the surface of
the fluid adheres to the glass or over the whole surface, it can
be concluded with probability that the cultivation is impure,
there being generally present a microccocus or a scum-forming
bacterium or a bacillus, and this can be easily verified]by micro-
scopic examination. During the first two or three days of
incubation, however, the fluid is not limpid, but more or less

1 These appearances are much more striking in neutral cultivations than in
those of acid or alkaline reaction. In the latter instances there is never the
same copious growth as in the first.

RELATION OF PATHOGENIC TO SEPTIC BACTERIA. 29

uniformly turbid, the growth not being limited to the bottom
layers; but soon this changes, and the characteristic nebulous
masses are visible at the bottom, while the rest of the fiuid
is perfectly clear. The first turbidity is due to the uniform
distribution in the fluid of the bacilli, isolated and in chains.
If the vessel in which the cultivation is carried on is kept
quiet (I generally keep the test-tubes in a beaker with a layer
of cotton wool at its bottom), the peculiar anthrax bacillus
growth retains its original naked-eye appearances for a con-
siderable time ; shaking the test-tube up after several days’
incubation destroys the coherence of the bacillus mass, and
this latter breaks up into small flaky particles, which, how-
ever, readily sink to the bottom, so that the supernatant fluid
again becomes limpid. As long as sufficient nourishing
material is present in the fluid the bacillus mass will of course
continue to grow in amount, and when this does not any
longer take place the fluid is ‘‘ exhausted.” Now, watching
the behaviour of the bacillus mass afterwards, i.e. after the
mass has ceased increasing, these important facts become
obvious: that the bacillus mass becomes gradually smaller ;
this diminution is in some instances so rapid and conspicuous
that at a first inspection it seems that the cultivation is not
the same, but that it might have been changed for another ;
but there can be no doubt about it if the inspection is made
a few days later. This diminution goes on till only a few
flaky transparent masses are left in the fluid. Below I shall
show what the reason of the diminution is and what the micro-
scopical appearances are. During this process of diminution
and disappearance of the bacillus mass the fluid remains
always perfectly limpid. But I may at once state that this
disappearance of the bacillus mass has nothing whatever to do
with spore formation, as might be at first supposed; for
we know from the researches of Cohn that in a bacillus
mass composed of long and convoluted threads, such as the
Bacillus anthracis forms in these artificial cultivations, the
formation of bright oval spores soon sets in. The newly-formed
spores become liberated, the bacillus threads become trans-

30 DR. E. KLEIN.

parent and disintegrate, while the spores sink to the bottom
to form a minute precipitate. There is nothing of the sort in
our cultivations. As a general rule in the flasks and test-
tubes with fluid neutral pork broth no spore formation takes
place. Whether the cultivation is carried on at an ordinary
temperature or at a temperature of 20°—25° or 382°—35° C.
here is no spore formation. I have test-tubes with neutral,
pork broth in which exceptional spores have been formed
in the threads. These test-tubes were of the following
nature: in one test-tube an enormous mass of bacillus
threads had made its appearance while the cultivation was
kept in the incubator at 22°C. for about three weeks, the
masses of threads were very loose and consequently occupied a
large volume, so large that the growth extended almost to the
surface of the fluid; here numerous spores were found in the
superficial threads. Another test-tube of exactly the same
material inoculated with exactly the same generation of
Bacillus anthracis, and kept under precisely the same
conditions, did not develop any trace of spores; here the mass
of bacillus threads formed a more dense growth, and kept its
place at the bottom of the test-tube far away from the surface
of the fluid. In another test-tube I found that the mass of
bacillus threads at first growing at the bottom of the fluid
after some time sent out some bundles of threads which grew
along the glass wall to the surface of the fluid. Here also
numerous spores were found in the threads. Buchner (l.c.,
p. 370) says: ‘‘ The physiological cause of the spore formation
lies in the commencing deficiency of nourishing material.”
This is proved to be incorrect by our cultivation. The ex-
haustion of the nourishing fluid is long apparent and no spore
formation occurs, and vice versa, spore formation may be
observed long before there is any sign of exhaustion of the
nourishing matter, so much so that spores appear early in the
culture and again become converted into bacilli, but no spores
may be formed in this new generation. Pasteur says that he
prevents the formation of spores in the Bacillus anthracis
cultivated in chicken broth, by keeping the cultivation exposed

RELATION OF PATHOGENIC TO SEPTIC BACTERIA. Sl

to a temperature of 42°—43° C., which is still suitable for the
growth of the bacilli, but not for the spore formation ; in many
instances this is unnecessary, because even at ordinary tem-
peratures no spores are formed. Cohn and Koch maintain
that for the formation of spores in the bacillus of anthrax as in
other bacilli, a certain degree of warmth, a certain degree of
moisture, and a sufficient supply of air are indispensable for
the formation of spores. There can be no doubt that this is
so; Bacillus anthracis does not form spores in the body of
an animal, as I can fully confirm Koch against Pasteur, who
makes this assumption, and as I have repeatedly convinced
myself by direct and systematic observations, to be mentioned
in my next Report; but I cannot admit, if it is said that given
those three conditions, viz. a certain degree of warmth and
moisture, and a sufficient supply of air, supposing other things
unchanged, the Bacillus anthracis must of necessity form
spores. This is by no means the case, for I have seen nume-
rous cultivations in which these conditions were present, but
no spore formation ever occurred, although the bacillus went
on increasing in numbers in a most satisfactory way. In the
cultivations of Bacillus anthracis in neutral pork broth in
test-tubes or flasks, if they are kept quiet, no formation of
spores ever occurs. In these instances the absence of a suffi-
cient amount of air is no doubt the cause, as I shall show
below, the growth taking place at the bottom of a fluid which
had been well boiled for the sake of sterilisation. Dr. Loffler
(“aus dem Kaiserl. Gesundheitsamte,” 1881, p. 134) also
suspects that in Pasteur’s cultivations it is perhaps not the
temperature of 42° or 43° C. which prevents the spore forma-
tion, but the immersion of the bacilli in the fluid, for the
bacilli form spores at this temperature when cultivated in flat
dishes (see below).

Apart from spore formation, inoculation with Bacillus
anthracis of gelatine pork in test-tubes or flasks (and kept
solid) yielded slightly different results, according as the inocu-
lation was established on the surface of the gelatine or in the
depth. In the first case the growth proceeds with rapidity,

32 DR. E. KLEIN.

the bacillus forming fluffy masses of convoluted filaments;
where these appear the gelatine becomes liquid; and as the
growth gradually extends downwards, deeper and deeper layers
of the gelatine material become liquefied, the bacillus growth
of course occupying the deepest layer of the liquefied material ;
the liquefied layers above remaining perfectly limpid. In this
way the growth gradually comes to lie deeper and deeper, and
when the deepest layer of the gelatine has become liquefied,
the bacillus mass is at the bottom of the test-tube. When this
stage is reached the growth does not differ in any respect from
one in a test-tube of pure pork broth. The same changes, de-
scribed above, of diminution and gradual dwindling away of
the bacillus mass is also here noticed.

Differing, however, from the bacillus growth in pure pork
broth, the bacillus growth on the surface of the gelatine pork
is capable of spore formation, as long as the growth is still
close to the surface ; but as the superficial layer of the gelatine
becomes liquefied by the bacillus growth, this latter gradually
takes a deeper position, and since the spores previously formed
again germinate into bacilli, a time arrives when no more
spores are formed in the bacilli. When a sufficient amount of
the gelatine pork has become liquefied so that the bacillus
mass being placed at the lowest part of this liquefied portion
is away from the surface, no more spore formation is observed ;
the bacillus threads, however, continue to increase in length
and numbers till all the nourishing material is exhausted. But
I have some instances where no spore formation took place
even at the commencement; that is the case when the inocula-
tion of the gelatine pork in test-tubes takes place (though it
be at the surface) at a point between the gelatine and the glass.
Here the growth increasing burrows itself at once into the
depth and liquefies the gelatine on one side in the shape of a
pit or channel passing downwards, and as the bacillus mass
occupies the deepest position in this pit of liquefied gelatine it
becomes removed from the surface, while yet only composed of
bacillus threads. And the growth proceeding into the depth
and from here into the sides it may happen that the whole

RELATION OF PATHOGENIC TO SEPTIC BACTERIA. 33

gelatine pork becomes liquefied without there having occurred
a trace of spore formation in any part of the gelatine material.
But I have seen a cultivation in which spore formation never-
theless appeared subsequently, although the bacillus mass has
become deeply placed in the fluid gelatine. In this case the
same process occurred as is mentioned above to have happened
in a test-tube of pure pork broth, viz. bundles of threads
grew along the glass wall of the test-tube towards the surface
of the liquid, and having reached this spores made their
appearance.

Spore formation can be, however, easily procured and kept
up in gelatine pork in this manner: A flask containing gelatine
pork, for about a quarter or a half of its volume, is inoculated
with the Bacillus anthracis in the middle of the surface of
the solid gelatine material. The growth naturally spreads
from here to all parts of the surface, since growth in this
direction is much easier than into the depth owing to the
resistance offered by the solid material. The layer of gelatine
on which the growth takes place becomes liquefied, and thus
the growth passes downwards. Owing, however, to the large
surface presented by the gelatine material, the bacillus threads
show very copious spore formation, and this spore formation is
kept up by the bacillus threads for a long time, since even
when the superficial layer of the gelatine has become liquefied
the bacillus mass is still near a very large surface of air. This
superficial fluid layer can be easily drawn off with a glass
pipette drawn out into a thin tube at one end, which is intro-
duced into the flask through the cotton-wool plug, in the same
manner as for the purpose of withdrawing a single drop, or of
inoculating it in the first instance with some seeds, a method
that has been minutely described above. This liquefied mass
thus drawn cff teems with bacillus threads and spores, multi-
tude of them being quite isolated. It can be easily discharged
into a sterilised test-tube plugged with cotton-wool, without
contamination with air organisms, and kept here ad infinitum.
The fluid mass being nearly or quite exhausted of the nourish-
ing parts, at any rate for the Bacillus anthracis and its

VOL, XXIII,—NEW 8ER, c

34 DR. E. KLEIN.

spores, is naturally unable to supply the spores with the pabu-
lum necessary for germination, and hence these spores remain
as such in the fluid. These relations are perfectly in harmony
with all that Cohn has taught (‘Beitrage zur Biol. d. Pfl.
II.,’ Band ii.), about the behaviour of the spores of other
bacilli, notably the spores of the hay bacillus. Such spores I
have kept in the above test-tube as a sort of stock, both for
the production of fatal anthrax in animals as well as for the
establishment of new artificial cultivations.

A new layer of liquefied gelatine teeming with spores is
gradually formed in the above flask, owing of course to a con-
tinuation of the growth of the bacillus threads and spores left
behind, and this layer can be drawn off in the same manner as
the former; thus liquefied masses teeming with spores can be
obtained and drawn off in succession, until a thin layer of the
gelatine pork is left in the flask, in which, owing to the
enormous surface, abundant spore formation takes place in the
bacillus threads, and for the reasons above stated, many of
these spores are retained as such. If we start at the outset
with only a thin layer of gelatine pork or pure pork broth at
the bottom of a flask, and if we inoculate this with Bacillus
anthracis, we also obtain here, after a certain progress of the
growth of the bacillus threads, a copious crop of spores. This
has also been observed by Dr. Loéffler, as mentioned above.
Many of these spores remain naturally as such in the fluid ad
infinitum. We have, then, several methods by which we can
with certainty obtain a crop of spores and preserve them ad
infinitum. All these observations prove ina most definite
manner that for the formation of spores in the Bacillus
anthracis a rich supply of air is required, and unless the
bacillus threads are well exposed to the air, no spore formation
takes place in them. Thus the statement of Cohn and Koch
are fully borne out by my observations.

If on the other hand the inoculation of the gelatine pork in
test-tubes or in flasks takes place in the depth, that is to say,
if the seed is deposited at the outset at the bottom of the solid
gelatine mass, the growth proceeds slowly owing to the re-

RELATION OF PATHOGENIC TO SEPTIC BACTERIA. 35

‘ sistance of the solid material, but in the same manner as is the
case in the pure pork broth. Like a beautiful bed of more or
less distinctly dendritically branched weeds the masses of
bacillus threads springing up at the bottom of the vessel rise
into the superimposed layers. The appearance produced
hereby is very fine after the growth has made some progress ;
we perceive the growth to rise perpendicularly from a common
bed at the bottom of the vessel like a forest of dendritically
branched plants. The gelatine of the deepest layer becomes
of course here first liquid. This growth does not yield spores
at any time owing to being far away from the surface, and it
always remains in the state of masses and convolutions of
bacillus threads.

The following confirms in a marked manner what has just
been said about the spore formation. In a flask filled to a
third or fourth of its volume with solid gelatine pork, Bacillus
anthracis is introduced on to the middle of the surface, as in
the case above described. Masses of bacilli soon spread over
the surface; the superficial layer of the gelatine becomes
liquefied by the growth, and this liquid layer teems with
spores. Now, I decant this liquid layer, and having plugged
the flask again, subject it to boiling. What will happen ?
By the process of decanting a quantity of the growth is
removed, but a great deal (spores and threads) is still left
behind on the surface ready for fresh growth. Next, heating
the gelatine mass, and thereby making it liquid, of course all
growth (spores and threads) sinks to the bottom of the flask.
But boiling the mass for about a minute or so does not kill all
living matter. The threads of bacillus are indeed necessarily
killed by the boiling, but not the spores. This is proved by
the fact that on allowing the gelatine to cool again it becomes
solid, and now all particulate matter is kept enclosed at the
bottom of the flask ; and a new and beautiful growth of typical
Bacillus anthracis growing from the spores soon makes its
appearance at the bottom of the flask, while the surface of
course remains free. One of the finest growths of Bacillus
anthracis threads in the shape of a forest-like mass of per-

36 DE. ELEKLEIN.

pendicularly ascending branched minute plants at the bottom
was obtained in this very manner ; but under these conditions
no new spore was formed—the growth of the threads did not
reach the surface.

In cell specimens of the kind mentioned above the Bacillus
anthracis grows very well both in the neutral pork broth as
well asin the gelatine pork. In the latter, when kept solid,
i.e. at a temperature of about 20—25° C., the progress of the
bacillus can be readily watched. The change of the bacilli
into the very characteristic homogeneous-looking long threads
forming bundles twisted round one another in a spiral manner
so as to resemble a cable; the extension of the threads in all
directions ; the appearance of spores and their full develop-
ment, come out here with the same beauty as in the previous
cultivations. One drawback to the cultivation in cell speci-
mens is the possibility of contamination with air organisms, as
has been mentioned before. In many instances the growth of
Bacillus anthracis proceeds all right for two or three days,
even to the formation of spores, but then an unpleasant crowd
of moving bacilli, or what is equally common, zoogloea masses
of innumerable micrococci cover everything in the field, includ-
ing the Bacilli anthracis. But I have had a good many
specimens in which the growth of Bacillus anthracis re-
mained free of contamination. In such cases it is noticed that
after some days the gelatine becomes also here liquefied.
While in some specimens active spore formation is observed,
in others, kept and established under apparently the same con-
ditions, there never is a trace of real spores, or at the utmost
there is a sort of abortive or imperfect spore formation. In
the latter cases the whole growth of bacilli in the preparation
gradually disappears, and dwindles down to an insignificant
number of hyaline threads, just as was the case with the culti-
vations in the test-tubes and flasks. But whether the growth
leads to the formation of spores or not, there is always, already
in the early stages when the bundles of bacillus threads are
yet few and not very long, this fact to be noticed, viz. that in
many threads there are a good many shorter or longer spaces

RELATION OF PATHOGENIC TO SEPTIC BACTERIA. 37

in which nothing but a hyaline sheath or tube is noticeable,
the highly refractive contents or the protoplasm within the
tube being wanting; as the growth proceeds the number of
such threads with empty spaces in their sheath increases, and
whole threads of immense length may be found in this condi-
tion, i.e. in the state of hyaline tubes or sheaths from whose
interior the protoplasm has altogether disappeared. This is
always noticed in the growth of the Anthrax bacillus
threads in cell specimens; samples taken out at any stage from
the cultivation in test-tubes or flasks show the same condition,
viz. there are always present longer or shorter threads, which
either entirely or partially have become barren of the proto-
plasm inside the sheath. In cell specimens, it is possible to
ascertain that the growing ends of the threads which may be
found in the peripheral part of the drop as straight filaments
with a rounded end, are always full of protoplasm, and that the
deficiency in protoplasm commences at some distance from the
end. By-and-by the greater number of threads may thus lose
altogether their protoplasm, and hereby become quite trans-
parent and almost lost to sight, but a careful inspection can
still detect their presence. Some appear ultimately to break
up altogether. This change, viz. the disappearance of the
protoplasm in the threads from place to place, is associated,
generally but not always and in all places, with the appearance
of irregularly sized granules in the tubes, these disappear
gradually, becoming evidently dissolved and absorbed, and the
then sheath appears at such a place or places quite empty,i. e.
without containing any solid protoplasm. These granules are
not spores, as I shall show below. I consider it merely a form
of degeneration or death of the protoplasm. Another change
in the cell specimens is the appearance of spherical corpuscles,
either isolated or in close rows or chains; in this latter case we
have a thread of regular varicose appearance, not unlike a
chain of torule. The size of these spherical corpuscles is in
their best development that of a human red blood-corpuscle,
and in aspect are identical with the gonidia of an oidium or
the cells of torula, i.e. within a cell membrane they contain

38 DR. E. KLEIN.

clear contents and in this a minute nucleus. They become
elongated and by fission divide into, or by gemmation produce,
two new spherical corpuscles. The growing ends of the
threads seen at the marginal part of the specimen sometimes
are connected with such a chain of spherical torula-like cor-
puscles, and in this respect the appearances bear a striking
resemblance to the formation of gonidia by mycelium threads.
This change also has nothing whatever to do with the forma-
tion of spores. It can be ascertained to exist always also in
the cultivations in pure pork broth and gelatine pork in the
test-tubes and flasks. In some cultivations in the neutral
pork broth I have met with it very extensively already after a
few days’ incubation. With great distinctness and profusion I
have seen it in cultivations in gelatine pork carried on at the
temperature of the room.

The observations which I have made on the life-history of
the Bacillus anthracis differ in some respects from those of
previous writers. Starting with the Bacillus anthracis of
the blood, introduced into the cultivations of neutral pork
broth or of the mixture of this broth with gelatine, it is invari-
ably the rule that, as noticed by other observers (Koch, Pas-
teur, Buchner, and others), the bacilli grow out sooner or later
into long homogeneous-looking threads which form bundles,
the individual threads coiling round one another in the manner
of the wires of a cable. But there are always some short
bacilli, or chains of them and short threads; especially in the
former it is noticed in the fresh state and after staining, that
their ends are not so blunt as is usually represented, but that
they are slightly rounded ; and the same rounded appearance
is also noticed on the ends of the threads, which are undoubted
anthrax bacillus threads. In all specimens of gelatine pork
above described the rounded conditions of the ends of the
threads is easily perceived.

In the first few days the cultivations of neutral pork broth
invariably show, as mentioned above, a uniform distribution of
shorter or longer bacilli, isolated and in chains. ‘These gradu-

RELATION OF PATHOGENIC TO SEPTIC BACTERIA. 39

ally lengthen and then of course by their weight settle down
at the bottom of the fluid whence they grow upwards into the
characteristic long convolutions. I presume the uniform dis-
tribution of the bacilli and the general turbidity of the fluid
caused hereby in the early days is due to the bacilli following
Brownian molecular movement, as well as to their being able
to float in the fluid, but when they grow out into long threads
their weight and the cessation of molecular movement draws
them down to the bottom of the fluid, which for this reason
then becomes clear. That the isolated and short chains of
bacilli causing the general turbidity of the cultivation in the
first few days are really anthrax bacilli as much as the typical
long threads afterwards formed, is shown by the fact that if
with these latter a cultivation is started it presents the turbid
appearance in the first few days, and secondly, at any stage the
smallest quantity of the cultivation kills with typical anthrax
guinea-pigs and rabbits.?

With Abbe’s condensor and Zeiss’ oil immersion objectives
it is possible to discover in these threads, already in the fresh
and living state from place to place, a differentiation of a thin
sheath forming the tube, as it were, and a protoplasmic con-
tents, and this protoplasmic contents appears subdivided into a
single row of short, almost cubical blocks or cells. In many
places this subdivision of the protoplasm into cells or even the
differentiation into sheath and protoplasm is not distinct in the
fresh state, but comes out with greater or lesser distinctness
after staining or after certain reagents. Thus, for instance,
careful staining them fresh with anilin dyes (gentian violet,
methyl violet, methyl blue, Spillers’ purple, &c.) brings out in
many places this differentiation into sheath and protoplasm,

1 These peculiarities of the early bacillus growth may or may not be con-
nected with the ability of that growth, which is not possessed by later stages
of the same growth, to kill mice that are inoculated with it. However that
may be, these early peculiarities have no relation to spore formation. Spores
have, indeed, nothing in common with these rounded ends and cubical cells,
which stain in a way that spores do not stain, and have a quite different shape
and refractive power.

40 DR, 4B. KGEIN:

and the subdivision of the protoplasm into square or rather
cubical individuals or cells. Adding a nearly concentrated
solution of acetate of potash to the fresh preparation brings
out these appearances also very well. Still more, and in fact
with marvellous distinctness, dues it come out in dried and
stained specimens (after Weigert and Koch). Watching the
bacilli of the spleen or any other organ while drying under the
microscope, the gradual differentiation into sheath and cubical
cells can be followed very readily. I have made an endless
number of stained specimens of bacillus threads of my cultiva-
tions in pork broth and gelatine pork, and have invariably
found the same appearances, provided the specimens be not
overstained, or if so, well washed with alcohol, viz. the whole
protoplasm of each thread is subdivided into a single row of
cubical cells, stains well with the anilin dye, and distinct from
the general sheath of the thread. Koch has pointed out that
the Bacillus anthracis shows in dried and stained speci-
mens a very characteristic subdivision into shorter or longer
rod-like structures, and by this alone Bacillus anthracis
can be distinguished from other bacilli. He gives in his work
(‘ Cohn’s Beitrage II,’ Bnd. iii) a photograph to illustrate this
point. In this illustration the subdivision of the protoplasm
is not by any means numerous, far less than in my case, for I
find the individual members not rod-shaped but cubical. It is
true here and there it is seen that instead of a cubical we have
an elongated or rod-shaped cell, but in some of these we can
clearly discover a slight constriction in the middle, a sign of
commencing division into two. The independence of the
common sheath of the thread and these cells Koeh has not
noticed.

But also in the bacilli of the blood and spleen of mice,
guinea-pigs, and rabbits, dead of anthrax, I have noticed pre-
cisely the same distinction into common sheath and the sub-
division of the protoplasm into cubical cells, or when the cells
are elongated a middle constriction was noticeable. Accord-
ingly the length of a bacillus, viz. whether a longer or shorter
rod or a longer or shorter thread, depends entirely on the

RELATION OF PATHOGENIC TO SEPTIC BACTERIA. 41

number of cubical protoplasmic cells contained within the
common sheath. The cells may be seen aggregated into
twos or into fours, that is, forming longer or shorter rod aggre-
gations.

Bacilli obtained fresh of the spleen of an animal dead of
anthrax show in some instances an absence of the protoplasmic
contents within the sheath; this may be only limited to a
small spot or may involve the greater part of the length of the
bacillus. In sections made of hardened or fresh organs of an
anthrax animal, after staining with logwood, or, still better,
with anilin dyes, the same appearance may be met with, viz.
limited deficiency of protoplasm in some bacilli. Koch (1. c¢.,
p- 40, and Plate v, figs. 29 and 30) mentions this appearance
of stained bacilli from the spleen of an anthrax rat.

On a former page I have described a similar local deficiency
of the protoplasm in bacilli and bacillus threads in the arti-
ficial cultivations at all periods of growth, even at the earlier
stages, andin these circumstances the deficiency extends in
some threads over long distances, and in consequence only the
hyaline transparent sheath of the original thread is left, and
this ultimately may also become broken up.

It appears to me very probable from numerous observations
that in every bacillus at some period of its growth, one, two,
or more consecutive cells may cease to grow and to multiply.
These cells die if spore formation does not occur in them, and
their death is indicated by a granular disintegration of their
protoplasm and a final solution and absorption of it. In this
case the sheath of the bacillus thread remains empty at this
place. Such bacilli and bacillus threads are thicker than the
unaltered ones.

The division and gemmation of the above-mentioned torula-
like corpuscles or gonidia leads to the formation of chains, at
first entirely composed of torula-like gonidia; by active divi-
sion of these gonidia the chains rapidly elongate ; in a further
stage the gonidia are transformed into oval elongate cells,
which are thinner than the original gonidia, and ultimately

42 DER. .H. KLEIN.

they change into rod-like cells, again thinner than the oval
cells. When the latter stage is reached we have already to do
with the typical thread of an Anthrax bacillus. In some
such threads there are seen numerous places in which the pre-
ceding stages of oval cells and of spherical gonidia can be easily
recognised. We have, then, here before us a new form of
growth of the Bacillus anthracis very similar to that of an
oidium growing in a fluid.

The diminution of the bacillus mass in the artificial cultiva-
tions in test-tubes and flasks described on a previous page is
due to the degeneration and disappearance of the protoplasmic
cells in the threads, so that at first the transparent sheaths are
left, and they also break up ultimately. This degeneration takes
place chiefly on the plan of a gradual granular disintegration
of the cells within the sheath. In the first stage of this pro-
cess, and especially if the preparation has been stained, it is
noticed that instead of the cubical mass of protoplasm repre-
senting one cell, we find either one large granule or a delicate
dumb-bell joined by a shorter or longer thin pale bridge.
These appearances I have seen in many places in undoubted
Anthrax bacillus threads; of an accidental admixture
there can be no manner of talk, since the general sheath passes
yninterruptedly over all unaltered and altered cells. In a
further stage of disintegration the granules dwindle down, and
are ultimately altogether lost. Koch figures (‘ Unter u. path.
Organismen,’ Plate vii, fig. 39) thin bacilli showing a similar
appearance to that just Sea ei In Koch’s case they were
not Anthrax bacilli, and Koch does not decide whether
this appearance means spore formation or not. I am quite
confident it has nothing whatever to do with spore formation,
although I at first thought this to be the initial stage of it (see
Cohn and Koch); but in my case there is at no time to be
seen in them a trace of a bright oval spore. The whole pro-
toplasm of a thread may give origin to these granules; they
become smaller and smaller and more numerous, and
irregularly distributed in the sheath, and ultimately altogether
disappear.

RELATION OF PATHOGENIC TO SEPTIC BACTERIA 43

In some instances I can see something like a thin septum
stretching between some of the cells and fixed to the mem-
brane of the common sheath, but I cannot be quite certain
about a septum being present between each two cells. I think
that the sheath of many a bacillus, be it short or long, be it a
straight thread or a curved one, is traversed by such septa at
relatively few places; in many places it is a continuous mem-
branous tube in which the protoplasmic cells lie in a single
row. In threads in which the above-mentioned granular de-
generation occurs the presence of septa can be easier ascer-
tained than in perfect bacilli, but one must not confuse them
with transverse discoid débris within the tube. In some such
tubes it is seen that many compartments contain one cell ; others
contain two cubical cells or one oblong, and still others there
are that contain three cells, one oblong and two cubical ones.
In threads in which the above-mentioned nodose swelling up
of the cells has taken place there are distinct signs of septa
between the individual cells, especially where the degeneration
comprises a whole row of cells.

Comparing the Bacillus anthracis of heart’s blood or
spleen of a mouse, guinea-pig, or rabbit, dead of the disease,
with the bacilli and bacillus threads grown in the cultivations
of neutral pork broth, or in a mixture of this and gelatine, it
is found that the organisms are almost twice the thickness of
those taken from the animal. I do not refer to the bacilli and
threads in which either granular degeneration or the torula-
like swelling of its cells or spore formation is going on, for
these are naturally thicker, but I refer to bacilliin which un-
altered protoplasm is contained within the sheath.

Ewart (‘Quarterly Journ. of Micr. Scien., April, 1877)
maintains to have observed a transition of the ordinary non-
moving Bacillus anthracis into flagellate-moving bacillus.
I can only say with reference to this, that in all my observa-
tions, whether conducted in test-tubes or in cell specimens,
there was never anything of the sort observable. Ewart took
no precaution whatever against contamination with other

44 DR. E. KLEIN.

bacillus, and therefore his observations lose all value, since it
is probable that he had before him an ordinary flagellate
bacillus.

Buchner (I. ¢., p- 394) also claims to have seen a transition
of a non-moving typical Bacillus anthracis into a flagellate
bacillus, but in Buchner’s case this is supposed to have come
about in a gradual manner after more than 1100 generations.
Notwithstanding the imposing number of generations, I never-
theless doubt the reality of this transformation, since Buchner’s
cultivations are open to the objection that they were contami-
nated with an air bacillus. Besides, Buchner, in connection
with this very transformation, makes certain statements as
regards the influence on this transformation of the acid re-
action of the hay infusion in which the bacilli were cultivated
and transformed into flagellate innocuous hay bacilli, state-
ments, I say, which I know to be incorrect, as I shall show
lateron. He says, for instance (l. c., p. 392), “ that the slight
acidity of hay infusion prevents altogether the growth of the
true Bacillus anthracis.” This statement is to me unin-
telligble, since I have seen Bacillus anthracis starting off
into a very good growth in acid hay infusion, as well as
in other acid nourishing fluids. This statement of Buchner’s,
if it is to be accepted at all, must be accepted to mean some-
thing else than what Buchner infers, viz. that the true
Bacillus anthracis in the acid hay infusion used by Buch-
ner had no chance against the hay bacillus growing in it, and
of which Buchner had not quite got rid previous to the inocu-
lation with Bacillus anthracis.

I have mentioned above some of the conditions under which
spores were formed in my cultivations, and I wish now tostate
the manner in which this takes place. Examining in the first
state the bacillus of such cultivations in which spore forma-
tion is just commencing, it will be found that the protoplasm
of the bacilli and bacillus threads appears slightly granular.
Where no spores appear it is uniform in aspect. In the
granular ones are seen here and there bright, glistening,

RELATION OF PATHOGENIC TO SEPTIC BACTERIA. 45

spherical, or rather cubical and elliptical, or rather rod-shaped
spores. The spores are slightly truncated at their ends, and
slightly convex where in contact with the sheath. Staining
such specimens with Spiller’s purple it will be found that the
granular as well as the homogeneous protoplasm stains readily
and deeply, whereas the spores, both the spherical as well as
the oval ones, remain unstained, and therefore contrast well
with the rest of the bacilli. This same relation is exhibited
by specimens stained with other anilin dyes, but best with
gentian violet and Spiller’s purple. The different parts of the
bacillus threads show a great difference with respect to the
number of spores. In some places there is for long distances
a single cubical or oblong spore contained in the thread; in
others they are more numerous; and still in others they
follow one another as numerously as the elementary cells. It
is found that wherever a spore appears it is at once either
cubical or elongated, and conspicuous by its glistening appear-
ance and remaining unstained. The cubical ones when ripen-
ing become elongated, but always remain of the bright
appearance ; it is further true that each cubical spore belongs
to an elementary cell, but where this latter has become elon-
gated and slightly constricted, preceding division, as mentioned
on a former page, we may find two such pores contained in
it. Insome places the cubical spore remains single in the
oblong cell. The elongated spores are asa rule placed parallel
to the long axis of the bacillus; but in some places I have
seen one or the other spore placed in a diagonal direction.
The elementary cell containing a spore still possesses a trace of
protoplasm around the spore; but this remnant of proto-
plasm sooner or later breaks altogether away, and the spore is
free of it. I have, however, seen spores which after having
left the sheath of the thread, still showed at one end a trace
of the protoplasm.

According to the facility with which spores are formed in
a cultivation we find the number of spores formed in a bacillus
thread varying. In some threads every cell for some distance
may develop a spore, in others numerous cells remain always

46 DR. Ei SEIN,

without spores, and their protoplasm crumbles down into a
granular débris. Under all conditions, however, the thread
becomes much thicker, the sheath swells up, and gradually is
lost as such. In some cell specimens I noticed an abortive
formation of spores ; these appeared as irregularly distributed
spherical small spores, which never grew into the typical fully-
formed large elongated spores.

The conclusion we then arrive at from all these observations
is this:—Under most favorable conditions every elementary
cell is capable of forming a spore ; these spores are bright and
glistening, and do not stain (see Koch). At first they are
spherical, afterwards larger and oblong. If the cell is an ele-
mentary or cubical one it forms one spore; if it is elongated
and constricted, i.e. before dividing, it may form two spores
not all the protoplasm of the cell is involved in the formation
of the spore, a trace of it is left around the spore, but sooner or
later crumbles away as a granular débris. If the conditions
are not so favorable only a limited number of cells form spores,
in the rest the protoplasm degenerates into granular débris,
and under unfavorable conditions, especially in the absence of
a sufficient supply of air no spore is formed in any of the cells.
When spores are formed they escape after the sheath breaks
down.

Ewart (l.c.) maintains to have observed a division of the
spores after they had become freed of the bacillus sheath.
This statement also requires confirmation. The above-men-
tioned couples of spherical spores Ewart also noticed, but they
are not due to a division of spores as Ewart maintains, but are
developed as such in an oblong dividing cell.

I now enter one of the most important parts of this research,
viz. the results of the inoculation of rodent animals with
the Bacillus anthracis of the artificial cultivations described
above.

At the outset I wish to state the manner in which the inocu-
lations were carried out. The animals used were white and
tame brown mice and offsprings of both; further, guinea-pigs

RELATION OF PATHOGENIC TO SEPTIC BACTERIA. 47

and rabbits. As a rule the mice were inoculated into the
subcutaneous tissue of the tail, the guinea-pigs and rabbits
into the inguinal skin ‘or subcutaneous tissue, or into the skin
of the ear-lobe. The infective material—blood of an anthrax
animal or bacillus of a cultivation—is collected in a capillary
tube freshly drawn out, and is then blown out into a small in-
cision made into the true skin or subcutaneous tissue, accord-
ing as desired, with a sharp blade that has been before per-
fectly disinfected in the gas blowpipe. By this method of
inoculation I always made certain of not getting any contami-
nation by the instruments—syringe and canula—that may have
been used in previous inoculations. In some cases I used also
Pravaz syringes, viz. when I inoculated with blood of anthrax,
and expected fatal anthrax; for in this case, even if the syringe
should not have been cleaned of anthrax particles of former
inoculations it did not matter. Knowing the great difficulty
of thoroughly disinfecting Pravaz syringes, heating not being
available, [ did not practise as a rule inoculation by means of
the syringe. In the case of guinea-pigs and rabbits a syringe
can be always dispensed with, since a capillary pipette drawn
out to a fine point and charged with the infective material is as
easily inserted to any distance into the inguinal subcutaneous
tissue as the canula of a hypodermic syringe. A minute in-
cision having previously been made, the capillary pipette is
emptied of its contents as usual i.e. by blowing into the near
end of.it. But even into the subcutaneous tissue of the tail of
a mouse a capillary pipette drawn out into a fine point can be
easily advanced for a distance sufficiently long for safe inocula-
tion. In all my inoculations with blood and with fresh cul-
tures I have used only very minute quantities of the infective
material, a portion of a droplet to adrop, and I found that, as
a rule, the quantity of the material introduced was, if other
conditions were equal, seldom a matter of any importance.
Buchner and Greenfield, speaking of the early stages of suc-
cessive cultivations, maintain to have had to introduce in some
cases larger quantities of the same material than in others, in
order to produce an effect, owing to the activity of the material

48 DR. E. KLEIN.

having become diminished by cultivation in the former and
not in the latter. These statements may be, and probably are,
best explained after Koch, by the assumption that the original
cultivation had become contaminated by another bacillus, the
Bacillus anthracis remaining in the minority is gradually
overgrown after a certain period or after a certain number of
cultivations by the contamination of bacillus, and hence larger
quantities of the fluid are to be used to get hold of one or the
other stray Bacillus anthracis left therein, and after some
more transfers the original number of Bacillus anthracis
had become so much diminished that perhaps even a larger
quantity contains no other than the contamination bacillus. In
my cultivations I never noticed such a condition, i.e. I never
found reason for supposing the anthrax bacillus to undergo
change in its virulence, otherwise than as there might be ques-
tion of spore formation on the one hand, or of degeneration
(that rendered the bacillus completely inert) on the other hand.
I have, indeed, on occasion found an exceptionally large quantity
of inoculating material to be required; but this circumstance
has always appeared to me perfectly well explained through
the small number of really active bacilli existing in the par-
ticular material. This has taken place, e.g. in cultures of pure
Bacillus anthracis without spores, that had been kept for
some weeks (see below), and where there had been a gradual
diminution of the number of active anthrax bacilli.

A point of importance which I wish to mention here refers to
the time of death of rodents after inoculation with anthrax. As
a rule they die within forty-eight hours from the time of inocu-
lation ; some within twenty hours, others in thirty or thirty-six
hours, and afew others between this and forty-eight hours.
Few animals survive the third day, although I have seen mice
and guinea-pigs die after five days of typical anthrax. A given
cultivation used after the same method and in the same quan-
tity for the inoculation of several mice will kill some of them
more rapidly than it will kill others, and the same is true if
guinea-pigs are the animals under experiment. I have seen
animals (mice and guinea-pigs) die within twenty hours after

RELATION OF PATHOGENIC TO SEPTIC BAOTERIA. 49

inoculation with infinitesimal doses of artificially cultivated
Bacillus anthracis. The presence of spores makes no
difference to the rapidity of death. A guinea-pig inoculated
with spores that had been first of all frozen with ether spray
and then allowed to thaw died within twenty-four hours,

Observations on Inoculation.

It is not necessary for me to enter here into a description of
the symptoms of anthrax in rodents, since these are well known
through the various descriptions already existing, and 1 may
refer the reader especially to those given by Koch in his several
writings on anthrax.

But one fact I must mention here, viz. the great irregularity
presented by the spleen in the animals (mice, guinea-pigs, and
rabbits). In some instances the spleen is very much enlarged,
in others it is not enlarged at all, but in all instances it con-
tains bacilli, though the number of these varies considerably
without, however, standing in any relation to the duration of
the illness.

1. Inoculation with blood of an animal dead of anthrax :

The blood was derived from the heart of mouse, guinea-pig,
or rabbit, either quite fresh or after one, two, three, to eight
days. In all instances death followed after the introduction of
infinitesimal doses. Before inoculation I ascertained that the
blood contained the bacilli in sufficient numbers to make it pretty
certain that some bacilli will be present even in so minute a
quantity. When using infinitesimal doses it is necessary to
bear this in mind for the following reasons :—Supposing the
point of a needle is well steeped into the blood of a guinea-pig
or rabbit dead of anthrax, and then with this needle the skin
of a mouse, or guinea-pig, or rabbit is pricked down to the sub-
cutaneous tissue, out of ten such inoculations the chances are
that all ten will be successful ; but supposing the point of the
needle be well dipped into the blood of a mouse dead of anthrax
for some hours, and the inoculation be performed as above, out
of ten such inoculations the chances are that there will be

VOL, XXIII, —-NEW SER, D

50 DR. E. KLEIN.

several failures. The reason lies in the peculiar way the bacilli
are distributed in the blood fluid, being now most of them col-
lected in masses, owing to being held together by a granular,
imperfectly-coagulated fibrine. Dipping the point of a needle
into the blood, it may chance that the point of the needle does
not take up one of these coagula, they being sometimes large
and far between, and in this case the inoculation will be un-
successful. A similar connection, i.e. the aggregation of the
bacilli by granular coagula, may be also observed in guinea-
pigs and rabbits some time after death, but not to such an
extent as in mice. I well remember to have been rather
puzzled one day about the inexplicable cause of death of four
of my mice that had been inoculated with anthrax, but appa-
rently did not show any bacilli in the blood. The animals had
been dead for some hours (less than twenty), and specimens of
blood of the jugular vein and heart withdrawn with a capillary
pipette, and used for microscopic preparations, did not reveal
the presence of the bacilli. I then made specimens of the tissue
of the spleen, which organ was only slightly enlarged, and found
it teeming with the characteristic anthrax bacilli. I examined
the blood again, and especially collected from the cavity of the
heart blood with the blade of the knife, so as to get out not only
fluid blood, but coagula as well, and then I met a number of
large coagula crowded with the bacilli. These peculiarities, in
fact, account for several cases in which I thought at first to
have to deal with animals refractory agaiust anthrax ; but on
inoculating them again with guinea-pig’s blood containing uni-
formly distributed bacilli they all succumbed. The older the
blood, and the longer it has remained within the body of the
dead animal, the less chance there is of its retaining the bacilli
in a living condition, and the greater also the chances of these
bacilli having disappeared and other saprogenic bacilli having
made their appearance. These points have been well ascer-
tained by Koch, and I can fully confirmthem. I shall have to
return to this point in a later report, in which I shall give the
results of a systematical inquiry into this death of the bacilli
in the organs of an animal dead of anthrax. In some instances

RELATION OF PATHOGENIC TO SEPTIC BACTERIA. ail

I have had opportunity of using grey mice for my inoculations,
and also bastards of white and grey mice, but did not notice
any refractory power possessed by the former against anthrax
blood. Some observers have noticed a certain amount of re-
sistance offered to the anthrax virus by the grey mouse, but
in these instances they were wild mice, whereas in my cases
they were born in captivity and tame.

At the seat of the inoculation with blood in mice the place
is always more or less marked as slightly discoloured and a
little tumid, but since death as a rule occurs within thirty-six
hours there is not much chance of inflammation. After the
inoculation of mice with blood, and the same applies to other
anthrax material, as well as to artificially-cultivated anthrax,
the animals appear in perfect health until a short time, some-
times half an hour or an hour before death; they are very
lively and feed well; nothing in their condition reveals the
presence in them of the seed of death, when suddenly they
become quiet, their movements become impeded, their breathing
rapid, and their temperature begins to sink; and after half
an hour, or even less, sometimes more, up to two hours, they
are dead.

So remarkably sudden is this change from apparent health
to a sickness that is rapidly and surely fatal, that this form of
the disease has in Germany received the name of fulmini-
render Milzbrand, Teufelsschuss (R6ll, ‘ Lehrbuch d.
Path. und Ther.,’ &c., Wien, 1876, i, p. 493).

In guinea-pigs I noticed on inoculation of blood or any other
anthrax material and artificial cultivations into the subcuta-
neous tissue of the inguinal region, as a rule before the day is
over, a distinct cedematous swelling; this increases gradually
till the animal dies, which in the case of blood happens gene-
rally within thirty-six or forty-eight hours. Death is also here
rapid, but not so rapid as in mice, since the animal becomes
quiet and weak several hours before death takes place. On
post-mortem examination the subcutaneous tissue of the in-
guinal region and abdomen, especially on the side of inocula-
tion, shows much cedematous swelling. Near the seat of the

cr

52 DR. E. KLEIN.

inoculation there are a few hemorrhagic spots, but the cedema
is generally free of blood; cutting into the tissue a quantity of
clear serum flows out that contains only very few bacilli, if
examined soon after death. Later on their number is much
increased. In some cases, especially of prolonged illness, this
cedema is so extensive that the subcutaneous tissue of the in-
guinal region, abdomen, and chest is uniformly infiltrated with
the serum. This cedema I have not missed, with very few ex-
ceptions, in the cases of inoculation of Bacillus anthracis
and their spores, no matter whence derived, if the inoculation
is made into the subcutaneous tissue of the inguinal or abdo-
minal region, but it was absent if that of the ear-lobe was the
seat of inoculation. That in our case we have to do with
veritable anthrax there can be no manner of doubt from the
symptoms, the nature of the bacilli, and their distribution in
the organs. Pasteur mentions a similar appearance of cedema
in some of his sheep.

In rabbits the cedema is not so frequent as in guinea-pigs, nor
is it so pronounced, but nevertheless I have met with it in several
instances of inoculation into the subcutaneous tissue of the
inguinal region.!

It must be, however, understood that this symptom of sub-
cutaneous oedema in guinea-pigs after inoculation into the
subcutaneous tissue of the inguinal region occurs not only after
the inoculation with blood or with tissues, but equally distinct
with artificially cultivated bacillus and its spores.

Inoculation into the corium itself of the inguinal region of
guinea-pigs produced only very slight cedematous swelling about
the point of inoculation ; in rabbits such an inoculation is not
associated with oedema.

2. Inoculations with artificially cultivated Bacillus
anthracis.

1 What the meaning of the statement of Wernich’s (‘ Central f. med. Wiss.,’
No. 12, 1882, p. 217) is, that rabbits, ‘‘ although not absolutely refractory,
nevertheless are very little susceptible” to anthrax, I cannot comprehend,
since I have never found a rabbit escape death after inoculation with anthrax
blood or artificially cultivated active Bacillus anthracis.

RELATION OF PATHOGENIC TO SEPTIC RACTERIA. 53.

The experiments which I wish to mention here were made
with Bacillus anthracis derived from the mouse, guinea-pig,
or rabbit, killed by inoculated anthrax, after this bacillus had
been cultivated in the neutral pork broth or the neutral gela-
tine pork above described. In the following I shall, of course,
only take into account the cultivations which from the unaided-
eye aspect, the microscopic examination, and the experimental
results of inoculation with them, are to be considered as undoubt-
edly pure cultivations of Bacillus anthracis. The first
remove of Bacillus anthracis from the anthrax animal
will be considered as the first cultivation ; the remove from this
into a new cultivation, the second cultivation; from this,
again, a third cultivation ; from this, again, a fourth, and so
on. As arule, as soon as a cultivation showed a good crop of
the bacillus, a next cultivation was established from it by
transferring into the sterile nourishing material an infinitesimal
part of a drop of the parent cultivation. In some instances,
especially when the cultivations were kept at a temperature of
32—35° C., there was a copious growth of bacillus. obtained
already after two or three days; in other instances, if the in-
cubation was carried on at 20—25° C., I generally waited about
six or seven days before utilising the cultivation for the estab-
lishment of a new cultivation.1 To term the cultivations
** generations,” as Buchner does, seems to me altogether arbi-
trary; his “1500 generations” are no more in reality 1500
generations than they are “ 150 generations,” as Koch is only
too leniently inclined to admit (l.c., p. 24). A ‘‘ generation”
could really only be called a new crop of bacilli produced from
spores of bacilli. If a bacillus grows out into a long con-
voluted thread or threads, i. e. if one or several elementary cells
continue to divide till they have formed a chain of enormous
length, then we have no more right to call this chain a full
generation than we have to consider the initial cell of the
bacillus thread as the parent, the second cell derived from this
as the first generation, the third cell as the second generation,

1 For convenience’s sake, I shall speak of the days of exposure of a cultiva-
tion to a constant temperature in the incubator, as being days of “ incubation.”

54 DR. E. KLEIN.

and so on, for then we could no doubt get millions of genera-
tions in the very first cultivation that we established with the
bacillus taken from the blood. But even if in a cultivation
spores are formed this cultivation need not represent one gene-
ration only, because the spores first formed may, and as a rule
do, germinate into new bacilli; these or their offsprings again
form spores, then germinate into bacilli, and so on within the
same cultivation as long as nourishment is present. But in
Buchner’s case there must have been cultivations in which
spores were never formed. In whatever way we look at it we
cannot fix the meaning of the term “ generation,” and therefore
we cannot speak of a cultivation as a “ generation.”

At the commencement of my experiments I used white mice
for testing the activity of fluids and cultivations containing
Bacillus anthracis, since, as is well known from Koch and
others, these animals are very susceptible to anthrax. Inocu-
lation with a given cultivation of blood bacillus in neutral pork
broth kills mice, guinea-pigs, and rabbits, when used during
the first few days of incubation; but soon a difference sets in,
for after the first few days inoculation of mice with the same
cultivation proved fatal only in a certain percentage of cases,
and after several days more a good many mice remained per-
fectly unaffected by the cultivation. My cultivations were
typical and perfectly pure, as I ascertained by microscopic ex-
amination and further experiments ; there were the typical con-
voluted cable-like bundles of the threads ; on staining they were
beautifully “cellular” in structure; inoculations of new cul-
tivations which I made came up very finely; and inoculations
with infinitesimal doses in guinea-pigs and rabbits produced
typical anthrax. The above mice were inoculated, some by
Pravaz syringe, others by deep incision, and placing into this a
drop of the cultivation. ‘The inoculation was repeated in some
instances, but without effect ; the mice remained perfectly free
of illness. This result was obtained with the cultivation, in
some cases within a week, in others after a longer period.
How could this be explained? Were all these mice refractory to
anthrax. Were they refractory only to the artificially-cultivated

RELATION OF PATHOGENIC TO SEPTIC BACTERIA, 5D.

Bacillus anthracis? Were they attacked by it, but did not
die, as was the case in Pasteur’s experiments with the “ vac-
cine?’ Had the cultivated bacillus lost its virulence or atto-
gether its specific effect after the first remove from the blood,
and thus putting the cultivations of Buchner and Greenfield
altogether in the shade? These were questions that presented
themselves for solution.

As I mentioned above, of the purity of the cultivation I had
no manner of doubt; for this I had the most cogent reasons.
I must state in connection with this, that I refer not only to
the use for inoculation of the cultivations in fluid pork broth,
but also to cultivations carried on in gelatine pork in a micro-
scopic cell specimen above described, where from day to day I
could follow the increase in number and length of the bacillus
threads. Such cultivations were also used for the inoculation
of mice, and if containing no spores, proved without effect.
Another point I must not omit to mention, the method or
rather methods of inoculation used with mice were perfectly
reliable, since really active material introduced in the same
manner proved efficacious.

That the mice were not refractory to anthrax—[it would have
been a most extraordinary thing if I should have happened to
get hold of one refractory mouse after another; I have not
found a mouse that was really refractory! to anthrax, if the

1 In several instances of mice I have noticed what seems to denote a certain
resistance offered to the anthrax virus on the part of white mice, viz. that
some of my animals did not become affected by the virus the first or even the
second time they were inoculated with it. Thus I noticed some that resisted the
action of the cultivated Bacillus anthracis of a first cultivation; then they
remained also unaffected by the introduction of typical anthrax bacillus threads
of a second cultivation; and on a third time being inoculated with blood
bacillus remained nevertheless alive. They succumbed, however, on a fourth
inoculation with bacillus spores of an artificial cultivation. Another mouse
remained unaffected after the introduction of anthrax blood filled with bacilli,
but succumbed to the influence of an artificial cultivation of anthrax bacilli filled ©
with spores. The inoculation in these cases was carried out in the way described
above, and I have no doubt that the material was properly introduced into the
subcutaneous tissue of the tail. Not seeing any reason to accept in the first
instances a refractivity of these particular animals to the anthrax virus,—for

56 DR. E. KLEIN.

proper material is used ; every one of them inoculated once or
twice with it died]—was proved by the fact that they one and
all succumbed to anthrax afterwards when inoculated with a
different but active material. And this latter circumstance
proves at once that they had not been “vaccinated,” in the
sense of Pasteur, or any other sense, by the first inoculation,
that in fact the first inoculation produced absolutely no disease.
But, secondly, had the bacillus as such lost its virulence by
being taken from the blood and cultivated in an artificial
medium? Not in the least, because the very same bacillus
killed mice in the first few days of the cultivation, and later
on it killed guinea-pigs and rabbits within forty-eight hours
by typical anthrax, and blood of these animals killed without
fail within thirty-six hours. What is more than this, the very
same cultivation which failed to kill a mouse or mice at one
time, killed them without fail at another, provided the bacillus
had in the meantime had the opportunity to form spores. And
this fact, viz. the presence of spores in the cultivation, is of the
utmost importance in respect of the fatal efficacy of the arti-
ficially cultivated bacillus on the mice.

Pasteur, as mentioned above, produced a certain incapability
of the bacillus to kill sheep by growing it at 42°—438° C. By
these means he maintains that he can prevent the bacilli from
forming spores which prove fatal tosheep when inoculated. I
have mentioned above that in my cultivations in neutral pork

they succumbed to it ultimately, thus proving that they were susceptible to
the virus,—it remains as the most probable explanation to assume that the
virus, although locally introduced, was for some* unknown reasons not carried
into the general circulation. That in our instances it was the resistance offered
by the tissue of the tail to the life of the Bacillus anthracis, which prevented
the development of the disease, is not a probable reason, since there exists no
real resistance to the anthrax bacillus, of either mice, rabbits, or guinea-pigs
_to prevent the fatal result generally produced after such incubations.

A similar negative result after first inoculation I have noticed also in a few
of my guinea-pigs, where the fluid had been introduced into the subcutaneous
tissue, and also in a sheep. But in both cases a second inoculation with the
same virus produced positive results. The virus was introduced during the
first inoculation in sufficient quantity very safely into the subcutaneous tissue.

RELATION OF PATHOGENIC TO SEPTIC BACTERIA. 57

broth, in which the bacillus mass remains quiet at the bottom of
the vessel, no spores are formed, and it is such a cultivation
which proves inactive on mice only. In Pasteur’s case the sheep
inoculated with such bacilli (prevented from forming spores) are
not killed by anthrax, but ‘‘ vaccinated,” and protected against
the most virulent anthrax material. I have not yet succeeded
in discovering the method employed by M. Pasteur (and the
details of which he has not published) for the production of
*‘ vaccine” protective against anthrax ; and I can only say that
in the case of mice there is no such diminution of virulence as
Pasteur has obtained in the cultivation with which he inocu-
lates his sheep. The mice not killed or even injured by the
pure bacillus threads of our cultivations succumb without fail
to an inoculation with spores or blood bacillus, or to an inocu-
lation with the early stage of a new cultivation of bacillus de-
rived from the former cultivation. This inefficacy of the bacillus
of the cultivation on the mice, after several days’ cultivation,
must be borne in mind when judging of Buchner’s results
above quoted. Buchner (l.c., p. 384) finds the greatest irre-
gularity in respect of the supposed deterioration in virulence
of the cultivations, for while in one series, the third and fourth
cultivation is inactive; the fifth active if used in large quan-
tity; in other series other results are observed. Granted that
Buchner had pure cultivations, of which, however, there is no
sufficient evidence—see Koch (l.c., p. 25)—these irregular
results, I think, might be explained by the assumption that
the active cultivations were fresh or contained spores, the
inactive ones were of some age and had no spores, Buchner’s
cultivations being carried on in a fluid medium, and being
used solely on white mice. More difficult is it to explain
Greenfield’s statements. He speaks of mice, guinea-pigs,
and rabbits as all giving identical results under all circum-
stances, and this as if the identity of result were matter of
course and of necessity. I do not propose to comment on his
statements.

As has been already indicated, a given cultivation of Ba-
cillus anthracis, although speedily becoming inactive on

58 DR. E. KLEIN.

some mice, proved under all conditions and for a considerable
length of time, fatal to guinea-pigs and rabbits, no matter
whether spores had developed in it or not. This different be-
haviour of mice on the one hand and guinea-pigs and rabbits
on the other, towards an artificial cultivation of Bacillus
anthracis without spores, came indeed after a while to be a
useful means to decide whether a given cultivation of Bacillus
anthracis, after several days’ incubation, contained spores or
not. I have so often repeated the following experiment that I
am confident it can serve as a typical one. A sample of a cul-
tivation of Bacillus anthracis in neutral pork broth, which
appears to the unaided eye a typical growth, and in which cul-
tivation the bacillus mass is left quiet at the bottom of the test
tube or flask for a week or two, when examined under the
microscope does not contain any spores. Inoculate with it
half-a-dozen mice and half-a-dozen guinea-pigs or rabbits. All
or most of the mice will probably remain well, all the guinea-
pigs and rabbits die within forty-eight hours. Allow the
bacillus of the above cultivation to form spores, by sowing
them on to gelatine pork, and keeping them well exposed to
the surface, or establish a new cultivation in neutral pork
broth, and now inoculate the above six mice, or as many other
mice as you like, with this new cultivation in its early stage or
with the above spores, every one of them will be probably dead
within thirty-six or forty-eight hours.

The conclusions to be drawn from this seems to me obvious.
Mice, unlike guinea-pigs and rabbits, are insusceptible to the
Bacillus anthracis when cultivated artificially in neutral
pork, after this cultivation has been kept for some time, pro-
vided no spores are formed in the bacilli, But no immunity
of any kind is by such inoculation conferred on the mice.
Since mice are very susceptible to the Bacillus anthracis
of the blood and tissues of an anthrax animal in which noto-
riously no spores occur, and since they are (equally with
guinea-pigs and rabbits) susceptible to the spores of the arti-
ficially cultivated Bacillus anthracis and to the bacillus of
a fresh cultivation, it seems to me it follows from the facts, as

RELATION OF PATHOGENIC TO SEPTIC BACTERIA. 59

a necessary conclusion, that this insusceptibility must depend
both on the mice as well as on a change in the bacillus.

I have made several series of observations, to be detailed at
a future period, by cultivating Bacillus anthracis in acid
pork broth, and to my great surprise the first cultivation, and
also sometimes the second cultivation, of blood Bacillus an-
thracis in this acid pork broth formed spores, and consequently
killed all mice when inoculated into them in infinitesimal doses ;
but as cultivation was carried on into a third and fourth, the
bacillus, although still copiously and typically growing during
the first four or five days, nevertheless did not form spores at
any time. As a consequence it did not prove effective on
many mice; but it proved fatal on guinea-pigs and rab-
bits when inoculated into them in minimal doses.

What has been said in the foregoing paragraphs, respecting
the effects of inoculating with a given cultivation of Bacillus
anthracis in neutral pork broth and in gelatine pork, applies
not only to the first or the second cultivation, but also applies,
in exactly the same manner and to exactly the same degree, to
the third, fourth, fifth, sixth cultivation, and even (as I have
proved by my own observation) to the twentieth or thirtieth
cultivation of the bacillus in like nourishing material. As
each new sample of sterile material is inoculated from a former
sample, a typical and copious growth of bacillus threads takes
place init. If the material be pork broth, spores will not be
formed in it as long as the growth takes place undisturbed
below the surface of the liquid. And each successive sporeless
cultivation will after a few days (and always within a week or
two) lose its power to kill mice, though it will retain for about
two months the same power that preceding cultivations pos-
sessed of killing guinea-pigs and rabbits when inoculated into
them.

If any of these cultivations of Bacillus anthracis in
neutral pork broth or gelatine pork are kept for several weeks,
it will be noticed, as described above, that the mass of bacilli
gradually diminishes in a conspicuous manner. On a former

60 DR. E. KLEIN.

page I have pointed out that already while active growth is
going on in the cultivation, some threads undergo degenera-
tion, and when the pabulum in the cultivation is exhausted,
this degeneration gradually extends over the whole growth.
As a rule, as pointed out before, if during active growth the
bacillus mass has been kept at the bottom of the fluid, no
spores are formed, and therefore degeneration after the exhaus-_
tion of the pabulum gradually destroys every active particle of
the growth. Thus it happens that the cultivation, taken as a
whole, gradually loses its virulence, inasmuch as with the pro-
gress of the degeneration extending over greater numbers of
bacilli, larger doses must be injected into guinea-pigs and
rabbits to produce fatal result. But this must not be taken as
identical in meaning with what Buchner calls a diminution in
virulence of a cultivation. According to Buchner larger quan-
tities have to be used of a later generation (ceteris paribus)
to produce the same result as with a former generation,
because, he tells us, the Bacillus anthracis is gradually
changing its nature, becoming gradually converted into an
innocuous hay bacillus.

In our case the diminution in virulence of a cultivation is
entirely due to a diminution in the number of active bacilli,
and not to any progressive weakening of the potency of each
several bacillus. Wherefore the greater the number of bacilli
destroyed, the fewer undestroyed or active bacilli will be found
in a given quantity, or what comes to the same thing, a larger
quantity of material must be used in order to meet with an
active bacillus.

While there exist any living bacilli in the cultivation it is
possible to start new cultivations, which when used in the
early stage of the new cultivation or when allowed to form
spores kill without fail all rodents. I have made several
experiments in this respect, and I have invariably obtained the
same results.

A first cultivation, which promptly killed guinea-pigs during
the first fortnight when used in infinitesimal doses, failed to
kill a guinea-pig when injected into the subcutaneous tissue

RELATION OF PATHOGENIC TO SEPTIC BACTERIA. 61

after one month; a sample examined under the microscope
showed hardly a trace of a well-preserved bacillus and no
spores. I inoculated a test-tube of pork broth with it and
produced a beautiful growth of typical Bacillus anthracis;
some of these extended up to the surface and formed a copious
crop of spores. This killed a guinea-pig three weeks after it
was established. In the above case it was evidently only a
chance of our missing to have an active bacillus in the samples
which we used for the inoculation of the guinea-pig; but more
luckily we got one in the sample used for the inoculation of
the pork broth in the test-tube.

The above guinea-pig which escaped anthrax was not
immune against the introduction of the active virus, since it
succumbed to an inoculation afterwards with bacillus and
spores of a cultivation in gelatine pork.

All these observations seem to point out that there are two
conditions to be borne in mind: (a) a peculiarity possessed by
mice and not possessed by guinea-pigs and rabbits; and (b) a
peculiar change that the bacilli of the artificial cultivation
undergo as the duration of incubation advances. ‘As regards
the first of these conditions it is known through Chauveau
that Algerian sheep are altogether refractory against anthrax,
and consequently these Algerian sheep possess a peculiarity
not owned by the French sheep.

According to Pasteur the influence of the air (oxygen) on
the artificial cultivations of the micrococcus of fowl cholera,
and the artificial cultivations of the Bacillus anthracis has
a deleterious effect, inasmuch as it gradually weakens and
ultimately altogether destroys the activity of the respective
organisms. As I have no experience of the micrococcus of
fowl cholera, I cannot say anything about it; but of the
Bacillus anthracis I can say something from my own
observations, and I will undertake to offer to this theory of
Pasteur, viz. of the deleterious influnace of the oxygen of the
air on the Bacillus anthracis, an unconditional opposition,
If in a cultivation we meet with a copious production of typical

62 DR. E KLEIN.

and beautiful anthrax bacilli and threads thereof; and if we
find that owing tothe absence of sufficient oxygen these bacilli
fail to produce spores; and if we further find that after a
certain time the bacilli undergo degeneration, and not being
able to form spores owing to the absence of sufficient oxygen,
they generally all disappear from the cultivation, I think we
are justified in concluding that the conditions are exactly the
reverse of what is postulated by the theory of Pasteur; in con-
cluding, viz. that it is the want of sufficient oxygen which
destroys the bacilli. If oxygen had been present in sufficient
quantities, the bacilli would have formed spores and the culti-
vation would have preserved its full virulence for an indefinite
period.

Pasteur, as mentioned above, maintains that his cultivations,
kept without spores, gradually lost all activity. “If we
examine the virulence of the culture at the end of two days,
four days, six days, eight days, &c., it will be found that long
before the death of the culture the microbe has lost all viru-
lence, although still cultivable.” My observations bear out to
a certain extent this statement of Pasteur, inasmuch as the
cultivation lost its power to kill mice before it lost its power to
kill guinea-pigs and rabbits. But as regards guinea-pigs and
rabbits it does not hold good; for in their case complete want
of power to kill has appeared in my experience to be the same
thing as want of power to grow in a cultivation. Pasteur
further states that the animals inoculated with the mitigated
virus remain immune against further attacks of anthrax. It
is evident that Pasteur’s process of cultivation must in some
way have differed from my own, or that his assertion for
“animals” generally is too broad, for as regards the mice of
my experiments there is no immunity of any kind conferred on
them.

Pasteur in his cultivations, found, that owing to the diminu-
tion of virulence, as time went on, he could at will choose for
inoculation a fluid of less and less virulent effect, from one that
would produce a fatal effect to one that would have only a
slight or local effect. But, says he, sheep inoculated with such

RELATION OF PATHOGENIO TO SEPTIC BACTERIA, 63

a cultivation, which, owing to having been kept for a certain
length of time, produced no fatal effect, are ‘‘ vaccinated ”’ and
protected from anthrax in a virulent form.

As regards my mice, guinea-pigs, and rabbits, I have not
found anything of the sort. Either the inoculation with my
cultivations is fatal or it is not; in the latter case it has no
effect whatever, and does not at all protect against active
virus; in the former case it is always fatal. The inactivity on
mice of a cultivation may be due to the absence of spores, or
to the age of a cultivation—Pasteur’s statement of a diminu-
tion in virulence in two days and four days, does not quite
cover my facts—or the bacillus mass in a cultivation, not
being able to form spores and gradually degenerating and
dwindling away and becoming macerated into a granular
débris, loses after a time altogether its power to infect mice,
guinea-pigs, or rabbits, or to start new cultivations.'

These latter conditions come out especially strikingly in

1 As regards the slight effect (constitutional disturbance and rise of tem-
perature) produced in cattle after inoculation with anthrax blood of rodents
(Sanderson and Duguid), or with artificially cultivated Bacillus anthracis(?)
of a rodent (Greenfield), as well as the non-fatal effect produced on sheep by
Pasteur with his vaccine, we have to deal with peculiar conditions, not solely
due to a diminution of virulence of the bacillus, but chiefly to some pecu-
liarity (breed appears to be one of such peculiarities) of the animal inocu-
lated. These cases are comparable in a certain sense to those mild cases of
other infectious maladies, which not occurring more than once during the
lifetime of an individual, would naturally confer immunity on this individual
against a second attack. Thus, a person once having had a mild attack of
scarlatina, measles, &c., very likely remains free from a second attack. In
cases of scarlatina the differences in the severity of the attacks are due to
differences of the source of the virus [i. e. differences of the nature of the
virus], as well as to differences of the individuals attacked,—cases of varying
severity being derived from the same source, i.e. the same virus. The same is
also noticed in the eases of anthrax produced by the Bacillus anthracis ;
the bacillus of some cultivations is altogether ineffectual on mice, deadly on
guinea-pigs and rabbits, while it appears to produce, according to Pasteur,
only a slight effect on sheep. Now, no one could say this difference is due
entirely to a change of the bacillus, since it is equally due to the difference
of the individual. Again, the non-fatal result with the blood bacillus of a
guinea-pig, dead of anthrax, produced in a cow contrasts strongly with the

64 DR. E. KLEIN.

cultivations of the Bacillus anthracis carried on in acid
pork broth. I have made a sufficient number of observations
to state this positively, and I have seen such a cultivation
losing its infective power both for animals and for new cultiva-
tions after five days, no other organism making its appearance
in it, and the original mass of Bacillus anthracis having
altogether broken up.

The important statement by Pasteur that “each of these
conditions of attenuated virulence may be reproduced by
culture,” is not borne out by my observations, since every one
of the cultivations containing only anthrax bacilli but no
spores, and incapable of producing any effect on mice, is in-
variably capable of starting a new cultivation proving fatal to
all rodents when used fresh.

Ihave before me a fourth cultivation of Bacillus anthracis
in neutral pork, which had proved fatal to guinea-pigs and
rabbits. It had never any spores, and the days for its activity
on mice had passed. After the lapse of two months it was
again examined, and there were found in it bundles of dege-
nerated bacilli, as wellas a few good bacilli. Inoculated into
a guinea-pig in minute doses it proved without result, but it
started a good and copious new cultivation of typical anthrax ba-
cillus threads, which killed a guinea-pig with typical anthrax
in twenty hours. Pasteur maintains that if a cultivation is
weakened in activity by keeping it for some days, it is capable

invariably fatal result produced by the same bacillus on mice, guinea-pigs, and
rabbits.

It is very curious to find that Greenfield talks (‘ Veterinarian,’ 1881) of a
certain immunity against fatal anthrax conferred on cattle by his artificial
cultivations, although these animals showed considerable illness after a further
inoculation with blood of man or guinea-pig dead of anthrax. He had already
learned that cattle do not die after blood inoculation, even when not inocu-
lated previously with any artificial cultivation. If he had inoculated his cattle
with the blood of a guinea-pig or man (woolsorters’ disease), without pre-
viously inoculating them directly with artificial cultivations, the result would
have been precisely the same This appears to me to furnish decisive evidence of
Greenfield having had to deal, not with cultivations of Bacillus anthracis,
but with some other harmless bacillus.

RELATION OF PATHOGENIC TO SEPTIC BACTERIA. 65

of starting a new cultivation, whose activity is also weakened,
that is to say, the bacillus having become modified by time,
transmits to its offspring this acquired mitigation. In the
case of the cultivations of Bacillus anthracis in neutral
pork or gelatine pork there was nothing of the sort. As
long as a cultivation, no matter which, contains living Bacillus
anthracis, it is capable of starting a new cultivation, and
this as well as its parent is capable of killing guinea-pigs and
rabbits.

All that has been said of the first, second, third, fourth,
fifth, and sixth cultivations of Bacillus anthracis in neutral
pork broth holds good for the tenth, eleventh, twelfth, thir-
teenth, and so on cultivations. In no instance have I seen,
with reference to its infective power on mice, guinea-pigs, and
rabbits, any difference of behaviour from that mentioned of
the previous cultivations.

It is altogether impossible for me to understand how Green-
field (‘ Veterinarian,’ 1881) could have come to the conclusion,
that once arrived at the eighth cultivation, he already knew that
no fatal effect could be produced with it. He has tried the
effect of his cultivations on mice, guinea-pigs, and rabbits ;
but with pure cultivations of anthrax bacillus, the result is
to Some extent the reverse, since guinea-pigs and rabbits are
killed by any cultivation, provided there are living anthrax
bacilli in it.

The conclusion, it seems to me, forces itself on us, that
Greenfield’s like Buchner’s cultivations were impure, and the
further away from the earlier cultivations the smaller the
number of the anthrax bacilli, until the contaminating
innocuous bacillus gets altogether the mastery in the culti-
vations, and then the anthrax bacilli gradually disappear
altogether.

Of the power of resistance spores are capable of, an idea
may be gained from the following facts:

I have tried to ascertain whether the spores of Bacillus
anthracis in my cultivations become killed, like the bacilli
themselves, through boiling or freezing. As regards the first

VOL, XXII], —NEW SER. rE

66 DR. E. KLIN.

process, boiling a minute or two does not destroy the life of
the spores. I have thus treated, as mentioned on a former
page, spores contained in a flask of gelatine pork, and have
obtained afterwards from them a copious crop of bacilli proving
fatal to guinea-pigs and rabbits. I have similarly exposed in
a capillary pipette fluid full of spores to the influence of ether
spray, and having thus kept the fluid well frozen for several
minutes, have injected it into the guinea-pig and rabbit with
fatal result. I then subjected spores in the same manner to
repeated freezing, each time for several minutes, the freezing
being also carried out by the ether spray; but these spores
nevertheless retained their full virulence. Before forty-eight
hours were over the inoculated animals were dead of anthrax.
I then placed a capillary tube filled with spores in a mixture
of ice and salt, and kept it here for one hour exposed to a
temperature of 12° to 15° C. below freezing point ; after thaw-
ing the material was injected into the subcutaneous tissue of
a guinea-pig. This animal died of typical anthrax during
the third day. There was, however, no cedema. about the seat
of inoculation.

Such a low temperature, viz. 12° to 15° C. below freezing
point (or 21°—27° Fahr. below freezing point), does not occur
in the soil of middle Europe or of these kingdoms, even in
the depth of the coldest winter, and therefore spores of
Bacillus anthracis formed in the soil from the Bacillus
anthracis that happens to be growing there in a suitable
nourishing material (vegetable infusion, &c.), are practically
indestructible.

As an addition to our knowledge of the mode of propagation
of anthrax in animals the following facts may not be value-
less. In several instances I found that of a mouse that had
died of anthrax a great portion had been eaten by its fellow-
companion not inoculated with anthrax ; its neck, heart, lungs,
and liver had all been swallowed up, but, nevertheless, the
mouse that had thus feasted on anthrax remained perfectly
well. Koch (‘Die Aetiologie d. Milzbrand,’ p. 18) is very
strong on the communicability of anthrax through simple in-

RELATION OF PATHOGENIC TO SEPTIC BACTERIA. 67

gestion from the alimentary canal in animals and man. The
above facts do not support this theory as regards mice, for no
mouse would escape inoculation with an infinitesimal dose of
blood of anthrax. They are, however, well in harmony with
those of Pasteur and Toussaint, who maintain that in the case
of sheep the production of anthrax in these animals by in-
gestion is in reality due to an inoculation with anthrax into
the mucous membrane of the mouth, owing to small wounds
and abrasions produced there by the material mixed with the
food.

The important question as to the preservation of the activity
of the Bacillus anthracis within the body of the dead
animal, although commenced, is not advanced enough to be
here reported, and therefore must be reserved for a further
report. Also the observations on the changes the Bacillus
anthracis undergoes when cultivated in acid and alkaline
fluids. Some of these observations have been already touched
upon, but others of equal importance cannot find room here
to be discussed. But this much I will state already now, that
I have ascertained that, contrary to what Koch and Buchner
have found, Bacillus anthracis is capable of growing in
acid fluids, and that it most undoubtedly behaves in a different
manner, both as regards size, mode, and time of degeneration
and formation of spores, in alkaline and acid cultures, from
what it does in neutral nourishing fluids.

In conclusion, and as concerns the more general question
which has been held in view throughout the present report, it
is to be observed that the theory of the transformation of a
pathogenic organism into a non-pathogenic septic organism,
as expressed by v. Nageli and regarded by Dr. Buchner as
being capable of direct proof, is not supported by my own in-
vestigations into the characters of the Bacillus anthracis,
which has appeared to retain its full power to produce specific
disease as long as it has retained any power at all. And in
regard of that particular bacillus, it is further to be noted

68 DR. E. KLEIN.

that its behaviour towards sheep in Pasteur’s hands is not
the same as its behaviour towards rodents in my own ex-
perience; so that we must remain unable to accept, asa general
proposition, the general view of attenuation that Pasteur
would propose.

TONGUE OF PERAMELES NASUTA. 69

The Tongue of Perameles nasuta, with some
Suggestions as to the Origin of Taste Bulbs.

By
Edward B. Poulton, M.A.

With Plate I.

I am indebted to the kindness of Professor Moseley for the
opportunity of working upon the tongue of this little-known
animal. I had expected to meet with interesting details in
this investigation, but I hardly ventured to hope for characters
with so important a bearing on general development as, it
appears to me, are to be found in this organ, especially when
such suggestive structures are combined with so much highly
specialised peculiarity.

The animal to which this organ belonged was caught in the
summer of 1874, and its capture is described in the “ Notes
by a Naturalist on the ‘Challenger,’” page 269. Professor
Moseley had hardened the back part of the tongue in chromic
acid, and since that time it had been kept in strong alcohol.
This treatment was so successful that the cells came out in my
sections fully as well as in recently hardened structures. The
following is a description of the obvious characters of the piece
of tongue when it came into my possession. The length was
18 millimeters, including the limits of the papillate surface
behind, but cut transversely across this surface in front. The
width was 12 mm. and the thickness 9mm. There is a de-
scription of the whole tongue in vol. vi of the ‘ Memoirs of the
Wernerian Natural History Society,’ by Dr. R. E. Grant (in
a paper dated January 26th, 1828, on the “‘ Anatomy of Pera-
meles nasuta’’). He states that the tongue is very long,
flat, narrow, and rather thin; of equal breadth from the root

70 EDWARD B. POULTON.

to near the extremity, which has an elliptical form. Its length
is 3 inches from root to apex, and it is quite free for nearly 2
inches from the frenum, and thus capable of great freedom of
motion.

I thus had rather less than a quarter of the length of the
tongue, but probably including all the most interesting details.
At the posterior part of the papillate surface are three large
circumvallate papillz, each situated at the angle of an isoceles
triangle with the base directed forwards, and 4°75 mm. in
length (measured from the centres of the papille), the sides
being 2°5 mm. in length. The papille themselves are about
1 mm. in diameter, and are encircled by a peculiarly deep and
narrow trench.

The general surface in front of these three papille is densely
covered by very small papille, of a type which I believe to be
entirely peculiar. Each papilla is crowned by a circlet of 8—
10 fine, long, bristle-like filiform papille, whose points are
directed backwards, thus causing a slight roughness to the
finger when drawn from behind forwards. Thinly scattered
among these excessively numerous papille are others of the
“ fungiform’”’ type, of which about twenty were present on this
piece of the tongue, but as they were more thickly placed in
the front part they are no doubt commoner on the organ in
front of this piece. They are chiefly arranged as an irregular
single line on each side, beginning about 7°5 mm. in front of
the circumvallate papilla on the same side; but they also occur
on the upper surface, about 11°5 mm., in front of two anterior
circumvallate papille. Their appearance is quite normal.
Beneath the lateral row of fungiform papille (of which the
papille are separated by the small peculiar papille) is a row,
from two to three deep, of large filiform papille of ordinary
appearance, and beneath these the papillate surface ends
abruptly in a perfectly smooth epithelium. These rows of
filiform papille are continued backwards and upwards until
they meet at the circumvallate papillae, between and around
which they are thinly scattered, and are also longer than else-
where.

TONGUE OF PERAMELES. NASUTA, 71.

There is a slight trace of a median raphe, in the form of a
shallow groove, in the anterior part. Dr. Grant also mentions
a close covering of minute papille, no doubt referring to those
of peculiar type to be described further on; and he speaks of
others of larger size more thinly scattered (probably the fungi-
form papille), and describes the arrangement of the three
circumvallate papille. At this latter part of the tongue he
mentions the orifices of minute ducts, which I was unable to
detect (except deep in the trenches round the papille),
although I made horizontal sections in this part. He states
that the circumvallate papillae of the opossum are similar in
form and arrangement. The slight groove which appeared in
the anterior part of the upper surface of my specimen he de-
scribes as running the whole length ; and he also mentions that
there is a median ridge on the lower surface from the apex of
the tongue to the frenum, bordered by a fold on each side, and
a corresponding groove on the floor of the mouth beneath, with
a ridge on each side. The roof of the palate is covered by a
black cuticle, and is traversed by about fourteen transverse
elevated ridges slightly curving forwards; and as Dr. Grant
suggests, of importance in grinding down the hard coverings of
insects when the sharp points of the papillee are rubbed against
them.

I have now mentioned the details of this organ given by Dr.
Grant, and I have been careful to include all points, since
this animal seems to be very slightly known in Europe. The
structure of the abundant peculiar papille renders it almost
certain that the animal is insectivorous. Yet Gould, in the
‘Mammals of Australia,’ says that its ‘‘ food consists of bulbous
and other roots, obtained by its powerful fore feet and claws,”
but he adds that there is very little information known respect-
ing it. Waterhouse, however, quotes Dr. Grant as the autho-
rity for the insectivorous habits of P. nasuta.

In the paper above-mentioned Dr. Grant says that the feeces
were composed of insects, together with some tufts of woolly
hair and some vegetable matter, probably taken in accidentally.
The stomach and small intestines also contained a little hair,

72 EDWARD B. POULTON.

sand, and vegetable matter, with no regular food. The simple
character of the digestive tract and short cecum is also evidence
that the animal does not subsist on vegetable food. E. Geoffroy
first named and described the genus Perameles from this
species and another in the ‘Annales du Muséum d’Histoire
Naturelle’ (1804) ; and in describing this species he infers that
it is insectivorous, from the characters of the teeth, and sug-
gests that this food is obtained by digging. This latter habit
probably explains Gould’s mistake and that of the colonists, for
I believe that in Australia the bandicoots are generally believed
to be plant-eating.

I have now mentioned all the points bearing upon my sub-
ject that I can find in any former writings upon this animal.
The structures J am about to describe admit of a simple classi-
fication under two heads. First, those structures which are
probably concerned with taste—the circumvallate and fungi-
form papillz, together with some suggestions as to the origin
of taste bulbs. Secondly, those structures of mechanical or
tactile use—the papille of peculiar type and the filiform
papille.

I. Gustatory Structures.

The Circumvallate papilla—The surfaces of these
papille are circular, and a little more than 1 mm. in diameter.
The sides (protected by the trench) are vertical for °33 of a
mm., and then incline inwards for about the same distance,
making an angle of about 36°, with the vertical side above.
Below this the side of the papilla turns inwards for a
short distance almost horizontally, and by this and the inward
slope above, the diameter of the papilla is only about °5 of a
mm. at its base. The taste bulbs are arranged only upon the
side inclining inwards and downwards, and are thus peculiarly
protected. The depression of the epithelial surface forming
the outer wall of the trench follows the sides of the papilla
with a curved and smooth outline in vertical section, but seen
to be vertically ridged by horizontal sections. The papilla
itself is not similarly ridged. The outer wall of the trench

TONGUE OF PERAMELES NASUTA. 13

is prolonged upwards very nearly to the level of the surface of
the papilla, and as this is rather above the level of the tongue
the papilla is surrounded by a slight ridge. The trench is
very narrow, and its depth and relation to the shape of the
papilla is shown in fig. 1, which is magnified 24°5 diameters.
Thus, in shape this papilla is peculiarly specialised in the
way of protection; its minute structure will also be found to
be highly specialised in many points. Glands are very
abundant within the bodies of the papille, between the three
papille, and for a considerable distance around them. They
are almost entirely of the granular “serous”? type, which
Klein points out as always accompanying taste bulbs. Their
structure is exactly as described by Klein. Their ducts open
into the trench, especially in its deeper part, and are very
numerous. In one horizontal section I counted twenty-six
ducts, probably all separate, and at a lower level I calculated
that there must have been at least forty at one horizon, while
the bottom of the trench is completely surrounded by thickly
crowded gland-ducts radiating inwards (see figs. 1 and 2 for
vertical sections). The body of the papilla, as usual, bears
secondary papille on its upper part, the depressions between
which are filled up to one level by the epithelium.

The most remarkable structure, and as far as I am aware
one hitherto undescribed, is a large and distinct ganglion in
the form of a thick axial column making up a great part of
the bulk of the papillary body. It is surrounded by a clearly
defined connective-tissue capsule, which enters and supports
the nervous elements. Above, the ganglion breaks up into
branches, which stream outwards towards the sloping side of the
papilla containing the taste-bulbs. The gradual collection of
the scattered branches above and to the sides into the dense
and compact ganglion centrally and below is especially well
seen in successive horizontal sections. The nerve-fibres appear
to be almost entirely non-medullated, but they possess a distinct
sheath of Schwann. Among the nerve-fibres occur primitive
nerve-fibrils. ‘The nerve-cells are few in number, very large
and distinct (they are indicated even in the lowly magnified

74 EDWARD B. POULTON,

fig. 1), and always situated at the base of the ganglion close
to the fibrous capsule. In vertical sections two to six cells
appear in one section. Ina horizontal section across the base
of the ganglion about twelve cells are seen, and many others
also present appear to be belong to small ganglia in the
course of nerves entering the larger ganglion. The nerve-
cells are very large and fusiform, with branched ends (the
branches losing themselves among the nerve-fibres of the
ganglion). The large oval nucleus, containing a distinct
nucleolus, is coarsely granular in appearance, and does not
stain in hematoxylin, although the finely granular cell-body
stains deeply. Large nerves with non-medullated fibres are
distinctly seen entering the base of the ganglion, and in a
deep horizontal section a ground-plan, as it were, of the gan-
glionic nerves is seen, as they radiate outwards from the base
of the ganglion. These nerves contain isolated nerve-cells,
and also small groups, in their course, the cells exactly resem-
bling those of the ganglion, except that they become polyhedral
when crowded together

It thus appears almost certain that nerve-cells are inter-
calated in the course of sensory impulses from the peripheral
organs to the nervous centres. ‘This is of interest in bringing
these terminations into closer connection with the related
terminal organs of sight and hearing, where ganglion cells
similarly intervene. If this be a true correlation, it seems
likely that ganglion cells will be found generally in the nervous
masses of the large gustatory papille, now that attention is
directed to their existence. Microscopic ganglia on the nerve-
branches have been described, and a nerve is_ generally
figured in the axis of the papilla, but a true, large, compact
ganglion making up most of the papillary body is, I think, as
yet unmentioned.

Beneath the epithelium containing the taste bulbs the
nerve-fibres (in the course of or between which there appear
to be many nuclei) are very numerous, and are cut in all direc-
tions owing to their irregular and sinuous course. There is
no doubt of their connection with the. bulbs, but it would

TONGUE OF PERAMELES NASUTA. 75

probably need the fresh tissue to trace the actual union. The
connective-tissue matrix, in which these fibres ramify, is less
dense than the mucous membrane of the organ generally,
although derived from the latter. This mucous membrane is
peculiarly firm and tendinous in appearance, with its fibres
arranged transversely to the long axis of the organ and con-
taining many interfascular spaces. Striated muscle fibres
appear to terminate very abruptly in it. The ganglion and
nervous elements are shown in fig, 2, which is taken vertically
through the base of the papilla.

The taste bulbs of the circumvallate papille are fairly
numerous. They are arranged in a zone of seven or eight
tiers, exactly filling up the overhanging side of the papilla. The
calculation of the number of bulbs in a tier from horizontal
sections cannot be very exact. In one semicircle towards the
upper part I counted fifty bulbs; in the lowest there seem to
be not less than forty, although here they are somewhat larger.
Thus, allowing the mean, ninety in each tier, and allowing
for eight tiers, the number of bulbs in each of the three
papille becomes 720. The length of the bulbs appears to be
about ‘07 mm., but the lower are always larger, and the size is
somewhat irregular throughout. Their shape is often a per-
fect oval, but sometimes rather like that of a peg-top without
the spike, for there is hardly any neck, but in longitudinal
section the sides seem to meet almost at the surface of the
epithelium in a blunt point. The only representative of a
neck is the gustatory pore itself, which perforates a very thin
apparently homegeneous superficial layer, probably formed by
coalesced epithelial cells of the surface (see fig. 2). This
layer is also seen covering other parts of the tongue, and is
cornified, as it stains yellow in picrocarmine. The perfora-
tion of this thin layer is very often seen both in horizontal and
vertical sections. In some cases there was a distinct protru-
sion through the gustatory pore, but anything like a circlet of
hair-like processes projecting from it could not be identified.
The cells do not appear to be collected together into a distinct
basal pole, although they most distinctly converge at the apex.

76 EDWARD P. POULTON.

This contact of the cells of the bulb with a tolerably ex-
tended surface of mucous membrane at the base is still more
apparent in a simpler form of bulb to be described as occurring
on the fungiform papilla. Another indication of simplicity is
the persistence of papillary elevations of the mucous membrane
between adjacent bulbs. This was most distinct in horizontal
and vertical sections of the circumvallate papille, and was
especially marked at the demarcation between the lowest bulb
and the ordinary epithelium (of course in vertical sections).

The great irregularity of size and shape also appears to
favour the view that these bulbs are of a peculiarly undeveloped
and ancient type. This will be further considered in the dis-
cussion upon the origin of taste bulbs. No very distinct
separation of the cells of the bulbs into central and peripheral
could be made out, and it is possible that this structural
difference is not yet established, but this suggestion needs
confirmation by work upon a fresher tissue. The cells of the
bulbs stain very slightly in carmine or hematoxylin. Fig. 2
shows the structure and arrangement of the bulbs.

Thus these circumvallate papille in their shape and struc-
ture are peculiarly highly developed, notably in the abundant
glandular and nervous elements and the presence of nerve-
cells in the ganglion of the papilla. The overhanging sides of
the papillary body must also be regarded as marks of great
specialisation, carrying still further the protective function of
the trench. Yet combined with these extremely developed
structures are terminal organs of a lower type than have yet
been described.

The inference is that the former structures have reached
their high specialisation by assisting another form of terminal
organ, which has been comparatively recently replaced by the
bulbs. This probability will be further discussed after the fun-
giform papilla have been considered, for these latter possess
structures with an important bearing on the argument.

The fungiform papille.—These papille are entirely of
normal shape, appearance, and distribution. The only note-
worthy fact about their distribution is the collection into a line

TONGUE OF PERAMELES NASUTA. 77

on each side of the tongue. The peculiar papille which every-
where surround them always leave a little space immediately
round the fungiform papilla, in the centre of which it stands.
The shape is shown in fig. 3, which is a vertical section through
a papilla. The mucous membrane in the centre is of the ordi-
nary type, but less dense than that below, from which it is
prolonged.

A large non-medullated nerve occupies the axis, and is well
seen in vertical and horizontal sections. The epithelium re-
sembles that of the general surface of the organ, and, like it, is
penetrated by papillary upgrowths from the mucous membrane
below.

Taste bulbs are not very common on the fungiform papille
of the higher animals, and seem to be always isolated when
they are present.

It was therefore unlikely that they would be common here,
and I examined very many sections without meeting any
tracesofthem. At length I found some indications, and finally
the specimen shown in fig. 3. In this section (at the top of the
papilla amongst the diagrammatically-shaded epithelium) there
are two distinct bulbs of a very low order. They have not yet
anything of the bulbous shape, and the basal ends of their cells
are spread out over a wide extent of mucous membrane. In
this section they do not reach the surface of the epithelium, but
it is probable that in a section through the true longitudinal.
axis the surface may be reached or even perforated. They are
seen to be merely the lowest columnar cells of an interpapillary
process, greatly prolonged towards the surface, and it is in-
teresting to see one or two columnar cells outside the chief
mass, in both cases, also elongating and applying themselves to
the others. It is also noteworthy that the apex is not pro-
duced by any curve of the cells, as in true bulbs, but merely
by the cells being prolonged froma curved surface, and so, like
radii, meeting at acommoncentre. Three papillary upgrowths
distinctly separate the two developing bulbs from each other
and, from the surrounding epithelium.

There were no glands of the “ serous type,”

as far as I

78 EDWARD B. POULTON.

observed, near the fungiform papille, or indeed anywhere in
my sections, except in or around the area of the circumvallate
papille.

Mucous glands were, however, common in other parts and of
the type described (Klein, in the ‘ Atlas of Histology.’)

The origin of taste bulbs.—This low form of bulb, found
upon the fungiform papillz, suggested to me a possible expla-
nation of the manner in which taste bulbs have arisen in Mam-
malia. At the outset it seems probable that in Marsupials or
Monotremes we have the best chance of finding the course fol-
lowed in the development of these and other structures, as we
know them in the higher mammals.

For in these extremely ancient types, owing their existence
to isolation, with little rivalry to render structural advance and
complexity necessary in such forms, it is certain that long
halts will be made at stages long since left behind in the develop-
ment of other animals living on more warmly-contested areas.

There is little doubt that the gustatory terminal organs have
more in common with those of the general surface of the boly
than any other special sense. There is great structural resem-
blance, actual structural continuity, between skin and the oral
mucous membrane with its epithelium. Certain sensory ter-
minal organs are found in both, although it is probable that
they subserve the tactile rather than the gustatory sense in the
mouth.

Before the appearance of taste bulbs actually opening on the
surface of the epithelium, when the gustatory surface was less
specialised, the terminal organs (similar to those of skin or
modified to receive at first feeble gustatory stimuli) would be
placed as in skin, namely, in the papillary upgrowths from the
mucous membrane. In this position they would be nearer to
the gustatory stimuli than any other, without actual perforation
of the epithelium by the terminal organ, for, of course, the
layers of cells over a papillary process are far thinner than
elsewhere. It appears to me that then the serous glands were
modified from those of the general mucous type in the parts
where the terminal organs had been chiefly specialised. Simul-

TONGUE OF PERAMELES NASUTA. 79

taneously the nerve-supply would be increasing in amount and
advancing in complexity.

The chief taste areas will have also been sheltered by
becoming enclosed in folds, either of the circumvallate or
foliate type. Of course the exact order of these events cannot
be made out, nor is it of great importance. The important
point is that a time must have come when a more specialised
form of terminal organ, coming into closer relations with the
stimulus, was substituted for one of a more general type. It
is not necessary or possible to exactly define this point of time
relatively to that of the advancing accessory structures, but, as
before mentioned, the high place reached by these in the cir-
cumvallate papille, accompanied by a low type of bulb, renders
it probable that the latter was subsequently developed. It is
also noteworthy that these accessory advantages may have
been more necessary with a less advanced type of terminal
organ.

This substitution of a higher form of terminal organ seems
to have taken place by the growth of the columnar cells forming
the lowest layer of an interpapillary process. In this way the
cells would approach the surface, converging as they elongated.
There would also be a gradual concentration of the nervous
elements upon these new end organs and a corresponding
withdrawal from the papillary structures.

It seems to me that this stage is reached in the fungiform
papilla (fig. 3) above described. The deeply placed columnar
cells of the interpapillary process have elongated, and so come
into closer relation with the surface than the less deeply placed
cells of the papillary process ; that is to say, the former would
be more advantageous as terminal organs. This account of
the origin of taste bulbs explains one important difference be-
tween them and the other structurally related end organs, as
those of the olfactory region, or sacculi and ampulle, i.e. in
the fact that the gustatory cells are massed together in little
groups surrounded by protective cells, while the auditory cells
in the positions above mentioned and the olfactory cells are
isolated, each being separately protected by columnar cells.

80 EDWARD B. POULTON.

This difference, it appears, is simply due to the latter elongat-
ing from a tolerably plane surface, while the gustatory cells
have elongated from the curved surface of an interpapillary
process—approximately the segment of a sphere—and therefore
have met and penetrated the surface in a group. In the further
development of the bulbs the external columnar cells would
become protective, the axial cells-alone acting as end organs ;
the columnar cells would converge to form a basal pole, as the
nerve supply was limited to a small area in this region (being
connected only with the axial cells). In fig. 3 the basal con-
vergence has not commenced, and in the circumvallate bulbs
(fig. 2) it is not nearly as complete as in higher animals. In
the former case I do not think that there is yet any trace of a
division into protective gustatory cells in the bulb, and in the
circumvallate papille I am sure that the difference is not well
marked, even if begun. Finally, the papillary elevations would
disappear between the bulbs, and the latter would rest in the
cavities of an epithelium with a nearly plane surface below.
This, which is reached in the higher animals, is apparently
never attained in the tongue of Perameles nasuta. The
papillary upgrowths separating the bulbs give to them the ap-
pearance of interpapillary processes to a marked degree. In
fact, there is the greatest resemblance between the bulbs and
the interpapillary processes of the epithelium on the outer wall
of the trench—a resemblance so great as to suggest this expla-
nation of the origin of bulbs.

And yet indications of the ultimate disappearance of the
papille between the bulbs are seen in the fact that the papille
between the lowest tier of bulbs and the ordinary epithelium
below, are always far more marked than those between the
bulbs themselves (see fig. 2). I think that a comparison of
figs. 2 and 3 with the figures of taste-bulbs in ‘ Stricker’s Hand-
book,’ by Engelmann, or in Klein’s ‘Atlas,’ will at once suggest
an explanation of their origin similar to that which I have
given above. I must express my thanks to Mr. W. H. Jack-
son, M.A., for kind help and suggestions in working out the
above theory.

TONGUE OF 'PERAMELES NASUTA. 81

Il. Structures with Mechanical or Tactile
Functions.

The papille of peculiar type.—These papille are very
numerous, thickly covering the upper surface of the tongue
in front of the circumvallate papille, and certainly continuous
over the organ in front of the anterior limits of the piece in
my possession.

They are smaller and more thickly placed anteriorly, and
here I counted thirty-four on a square mm. of surface. Pos-
teriorly there were only sixteen on the same area. ‘These pa-
pille appear to be closely related to the compound filiform
type of other animals, differing in the regular ring-like arrange-
ment and the number of the secondary papille, and also in
certain points of minute structure.

Their appearance when examined as opaque objects is given
in fig. 10 (x 55°25), a and B; A representing an anterior, B a
posterior papilla. The summit of each papilla is surrounded
by a ring of fine hair-like papille, generally ten in number,
which sweep backwards, and must act very effectively in retain-
ing small insects.

The hair-like papillae are longer and finer anteriorly, and
form a more complete ring ; but even here the ring is most
developed posteriorly in each papilla, and tends to become in-
complete anteriorly with feebler secondary papille. This
arrangement becomes gradually more marked posteriorly, until
around the circumvallate papille the anterior part of each ring
finally disappears, while the posterior part becomes immensely
developed as a very thick, blunt, secondary papilla, with one
or two small ones on each side of it. (One of these depressed
forms of papilla is seen in vertical section in fig. 1, p.) Part
of this transition is seen in fig. 4. I believe that in front of
this piece of tongue the papillz are in a short distance sur-
mounted by a symmetrical ring of hair-like processes. The
same transition occurs from the centre to the side of the tongue,
but is far less marked. Here the fine processes bend upwards

VOL, XXIII.—NEW SER. F

82 EDWARD B. POULTON.

and backwards instead of backwards merely, and the superior
and posterior side of the ring is chiefly developed.

Within the ring the summit of the main papilla is concave,
the greatest depth being attained near the anterior part of the
ring (or inferior and anterior side in the case of the lateral
papille). This is because a far longer incline of central cells
leads up to the posterior hair-like processes than to the ante-
rior (see fig. 11, &c.), and this difference is especially marked
in the posterior part of the tongue, while it almost disappears
in front (fig. 4). Transverse vertical sections show that the
rings are developed equally on both sides of the antero-pos-
terior diameter, but as the hair-like papille bend sharply back-
wards they are only cut through near their bases (fig. 5). The
same bilateral structure is seen in horizontal sections (figs. 6
and 12). Each papilla is seated upon a single main papillary
upgrowth from the mucous membrane.

The relation of this to the secondary papille is best studied
in horizontal sections at various depths (figs. 6, 7, and 12). In
such sections the papillary process is almost circular in section
at the lowest level, but a very little higher there is an inter-
ruption at one point of the margin by a small ingrowth of cells,
which a little higher becomes so large as to convert the ori-
ginal simple involution into a ring, incomplete only where the
central mass of cells is continuous with the epithelium outside
the ring. These appearances are due to the primary papillary
involution growing upwards as a ring everywhere except ante-
riorly, and therefore the first appearance of the ingrowth of
cells indicates that this is the anterior part of the margin.
The mass of cells within the ring is convex below, and at the
same time slopes slightly downwards anteriorly, and is con-
tinuous with the wall of the ingrowth (see fig. 4, A end).
Therefore, this point is first'reached in horizontal sections from
below upwards. The extension of the primary process into
the ring is well seen in transverse vertical sections (fig. 5),
and less well in longitudinal vertical sections, for here the
section passes through the incomplete point of the ring or near
it, where the ring is less developed.

TONGUE OF PERAMELES NASUTA. 83

In the anterior papillz the ring is more perfect and the in-
complete point is much narrower. Hence longitudinal vertical
sections, which do not strike this point, show a well-developed
upward growth anteriorly (compare the ends of fig. 4 in this
respect, and the ring p” of fig. 7 with p’ of fig. 6). This
explains why the vertical longitudinal sections (figs. 8 and 11,
and the anterior [ B] end of fig. 4) do not often show the con-
vex surface within the ring sloping down to be continuous
with the anterior wall. Horizontal sections of course do
indicate this.

At higher levels the ring gives off small secondary papillary
processes for the hair-like papille. These are first met with
in the feebly developed anterior horns of the ring, and gradu-
ally extend backwards as the ring (becoming semicircular, finally
a semiiunar remnant) rises higher. Finally, the last semilunar
trace of the ring gives origin to the largest and most posterior
papillary upgrowths (fig. 6 shows all these changes very dis-
tinctly in passing from A to B). Although, strictly speaking, no
papillary upgrowth can take place in the exact anterior margin
of the ring (incomplete), it is common to meet a papillary
process almost at this point at a rather higher level (fig. 6, &c.).
This is due to a papillary process rising a little obliquely from
one side to the anterior point, and explains why two hair-like
papillz are almost invariably cut through in longitudinal ver-
tical sections (figs. 4 and 11), although the corresponding
papillary upgrowth for the anterior papilla is often wanting
(see posterior end of fig. 4). The upper surface of the papilla
within the ring of hair-like papillae corresponds to the under
surface within the ring-like extension of the papillary process
from the mucous membrane. The concave upper surface cor-
responds to the convex lower surface, and the downward
anterior slope of the latter to the upward posterior slope of the
former. From side to side there is a regular concavity above
corresponding to a regular convexity below (fig. 5). The
curves are, however, always more marked below than above,
as the great thickness of intervening cells tends to partially fill
up the hollows (figs. 5, 11, &c.). The inferior convexity is

84. EDWARD B, POULTON.

obviously an interpapillary process between the small secon-
dary papilla. It now remains to describe the minute charac-
ters of the cells of the papille and epithelium around, and the
relation of both to the hair-like papille.

When the mass of cells within the ring is cut vertically it is
at once seen to be divided into two chief layers, sharply
marked from each other. The upper stains deeply and the
cells appear homogeneous, the lower does not stain (or very
slightly), and the cells are extremely granular (figs. 4, 9, 11,
and also seen in horizontal sections, figs. 6 and 12).

The transition of characters met with in passing upwards
through the central mass of cells is very remarkable. The
mass may be divided into two chief parts (already indicated),
each of which may be further subdivided (see fig. 8 and de-
scription). From below these are (A) granular cells hardly
staining in picrocarmine or logwood.

(1) The lowest columnar and succeeding small polyhedral
cells. The columnar cells are shorter than in the rest of the
epithelium, and both kinds of cell more granular with less
distinct outlines. Indications of karyokinesis are not un-
common.

(2) The cells are fusiform and granular with indistinct out-
lines; the nuclei frequently have a vacuolated appearance.
This is commoner over the secondary papillary processes
beneath the hair-like papille than over the convex interpapil-
lary part.

(3) A very thick layer of cells with distinct outlines, fusi-
form in shape, attenuated below but much swollen above. The
contents are very large granules, arranged in groups or rows
in a fine granular matrix, staining slightly, and especially dis-
tinct at the margin, where a thin layer is generally free from the
larger granules. The nucleus is often indistinct, shrunken, and
sometimes absent, filled with large and smail granules. The
large granules are rounded or angular, and do not stain at
all. (See fig. 9, aand 4, for the outline of a swollen cell from
the upper part of this layer.) Attenuated cells often have a
single row of granules from end to end.

TONGUE OF PERAMELES NASUTA. 85

(4) A narrow layer of very attenuated cells, still granular,
and usually having no trace of a nucleus. This layer appears
to be more constant in the posterior papillee.

B. Deeply staining cells :

(5) A narrow layer of attenuated, deeply staining, finely
granular cells with no nuclei.

(6) Homogeneous, swollen, fusiform cells, very deeply
staining, and rarely containing faint traces of a nucleus, as a
stellate mass, commoner in the anterior papille.

The demarcation between A and B is extremely sharp, and
the cells which are at length formed at the surface, after such
varied changes, are wonderfully like those formed by the
simpler transitions of the ordinary epithelium outside the
papille. This epithelium is of the normal stratified type, and
its cells have very distinct nuclei. Traces of karyokinesis
can be distinguished in the lowest layer. The superficial
cells only differ from the highest cells within the papille in
possessing nuclei and in staining rather less deeply.

The hair-like papille are formed of cornified fibre-like cells
derived from three sources. First, from the small secondary
papillary processes on which each is situated. The rapid
transition from columnar and polyhedral forms to cells with
apparently vacuolated nucleus, and finally to vertical fibres in
the axis of the hair-like papillze, is well seen in a vertical section
through the side of a papilla. Secondly, the cells within the
ring of hair-like papille apply themselves to the latter on all
sides. The cells at all depths and of all the different kinds of
structure mentioned, seem to apply themselves to the hair-like
papille, either before or after the emergence of the latter at
the surface. Thirdly, the external cells of the regular strati-
fied epithelium of the tongue apply themselves to the outside
of the hair-like papillae. (These three sources are best seen
in fig. 8.) Thus the cells are received from different sources,
and one of these (the second) is extremely complex.

As to the explanation of the granular cells, it seems most
probable that the appearance is due to the development of the
corneous material. It appears that this material (or some

86 EDWARD B. POULTON.

necessary precursor of it), at first formed as large granules,
becomes gradually more finely granular, undergoing suddenly
a change which enables it to become coloured deeply by staining
fluids, and shortly after becoming entirely homogeneous. If
this be correct the granular cells must be looked upon as repre-
senting on a very large scale the stratum granulosum and
lucidum of skin, which are generally considered to be stages
in the development of the corneous condition reached at a
higher level.

The difficulty remains, that outside the papilla no trace
of the granular cells is seen, although very similar cells are
met with at the surface, both within and without the papille.
It also appears remarkable that the granular cells should be
applied at all stages of structure to the hair-like papille, and
that the latter should receive accessions from the outside where
there are no granular cells, and from its own secondary
papilla, where the granular cells are not much developed.

I have described the structure of these papille minutely,
as they seem to me to be an entirely new and hitherto unde-
scribed type, modified in a remarkable manner for the capture
of small insects.

Filiform papillaz.—These are very large and long (1:6
mm. in length), and entirely normal in appearance. Their
direction is backwards in the row along the sides (limiting
the papillate surface), upwards and backwards among the
circumvallate papille. They have no corneous investinent,
and a large non-medullated nerve may often be seen in the
axis. It is therefore probable that they are tactile rather
than mechanical in function, a conclusion which is also con-
firmed by the greater specialisation of the peculiar papille
for these very mechanical purposes. ‘The distribution of the fili-
form papille is also strangely antagonistic to the view that they
are mechanical, and certainly not opposed to the supposition
that they are tactile.

PLANT CELLS AND LIVING MATTER. 87

Plant Cells and Living Matter.
By

Louis Elsberg, ™.D.,
of New York.

To botanists biology owed its first knowledge of ultimate
structure and of living matter. The names “ cell” and “ pro-
toplasm” testify to the epoch-making researches of Schleiden
and Von Moh]. And in accumulation and classification of
further biological knowledge botanists have taken so prominent
a part that even those of us who are interested only in animal
morphology have had to keep some track of the labours of
Nageli, Pringsheim, De Bary, Hofmeister, Sachs, Prantl, Stras-
burger, and many others. It is all the more remarkable, there-
fore, that the investigations carried on during the past decade,
which have resulted in proving that all the so-called ‘ cells”
constituting animal tissues are interconnected by filaments of
living matter emanating from these “ cells,” seem to have
borne no fruit for the study of plants. It was in the hope of
being able to repay histological botany for some of the light it
has thrown on animal histology that I engaged in the researches,
the account of a few of which I am about to detail.

A small portion of a delicate blade of grass, cut off with a
pair of scissors, transferred to a slide together with a drop of
dilute glycerine (two parts of pure glycerine and one part of
distilled water), was examined with a power of 1200 diam. I
had at my disposal for these examinations two excellent immer-
sion lenses, made respectively by Tolles, of Boston, and Vérick,
of Paris. In some parts, in trichomes, stomata, air-vessels
&c., nothing more could be seen with such amplification than
with comparatively low powers of the microscope ; the epi-

88 DR. LOUIS ELSBERG.

dermal fields as well as the surrounding frames of cellulose
appeared structureless, or at most only very indistinctly granular.
The main mass of tissue enclosed by the epidermal system, the
parenchyma, presented blunt polygons separated from each
other by a shining narrow rim of cellulose, and containing
numbers of chlorophyll-granules. Some contained only very
few and very small such granules, surrounded by an extremely
delicate uncoloured reticulum, of which the filaments were of
about the same breadth as the points of their intersection. In
some polygonal fields there were a number of coarse chloro-
phyll-granules interspersed in a network, the threads of which
had points of intersection that were thickened so as to consti-
tute distinct though not green minute granules, while in other
fields there were so many coarse and smaller green granules
that they nearly completely filled up the polygon. Under all
circumstances, however, the granules, closely focussed, appeared
stellate, and were interconnected by means of delicate fila-
ments running in large numbers from each granule to all its
neighbours. If of small size a chlorophyll-granule appeared
homogeneous, of a comparatively higher lustre, and of less
intense green colour; larger granules exhibited an indistinct
reticular structure in their interior; the largest showed the
reticular structure very plainly, and not infrequently in the
centre a small shining body was observed sending radiating
spokes toward the periphery, inosculating with a thin wall that
enclosed the granule in toto. Toward the apex of the blade
the granules became fewer in number and smaller in size ; at
the apex there were no chlorophyll-granules.

In fig. 1 are represented chlorophyll-granules (cHL.) inter-
spersed in the reticulum (Rk), surrounded by the cellulose
frame (Cc).

These observations show that the vegetable living matter
enclosed by the wall of cellulose is arranged in the form of a
network, and that a similar reticular arrangement exists in the
chlorophyll-granules. It is well known that chlorophyll-
granules are themselves minute masses of the living matter of
plants, coloured green by a colouring matter, to which the

PLANT CELLS AND LIVING MATTER. 89

name chlorophyll is given. Living matter has been called by
Hugo von Mohl “ protoplasm,” by Lionel Beale “ bioplasm,”

Figk

Fic. 1.—Cells from blade of grass, showing —CH. Chlorophyll granules.
R. Reticulum of protoplasm, and C. Cell-wall.

and by me, because etymologically more correct, ‘ bioplasson.”’
I am no stickler for new names, but in scientific discussions
we should use, if possible, correct names; and of the four
synonymous designations, viz. living matter, protoplasm, bio-
plasm, and bioplasson, I therefore confine myself generally to
the first and last, although the term protoplasm is best known
and by others most used.

In the year 1873, in a communication to the Vienna
Academy of Sciences, entitled ‘ Phases of Living Matter,”
Carl Heitzmann first described, in Ameba, the youthful con-
dition of masses of living matter as being constituted by homo-
geneous granules, and advanced stages as being characterised
by vacuolation followed by reticulation. These statements
were confirmed as regards vegetable organisms in a paper on
‘“ The Structure and Growth of some Forms of Mildew,” in the
‘New York Medical Journal,’ November, 1878, by William
Hassloch, who says that the first visible form elements of the
plant are homogeneous granules, and the first appearing buds

90 DR. LOUIS ELSBERG.

compact projections, either globular or elongated, the first
differentiation consisting in the occurrence of a central vacuole,
while after a certain development has been attained the plant
protoplasm appears in the form of a network.

Many botanists have observed and described reticulated
living matter, not only when in its naked condition, as plas-
modium, as it is called, but also when enclosed in a cellulose
wall. Allow me to cite a few examples: Sachs has figured “a
cell of Zygnema cruciatum, with two stellate chlorophyll-
bodies which are suspended in the interior of the cell; they are
united by a colourless bridge of protoplasm in which lies a
nucleus; the rays which form the union with the parietal sac
are already nearly colourless in the middle. In each of the
two chlorophyll-bodies lies a large grain of starch (amplifica-
tion 550),” also “ forms of the protoplasm contained in cells of
Indian corn (Zea mais) ; A, cells from the first leaf-sheath of a
germinating plant, showing the frothy condition of the proto-
plasm, i.e. the many vacuoles separated by thin plates ; B, cells
from the first internode of the germinating plant; the proto-
plasm is broken up into many rounded masses in each of
which there is a vacuole (0); these are the so-called ‘ sap-
vesicles.’ ” Sachs has also figured “‘parenchyma cellsfrom the
central cortical layer of the root of Fritillaria imperialis,
longitudinal sections, A, very young cells, lying close above the
apex of the root, still without cell sap or vacuoles. B, cells of
the same description about 2 millimétres above the apex of the
root; by the entrance of cell sap the vacuoles s,s, s, have
been formed. c, cells of the same description about 7 to 8
millimétres above the apex of the root,” in one of which the
reticulum is very plainly seen. Bessey says “in the stamen-
hairs of Tradescantia Virginica the protoplasm forms a
rather thick layer over the inner surface of the cell wall, and
in some part of this layer the nucleus lies embedded. From
the nucleus, and from various parts of the protoplasmic layer,
there pass to the opposite side of the cell thicker or thinner
bands and strings, and gives a figure of the same after Hof-
meister. Prantl has figured Meristem cells of the stem of

PLANT CELLS AND LIVING MATTER. 91

Vicia faba in which filaments of living matter emanating
from the nucleus go to the peripheric layer of living matter,
and also hairs from the epidermis of ovary of Cucurbita, in
some of the compartments of which the reticulum is very dis-
tinctly shown with quite low power (x 100).

Heitzmann, the discoverer of the reticulum of living matter
and of its continuity throughout the entire animal organism,
states in his magnificent work just published, entitled ‘ Micro-
scopical Morphology of the Animal Body in Health and
Disease,’ p. 57, “ My own limited researches enable me to
assert that the granules of living matter in vegetable protoplasm
are, as a rule, united in the shape of a reticulum, in the same
manner as in animal protoplasm. Besides, the researches of
W. Hassloch elucidate the identity of both animal and vegetable
living matter in a satisfactory manner. I may add that all cells
of the vegetable organisms are uninterruptedly connected by
means of delicate offshoots piercing the walls of the cellulose.
The granules of amylum are transformed living vegetable
matter. The plant in toto is an individual and not composed
of individual cells.” But demonstration of this statement is
wanting. Low powers of the microscope, and even high
powers, show that a less or more thick cloak of cellulose sur-
rounds each plant “ cell,” and separates it from its neighbours.
The observations of the chlorophyll-granules and of the interior
of the polygonal cellulose frames of blades of grass herein de-
tailed, while they fully bear out the assertions of Heitzmann
and Hassloch as to the reticular structure, and perhaps even as
to the growth phases, at least so far as dimension is concerned,
of masses of living matter of plants, do not advance our know-
ledge much further. All my endeavours definitely to determine
whether the plant “cells” are interconnected or not were
unsuccessful with the means I employed in both transparent
specimens and in sections. The inspection, under all sorts of
circumstances, of the wall of cellulose, although it frequently
gave me the impression that it was faintly granulated, and
although delicate filaments emanating from the most peripheral
chlorophyll-granules were often seen tending towards the wall,

92 DR. LOUIS ELSBERG.

did not enable me to arrive at a conclusion concerning its
intimate structure.

Francis Darwin has discovered protoplasmic filaments pro-
truding from the cellulose investment of the glandular hairs
on the leaves of Dipsacus sylvestris (‘ Quarterly Journal of
Microscopical Science,’ 1877, p. 245). Previously, Hoffman
(“‘ Ueber contractile Gefilde bei Blatterschwammen,” ‘ Botan.
Zeitung,’ 1853, p. 857,and 1859, p.214) had described contractile
filaments projecting from cell walls in Amanita (Agaricus)
muscaria, and although De Bary has expressed the opinion
that these are not protoplasmic, Darwin believes them to be
so (‘ Quart. Journ. Mic. Sc.,’ Jan., 1878, p. 74). Later, W. J.
Beal (‘ American Naturalist,’ October, 1878, p. 643) described
threads, but does not say that they are protoplasmic, project-
ing from the end of hairs of several plants. Darwin has
observed filaments of living matter, emanating from the in-
terior of plant cells, pierce the cellulose frame. They pro-
truded from terminal cells only, and of course showed no
interconnection between neighbouring cells. Such intercon-
nection I can now demonstrate.

My first successful observations were made in specimens of
the flowers of flowering flax (Norimbergia gracilis), and of
the leaf and stem of the common india-rubber plant (Ficus
elastica), and were obtained as follows. The analogy between
epidermal layers, as well as other parts of a plant, and animal
epithelia, led me to the inference that reagents successfully
applied for elucidating the structure of animal epithelia might
serve for the same purpose in plants. Now, each epithelial
body is a nucleated, reticulated bioplasson mass, enclosed by a
continuous layer of bioplasson and separated from all its
neighbours by a cloak of cement-substance. ‘The cement-sub-
stance answers to the cellulose wall of plant cells, and as a
memento of Schleiden and his cell doctrine, I would advocate
not only the retention of the term cellulose, but its extension
to animal tissues, i.e. to take the place of the term cement-
substance. It is known to histologists that the cement-sub-
stance is traversed by numerous conical filaments which by

PLANT CELLS AND LIVING MATTER. 95

their discoverer, Max Schultze, were termed “thorns or
prickles.” It is also known that upon applying a 2 per cent.
solution of silver nitrate to fresh epithelia, the cement-sub-
stance assumes a dark brown hue, and appears perforated by
numerous light transverse lines; while if, on the contrary, a
one half per cent. solution of gold chloride be applied to epithe-
lium, the bioplasson reticulum in its interior assumes a dark
violet tint, the cement substance remains unstained, and in it
Max Schultze’s thorns, also coloured deep violet, appear very
plainly. Thus it has been proved that the wall of cement-
substance does not completely isolate the single epithelia, but
is pierced by bridges of living matter which interconnect all
epithelia into one continuous bioplasson mass.

I placed pieces of the flower of ‘‘ Norimbergia ” into a 2 per
cent. solution of silver nitrate for about half an hour, then
washed the specimens with distilled water and exposed them to
daylight. I found that nitrate of silver does not invariably
affect the cellulose alone, but sometimes stains also the “‘cell”-
contents ; a corresponding general tinction occasionally hap-
pens in the case of animal epithelia. Frequently, while the
cellulose wall on the inner surface of the flower was compara-
tively little coloured by the silver salt and dark granular pre-
cipitates filled the spaces between the radiating cellulose off-
shoots, the polygonal frames on the outer surface of the flower
were beautifully stained dark brown by the silver salt; and
examined with Tolle’s immersion lens, showed numerous
interruptions in their continuity, as represented in fig. 2,
exactly like the light-coloured transverse markings seen in
cement-substance of animal epithelia under similar circum-
stances. Usually the hairs were stained deeply brown; in
many compartments one or several light fields were seen, of
irregular shapes, freely branching; the periphery of such a
light-coloured field often looked serrated, and a reticulum pro-
ceeding from it pervaded the whole compartment. This
appearance is shown in fig. 3. In a number of instances I
observed that the septum separating two neighbouring com-
partments was marked by light-coloured lines, as represented in

94 DR. LOUIS ELSBERG.

fig. 3. The branching light fields were the smaller the nearer

Fiat,

Fie. 2.—Cells from the flower of No- Fie. 3.—Hair of flower of Norim-
rimbergia, stained with nitrate of bergia, stained with nitrate of
silver. silver.

the compartment was to the apex of the hair; at the end, the
whole hair, as a rule, appeared uniformly dark brown, or con-
tained in its interior an extremely delicate, light-coloured
reticulum only.

After a one half per cent. solution of gold chloride had been
brought to bear upon pieces of the flower for about forty
minutes, the wall of cellulose became more distinct although
not coloured by the gold salt. In the interior of the polygonal
fields, on the inner surface, a scalloped body had made its
appearance ; it was slightly retracted from the cellulose frame
and offshoots, bordered by a continuous delicate layer, and
filled with a very distinct reticulum in connection with a
central coarsely granular and also reticulated nucleus. The
bordering layer and the reticulum around the nucleus, as well
as the nuclear wall and the intranuclear granules and reti-
culum, were of a dark violet colour, just as in animal epithelia

PLANT CELLS AND LIVING MATTER. 95

(see fig. 4). On the outer surface the epidermal bodies

Fig &

Fig. 4.—Cells from flower of Norimbergia, stained with gold
chloride.

exhibited a distinctly reticular structure. The hairs showed
dark violet granules and clusters of granules in the interior of
the compartments; these granules had radiating offshoots
which formed a network, with frequently distinctly granular
thickened points of intersection, as represented in fig. 5.
There could be no doubt that this was the positive image of
the structure that was demonstrated by the silver staining in a
negative manner as depicted in fig. 3. In some, especially in
small hairs, the dark violet reticulum in the compartment was
very dense. Frequently, delicate violet filaments pierced the
transverse septa of neighbouring compartments and intercon-
nected the reticula and bioplasson formations in their interiors,
as seen in fig. 5.

But the most complete proof of the existence of living matter
within the cellulose walls of plant “cells”? I obtained in
sections of the stems of leaves of the common india-rubber
plant (Ficus elastica), a silver-stained specimen of which

96 DR. LOUIS ELSBERG.

is represented in fig. 6. The latex oozing out of the stem
proved to be composed of a viscid, as if mucous, colour-

Fic. 5.—Hair of flower of Norimbergia, stained with gold chloride.

less liquid, in which were suspended innumerable isolated
granules of a high lustre, somewhat similar to that of fat ;
gold chloride staining made the smallest granules appear dark
violet, while the larger were only indistinctly coloured, re-
taining their high lustre. ‘Transverse sections of the stem,
examined in dilute glycerine, showed chlorophyll-granules and
the reticular structure. The parenchyma of some specimens,
especially those treated with strong alcohol, plainly exhibited
the layer of living matter in the interior of the “ cell,” which
Von Mohl called “‘ Primordial utricle,” and sacs, more correctly
“protoplasmic sac ;” and in many cases the bioplasson mass
showed the reticular structure. Treatment of gold chloride
not only rendered the network of many bioplasson bodies
distinctly visible, but in some cases offshoots emanating
from such bodies were seen to penetrate more or less far into
the cellulose investment; what has been sometimes de-

PLANT CELLS AND LIVING MATTER. 97

scribed by authors, especially in growing tissues, as “ inter-
cellular spores ” and ‘ middle lamelle,” in the cellulose were

Tig 6.

ca, treated with silver nitrate.

revealed to be in a number of instances accumulations and fila-
ments of living matter wedged in between the “ plant cells,”
very much like the wedges of bioplasson and the medullary
elements which I have found to grow between animal epi-
thelia in cases of new growths (‘‘ Microscopical Study of
Papilloma of the Larynx,’’ ‘ Archives of Laryngology,’ March,
1880). Treatment with the solution of silver nitrate revealed
in the darkened substance of the cellulose light spaces occu-
pying the position of such wedges. These light spaces sent
off comparatively broad offshoots parallel to the inner surfaces
of the cellulose frame, and innumerable delicate light off-
shoots from both the central space and the broad offshoots
traversed the brown cellulose in uninterrupted connec-
tion with the delicate light reticulum seen here and there
within the so-called ‘plant cell.” The appearance of the
silver-stained cellulose frame in a portion of such a specimen
is accurately reproduced in fig. 6, and the results obtained in

VOL. XXIII.—NEW SER, G

98 DR. LOUIS ELSBERG.

these specimens I have verified by very numerous other ex-
aminations.

My researches demonstrate, and so far as I know, demon-
strate for the first time, that the frame of cellulose, analogously
to the cement substance of animal epithelia and the basis
substance of other animal tissues, is pierced by either single
filaments of living matter or a reticulum with more or less
large accumulations of living matter, interconnecting all
neighbouring tissue elements, and that the plant, therefore,
like the animal, is one continuous mass of living matter, with
interspaces which contain some non-living material.

The structure of plant tissue may be illustrated by the
structure of hyaline cartilage of animals. For many years it
was believed that cartilage consists of a homogeneous non-
living basis substance in which are embedded, at various
distances apart, isolated living cartilage corpuscles—cartilage-
“cells”? as they were called. ‘The more or less convincing
observations made by Heitzmann, and after him by Hertwig,
Thin, Prudden, Spina, and Flesch, have shown this to be a
mistake ; and the results which I obtained in the histological
examination of the cartilages of the larynx (published in the
‘ Archives of Laryngology,’ October, 1881, and January, 1882),
have proved beyond question that hyaline cartilage is a filigree
of living matter, in the meshes of which lumps of basis
substance are embedded. According to the former view car-
tilage could be compared to a pudding, in the dough of which
a certain number of raisins are embedded; in truth, it is like
a framework composed of larger and smaller raisins and bands
and strings of raisin substance, in the meshes or interspaces
of which blocks of dough are embedded.

Just so in the tissue of plants, the so-called plant “ cells”
are connected one with the other, and blocks of cellulose fill
up the interstices in the network of living matter.

Not to trespass too much upon the patience of the reader, I
must leave undetailed here the far-reaching consequences of
the “ bioplasson doctrine ”’ for the better understanding of the
relations and phenomena of plant life.

LIFE HISTORY OF THE LIVER-FLUKE. 99

The Life History of the Liver-Fluke (Fasciola
hepatica).

By

A. P. Thomas, M.A., F.L.S.,
Balliol College, Professor of Natural Science in University College, Auckland,
New Zealand, late Demonstrator in the Anatomical Department, Uni-
versity Museum, Oxford.

With Plates IL and IIL.

Ir has been known from very early times that the liver-rot
of various herbivorous mammals is a parasitic disease due to
the presence of very numerous flukes in the liver of the affected
animals. Amongst our domesticated animals the sheep is by
far the most frequent victim. The fluke disease is always
common in certain districts in England and in many parts of
the world; but in consequence of a succession of wet seasons
there was a serious outbreak of it in the winter 1879-80, and
it is estimated that in the United Kingdom 3,000,000 sheep
were then destroyed by it. Hence special attention was called
to the subject, and the research summarised in the following
paper was undertaken on behalf of the Royal Agricultural
Society of England.

For the sake of convenience the subject is divided into the
following sections:—I. Hisroricat. II. Mernops oF In-
vestIGATIoN. III. LirE-H1sToRY.

I. Historica.

From what was known of the course of development amongst
the digenetic ‘Trematodes, the nearest allies of Fasciola he-

100 A. P. THOMAS.

patica, there was reason to believe that here also an alternation
of generations existed, and that one or more molluscs served as
intermediate host for the asexual forms. Many attempts had
been made to discover the intermediate host by various eminent
biologists, including Professor Leuckart, von Linstow, Ercolani,
&e., but all had proved fruitless, and notwithstanding its im-
portant practical bearing the problem remained unsolved.}
Very many suggestions had been made as to the nature of the
intermediate host. Moulinié? had found in Limax cinereus
and Arion rufus (ater) sporocysts containing cercariz with a
rudimentary tail, and suggested that these might have some
connection with the liver-fluke. I met with this species in
Arion ater early in the course of my investigations, and was
able to disprove experimentally the conjecture that this was
the cercaria of Fasciola hepatica. Willemoes-Suhm? had
drawn attention to the fact that liver-rot was very prevalent in
the Faroe Islands, the molluscan fauna of which was restricted
to eight species, viz. Arion ater, A.cinctus, Limaxagrestis,
L. marginatus, Vitrina pellucida, Hyalina alliaria,
Limneus pereger, and L. truncatulus. Of these Limax
agrestis, our common grey slug, was by far the commonest
and most injurious, and he suggested that this slug might act
as intermediate host. Von Linstow* had mentioned Planorbis
vortex as being possibly the host. Weinland® had found the
liver of L. truncatulus infested with nurse forms. The

1 A statement has been published in several text-books, English and
American, to the effect that Cercaria cystophora inhabiting Planorbis
marginatus is the larva of Fasciola hepatica. This, of course, is erroneous,
and the mistake appears to have been copied from an abstract in the
‘ Zoological Record’ for 1872 of a paper by Willemoes-Suhm. The suggestion
really made in the original paper was that C. cystophora is the larval form
of Distoma lanceolatum. This species is known on the Continent as the
small liver-fluke, and is far less formidable than the larger, F. hepatica, the
true liver-fluke. It appears not to exist in England.

2 «Mémoires de ]’Institut Genevois,’ vol. iii, p. 267.

3 * Zeitschrift fiir wissentschaftliche Zoologie,’ 1873, vol. xxiii, p. 339.

4 * Arch, fiir Naturgeschichte,’ 1875, p 194.

5 Abstract in ‘ Archiv fiir Naturgeschichte,’ 1874, vol. ii, p. 423.

LIFE HISTORY OF THE LIVER-FLUKE. 101

cercariz had the habit of throwing off their tails and crawling
about by the aid of their suckers, and he thought that the
larvee of the liver-fluke might encyst upon plants. Kiichen-
meister had suggested certain slugs as possible hosts of
Fasciola hepatica. On April 7th, 1880, the ‘Times’
published a letter written by Dr. Cobbold to contradict the
statements made by Dr. J. Harley, who denied the existence
of any intermediate host. The letter contained the following
sentence :—‘‘ The investigations of the lamented Willemoes-
Suhm render it almost certain that Cercaria cystophora
infesting Planorbis margiuata is the higher larval state of
the small fluke (Distoma lanceolatum), and the still later
researches of Leuckart point to the mollusc called Lymnea
truncatula as the bearer of the cercarian stage of Fasciola
hepatica.” Seven days later, on April 14th, 1880, the late
Dr. Rolleston gave, in a letter to the ‘Times,’ reasons for
regarding the black slug, Arion ater, as the intermediate host
of Fasciola. Since that time Ercolani! has made a similar
suggestion for certain terrestrial molluscs. He found that larval
trematodes were of far more frequent occurrence in land-
molluscs than had hitherto been supposed, and this circum-
stance, together with the failure of the most illustrious hel-
minthologists to discover the genesis of F. hepatica, led him
to think that it was in this direction that success would even-
tually be obtained.

On the 2nd of the following June the Royal Agricultural
Society of England offered Dr. Rolleston a grant for an inves-
tigation into the life-history of the liver-fluke. Dr. Rolleston
being unable himself to undertake the work, mentioned my
name to the Society, and the research was begun by me on
June 7th.

On Dec. 22nd, 1880, I found ina Limneus truncatulus,
captured on an infected field at Wytham, near Oxford, on the
24th Sept., and since that time kept in an aquarium in the

1 « Tell’ Adattamento delle specie all’ ambiente,” ‘ Memorie dell’ Accademia
delle Scienze dell’ Istituto di Bologna,’ serie iv, tomo ii, 1881, pp. 241, 327.

102 A. P. THOMAS.

laboratory, a cercaria, which has since been proved to be the
larva of the liver fluke.

The reasons which led me to regard this as the cercaria of
Fasciola hepatica need not be explained here, as they are
given in another part of this paper. Ina report in the Royal
Agricultural Society’s ‘Journal’ for April, 1881, I described
this cercaria as a new species, and at the end of the description
wrote as follows :—‘‘ The structure and habits of this cercaria
render it possible that it may prove to be the larva of Fasciola
hepatica, but want of material has prevented my testing the
question by giving the cysts to rabbits. I intend, however, to
pursue this case further.”

On Dee. 12th, 1881, a paper appeared in the ‘ Zoologischer
Anzeiger,’ in which Professor Leuckart announced that he had
succeeded in infecting young specimens of L. pereger, but had
been unable to obtain the development of the expected cercaria.
He also made it known, for the first time, that the statement
made by Dr. Cobbold in the ‘Times’ for April 7th, 1880, was
founded on a private letter from himself, and that the announce-
ment that his researches pointed to L. truncatulus had
proved to be premature and incorrect, for on further examina-
tion of his snails he had found them to be not L. truncatulus
but L. pereger juv.

In the first number of the ‘ Archiv fiir Naturgeschichte’ for
1882, which, however, bears no further date, the same results
were given at greater length. Professor Leuckart stated that in a
number of specimens of L. truncatulus sent him by a friend
he had found three kinds of rediz. One of these contained
tailless distomes, which, he held, probably belonged. to the
developmental cycle of the liver-fluke. He considered this
supposition to be entirely justified until further results were
obtained. A-second form was not absolutely excluded from all
connection with the liver-fluke, but no such statement could be
made with respect to the third form. But in this third form I
at once recognised the cercaria found by me at Wytham, of
which a description had been published some eight months
earlier.

x q

LIFR HISTORY OF THE LIVER-FLUKE. 103

During the summer of 1882 I at length obtained L. trunca-
tulus for my experiments, and succeeded in infecting them with
the embryos of F. hepatica. Before the end of August the
development of the species was obtained up to the time when
the tailed cercariz were nearly mature, and, as I had by me
well-preserved specimens of the rediz and cercariz found at
Wytham, I was able to compare the two forms step by step,
and see that they were identical. A paper was drawn up for
the October number of the ‘ Journal of the Royal Agricultural
Society,’ giving these results, and was sent to the printer on
Sept. lst. A fortnight later it received a revision, which was
merely verbal, and the whole of the copies were printed off by
Oct, 2nd.

Separate copies of the paper were received for distribution on
the 24th Oct., but the journal was not published until nearly
the end of the month.

Inthemean time a paper by Professor Leuckart appeared in the
‘ Zoologischer Anzeiger’ for the 9th October. In this he stated
that he too had succeeded in rearing the cercaria of F. hepatica
in L.truncatulus, “the only true intermediate host,” and that
it had proved to be not the tailless form, but, on the contrary,
the third form mentioned above, which he had supposed early
in 1882 could have no connection with the liver-fluke.

On the 19th Oct. a résumé of my completed researches
was published in ‘ Nature,’ attention being called to the fact
that the cercaria of the liver-fluke was the one already dis-
covered by me in Dec., 1880, and described in April, 1881, as
probably belonging to F. hepatica, and that the necessary
proof had been furnished by myself, and also independently by
Professor Leuckart.

Il. Meruops or INVESTIGATION.

On commencing the investigation into the life-history of the
liver-fluke it was felt that where so many different molluscs
had been suggested as possible intermediate hosts, it was neces-
sary to examine the question anew, and not to be guided by

104 A. P. THOMAS.

the numerous conjectures already expressed, some of which
had very scanty evidence to support them. Work was begun
in two directions, first, by endeavouring to infect many of our
commoner molluscs, both terrestrial and fresh-water. In the
second place, numerous localities in the country around Oxford,
in which sheep were known to have incurred liver-rot, were
visited, the whole of the invertebrate fauna was carefully
studied, and many specimens brought home to be dissected
and searched for larval trematodes. Very numerous places were
examined in this way, but mention will only be made here of the
one which proves finally to have given the correct solution of
the problem. At Wytham, near Oxford, was a clearly cireum-
scribed area of infection. The fields, five in number, were
situated on the side of a hill far above the reach of river
floods, lying upon the Oxford clay. They were searched
thoroughly by day and by night, and the various invertebrate
animals found, including snails, slugs, aquatic insect larve,
crustacea, worms, &c., brought home and examined. Fresh-
water snails were very scarce on the ground; all that were
brought to light were two small specimens of Physa fonti-
nalis, a small Cyclas, and Limneus truncatulus in
moderate numbers. The last-named species was found in a
boggy spot in one of the fields. In one of them was discovered,
on the 22nd December, 1880, the peculiar and interesting form
of cercaria to which allusion has already been made. Its most
striking character was due to the presence of very coarsely
granular cells: arranged in lobed masses along each side of the
body. It was very active, but soon came to rest, encysting
itself upon surrounding bodies. The cyst was snowy white,
from the presence in its substance of the highly refractive
granules already seen in the granular cells forming the lateral
masses, which were thus shown to be cystogenous organs.

The other points in the structure of the cercaria were all
favorable to the supposition that I had here discovered the
long-sought cercaria of the liver-fluke. I had already found
in a lamb’s liver exceedingly minute flukes, smaller than any
yet recorded, one of them being only 1:1 mm. in length, i.e.

LIFE HISTORY OF THE LIVER-FLUKE. 105

ouly 1-30th part as long as the adult, and these immature
forms gave me valuable information as to the structures to be
expected in the cercaria, the relative size of its suckers, &c. It
was usually supposed that the sheep when grazing picked up the
parasites whilst they were still within the snail. This view
was upheld by so eminent an authority as Professor Leuckart
in a paper published in the beginning of 1882.! But I had
already collected evidence from independent sources, which
inclined me to the belief that the larva were picked up in
the encysted condition attached to the grass. Hence the pre-
sence of a special cystogenous organ in the cercaria, and its
habit of encysting on grass were highly suspicious. Further,
the suspicions raised by the structure and habits of the
cercaria were increased by the fact that its nurse-form was
the only one found upon the ground, although there was every
reason for expecting to find the larva of the liver-fluke, as an
infected sheep had been seen a few months earlier wandering
over the boggy spot from which the L. truncatulus was ob-
tained. I can testify that the sheep harboured numerous
flukes, for its liver was sent me for examination ; and there
could be no doubt that large quantities of fluke eggs had been
scattered all over the fields. My suspicions were accordingly
expressed in a report to the ‘Royal Agricultural Society’s
Journal’ for April, 1881.

During the summer of 1881 I was anxious to try infective
experiments with Limneus truncatulus, but was unfortu-
nately unable to obtain any specimens; the localities near
Oxford, where I had formerly found it, were searched in vain.
I went out repeatedly in quest of this snail, having on several
occasions the skilled assistance of my friend and colleague,
Mr. W. Hatchett Jackson, but we never found any other trace
of this species than the empty shells. It could not be dis-
covered in the localities given for it by Whiteaves in his
paper on the “ Mollusca inhabiting the Neighbourhood of
Oxford.’ My friends ata distance were appealed to, but were

1 * Archiv fiir Naturgeschichte,’ 1882, p. 80.
2 * Proceedings of the Ashmolean Society,’ 1857.

106 A. P. THOMAS.

unable to assist me. The comparative freedom from rot. of
sheep in the neighbourhood of Oxford last year was probably
due to the real scarcity of this snail.

In 1882, however, there were floods in July, and the waters
of the Isis brought down vast multitudes, probably from its
breeding haunts in marshy places up the river. So numerous
was it, that a single sweep of a small hand-net repeatedly
gave me more than 500 examples, and this was in a ditch
where last year not a single L. truncatulus could be found.

On obtaining the snails I had so long been searching for, I
exposed a number to infection by placing fluke embryos in
the vessel with them. The snails were speedily found to have
afforded a suitable place for the further development of the
embryos; indeed, infection was too successful, for very large
numbers of them died simply from exhaustion owing to the
excessive number of parasites they contained.

It may be well to mention here that infection experiments
have been tried upon the following species of molluscs, as well
as upon L. truncatulus, viz., Limneus pereger, L. palus-
tris, L. auricularius, L. stagnalis, Physa fontinalis,
Hpdorbie marginatus, P. carinatus, P.vortex, P.
spirorbis, Bythinia tentaculata, Paludina vivipara,
Succinea amphibia, Limax agrestis, L. cinereus, Arion
ater, A. hortensis. None of these could be infected, with
the partial exception of Limneus pereger. With regard
to this last species I can corroborate Professor Leuckart’s
statement that the youngest specimens only of this species can
be infected, and that even here development does not proceed
beyond an early stage.

Although it appears that L. truncatulus is the only Eng-
lish molluse which can serve as intermediate host to the liver-
fluke, it is quite possible that elsewhere some other mollusc of
similar habits may be victimised. L. truncatulus has a very
wide geographical distribution, but so, too, has the liver-fluke ;
and if the latter has the wider distribution, there must of course
be some other intermediate host. | Leuckart! states that

'*Die menschlichen Parasiten,’ p. 531,

LIFE HISTORY OF THE LIVER-FLUKE. 107

Fasciola hepatica is reported from Australia, and a similar
assertion has been published in the ‘ Veterinarian.’ According
to Hutton! and Wallace,’ the genus Limnzus does not exist
in Australia. Hence if both these statements are correct there
must be another intermediate host. The liver-fluke is also
found in North America, where the genus Limnzus occurs
indeed, but not the species L. truncatulus. Sheep-rot is
also found in the Shetland Islands, where, according to Forbes,?
the genus Limnzus is represented by the species L. pereger
alone. It is, however, quite possible that in the last case L.
truncatulus has been overlooked on account of its minute
size.

Ill. Lire-History.

A. First generation—l. Egg.—The eggs of the liver-
fluke occur in very large numbers in the contents of the bile
ducts and gall-bladder of the infected animal. They give a dark
brown colour and sandy appearance to the bile, and in some of
the smaller terminal ducts often form a stiff brown mass, com-
pletely plugging up the lumen. They pass with the bile into
the intestines, and may be found abundantly in the droppings
of animals suffering from liver-rot.

The egg is an oval body, with a smooth, transparent,
yellowish-brown chitinous shell. The average size may be
said to be 0:13 mm. in length by 0:08 mm. in breadth, but
the dimensions vary greatly, the length from 0:105 to 0:145
mm., and the breadth from 0:066 to 0:09 mm. The ante-
rior end is a little more rounded than the posterior, and a
slightly serrated line running around it marks off a circular
segment, forming an operculum 0:028 mm. in diameter.
‘The opposite end is frequently a little thicker, and slightly
roughened.

The number of eggs produced by a single fluke is

1<Transactions ’ of the New Zealand Institute, vol. v, p. 18.

?* Geographical Distribution of Animals,’ vol. i.
3° Brit, Ass. Report,’ 1859, p. 127.

108 A. P. THOMAS.

exceedingly large, and its fecundity has been underrated.
In one case I obtained 7,400,000 from the gall-bladder
of a sheep suffering from the rot, and, as the liver contained
about 200 flukes, this gives an average of 37,000 eggs to each
fluke. And these eggs were found in the gall-bladder alone ;
the liver must have contained at least as many more, and eggs
had been passed copiously by the sheep for several months.
The number of eggs produced by a single fluke may be safely
estimated at several hundred thousands.

When first formed, the egg includes a single germinal cell,
supplied by the germarium, and fertilised by a spermatozoon,
and a considerable number of secondary yolk-cells supplied by
the vitellaria, which serve as food to the growing embryo.
Segmentation of the ovum takes place during the descent of
the egg through the oviduct, but no further development takes
place so long as the egg remains with the body of the host.
Fig. 1 on Plate II represents a fluke egg in the condition in
which it is found within the bile ducts of the sheep; the
embryo is represented by a pale spherical mass of delicate
nucleated cells, and is situated near the opercular end of the
shell. It is surrounded by the secondary yolk-cells, which are
filled with refractive spherules, both large and small, so that
the examination of the embryo is rendered very difficult.

The further development of the embryo can only take place
out of the body of the bearer of the adult fluke and at a lower
temperature. Eggs kept in an incubator, at the temperature of
the mammalian body, do not make any progress, whilst the
eggs kept at a lower temperature complete their development
in a few weeks. The conditions necessary for development
are moisture and a certain moderate degree of warmth. Light
I have found to exert no influence; eggs taken from the gall-
bladder and placed directly with water into an opaque vessel,
develop as soon as similar eggs exposed to light but otherwise
kept under the same conditions. A temperature of about 23°
C. to 26° C. is most favorable, and with this degree of
warmth the embryo is formed in about two or three weeks.
At a lower temperature development takes place much more

LIFE HISTORY OF THE LIVER-FLUKE. 109

slowly, and with an average warmth of 16° C. occupies two or
three months. During the winter no progress is made unless
artificial heat is supplied. |

All the eggs under the same conditions, however, do not
produce embryos in the same time, a certain number are hatched
out on every successive day for some weeks or even months,
and at the end of this time some of the eggs may remain in
the same condition as when just taken from the liver. No
explanation can be discovered in the eggs themselves of the
very variable time required for the development of the embryo,
but the fact is of much practical importance, for eggs scattered
over any damp ground may render it dangerous for a long period.

The granular character of the secondary yolk-cells render it
very difficult to follow in detail the growth of the embryo
whilst still within the egg, and as the matter is one of theo-
retical importance only, the examination of the development
by more elaborate methods has been postponed in favour of
matter of more practical interest. I hope, however, to have
the opportunity before long of observing the formation of
the layers in the embryo. All that can be seen in the egg by
direct examination is as follows. The embryo increases in
size, being nourished by the absorption of the secondary yolk.
The outlines of the yolk-spheres become more distinct and
the granules less numerous, whilst some of them appear to
coalesce or disintegrate (fig. 2, Pl. II). Within them the
outline of the embryo is visible, often showing one or more
annular constrictions. As the yolk-cells are gradually used up
the body of the embryo becomes larger and more plainly
visible, and comes to occupy the whole length of the egg (v.
fig. 3). Its surface is somewhat bossy owing to the projection
of the cells forming the outermost layer of the body. A papilla
appears at the anterior end, which is always directed towards
the opercular pole of the shell, and a little way behind a
quantity of dark brown pigment is produced, giving rise to a
double eye-spot, while the surface loses its bossy appearance.
Wave-like peristaltic contractions pass along the body from the
anterior towards the opposite end.

110 A. P. THOMAS.

In the last stage, when the embryo is ready to emerge from
the. shell, it lies slightly curved upon itself at one side of the
egg (v. fig. 4), the remainder of the space being occupied by
the fluid remains and refuse of the yolk-spheres. At the
anterior end, just beneath the operculum, is a quantity of
viscid mucus, which forms a sort of lining or cushion against
which the head-end of the embryo is pressed. Around the
body of the embryo may be distinguished a bright border,
which is formed by the cilia covering its surface ; these cilia,
however, can only in exceptional cases be seen in motion before
the animal quits the egg.

2. Free Embryo.—The embryo is now ready to come forth;
its movements become more marked, and at length a vigorous
extension of the body causes the operculum to fly open, asif
moved by a spring. The cushion of mucus pours out, the
embryo thrusts the fore part of its body out of the shell, the
cilia begin to move instantly the water touches them, and the
animal, after a short struggle, succeeds in drawing the whole
of its body through the narrow opening of the shell, and glides
away with ease and rapidity through the water. Although
light has no influence in accelerating development, the embryo
itself is very sensitive to it. Thus they congregate especially
on the light side of a vessel containing them; and I have
repeatedly observed that, although on removing a vessel of
egos from the darkened incubator in which they were being
hatched, not more than two or three embryos could be seen,
yet after it had stood in a window for twenty minutes the
water was quite nebulous from their presence.

The form of the free-swimming embryo is an elongated cone,
with rounded apex (fig. 5, Plate II), its average length 0°13
mm., its breadth at the anterior end 0°027. The broader end
or base of the cone is directed forwards, and in its centre is a
short retractile head-papilla. The whole of the surface, with
the exception of the head-papilla, is covered with long cilia,
which are borne by an outer layer of flattened ectoderm cells.
These cells are arranged around the body in transverse rings,
usually five in number, though occasionally six may be

LIFE HISTORY OF THE LIVER-FLUKE. ca

counted ; they vary in length from -025—0°35 mm.; each has
a very small nucleus, 0:003 mm. in diameter. The cilia are of
the same length (‘012 mm.) over the whole of the surface, but
on the cells of the anterior ring they are more numerous, and
hence more conspicuous. This first row is.composed of four
or sometimes five cells, arranged round the papilla, and these
are thicker than the other cells belonging to the same layer,
often forming ear-like projections at the side of the embryo,
and resembling epaulets. ‘The second ring contains five or six
cells, the next two rows, each four, as a rule, whilst the last
ring is formed by two cells only. In the last two rings the
cells are of greater length than in the others. Seen in a
surface view the cells of this outer layer are polygonal and
sometimes hexagonal. They overlap one another at their
edges, and it is probably owing to this fact that the outlines
appear double in silver nitrate preparations. In the small
number of examples possessing six rows of cells, the second
and third rows are formed by smaller cells.

Beneath the ciliated cells the body wall is formed by a
granular layer, the cellular nature of which is not easily made
out. In favorable preparations, however, nucleated cells can
be seen slightly projecting on the inner surface. In the outer
parts of the layer are situated both transverse and longitudinal
muscle-fibres. ‘The longitudinal are more feebly developed than
the transverse, and are only seen with difficulty. The double
eye-spot belongs to this deeper layer; it has been figured as
having a form of the sign of multiplication. This, however, is
not the case, for it is really double, and has commonly the form
of two crescentic masses of dark pigment, placed with their
convex sides turned towards each other, and in contact near
the anterior horns. On closer examination it is seen that each
eye-spot is composed of a cell in which the pigment is arranged
at one side in a crescent, the hollow of which is filled up by
refractive material which will act as a rudimentary Jens. The
body wall also contains numerous yellowish refractive granules,
especially just behind the eye-spots, and to it belong the two
ciliated funnel-shaped spaces of the excretory system. These

J12 A. P. THOMAS.

are situated one on either side of the middle of the body, in
each is a large cilium carried by a nucleated cell, and usually
directed forwards. The cilium is connected with the cell by a
disc at its base, it is tongue- or flame-shaped, and is con-
stantly in motion, waves passing along the cilium from the
base to the tip, and hence towards the apex of the rather
narrow infundibulum. Just behind the head-papilla is a
granular mass, which reacts with staining fluids differently from
the adjacent tissues. This, from comparison with other tre-
matode embryos, would seem to be a rudimentary digestive
tract. Behind, the rest of the body cavity is occupied by
delicate round nucleated cells—the germinal cells.

The embryo is exceedingly active, and with head-papilla
retracted swims swiftly and restlessly through the water, not
unlike some of the larger infusoria, though more rapidly.
Sometimes it goes directly forwards, and then rotates on its
longitudinal axis, just turning a little from side to side, as if
searching for something. At other times, by curving its body,
it sweeps round in circles, or, curving itself still more strongly,
spins round and round without moving from the spot. When
the embryo, in moving through the water, comes in contact
with any object, it pauses for a moment, and feels about as if
trying to test its nature, and, if not satisfied, darts off hastily
again. Butif the object be a Limneus truncatulus it at
once begins to bore. Prof. Leuckart has said of the head-
papilla of the embryo, that “it seems to have the function of
a tactile organ.’ But I have no doubt that it has the function
assigned to it in my former papers, viz. that it isa boring-
organ.! The papilla is ordinarily short (about ‘006 m. in
length), and the end is quite blunt, or may have a slight
depression in the middle. A differentiation in the tissue of
the head-papilla is visible in the form of a delicate rod-shaped
structure, occuping the axis, certainly not distinct enough to
be called a spine, though the papilla seems to possess consider-
able rigidity. It is particularly evident in preparations of
embryos killed with osmic acid and stained with picro-carmine.

1 «Roy. Agric. Soc. Journ.,’ 1881, p. 7.

LIFE HISTORY OF THE LIVER-FLUKE. Uns

As soon as the embryo begins to bore the head-papilla
becomes longer, conical, and pointed. The embryospins round
on its axis, the cilia working vigorously and pressing the
embryo against the surface of the snail. This pressure is in-
creased by the body of the embryo being alternately drawn up
and then suddenly extended. As the papilla sinks further into
the tissues of the snail, it becomes longer and longer until it
may reach five times its original length (Plate II, fig. 6), and
the tissues of the snail are forced apart, as if by a wedge,
leaving a gap through which the embryo squeezes its way into
the snail.

The embryo appears to exert an instinctive choice in select-
ing the host into which it enters. It is conceivable that the
tissues of Limneeus truncatulus are softer than those of
other molluscs, and that the embryo is able to bore its way
into the former whilst it cannot do so into the latter. But
from the greater eagerness which the embryo exhibits when
placed ona slide with Limnezus truncatulus I do not think
that this is the correct explanation. Moreover, the tissues of
such snails as young specimens of Physa fontinalis or Lim-
naeus palustris appear to be quite as soft as those of L.
truncatulus, and yet if a quantity of embryos are placed in a
vessel containing equal sized examples of the three species just
named, it is found, on subsequent examination, that whereas
each L. truncatulus may contain fifty or more intruders, the
other snails are quite free from them. The most probable
explanation seems to be that there is some difference in the
nature of the secretion of the surface of the body in these
snails, which is sufficient to serve as a guide to the instinct of
the embryos.

But, although the embryo instinctively chooses the snail in
which its further development is possible, it does not always
make an equally happy selection of the part of the snail into
which it enters. I have found as many as a dozen embryos
embedded in the substance of the foot of a L. truncatulus,
such a place of course is unfavorable to further development,
but they may remain alive there for two or three days. Once

VOL. XXIII,—NEW SER, H

114 A. P. THOMAS.

only I found a full-grown sporocyst in the foot of a snail; the
survival of this one in so unsuitable a place was probably
owing to its having accidently forced its way into a connective
tissue space or into a venous sinus. The most natural situa-
tion for the development of the embryo seems to be the pul-
monary chamber, and this organ is, of course, from its position
and the thinness of its walls, most easily accessible to the
embryo. Other embryos, however, may be found in the body
cavity of the snail.

The average maximum duration of the embryo’s free and
active life in water is only about eight hours, though occasionally
one may live over night. During the last portion of the time its
movements become slower, and it will then in desperation often
endeavour to bore into any object which presents itself, even
into its own empty egg-shell. If an embryo has not succeeded
in finding a host, its motion becomes gradually feebler, and at
length ceasing the body assumes an oval or elliptical shape ;
the outer ciliated cells absorb water and swell up into round
vesicles, and the whole body disintegrates. Ina feebly alkaline
solution of peptone I have kept them alive for three days.
The cilia were not lost until the third day, though their
motion became very sluggish ; the embryos increased a little in
size and remained alive even after a number of the ciliated cells
were detached.

3. Sporocyst.—Arrived within the suitable snail the embryo
undergoes a metamorphosis, loses its organs of locomotion, and
degenerates into an inactive sporocyst. The outer layer of cili-
ated cells is lost, whilst the embryo changes in form. The
ciliated cells absorb water and appear as round or hemispherical
vesicles with the cilia standing out perpendicularly from their
surface (fig. 7). During the metamorphosis embryos may
have various irregular shapes, but sometimes retain a less
elongated conical form, even after they have lost the ciliated
cells. The conical form is, however, soon lost, and the embryos
take an elliptical shape such as is shown in fig. 8. The eye-
spots of the embryo become detached from one another and
lose their crescentic form; but they, as well as the head-

LIFE HISTORY OF THE LIVER-FLUKR. 1d

papilla, persist, showing the identity of the young sporocyst
with the embryo of the liver-fluke. After the change in form
has taken place the length is only about ‘07 mm. The rudi-
mentary digestive tract remains for a time, but later on is no
longer distinguishable. The growth of the various larval forms
of trematodes depends very much on the temperature to which
they are exposed. During warm summer weather it is very
rapid, and in the case of Fasciola hepatica, the sporocyst .
may reach its full size before the end of a fortnight; in autumn
development to the same stage takes a period of double the
length. The sporocyst commonly preserves the elliptical
shape untilit reaches the length of ‘15 mm., after this time the
growth is most rapid in the longitudinal direction and the
form becomes sac-shaped. The contents of the sporocyst are
formed by a number of very clear rounded cells, some of which
are the germinal cells of the embryo or cells derived from
them by division, others are formed by a proliferation of the
epithelium lining the cavity of the sporocyst. If the sporo-
cyst be contracted, these cells seem to fill up the whole of the
space, and the cells which are still attached to the body wall,
and form part of its inner surface, cannot be properly distin-
guished from those which are lying free. But if a sporocyst
be chosen for examination which is not in a state of contrac-
tion, cells of various sizes, with very large nuclei, may be séen
projecting here and there from the inner surface ; sometimes in
a single layer, at other times in rounded heaps, two or three
cells deep. It is very difficult to follow the earliest stages in
the formation of the spores within the sporocyst, but by the time
the sporocyst has reached the size of ‘2 mm., there are always
indications that the contents are becoming arranged in separate
balls of cells—the germs of the next generation.

The sporocyst continues to increase in size, and ultimately
reaches the length of -5—°7 mm. (Plate Iil, fig. 10). On the
outer surface is a structureless cuticle, and beneath this is the
thin layer in which the external circular and internal longitu-
dinal muscle-fibres are often the only structural elements which
can be distinctly observed; but in some cases, though not in

116 A. P. THOMAS,

all, there is visible beneath the cuticle a finely granular layer
in which the muscle-fibres appear to be embedded.

It appears probable that some of the most superficial cells of
the body, or at least portions of them, are converted into
muscle-fibres, whilst others undergo more or less degeneration.
These muscle-fibres are more feebly developed in the sporo-
cysts than in the rediz which form the next asexual genera-
tion, and in accordance with this feeble development the sporo-
cysts are exceedingly inert, and rarely show any movements.
In the redia active movements are necessary in order that the
digestive tract, which is present, may be filled with food, and
the muscular system accordingly reaches a greater development.
In the sporocyst the digestive tract is altogether rudimentary,
and as no exertion is required to procure the nourishment, the
degeneration of the muscle-fibres has to some extent followed
that of the enteron. In those sporocysts, however, in which
the power of performing active movements is useful for some
other reason, the muscles may retain a high degree of develop-
ment; as is, for instance, the case in the sporocysts of Cer-
‘ caria limacis, which, as soon as their included cercarie are
sufficiently matured for transference to the ultimate host, bore
their way out, through the thick integument of the slugs (Arion
ater and Limax cinereus), which serve as intermediate hosts,
and are then left behind in the mucous track ofthe slug.

Immediately following the layer of muscle-fibres is an epi-
thelium, which lines the cavity of the sporocyst, and forms the
greater part of the thickness of the body wall. It is composed
of cells of very various sizes, round or polygonal in form, and
containing large nuclei (fig. 11). The layer in most places is
only one cell deep, though adjacent cells may overlap one
another ; but in places, and especially in the less mature sporo-
cysts, itis two or three cells deep. The excretory system is
lodged in the body wall of the sporocyst; on each side may be
distinguished an irregular group of about half-a-dozen ciliated
infundibula. They have the same structure as the two described
above as being present in the embryo, and are always found in

the middle third of the length of the body. No clearly defined

LIFE HISTORY OF THE LIVER-FLUKE. Tt7

or regular canals can be distinguished, but the ciliated infundi-
bula appear to communicate with an extensive system of
irregular lacune between the cells of the body wall.
Numerous yellowish refractive granules occur in_ the
tissues of the sporocyst, or in the body cavity; they are found
within the cells, but are especially numerous between them, or
on their surface. They are also present in large numbers in
the lacunz, and there frequently exhibit molecular motion,
thus showing that the lacune contain a fluid of some tenuity ;
occasionally a whole group of them may be seen to move en
masse, or by careful pressure on the cover-glass may be made
to travel for a short distance along the lacunar passages. No
external opening of this system of passages can be seen, nor
any communication with the body cavity be clearly proved.
In the ciliated infundibula, however, there is present in the
lower wall of the space and close to the base of the flame-
shaped cilium, an elliptical structure, closely resembling that
which Fraipont has described in the ciliated infundibula of
certain other trematodes, as an opening into the body cavity.
This interpretation has more recently been called in question,
but whatever the real nature of this structure may be, and it
is difficult to see what else it can be, there can be no doubt
of its presence in the sporocysts. The yellowish granules
described appear to be excretory products formed within the
cells of the sporocyst and then ejected. They are partially
soluble in acids, leaving an organic basis.

Wagener’ found in the sporocysts of Cercaria macrocerca
(the larva of Distoma cygnoides) vibratile lobules (Flim-
merlappchen) or ciliated spots, which he did not describe in
detail. Thiry* described in the same sporocysts a system of
vessels, the branches of which had ciliated terminations open-
ing into the body-cavity. The vessels, however, were very
pale and difficult to follow, and only in one instance, where the
animal was exceptionally transparent, did he plainly see the
whole system with its branches. The ciliated ends, on the

' ‘Beitrage z. Entw. d. Kingeweidewiirmer,’ p. 65.
* ‘Zeitschrift fiir wissentschaftliche Zoologie,’ x, p. 272.

118 A. P. THOMAS.

other hand, were present in almost all fairly developed ex-
amples ; in form they most closely resembled the ciliated open-
ings in Clepsine complanata, as figured by Leydig. The
vessel was opened on one side and expanded into a two-horned
lobe, covered on its inner surface with cilia. In short, the
ciliated opening was described as similar to the inner ends of
the segmental organs of various annelids. I have never found
the sporocysts which Thiry studied, but it seems improbable
that there should be any great difference in the structure of the
ciliated ends in question in the various species of sporocysts.
The isolated cilia really present are very large and their motion
peculiar, so that it is not difficult to understand that any one
who had before his mind the segmental organ of the earthworm,
and was not prepared to see a large isolated cilium within an
infundibulum, might take the waves passing along the cilium
for waves travelling over a series of small cilia. There can be
no doubt, therefore, that the ciliated infundibula have essen-
tially the same structure in the asexual generations (sporocysts
and rediz) as in the adult sexual trematodes.

Amongst the digenetic trematodes the reproduction of sporo-
cysts by sporocysts takes place, either by transverse fission,
which may be continued through several generations, as in the
case of Cercaria limacis, or by the formation of sporocysts
within the parent, or both methods may occur in the same
species (e. g. Cercaria chlorotica, &c.). But the only way in
which the sporocysts are multiplied in the case of Fasciola
hepatica is by transverse division, and this is of far less
frequent occurrence than in some other species of trematodes.
There appears to be a great and invariable increase of the
nurse-forms amongst the Distomide, and in Fasciola
hepatica the multiplication is effected by the production of
numerous broods of the more highly organised rediz. Fission
does, however, sometimes occur, and usually at an early stage
in the growth of the original sporocyst. A constriction
appears about the middle of the body, and becomes deeper
and deeper (fig. 9), and finally the two halves are completely
severed. One of these contains the remains of the head-

LIFE HISTORY OF THE LIVER FLUKE. 119

papilla and the two separated eye-spots, whilst the other is, of
course, without the signs of any such structures. Hence sporo-
cysts can be found, which even at an early stage show no trace
of head-papilla or eye-spots; and in the majority of adult
sporocysts these organs have degenerated entirely.

B. 1. Development of Redia withinS porocyst.—lIt has
been mentioned above that the germinal cells which give rise to
the redize are in part already present in the embryo, but that
they gain an increase in their numbers by the proliferation of
cells lining the body cavity. The earliest stages in the develop-
ment of the spores cannot be so well distinguished in the sporo-
cyst as in the redia. The cells within a sporocyst having the
length of about °2 mm., begin to show an arrangement into
rounded masses or solid morule. One side of the morula then
becomes flattened, and the cells here then appear to be invagi-
nated, producing a gastrula, whilst the surface becomes smooth,
and its outline first round and then oval. The cells forming the
opposite sides of the archenteron are in contact, so that there is
no archenteric cavity. Each spore may now be seen to be sur-
rounded by a delicate membrane, and as it increases in size its
form becomes more nearly quadrate. At one end a number of
cells are separated to form a spherical pharynx leading into the
blind digestive tract, which now extends a little beyond the
middle of the body. A little behind the pharynx, the body
shows a slightly raised annular ridge, whilst more posteriorly
two short blunt processes are formed. Germs, as described
above, and in various stages of development, are found in each
mature sporocyst ; there is usually one redia (or less frequently
two), nearly ready to leave the sporocyst, with two or three
germs of medium size, and several small ones. Owing to the
varying size and shape of the included germs, the sporocysts
have frequently a very irregular outline.

As soon as the redia is ready to issue from the sporocyst,
which is usually the case by the time it has reached the length
of :°26 mm., it shows active movements, which increase in
strength until at length it succeeds in rupturing the wall of the
sporocyst, and as this state of contraction is continued, the

120 A. P. THOMAS.

wound produced by the forcible exit of the redia is kept closed
until it has healed up. Meanwhile the development of the
remaining germs proceeds. Many of the nurse-forms of
trematodes are known to possess a special birth opening for
the escape of the brood, and even amongst the sporocysts such
an opening is present in those possessing a filiform shape; and
I have myself observed this structure in the sporocysts of
Cercaria gracilis. But in F. hepatica no such definite
opening can be detected in the sporocyst.

2. The Free Redia.—The sporocysts of the liver-fluke are
found in the pulmonary chamber of the snail, or less abundantly
in the body cavity. But the free rediz force their way through
the tissues of the host, and wander into the other organs, and
especially into the liver. They are usually found, with the
digestive tract quite yellow from the remains of the snail’s
liver-cells, with which it is filled. In thus forcing their way
through the tissues they necessarily inflict much injury on
their host, so much, in fact, that comparatively few snails survive
three weeks from artificial infection, and the majority, even of
these, die before the time when the cercariz are completely
mature. Thus, in the laboratory at any rate, the fluke-disease
is more fatal to the snail than it is subsequently to the sheep.

The redia increases in size until it may reach the length
of 13 mm. to 16 mm. It has an elongated cylindrical
form (Pl. III, figs. 12, 13), and at a little distance behind the
pharynx there is present an annular ridge or collar projecting
from the surface, the use of which will be explained below.
From this ring or collar the body tapers gently towards the
anterior end, which is abruptly truncated, and includes in its
centre the mouth. Behind the collar the body becomes a little
narrower, but then swells out gently again until it reaches the
middle of its length, from which point it tapers, at first almost
insensibly, and then more rapidly, the extremity being conical
with rounded apex. At a distance from the posterior end,
equal to about one fourth of the total length of the body, are
situated two short and bluntly conical processes, which serve
as rudimentary feet, and are no doubt of much service in
steadying the redia and preventing it from slipping backwards

LIFE HISTORY OF THE LIVER-FLUKE. 12a

whilst wandering through the tissues of the host. They are
not situated on opposite sides of the body, but are close
together on the same surface, and their bases may even be
connected by a low transverse ridge. They are directed out-
wards and somewhat backwards, their axes being usually
inclined at an angle equal to or rather less than a right angle.

The body-wall has a similar structure in both redia and
sporocyst, so that it will only be necessary to describe the
points in which a difference exists. The muscle-fibres are far
more strongly developed, especially in the anterior part of the
body, so that the rediz show considerable activity as compared
with the sporocyst. When the body is fully extended it may
have a length twice as great as when in a state of contraction.
If an example of the host be chosen, which has a clear and
transparent shell, and has had the greater part of its liver
consumed by the parasites, the rediz may be observed perform-
ing movements of elongation and contraction whilst still within
the living snail.

In the collar or ring mentioned above the muscle-fibres are
strongly marked, and havea peculiar arrangement. The trans-
verse muscle-fibres appear to lie directly under the cuticle, and
are closer together at the sides of the ridge then near its most
convex part. The longitudinal muscle-fibres, however, do not
follow the curve of the surface, but stretch across from one to
the other side of the base of the ridge. Sometimes, when the
ring is strongly marked, the longitudinal muscle-fibres as they
pass forwards may spread out in a fan-shaped fashion before
they are finally inserted in the cuticle (fig. 16). The extent to
which the ring projects above the rest of the surface of the
body varies very greatly according to the size and condition of
the redia, and may be altered from time to time by the con-
traction of the muscle-fibres. It is greatest in those which
show the most active movements, least in those which are the
most passive. The smaller or half-grown rediz commonly
show the greatest activity, and in one of these I have observed
the ring so enormously developed that the diameter of the body
was almost doubled at this point. Those rediz in which fully-

t

122 A. P. THOMAS.

developed cercarie are present are frequently very sluggish in
their behaviour, and the ring may then be relatively incon-
spicuous. In very young rediz the outline of the body appears
to present a slight process on each side anteriorly, and without
the most careful focussing it is often impossible to see that
these are simply the optical expression of the collar, the tissues
of which are still so delicate that the ridge is flattened above,
and therefore, owing to the transparence of its substance, not
readily recognised. The function of the collar is to maintain
the shape of the body and to produce a firm basis upon which
the neck of the redia can be moved. I have observed a redia,
whilst the whole of its body behind the ring was at rest, stretch
forth its neck in such a way as to sweep a considerable area in
front, and thus be enabled to reach conveniently the tissues of
the snail upon which it was browsing. When disturbed the
neck was retracted and the pharynx drawn back close to the
collar. But although the collar has thus a supporting function,
there is no thickening of the structureless cuticle in it, such
as could be termed a definite skeletal structure.

The excretory system is better marked in the redia than in
the sporocyst, and definite canals can be distinguished in the
body-wall. Sinuous longitudinal vessels, one on each side, have
been described in the rediz of several other Trematodes, and

1 Diesing (‘ Wien. Sitzungsberichte,’ vol. xxxi, p. 248) has described the
redia of Cercaria fallax as having two short processes situated anteriorly,
and two, of thrice the length, posteriorly. De Filippi (‘ Memorie della Reale
Accademia delle Scienze di Torino,’ Ser. ii, tomo xviii, p. 207) has described
the redia of Cercaria tuberculata as having four lateral processes, two
anteriorly and two posteriorly. I have met with a species which appears to be
identical with Cercaria tuberculata, and in the redia recognised a collar.
The same writer has figured (ibid., vol. xvi, pl. i, fig. 13) in the young redia
of Cercaria coronata four processes ; the two placed in front are slightly
smaller than the two posterior, but otherwise they are drawn as if exactly
alike. There can be no doubt that in all these cases the structures described
as anterior lateral processes are simply the projecting borders of the trans-
parent collar, seen perhaps in the flattened redia. From comparison with the
descriptions given by these distinguished observers 1 was led in my first paper
(‘Roy. Agricult. Soc. Journ.,’ 1881, p. 19) to similarly misinterpret the
corresponding projections in the young rediw of Fasciola hepatica.

LIFE HISTORY OF THE LIVER-FLUKE. 128

may also be distinguished here, though the main trunks are
less distinctly visible than their ramifications, and can rarely
be followed for any great distance. Hence it is impossible to
discover whether the system of vessels opens externally. The
branches begin with a long narrow infundibulum, in which a
flame-shaped cilium is constantly working, as described in the
sporocyst. The ciliated infundibula are arranged in two groups
on each side of the body ; the anterior group on each side lies ~
a short distance behind the collar, the posterior close to the
processes which serve as feet (fig. 15). The ciliated cells do
not all lie at the same level beneath the surface, so that occa-
sionally two of the infundibula may be seen lying across one
another, and sometimes the cells may le free within the body-
cavity, the end of the cell opposite the cilium being connected
with the wall by one or more processes (fig. 14).

The digestive tract is the characteristic structure of the
redia, and at once differentiates it from the simple sporocyst.
Quite at the anterior end of the body is the mouth surrounded
by projecting folds, which may be termed the lips. The trans-
verse muscle-fibres are especially well developed in the lips,
and assisted by the transverse muscles of the following part of
the body-wall, serve as a sphincter muscle in closing the orifice
of the mouth. The space within the lips is very small, and
leads almost directly into the pharynx, an elliptical muscular
organ by means of which the animal draws in and crushes the
tissues which serve as food. Its outer surface is formed by a
clearly marked limiting membrane, so that it is everywhere
distinguished with readiness from the mass of ill-defined cells
in which it is embedded, and its cavity is lined by a thickened
cuticle. To the pharynx immediately succeeds the digestive
sac, a blind tube of very simple structure. Its wall is com-
posed of a single layer of clear nucleated cells (fig. 12), sup-
ported by a basement membrane, and when it is distended the
cells are flattened out till they are little more than discs, in
which the nucleus causes a distinct swelling. The digestive sac
is seldom more than *3—"4 mm, long, and may be less than this,
but its length differs a good deal, not only in different indivi-

124 A. P. ‘THOMAS.

duals, but also according to the amount of food contained in
it. It reaches its full size early; indeed, in a redia not half
grown it may be as long as in a full-grown example.

The body-cavity is traversed in different directions by bridges
or trabecule of tissue, in which cells of various shapes, some of
them with long processes, can be distinguished. This tissue is
most abundant in the anterior part of the redia around the
pharynx and digestive tract, and here often contains fibres,
probably contractile. Its amount varies very greatly in differ-
ent specimens, and behind the digestive tract is sometimes alto-
gether absent. At other times it is so extensively developed
that the cavity of the redia appears to be divided up into a
number of imperfect compartments, in which the germs lie
loosely. De Filippi appears to have observed similar trabeculze
in the redia of Cercaria coronata, of which he says,' “ La
cavité du corps est traversée sans ordre par des brides.”

There is always a good deal of tissue around the pharynx
and beginning of the digestive tract, aud embedded in it may
be seen, in favorable specimens, a few large round cells with
clear protoplasm and large nucleus; each has a process or
duct (?) passing towards the angle formed by the junction of
the digestive sac with the pharynx. They are probably glan-
dular in function. At the side of the redia, a little behind the
collar, there is present a birth-opening (PI. III, fig. 13, v), which
permits the exit of the brood when ready to leave the parent.
Such an opening has been seen in a number of rediz of different
types, and probably exists in all.

The germs produced within the redia develop either into
daughter-rediz or into cercariz, and it appears to me that slight
differences exist between the individuals giving rise to one or the
other of these generations. A redia producing rediz is usually
smaller, but its pharynx and digestive sac are larger; for ex-
ample, two rediz were taken from the same snail, one producing
redize measured less than 1 mm. long, with a pharynx ‘117 mm.
and an intestine ‘44 mm. in length, whereas in the slightly larger
redia containing cercariz the pharynx measured ‘078 mm. and

1 Tbid., vol. xvi, p. 426.

LIFE HISTORY OF THE LIVER-FLUKE. 125

the digestive sac °24 mm. A further distinction lies in the number
of the progeny ; a mother-redia may contain from one to three
well-formed daughter-redie with a few germs in various stages
of growth; the highest total observed was ten. On the other
hand, in a well-grown redia producing cercariz, I have counted
a total of twenty-three.

The early stages in the development of the spores is the
same, whether they are destined to become rediz or cercariz.
Some of them may be formed from the cells which fill the body-
cavity in the very young redie, but the majority seem to be
formed in the following way:—Some of the cells lining the
body-cavity of the parent, especially those at the posterior end,
are greatly enlarged, and each of these germinal cells undergoes
segmentation, giving rise to amorula. Fig. 18 represents a
large number of germinal cells in the hind end of a young redia,
in which no morule were yet present. Similar cells may be
found in the mature redie (fig. 13 %'), for they retain the
power of producing more spores as the older ones reach their
full development and quit the parent. Hence we find in the
adult redia germs in all the successive stages of growth. Each
morula or germ is enclosed by a delicate membrane forming a
loose envelope. ‘The germs are usually detached from the body-
wall whilst still small and lie free in the cavity of the parent,
but occasionally they may remain in sitti in the body-wall
until they have attained a considerable size (fig. 13 w’). The
morula soon becomes flattened on one side (fig. 12 s), and the
cells of this area are then invaginated, giving rise to a gastrula
(fig. 12 m), whilst the germ again becomes round. The oppo-
site sides of the archenteron are in contact, so that there is
rarely any archenteric cavity, and as growth proceeds and the
cells become more numerous it is no longer possible to distin-
guish the cells of the endoderm, for the cells have the same
size and appearance. Nevertheless it appears to me probable
that the cells invaginated form the digestive tract, which
becomes visible at a later period in the development, rather than
any other cells in the germ. As the germ continues to increase
in size the surface becomes smooth and the outline oval,

126 A. P. THOMAS.

Further growth in size is accompanied by a change in shape,
and it then becomes possible to distinguish between the germs
destined to become rediz or cercarie. The growth of the
young redia within the redia agrees in every respect with the
development of the mother-redia within the sporocyst. The
growth of the cercaria follows a different line.

* Jt may be asked what determines the character of the
progeny, whether the germ shall become redia or cercaria.
My observations are not sufficiently extensive to definitively
decide the question, but it appears to me that the season of
the year is one of the principal determining causes. Redize
producing rediz were only found during warm weather, in the
cold months cercariz were always produced directly. Further,
it is a noteworthy fact that I found at the beginning of the
autumn a redia, containing a single daughter-redia in addition
to numerous cercariz and their germs (fig. 13). I am inclined
to think that the redia was producing rediz but that a fall of
temperature induced the formation of cercariz instead. The
explanation suggested is the more likely to be correct, since
such an arrangement would be highly advantageous to the
species.

C.1. The development of the cercaria within the redia.
—The earliest stages of the development have already been
described up to the time when the germ is an oval mass of
cells. As this continues to increase in size it assumes a more
elongated shape, whilst one end becomes rather more attenuated
than the other. The more slender end becomes slightly con-
" stricted off to form the rudiment of the tail, which as yet is
very stumpy. The remainder of the germ forms the body of
the cercaria, it becomes more depressed in shape, whilst cells
are separated at the anterior end to form an oral sucker, in the
midst of which opens the mouth, and in the centre of the
inferior surface to form a ventral sucker equal in size to the oral.
The digestive tract is now visible as a solid mass of cells.
Immediately following the oral sucker is the rounded pharyn-
geal bulb. Then comes a narrow cesophagus ascending slightly
towards the dorsal surface, and at a short distance in front of

LIFE HISTORY OF THE LIVER-FLUKE. 127

the ventral sucker it bifurcates to form the two limbs of the
intestine, which reach, one on each side of the ventral sucker,
to nearly the end of the body. The limbs of the digestive
tract are solid, being formed for the most part by single rows
of thick disc-shaped cells (fig. 13). The cells are finely
granular at this stage, and show out distinctly against the
clear spheroidal cells which surround the limbs and produce
concave impressions on the surface. At the sides of the body
refractive granules begin to collect in certain of the cells, which
are destined to assist in the formation of the cyst of the
cercaria, and may conveniently be termed cystogenous. At first
the granules are few and inconspicuous, but gradually become
more and more numerous until at length they may obscure the
nuclei, and render the cells opaque. Many of the cells in the
body of the cercaria are crowded with most remarkable rod-
shaped bodies closely resembling bacteria in size and shape
(fig. 20). They reach the length of ‘006 mm., and are often
arranged in rows side by side, whilst the long axes of nearly
all the rods in each cell have approximately the same direction.
Both Wagener and De Filippi appear to have observed similar
structures in the cercaria of Amphistoma subclavatum.
The former speaks of them as “ rod-shaped corpuscles,” and
the latter says that “their form may not inaptly be compared
to that of a shuttle or spindle, with thick walls, and truncated
at both ends. They are destined to disappear later.” These
bodies are not precisely like the narrower ones found in the
cercaria of the liver-fluke, but they are probably correspond-
ing structures.!

An adult redia generally contains a brood numbering about
a score ; amongst these there will be one, two, or three cercariz

1 Prof. Leuckart (‘ Zool. Anz.,’ Oct. 9th, 1882) has also found these curious
bodies (which had already been described by me in the ‘ Journ. Roy. Agric.
Soc.,’ for April, 1881), and as he was unable to find auy spines on the cuticle
of his cercariz, he suggests that the rod-like bodies are subsequently arranged
in bnndles to form the spines of the adult fluke. But I have found the
spines in the most mature cercariz, and can say that these rod-shaped bodies

have no connection with them, though I am unable to suggest any probable
explanation as to their nature.

128 A. P. THOMAS.

approaching complete development. On one occasion I counted
as many as SIX.

2. Free Cercaria.—As soon as the cercaria has reached
the limit of development within the redia, it escapes from the
parent by the birth-opening (fig. 13, v¥) and then by the aid of
suckers and tail, crawls or wriggles its way out of the host.
The free cercaria is very active, and its tissues so contractile
that the form and dimensions of the body are constantly
changing. When ina relatively quiescent condition, the body
has a depressed oval form (Plate III, fig. 19), its average size
is °28 mm. long and ‘23 mm. broad, though the largest may
measure over ‘3 mm. in length. The tail is more than double
the length of the body, and is exceedingly contractile. The
oral sucker is subterminal, the opening of the mouth being
directed downwards and forwards, and has a diameter of ‘06
mm.; the pharynx is ‘034 mm.in diameter. The ventral sucker
is situated slightly behind the centre of the ventral surface, and
is equal in size to the oral, or is sometimes a little larger. As
is the case with all the cercariz produced in rediz (with the
partial exception of Distoma Paludine impure armatum)
the cercaria has no head spine. In the most mature specimens,
and especially in such as have left the redia in the natural
course, and have not been disturbed by the dissection of their
host, the surface of the body is beset anteriorly with exceedingly
minute spines. But the most striking character is due to the
presence of the cystogenous cells, large nucleated cells so
crowded with coarse, highly refractive granules as to be
rendered quite opaque. They are arranged in two lobed masses
extending along each side of the body (Plate III, figs. 19 and 21),
from the level of the pharynx to the posterior end of the body.
Just in front of the ventral sucker is another group of these
cells, which is often large enough to connect the two lateral
masses, and behind the ventral sucker others are scattered.
Cells of the same kind, and showing a similar arrangement, are
found in Cercaria tuberculata (inhabiting Bythinia
tentaculata), a species which shows at first sight a remarkable
resemblance to the cercaria of Fasciola hepatica. I have,

LIFE HISTORY OF THE LIVER-FLUKE. 129

however, myself met with C. tuberculata, and from com-
parative measurements, as wellas the difference in the host, can
state with confidence that the species are quite distinct. And
even in an armed cercaria, recently found in Limnzus
pereger, I found cells, showing a similar arrangement, dis-
tinguished from the remaining cells of the body, not, indeed,
by coarsely granular contents, but by the possession of a pro-
toplasm of a finely granular nature. In this case also the more
granular cells are probably cystogenous.

The other organs of the body are much obscured by the
presence of the opaque cystogenous cells, but the contractile
vesicle of the excretory system, together with the principal
lateral vessels, one on each side, which contain small highly
refractive concretions, can be made out.

3. The Cyst.—When the snails infested with the larval
forms of Fasciola hepatica are kept in an aquarium, the
cercariz may occasionally be found swimming about in the
water, for the granular cells which render the body nearly |
opaque when viewed under the microscope by transmitted light,
give it a snow-white appearance by reflected light, and it is
thus rendered conspicuous for its size. The life as a free-
swimming animal, however, never seems to last long, for, on
coming in contact with the side of the aquarium or the water-
plants contained in it, the cercaria proceeds to encyst itself.
Numbers of minute snow-white cysts may thus be seen adher-
ing to the walls of the aquarium or to the dark-green leaves of
the water-plants. The way in which the cyst is formed can be
readily observed under the microscope, for when examined on
the glass slide the cercaria soon comes to rest, and assumes a
rounded form, whilst a mucous substance is poured forth all
over the body, together with the granules forming the contents
of the cystogenous cells already mentioned. The tail is
sometimes shaken off before the encystation begins, but, as a
rule, the tail remains in connection with the body during the
process, and continues to be energetically lashed from side to
side, until finally a more vigorous movement detaches it. The
whole process of forming the cyst is very rapid, and in a few

VOL, XX1II,—NEW SER, I

150 A. P. THOMAS.

minutes a layer of considerable thickness is formed, whilst its
substance begins to harden. The cysts, as already remarked,
are snowy-white, but the body of the included Fasciola is
quite transparent.

The habits of the intermediate host (Limnzus trunca-
tulus) are of much importance, as showing the manner in
which the cysts are distributed in places where they are likely
to be picked up by some herbivorous mammal, within which
they can attain the adult state. Limnezeus truncatulus
belongs to the group of fresh-water Pulmonata ; it is a common
snail, but one which is often very difficult to find on account
of its small size and peculiar habits. It has a very wide
geographical distribution, being found, according to Dr. Gwyn
Jeffreys, throughout Europe, in North Asia, Morocco, Algeria,
Madeira, and (doubtfully) in Guatemala. Several species be-
longing to the genus Limnezus occasionally crawl for short
distances out of the water, but in L. truncatulus this habit
is so much more strongly developed that the snail should be
termed amphibious. Indeed, it is oftener found out of the
water than in it. When kept in an aquarium it quits the
water, and as often as it is put back crawls forth again so long
as the necessary strength remains. It is said to breed on
the mud at the sides of ditches. To show how much it lives
out of the water I may briefly relate my own experience.
There were floods on the Isis in July last, and the waters
brought it down in vast multitudes, probably from its breeding
haunts in marshy places far up the river. It was extremely
abundant, and a single sweep of a small hand-net repeatedly
gave me more than 500 examples, and this was in a ditch where
previously I could not obtain a single L. truncatulus.

All along the margins of the ditches the ground was covered
by them, and they were found in numbers on the flooded ground
when the flood waters had retired. On returning a month
later to the same ditches I was unable to find a single example
alive in the water. There had been dry weather since the flood,
but early that morning heavy rain had fallen, and I found
numbers of specimens of L. truncatulus out on the gravel of

LIFE HISTORY OF THE LIVER-FLUKE. 13l

a path near the ditch, and these seemed to have crawled out
of the grass when revived by the rain. At the roots of the
grass, along the margin of the ditch, others were found in
abundance. Some few shells were quite empty, but the ma-
jority contained the dried remains of the snail, which had
shrunk far back into the spire of the shell. Most of these
appeared to be quite dead, but were, however, merely dormant,
for on placing them in water the tissues imbibed moisture
and assumed their normal bulk, and after a few hours the
snails had regained their full activity, and were seemingly none
the worse for their prolonged desiccation. ‘To test their power
of resisting drought I collected specimens of L. truncatulus
and placed them in an open vessel on a shelf in a dry labora-
tory, iu a position where the sunshine fell on them for an hour
or so daily. I found that rather more than 50 per cent. with-
stood twenty-six days of this treatment, and some few revived
after more than six weeks. That the snails can live on moist
ground quite away from any quantity of water for considerable
periods, is sufficiently proved by the fact that I have kept them
alive for eleven weeks on moist grass and moss, even when
infested with Fasciola hepatica.!

It is clear, therefore, that the species of snail under con-
sideration, when left on the fields by the passing away of a
flood, continues to wander and feed so long as the bottom of
the grass remains moist. It is equally clear that the numbers
so left are recruited from surrounding ditches and streams. A
drought may render the snail dormant, but, unless too long
continued, it revives at the first shower of rain. If there
are fluke-eggs on the ground and water in puddles or ditches
for them to develop in, the L. truncatulus will most certainly

' Sir Charles Lyell (‘ Life,’ vol. ii, p. 212), in speaking of Madeira, says
that Limnaeus truncatulus was unintentionally introduced by the Por-
tuguese thirty years before, and has spread so widely that it is now found even
in the pools and ruts in the roads, so that it must tave a mode of distribution
which needs investigation. It will be seen from the above account that the
terrestrial habits of this snail, and its power of withstanding drought, are
amply sufficient to explain its spread in Madeira.

182 A, P. THOMAS.

be infected with the larval forms of the liver-fluke ; and owing
to the habit this particular snail has of living so much out of
water, either on the banks of ditches or further away towards
the centre of the fields, if they are damp enough, the cercarie
will, on leaving their host, encyst on the grass in the places where
they have the best chance of being transferred to the herbivorous
mammals grazing on the ground. Having thus gained a suit-
able home they will attain the mature sexual condition, and
reproduce their species by means of ova, thus completing the
developmentat cycle.

Man himself sometimes serves as host to the liver-fluke, and
in this case the cysts are probably eaten with water-cress.

4. Growth of Sexual Fluke.—From observations,! which
I need not describe here, it appears probable that six weeks
elapse from the time of the entrance into the ultimate host
before the fluke begins to produce eggs. During growth the
body undergoes a very great change in form ; the posterior part,
which contains the reproductive organs, far outstrips the ante-
rior part (figs. 24—26). The ventral sucker shares in some
degree the greater growth of the hinder portion of the body;
in the cercaria the suckers are of nearly equal size, and the
same was the case in a young fluke 1:1 mm. in length. But
in specimens 2—3 mm. long, the diameters of the oral and
ventral suckers have usually the ratio of 1: 1:1, and in still
larger examples 6—8 mm. long, the ratio is 1:_1:2, whilst in
the adult the ratio is 1: 1:35, though there is much individual
variety.

The smallest fluke I have yet found in the liver of a sheep is
represented in fig. 23; the digestive tract, which in the cercaria
was simply forked, has already acquired a large number of
branches, though they are comparatively simple as yet. They
subsequently attain a much more complex form, owing to the
number of secondary branches. This branched intestine is
highly characteristic, and affords the principal reason for sepa-
rating the three species which constitute the genus Fasciola
from the species forming the distinct genus Distoma, none of

1 ‘Journ. R, A. §.,° 1881, p, 25.

LIFE HISTORY OF THE LIVER-FLUKE. 133

which have a branehed digestive tract. It is usually supposed
that the liver-fluke passes out of the sheep at the beginning of
the summer, ie., life lasts only about three quarters of a year.
But I have shown elsewhere! that the life of the liver-fluke
may extend beyond one year, and have found both digestive and
reproductive organ in full functional vigour in flukes at least
thirteen months old; the oviduct was filled with eggs, and
there was no indication of any exhaustion of the supply.

For an account of the economic aspects of the subject, in-
cluding the discussion of preventive measures, I may refer to a
paper in the forthcoming number of the ‘ Journal of the Royal
Agricultural Society.’

It gives me much pleasure to take this opportunity of thank-
ing Dr. Acland for kindly permitting me to use the Sanitary
Laboratory of the Oxford Museum for my experiments, and
Professor Moseley for kindly placing apparatus, &c., in the
Anatomical Department at my disposal.

1 Thid., p. 26.

134 W. F. R. WELDON.

Note on the Early Development of Lacerta
Muralis.

By

WwW. F. R. Weldon, B.A.,

Scholar of St. John’s College, Cambridge, Assistant Demonstrator in the
Morphological Laboratory of the University.

With Plates IV, V, VI.

THE following paper contains an account of some observa-
tions on the early stages in the development of Lacerta
muralis, begun during the summer of this year at the zoo-
logical station at Naples and completed in the morphological
laboratory at Cambridge. It relates chiefly to the mode of
formation of the germinal layers and to the early development
of the kidney.

On my return from Naples I found that in June last Pro-
fessor C. K. Hoffmann! had published an account of the mode
of formation of the germinal layers, and the results obtained
by him agree generally with my own. As, however, Professor
Hoffmann has published very few figures of the stages observed
by him, and as my observations lead me to differ from him in
one or two points of detail, it has seemed to me that it would
not be useless to give a short account of my own results.

The segmentation, which conforms to the ordinary mero-
blastic type, has already been fully described and figured by

1C. K. Hoffmann, “Contribution a |’Histoire du Développement des
Reptiles,” ‘Arch. Néerlandaises des Sciences exactes et Naturelles,’ t. xvil.

JARLY DEVELOPMENT OF LLACERTA MURALIS. 135

Kupffer and Benecke! and by Balfour.” Neither of these ob-
servers describes a segmentation cavity ; but Hoffmann ® states
that during the later stages of segmentation a cavity is present,
the floor of which is formed by the yolk, the roof by the lower
layer cells. ‘Towards the close of segmentation it disappears.

This cavity Hoffmann considers equivalent to the segmenta-
tion cavity of the Icthyopsida.

I have observed cavities similar to that described by Hoff-
mann, but I have been unable to satisfy myself that they were
not due to the action of the hardening reagents employed.
The cavity described by Professor Hoffmann differs strikingly,
as he himself points out, from the segmentation cavity of other
vertebrates, in the fact that its floor is never formed of lower
layer cells.

At the close of segmentation the blastoderm consists of a
superficial layer of epiblast cells, which is generally stated to
be a single cell thick; in my sec‘ions, however, the arrange-
ment is very irregular, the epiblast being in some places two
cells deep, in others more.

Beneath the epiblast is an irregular sheet of lower layer cells ;
this layer is in many places two or three cells deep, and the
cells of which it is composed are large, irregular, loaded with
yolk-granules, many having two or even more deeply-staining
nuclei.

In the centre of the blastoderm the epiblast cells become
more columnar than in the peripheral parts, and the lower
layer cells become slightly more regular in their arrangement.
An oval area pellucida is thus formed.

Hoffmann finds at this stage a marked thickening of the
lower layer cells at the posterior extremity of the blastoderm.

The posterior region of the area pellucida now becomes dis-

1 Kupffer u. Benecke, ‘Die erste Entwicklung am Ei der Reptilien.’
Konigsberg, 1878.

2 Balfour, “On the Early Development of the Lacertilia,” this Journal,
vol. xix.

5 Loc. cit.

136 W. F. R. WELDON.

tinguished from the anterior by the presence of the primitive
streak.

A median longitudinal section through an embryo with a
commencing primitive streak is shown in fig. 1. Anteriorly
the area pellucida is seen to be formed by an epiblastic layer of
irregular columnar cells and a sheet of lower layer cells, the
two layers being quite distinct. At a point (bp), however, the
position of the future blastopore, these layers are replaced by
a mass of closely-packed cells (pr), exhibiting no division
into layers, and forming the primitive streak, which may in
some cases at least extend backwards as far as the commence-~
ment of the area opaca.

The blastopore commences at the anterior end of this streak
as a pit, open above, and closed below. ‘This is shown in
fig. 2.

The floor of this pit presently breaks up, and the blastopore
assumes its normal condition, forming a communication between
the archenteron and the exterior, its anterior wall forming a
communication between the epiblast and the lower layer cells
(see fig. 3).

From this time a change in the character of the lower layer
cells takes place, beginning from the anterior wall of the blas-
topore, where they pass into the epiblast, and proceeding for-
wards. Instead of being large, irregular, full of yolk, as in the
previous stages, they become columnar, lose their yolk, arrange
themselves in a definite layer several cells deep, and take on
the characters of normal hypoblast. A median longitudinal
section through an embryo, in which about half the lower layer
cells are thus converted, is seen in fig. 4..

This process is evidently an invagination comparable to that
which takes place in an Elasmobranch. _ It especially resembles
the process described by Scott and Osborne? in the newt.

The first traces of mesoblast appear at a stage slightly earlier
than that represented in fig. 4. Fig. 5, which shows a portion
of a lateral section from the same series as that to which

1 Scott and Osborne, ‘On the Early Development of the Common Newt,’
this Journal, vol. xix.

EARLY DEVELOPMENT OF LACERTA MURALIS 137

fig. 4 belongs, shows the condition of the mesoblast shortly
after its origin.

The blastopore being funnel shaped, with its narrow opening
directed downwards, it appears in a lateral longitudinal section
as a pit, closed below, and from its closed extremity the meso-
blast grows forwards as a solid cap, separate from epiblast and
hypoblast.

Transverse sections show that the mesoblast is in connection
not only with the walls of the blastopore, but also with the
axial strip of invaginated hypoblast. Figs. 6—13 are selected
from a series of transverse sections of an embryo slightly older
than that represented in fig. 4, and show the relations of the
mesoblast. The figures are arranged in order from behind
forwards, fig. 6 being posterior. Figs. 6—9 pass through the
blastoporé, and a sheet of mesoblast, continuous with its walls,
is seen growing out on each side. In figs. 10 and 11, which
pass through the posterior embryonic region in front of the
blastopore, each sheet of mesoblast is seen to be free laterally,
but to be continuous near the middle line with the axial strip
of hypoblast, the cells of which will give rise to the notochord,
and are easily distinguishable from the more peripheral hypo-
blast cells by their more elongated forms and by being more
than one layer deep.

This mode of origin of the mesoblast, however, only holds
good for the posterior part of the embryo. Anteriorly (fig. 11)
the mesoblastic sheet loses its connection with the axial hypo-
blast and finally disappears (fig. 12), being replaced by
branched cells, which are budded off, partly from the axial,
partly from the lateral hypoblast. This mode of origin of the
anterior mesoblast has been overlooked by Hoffmann.

The account above given is obviously in complete accord
with the observations of Balfour,! who described a stage a
little later than that represented in figs. 6—13, with a widely-
open, neuro-enteric canal, and a sheet of mesoblast on each

' Balfour, “On the Early Development of the Lacertilia,” &c., this Journal,
IK,

138 W. F. R. WELDON.

side, which had separated from the axial hypoblast—all the
layers being, however, still fused in front of the blastopore.

The statement of Kupffer,! that the blastoporic invagina-
tion gives rise to a closed sac, the walls of which become the
allantois, is of course inconsistent with the truth of the above
observations ; but it has been already so abundantly disproved,
first by Balfour and afterwards by Stahl and Hoffmann, that
it is not necessary here to do more than refer to it in passing.

The actual mode of development of the allantois was first
figured by Balfour,’ a copy of whose drawing is reproduced
in the woodcut. The details of the process were worked out
by Strahl.%

I have nothing to add to the account given by these authors,
but I would call attention to a consequence of it which neither
observer has, to my knowledge, remarked.

It is obvious from the woodcut that, as has been shown in
detail by Strahl,* the allantois arises as a process of the
primitive streak, which projects at first backwards into the
body cavity.

Now, if this be the case, when the primitive streak is bent
ventralwards during the establishment of the tail fold, the
primitive streak must extend in the middle line from the
posterior extremity of the medullary canal, round the end of
the embryo, as far forwards as the point of connection of the
allantoid stalk, with head cut ; and therefore the proctodeum,
when it arises, must not pass through the primitive streak.

Therefore, if we adopt the view of Balfour, that the primi-
tive streak represents the position of the blastopore of other
gastrule, we shall be forced to conclude that, at any rate in
this group of Craniata, the anus is in the position of a part of

| Kupffer, ‘‘ Die Gastrulation an den Meroblastischen Hiern der Wirbel-
thiere und die Bedeutung des Primitiv Streif,’ ‘Arch. f. Anat. u. Phys.,’
1882.

* “On the Early Development of the Lacertilia,” &c., this Journal xix.

% Strahl, “ Ueber die Entwicklung des Canalis Myeloenteriens und der
Allantois der Eidechse,” ‘ Archiv. f.-Anat. u. Phys.,’ 1882.

4 Loc. cit.

EARLY DEVELOPMENT OF LACERTA MURALIS. 139

the blastopore—a supposition which simplifies our ideas as to
the origin of the vertebrate anus in general.

ME

3
o
be
6 20.
= ie)
r is}
© 3
900,990
i}
"50 COPEQO.S
22.000

O6-2080

Four TRANSVERSE SECTIONS THROUGH THE HINDER END oF a YOUNG
Emsryo or LaceRTA MURALIS (Balfour).

Sections a and B pass through the whole embryo, while c and D only pass through
the allantois, which at this stage projects backwards into the section of
the body cavity behind the primitive streak. me. Neurenteric canal.
pr. Primitive streak. Ag. Hind gut. Ay. Hypoblast. pp. Body cavity.
am. Amnion. sé. Serous envelope (outer limb of amnion fold not yet
separated from the inner limb ortrueamnion). a/. Allantois. me. Meso-
blastic wall of allantois.

The development of the kidney has been described by
Braun.! My observations lead me, however, to believe that
his account of the mode of origin of the segmental tubules
and of the Wolffian duct is erroneous.

* Braun, ‘Das Urogenitalsystem der einbeimischen Reptilien,” ‘Arb. aus
d. Zoolog. Inst. z. Wurzburg,’ Bd. iv, 1878.

140 W. F. RB. WELDON.

According to him, the first part of the urino-genital system
which appears is the Wolffian duct. He says: ‘‘In an em-
bryo of Lacerta agilis, barely 5 mm. long, in sections just
below the heart, I find the Wolffian duct lying close to the lateral
mesoblast plates, in a region belonging neither to these nor to
the protovertebrze, but lying between the two as a semicircular
mass of cells, sharply defined towards the ectoderm, but
' passing gradually into the lateral mesoblast; in the middle of
this cell mass is a lumen . . . . ” which he considers
to be the lumen of the Wolffian duct.

In the next stage described by Braun, a number of seg-
mentally arranged vesicles are present, which are for a short
time attached to the peritoneal epithelium, their cavities also
opening for a short time into the body cavity, but which
afterwards break away, form the well-known S-shaped tubes,
and communicate with the Wolffian duct.

From this account it is evident that Braun has not investi-
gated embryos less than 5mm. long. I have been fortunate
enough to obtain younger embryos, and have been led to some-
what different conclusions.

On the formation of the protovertebrz, each protovertebra
does not at once become completely separated from the lateral
mesoblast, but remains connected at a certain point with a
continuous solid ridge of tissue, generally in early stages
about two cells thick, which projects inwards from the peri-
toneal epithelium, thus forming an “ intermediate cell mass ”
comparable with the structure so called in birds.

Figs. 15 and 16 show the characters of this ridge in an
embryo of about seven protovertebre ; fig. 15 is taken from
a vertebral region, and shows the ridge (?. c. m.) connecting
the protovertebra with the peritoneal epithelium; fig. 16 is
from the next intervertebral region, showing the ridge pro-
jecting freely inwards from the peritoneum. In fig. 16 traces
of a prolongation of the body cavity into the intermediate
cell mass may be observed. In an embryo with ten pro-
tovertebre this cell mass, without losing its connection with
the protovertebre, swells up and becomes semicircular in

EARLY DEVELOPMENT OF LACERTA MURALIS. 141

section, the convexity being directed outwards; this condition
is shown for vertebral regions in fig. 17, for intervertebral in
fig. 18.

At a stage with eleven protovertebre, the vertebral portions
of the intermediate cell mass, behind the fourth protovertebra,
acquire a circular lumen, which is bounded by a single layer
of columnar cells ; this condition is seen in fig. 19. In fig. 20,
which represents a section passing through the end of the
same protovertebra as that from which fig. 19 is taken, the
lumen is smaller; in the intervertebral region behind the
lumen altogether vanishes, and the solid, swollen cell mass pre-
sents an appearance exactly like that seen in the preceding
stage (fig. 18).

There is thus formed a series of cavities in the continuous
intermediate cell mass, each situated opposite a protovertebra,
and having its walls continuous both with the protovertebra and
with the peritoneal epithelium. These cavities are separated
from one another by the solid intervertebral parts of the inter-
mediate cell mass.

In embryos with eleven protovertebre there are five of these
vesicles, opposite the fifth to the tenth protovertebra, the last
two somites being as yet without them. In these last somites
the intermediate cell mass is swollen and solid, as in the
anterior region of an earlier embryo.

These cavities are, as will be seen from their subsequent
history, the segmental vesicles described by Rathke and subse-
quent writers.

They have hitherto been described entirely separate from
one another, and have been supposed (Braun., loc. cit) to arise
as invaginations of the peritoneal epithelium.

When twelve protovertebre are present the Wolffian duct
begins to appear as a solid cord of cells, splitting off in the
intervertebral region only from the intermediate cell mass,
and passing, in the region of each protovertebra, into the wall
of a segmental vesicle. |

Figs. 21—23 represent three sections through about the
sixth and seventh somites of an embryo with twelve proto-

142 Ww. F. BR. WELDON.

vertebre. Fig. 21, the most anterior, passes through a verte-
bral region, and shows the segmental vesicle, with its lumen ;
the section passes through the attachment to the peritoneum
(which in the vertebral regions is becoming smaller), but not
through the connection with the protovertebra. The next
section (fig. 22), through the commencement of the interverte-
bral region, shows the solid cell mass, with a few cells (w. d.)
split off from its outer portion. These cells are the rudiment
of the Wolffian duct. In the next protovertebral region this
cord ceases to be visible. Fig. 23 shows a section through the
commencement of the next protovertebra, passing through the
solid wall of the corresponding vesicle, which has no trace of
the duct.

These cords of cells are present at this stage in four inter-
vertebral areas, behind protovertebre five to eight inclusive.

With the formation of the thirteenth protovertebra the solid
rudiment of the Wolffian duct becomes more distinctly split off
in the intervertebral regions, while opposite the protovertebre
it appears as a solid appendage of the wall of the segmental
vesicles, with which it is perfectly continuous.

At the same time it extends backwards into the ninth inter-
vertebral region.

In an embryo with fourteen protovertebre there are eight
segmental vesicles with a lumen opposite the protovertebre
five to twelve inclusive. All these have the Wolffian duct asa
solid knob on their outer wall, while in the corresponding inter-
vertebral regions there appears a distinct lumen in the duct,
which is more or less completely split off from the rest of the
intermediate cell mass.

The relations of the duct and vesicle in an embryo with
fourteen somites are shown in fig. 24, from the second segmental
vesicle of suchanembryo. In this figure the segmental vesicle
(s.v.) is seen to have a large lumen, and the solid Wolffian
duct (w. d.) appears attached to its outer wall.

In fig. 25, from the next intervertebral region behind fig. 24,
the Wolffian duct has a large lumen, and is attached to the
solid intervertebral cell mass.

EARLY DEVELOPMENT OF LACERTA MURALIS. 143

A section through the next protovertebra would repeat the
features shown in fig. 24. |

On the appearance of the fifteenth protovertebra the lumen
of the Wolffian duct becomes continuous throughout the region
of the first eight segments, and at the same time it acquires a
communication with the cavity of each segmental vesicle in its
course.

The first eight segmental tubules are therefore differentiated,
continuously with the Wolffian duct, from a ridge of cells, con-
tinuous at first along its entire length with the peritoneal
epithelium, and at certain points with the adjacent peeks
vertebre.

With regard to the tubules behind the eighth, they are
developed from the intermediate cell mass in exactly the same
way as those in front; but the Wolffian duct, instead of arising
continuously with them, grows backwards as a free projection
of the above-described portion, without coming into relation
with adjacent structures. It is at first solid, but afterwards
acquires a lumen, and becomes connected with the segmental
vesicles in order from before backwards.

On the subsequent behaviour of the tubules and on the
development of the metanephros I have no observations.

The most interesting feature in the preceding account of the
early development of the lacertilian kidney is the close resem-
blance which it shows to exist between the process of develop-
ment in that group and the process which has been shown by
Sedgwick! to exist in birds and Elasmobranchs. In both these
groups Sedgwick has shown that the segmental tubules arise
from a continuous cell mass connected with the peritoneal epi-
thelium and with the mesoblastic somites, which cell mass is
present from the very beginning of the process of mesoblastic
segmentation.

In the anterior part of the Wolffian body of the chick

1 Sedgwick, ‘‘ Development of the Kidney in its relation to the Wolffian
Body in the Chick,” this Journal, vol. xx; and ‘‘ Karly Development of the

Wolffian Duct and anterior Wolffian Tubules in the Chick,” &c., this Journal,
vol, xxl.

144. W. F. R. WELDON.

Sedgwick has shown that the Wolffian duct and segmental
tubules arise continuously by differentiation of the cell mass.
In the chick, as in the lizard, the independent origin of duct
and tubules in the posterior region is probably a secondary
character.

In conclusion, I wish to express my gratitude to the authori-
ties of the Zoological Station at Naples for their kindness to
me during my visit, and to Mr. Sedgwick for the advice and
assistance which he has given me since my return to
Cambridge.

R. VON WILLEMOES-SUHM. 145

On a Crustacean Larva at one time supposed to
be the Larva of Limulus.

By the late

R.von Willemoes-Suhm, Ph.D.,
Naturalist on board H.M.S. “ Challenger.”

With Plate VII.

(Published by permission of the Lords Commissioners of the Treasury.)

[THE manuscript and figures of Dr. v. Willemoes-Suhm
relative to a Crustacean larva, which he at one time was led to
regard as the larval form of Limulus molluccanus, have
been placed in my hands for publication. I am informed
by Professor Moseley that before his death Suhm was definitely
of opinion that the larva in question belonged to a Cirrhipede.
In the mean time a reference to its supposed connection with
Limulus had been published in Suhm/’s ‘ Challenger Briefe,’
In consequence of this, as well as in view of the intrinsic
interest of the form, I think it well to publish Suhm’s draw-
ings and description in full. In the description I have in
brackets made such additions as are consistent with Suhm’s
later interpretation of the larve.—E. Ray LANKESTER.]

On her way from the Moluccas to Hongkong, and when
sailing from this port to New Guinea, H.M.S. “ Challenger”
went twice through the Philippines, and both times touched
at Zamboanga. This is a port in the south-western part of
Mindanao, on the Straits of Basilan, so called from the large
island which, with numerous small ones, forms the opposite
side of the straits. There a very intelligent native brought a

VOL, XXIII, —NEW SER. K

146 LARVA OF LIMULUS.

living Limulus rotundicauda, Latr, which, according to
Milne-Edwards, is also found in the Moluccas. The man was,
of course, asked to bring more of them, but declared he could
not do so unless we went over to the other side of the straits
to Basilan or Malamani, where his specimen had been caught.
This was done afterwards, when the ship had to go there in
order to take in coals; but during the short time of our stay
no king-crabs could be procured, though I offered high rewards,
and some men, who said they knew where to find them,
were sent out, but never came back to bring anything. It
appears that this Limulus inhabits the shallow water round the
small islands, a great many of which are to be found in these
straits. The native who had brought me the first specimen
was further asked whether he knew what they did with their
eggs, and he at once pointed to the swimmerettes as the place
where these were attached. It appears also that L. molluc-
canus carries its eggs about, for it is stated that the animal
with its eggs is frequently brought to the market of Batavia,
where they are both eaten. [If this is the case the American
king-crabs differ in habits very much from tleir eastern cousins,
for, according to Lockwood, they deposit their eggs in a sand-
hole, where these are fecundated by the male, and then left to
themselves. Unfortunately the shortness of our stay in Zam-
boanga did not allow me to investigate this question more
thoroughly, nor did I attach at first so much importance to the
matter, thinking the development of the eastern king-crab
would be very much the same as that of the American one.
But to my greatest astonishment I found one day among the
surface animals brought up by the towing-net from behind the
ship, Nauplii and larve in different stages, which [at the
time appeared to me to] clearly belong to Limulus. The
next days my attention was, of course, entirely directed to
them, and I succeeded in getting the whole series of stages
from the newly-hatched Nauplias to the larva, which shows
already under its skin the abdominal and the first traces of
the thoracic appendages. In the whole these larve were
rare, at night commoner than in daytime. Altogether I

R. VON WILLEMOES-SUHM. 147

think they had been seized by the current and carried away
from their breeding places, otherwise they would have been
more common. I found these larvze not only in Zamboanga,
but also during the night which we spent in the narrow channel
between Basilan and Malamani. At our return to the former
place another current had set in, bringing in pelagic animals
and sweeping away all the small larve.

After this we took to determining the surface animals, which
had been kept in spirit during the time when the [supposed]
young Limuli were found, and among them we got some more
larve, so that satisfactory drawings, and also some prepara-
tions showing the Nauplius with only one eye and the larve
with the additional large lateral eyes, could be made. (See
Plate VII.)

The Nauplii [in question] are easily recognisable from the
position in which they hold their antenne, which are never
carried in an upward position, but always either ata right angle
to the body, in a horizontal position or directed backwards.
As soon as the animal rests or is touched by the covering glass
all the antenne take the latter position, which makes it rather
difficult to the observer to make them out. Another very
characteristic point in all the Nauplii, except in the very
earliest stage, 1s the pointed caudal portion, [which erroneously
suggested] the future spine [of a Limulus]. Even before
the large lateral eyes have come out, and before the divi-
sion of the body into three regions is complete, i.e.
after the first moult, this [larva supposed to be a] young
Limulus is very easily to be distinguished by the cordiform
shape of its body, its peculiar antennz, and the pointed
abdomen. This guided me with great certainty, for it is at
first by no means easy to find out the [supposed] small king-
crabs among the enormous number of larve which swarm on
the surface near these tropical shores.

In the first stage, the one in which I imagine the (supposed)
young Limulus to leave the egg, we see a somewhat oval-shaped
embryo filled with yolk and granulations, with one central eye
and three pairs of appendages (Pl. VII, fig. 1). I am,

148 . LARVA OF LIMULUS.

however, not quite sure whether this first stage really belongs
to the larve, which I am now going to describe, as I did
not see it leaving the egg, but consider this to be the case
from the form and position of its antenne. In the second
stage, which undoubtedly belongs to this series, the body
is divided into a thoracico-abdominal portion [the shield
being not yet formed] and the jointed spine, the whole
Nauplius being of an elongate heart shape (fig. 2 a, 2 6b).
The inferior part of the body, the future abdomen, is divided
into nine segments, separated by a ridge, which gives their
edges a spiny appearance. The last of these ends in two large
lateral spines. [What seemed to correspond to], the spine of
Limulus in this stage consists of seven segments, the last of
which ends in a point. Of appendages there are three pairs,
the first being just opposite the Nauplius eye, and showing no
joints, with three hairs at the top. The second consists of five
joints and has a two-jointed flagellum. ‘The third pair finally
has about the same number of joints and very nearly the same
length, also showing a small flagellum. There are as yet no
parts of the mouth, nor does the interior, which is filled with
large globules of yolk, show as yet any differentiation. Above
the Nauplius eye, which is very conspicuous, we remark a
small lens. In this stage the embryo has a length of about
0:14 mm., and it is the one in which it was most commonly
met with at Zamboanga.

In the next stage, the third, it has grown a little in size, but
shows as yet very few differences from the former stage, except
that the walls of the intestine begin to form and show contrac-
tions, without as yet communicating with the anal opening
(fig. 3).

After this, one of the most important changes goes on, for
with the fourth moulting the shield appears, and the body up
to this period, consisting of two divisions only, shows three
parts simulating those of a full-grown Limulus, the head and
thorax with its shield, the abdomen, and the jointed spine,
The Nauplius has now a length of 0°36 mm. (figs, 4a, 46).

The shield is rounded at the edges, just as Milne-Edwards

R. VON WILLBMOES-SUHM. 149

describes it in the young Limulus which he figured in ‘ Cuvier’s
Régne animal.’ Only once I saw the edges ending in a rounded
point, a case which I think was due to a folding in of the skin,
and therefore could not be taken as evidence. In all the sub-
sequent free swimming stages the edges of the shield (the
diameter of which is 0°20 mm.) were round. The abdomen
shows its nine larval segments, and under the thin chitinous
covering you distinguish already clearly six newly formed seg-
ments, with lamellar appendages on all of them. The latter
can be best observed in a side view, when also the hairs or
their ends are already to be remarked. They consist of two
joints, the terminal being the shortest.

The [region which simulated the] spine [of an adult Limu-
lus] being 0°056 mm. long [about one sixth of the whole length]
now shows eleven larval segments, with slightly serrated edges.
The terminal point has about the length of three of the pre-
ceding segments.

The appendages are the same as in the former stages—the
three nauplial antenne. The eye also is still the simple Nau-
plius eye with a lens, but in one case we saw two lenses above
the black spot, perhaps the earliest trace of the later subdivi-
sion of the central eye into two. To the right and the left of
the eye there are still large globules of yolk, filling the cara-
pace. The organs of digestion show some changes, for the
upper part of the intestine is widened (the future muscular
stomach), and the position of the mouth, with which there is
as yet, however, no communication, is indicated by a chitinous
under lip. The anus is formed and communicates with the
intestine by a short rectum. It opens between two large
spines at the base of the last abdominal segment. Some yolk
globules are as yet left in the thorax, and some others are in-
cluded in the intestine.

The fifth stage is signalised by the appearance of the two
large lateral eyes, which are situated a little below the base of
the first antenne. They consist of intensely black pigment,
and a circle round them indicates already the growing con-
nection between these large globular spots and the surface of

150 LARVA OF LIMULUS.

the carapace. In this stage the six segments of the thorax
shine still more clearly through the chitinous covering, and in
a side view one gets a most perfect idea of the swimmerettes.
The animal remains, however, as far as its appendages are
concerned, in the Nauplius stage, and in this it remains as long
as it swims on the surface, for in the last stage, in which three
pairs of thoracic legs were visible on both sides of the mouth
under the larval skin, no abdominal feet had as yet disen-
gaged themselves. The animal has grown a little, its spine is
somewhat longer and more pointed, but it still must be con-
sidered as a larva.

[It will be interesting to know precisely to what animal
Suhm’s Philippine larva above described belongs. It differs
from all known Cirrhipede larve in the structure of the tail.
There is no ground whatever for reviving the view, discarded
by Suhm himself, that this larva belongs to a Limulus.—
HK. R. L.]

ON PLASMOLYSIS. 151

On Plasmolysis and its bearing upon the Rela-
tions between Cell Wall and Protoplasm.

By

F. 0. Bower, M.A.,

Lecturer on Botany at the Normal School of Science, South Kensington.
(From the Jodrell Laboratory, Royal Gardens, Kew.)

With Plate VIII.

Ir is not surprising that, after devoting their efforts for so
long to the study of the nucleus, botanists should again turn
their attention to the cell wall and its relation to the proto-
plasm. It was only to be expected that by the application of
those accurate methods of study, elaborated during investiga-
tions of the nucleus, to the formation and origin of the cell
wall, new results would be obtained. Such expectation has
been amply justified by the works of Dippel, Schmitz, and
Strasburger. The mode of increase of substance of cell walls
by apposition, and more especially the mode of formation of
walls in the first instance in ceJl division by the lateral coales-
cence of “ microsomata,”! leads naturally to the supposition
that if there be such a genetic connection between the cell
wall and the protoplasmic body, it would also be possible to
demonstrate that the physical connection between them is very
close. Further, the idea that cells may be connected with one
another by delicate threads of protoplasm, which keep up a
protoplasmic continuity through their cell walls, also presents
itself as a natural corollary on these observations.?2. Such con-

? Strasburger, ‘Ueber den Bau und das Wachsthum der Zellhaiite,’ p, 174.

2 Cf., Strasburger, l. c., p. 246.

152 F. O. BOWER.

tinuity has actually been observed by Tangl' in the endosperm
of certain seeds, and by Gardiner® in the pitted cells of the
pulvinus of Mimosa, Robinia, and Amicia. Again, Fromman®
states that he has been able to observe, in various cases, a
continuous network extending from the protoplasm into the
cell wail.

Such observations as these do not harmonise with the view at
present held of plasmolysis, which is derived mainly from the writ-
ings of H. de Vries.* According to his descriptions of the process
of plasmolysis (1. c., pp. 37—39, and esp. pp. 47, 48), the proto-
plasmic body would appear to separate with a ‘* smooth surface”
from the cell wall on treatment with the plasmolysing solution,
and, when the solution is strong enough, to be completely iso-
lated from the cell wall. Hence is: derived the idea, which is,
it is true, more often tacitly understood than directly expressed
in words, that the smooth surface of the protoplasmic body is
merely in apposition with the cell wall, and not more closely
connected.

The observations detailed below will tend to show that results
obtained by plasmolysis do not disagree with those obtained by
the direct observations of the above-named authors, i.e. that
the connection of the protoplasmic body with the cell wall is
very close.

Before entering upon the description of my own observations’
it would be well shortly to review the chief sources from which
our present information on the subject is derived.

V. Mohl, in his treatise on the vegetable cell, speaking of
the “ primordial utricle,” remarks that “it usually adheres
firmly to the cell wall.”® His results were, however, obtained
for the most part by treatment with acids, &c.

1 Pringsh., ‘ Jahrb.,’ vol. xii, p. 170.

= «Quart. Journ. Mier. Sci.,’ Oct., 1882; ‘ Roy. Soc. Proc.,’ Nov. 11, 1882,

3 *Beob. iiber Structur und Beweg. d. Protoplasma der Pflanzenzellen.’
Jena, 1880.

4 «Unters. iiber die Mechanischen Ursachen der Zellstreckung.’ Leipzig,
1877.

5 V. Mohl, ‘ Vegetable Cell,’ English translation, p. 37.

ON PLASMOLYSIS. ie

In an early work by Pringsheim! a description is given of
the process of plasmolysis, in which (pp. 12, 13) he compares
the separation of the protoplasm from the cell wall with the
separation of a sticky substance from a membrane to which it
had hitherto adhered. He further notices the way in which
the protoplasm remains here and there adherent to the
cell wall, while sometimes, though separated almost entirely
from the wall, it remains connected with it by isolated threads
of protoplasm. He goes on to describe how these strings, after
undergoing various changes of form, finally break off (cf. his
Taf. ii, figs. 16—21).2 Naegeli (‘ Pflanzenphysiologische
Untersuchungen,’ 1855, Heft 1) also observed and described
strings of protoplasm which connect the contracted proto-
plasmic body with the cell wall in plasmolysed cells. He
observed them in various instances (epidermis of petals, Spiro-
gyra, &c.), but did not recognise their appearance as of general
occurrence (cf. his Taf. i, 23; Taf. ii, 2—6; Taf. iti, 4, 5, 12).
He also notes in Spirogyra that strings are often attached at
corresponding points on opposite sides of the wall, but leaves
it an open question whether this is significant or not.

Hofmeister? describes the appearance of the contracted pro-
toplasm of cells with large vacuoles (p. 8, &c.) as lying free in
the cell cavity, but makes no mention of any connecting pro-
toplasmic strings as of general occurrence, though (p. 15) he
notices the occurrence of such strings connecting the con-
tracted protoplasm of cells of certain Alge with the terminal
walls.

Itis to H. de Vries* that we owe the most extended treatment
of the subject of the action of dehydrating reagents upon the

' «Bau und Bildung der Pflanzenzelle,’ 1854.

? From his description and figures, I conclude that, Pringsheim has only
seen the coarser strings to be described below. As I there point out, how-
ever, the difference between these and the finer strings, which appear to have
escaped his observation, is only one of degree.

2 «Die Pflanzenzelle,’ 1867.

4 *Unters. itber die Mechanischen Ursachen der Zellstreckung.’ Leipzig,
1877.

164. F. O. BOWER.

living cell) We may leave on one side the very valuable
conclusions as to the connection between turgescence and
growth, which he obtained by the use of plasmolysis, since
these fall outside my present subject. It is unfortunate that
the importance of these conclusions made him lose sight of the
structural details, which had already been partially observed
by Naegeli and Pringsheim. He even quotes (p. 38) the de-
scription of the latter word for word, though in the text he
repeatedly ignores his results, speaking of the contracted pro-
toplasm as free on all sides (“ allseitig frei,” pp. 9, 38, &c.).
His figures (p. 35) also represent the contracted protoplasm as
completely disconnected from the wall, with which it was
originally in contact.

The results of these earlier observations being thus but little
taken into account in what is certainly the most important of
the more recent works on this subject, it was only natural that
for a time no further advance should be made. The description
of de Vries and his figures were adopted in text-books subse-
quently written, and, as far as I know, there has been no
further statement on this subject! till Gardiner, in a notice
communicated to the Royal Society (Nov. 11, 1882), described
observations on plasmolysis of cells of the pulvinus of
Robinia pseudacacia, which were made in connection with
his work ‘‘On the Continuity of Protoplasm in the Motile
Organs of Leaves.”® He also extended his observations to
pulvini of a number of other plants, and also to stems and
roots. He found that in a very great number of cases strings
of protoplasm connect the contracted protoplasmic body with
the cell wall. His attention was naturally attracted to the
relation of these strings to the pits, and he found that “in
several well-defined instances many threads do go to pits, and
also that in two adjoining cells many threads.on different sides
of the common cell wall are exactly opposite one another.”

Before these observations of Gardiner were published, and

1 The matter seems to have been entirely overlooked by Pfeffer in his

‘Qsmotische Untersuchungen,’ and in his ‘ Pflanzenphysiologie.’
2 © Quart. Journ. Mier. Sci.,” 1882.

ON PLASMOLYSIS. 155

quite independently of them, I had already arrived at conclusions
in the main similar to them, as the result of observations on
the plasmolysis of the prothalli of ferns, which were instituted
with a very different object, viz. that of finding whether plas-
molytic contraction of the protoplasmic body would be a good
method for preparing the apical region of the prothallus, so as
to show the form and arrangement of the individual cells. It
was impossible to overlook the fact that strings of protoplasm
are very universally to be seen connecting the contracted
protoplasm with the cell wall in cells of prothalli thus
prepared.

For the reasons which determined the choice of De Vries
(1.c., pp. 7—13) Ihave adopted as the dehydrating agent solutions
of common salt of varying strength, from 1 per cent. to 10 per
cent., according to the requirements of the object under treat-
ment. It has been my practice to use as weak a solution as
will suffice to bring about the desired result, and it will be seen
that in the majority of cases solutions varying from 2 per cent.
to 5 per cent. have proved strong enough. As changes in the
appearance of the protoplasm follow slowly upon its contraction,
the time at which certain appearances are presented is usually
given. The following are the details of the experiments :

PROTHALLUS OF NEPHRODIUM VILLOSUM AND ASPIDIUM
FILix-MAS. |

On treating a prothallus of either of the above species (others
have not been examined) with a 2 per-cent. solution of common
salt, the protoplasmic body in each cell is seen to separate itself
gradually from the cell wall, the process beginning as a rule
at the corners of the cells. The contraction goes on slowly for
a considerable time, and usually results in the protoplasmic
body assuming a more or less regular spherical form, as has
been frequently described by former writers. When stronger
solutions are used the contraction is more rapid but usually
less regular.

When the protoplasm first contracts in this way there is

156 F. O. BOWER.

often little or no visible connection remaining between it and
the cell wall. Frequently, however, there is to be seen from
the first a faint silky striation in the space between the proto-
plasmic body and the cell wall, running in a radiating manner
between them, This is in most cases extremely delicate, and
even with Zeiss, obj. F, it is sometimes impossible to define the
appearance as any distinct system of lines. Again, in other
instances coarser threads, the outlines of which can readily be
made out with high powers, are seen from the first to maintain
a connection between the cell wall and the contracted proto-
plasm. On these coarser threads are often to be seen nodal
thickenings, similar to those described by Gardiner. Though
the above difference is easily recognised under the microscope,
there can be little doubt that the appearances are merely
phases of one and the same phenomenon.

In those cases where there is at first no visible trace of a
connection between the protoplasm and the cell-wall, there
usually appears, after the lapse of a short time, a striation of
the intervening space similar to that which may often be
observed from the first ; while in the latter case the striation
becomes more obvious, and after a short time (e. g. quarter or
half an hour) it may be seen that it is due to the existence of
numerous very delicate threads, which extend from the proto-
plasmic body to the cell wall. Some idea of the fineness of
these threads in the first instance may be gained from the fact
that they cannot be individually defined even with high power
(F., Zeiss). Fig. 1 represents cells as they appear about a
quarter of an hour after plasmolysis ;' the threads, being tense
at first, appear quite straight. Some time after plasmolysis
has taken place, and the strings have become more obvious,
they may be seen to be executing rapid and more or less irre-
gular vibratory movements; these show that they are not then
very tightly stretched. The strings run not only to those
walls which separate contiguous cells, but also to the free
marginal walls as represented in the figure, and further, as
may be ascertained by careful focussing, to the walls which

1 Compare Pringsheim’s fig. 16, Taf. iii, 1. ¢.

ON PLASMOLYSIS. 187
form the upper and lower surfaces of the prothallus. As far as
I was able to judge, they run as a rule in just as large
numbers to the free wails as to the walls separating con-
tiguous cells. 7

It may often be seen that strings appear to cross one
another, as in the lower cell in fig. 1. This appearance may
be explained by reference to the protoplasmic body, which will
in such cases be found to have contracted irregularly. It may
also be seen that strings, which thus cross one another, are
not in the same plane, a conclusion which might easily be
drawn from fig. 1. Where, as in other cells of fig. 1, the con-
traction goes on more regularly, such crossing of the strings is
not seen.

Remembering Strasburger’s observations on the formation of
the walls in cell division, as well as the results obtained by
Tangl and Gardiner, it was of course a matter of interest to
observe whether these strings in two contiguous cells are
Opposite to one another, and thus point to a direct continuity
of protoplasm through the walls, or whether this is not the
case. In many instances it does appear that the strings on
opposite sides of a wall are attached at corresponding opposite
points ; in a much greater proportion of cases, however, they
appear to have no relation to one another, but to be distributed
quite independently over the walls. It should be remembered
in connection with this that the strings run with equal fre-
quency to the free walls, and to those separating contiguous
cells.

Such connection of the contracted protoplasm with the cell
wall, as that above described, is found to exist in the cells
throughout the prothallus. It has been observed in the cells
at the extreme growing point in young prothalli, and also in
the root hairs at points close to their apex. In such cells,
however, the phenomenon is not so well marked as in cells of
medium age, the threads being of finer texture.

That these connecting strings consist of protoplasmic sub-
stance can hardly be doubted from their mode of origin and
their properties to be detailed below; the application of re-

158 F. O. BOWER.

agents to them is, however, a matter of difficulty, as, under the
action of reagents which injure living protoplasm, they assume
a ropy appearance, and often break away, while they refuse to
take up neutral colouring matters. It has been ascertained
that they stain slightly brown with iodine solution, while they
give a characteristic reaction with gold chloride.!

It has been stated above that the strings, which are at first
as a rule extremely thin, become more obvious a short time
after plasmolysis, there being usually a marked change in the
first quarter of an hour (fig. 11). This is due to an increase
in thickness of the strings, which might be produced by either
of two processes, or by both simultaneously—(1) by the supply
of fresh substance from the main protoplasmic body; (2) by
the lateral coalescence of two or more originally separate
strings.

Exact observation shows that the first process does take part
in the change. It has been noted above that nodal swellings
are sometimes to be found on the threads. By fixing the
pointer of an indicating eyepiece upon one of these swellings,
on a thread of a recently plasmolysed cell, and watching it for
a period of a quarter of an hour or more, it has been seen and
verified in a number of instances that the nodal swelling moves
slowly from the main mass of protoplasm. Since the motion
is, as far as my observations go, always from the main mass of
protoplasm, we have thus an indication of the supply of fresh
substance from it to the threads, which may account for the
increasing prominence of the latter. The lateral vibratory
motion, which is seen in the strings some time after plasmo-
lysis, but is not so marked or is absent immediately after the
contraction, has been alluded to above. From these move-
ments it is inferred that the strings, though apparently
tightly stretched at first, become gradually slacker as time
goes on, a conclusion which harmonises with the observations

' The method adopted was as follows: after plasmolysis with 3 per-cent
salt solution treat with a solution containing 3 per cent. salt and 1 per-cent.
gold chloride, then wash with water and expose to the light in very dilute
acetic acid,

ON PLASMOLYSIS. 159

on the movement of the nodal swellings from the main mass
of protoplasm. It may then be inferred that fresh substance
is derived from the main mass of protoplasm after the original
plasmolytic contraction.

The question still remains, whether the increase in promi-
nence of the strings may not in part be due to lateral coale-
scence of originally separate strings. I have no direct
evidence that such coalescence does occur. Branching strings,
such as those represented in figs. 2, B, and 3, A, B, are often to
be found, which might appear to give colour to the idea that
the branches had originally been separate, and had subsequently
coalesced. It is, however, as far as my observations go, a
universal rule that the branching is in the direction of the
cell wall. This being the case, and taking into account the
process of drawing out of fresh material from the main mass
of protoplasm as above described, the following is a more pro-
bable explanation of such branchings. That two strings (or
more), originally separate but attached to the main body of
protoplasm at points very close to one ancther, had drawn out
from that body a common string on which they appear as
branches. Direct evidence that such a process does take place
is afforded by such objects as are represented in fig. 3, aand B.
In a are seen numerous strings, branched and unbranched, as
they appeared twenty minutes after plasmolysis. 3B represents
the same cell half an hour later; only one of the most promi-
nent branched strings is drawn; on comparing it with the
corresponding string in A it will readily be seen that the
change of appearance points to a process such as that above
suggested. The instances of branching, which are represented
in the figures, are only the last and roughest examples of the
process above described. On examining cells of prothalli
soon after plasmolysis with a high power (Hartnack, 13),
it was seen that not uncommonly strings, which appeared
single throughout the greater part of their length, branched
close to the cell wall, and were thus attached at a number
of points.

It would appear, then, that the change, which gradually

160 F. O. BOWER.

comes over the strings after plasmolysis, is due, at least in a
great measure, to a drawing out of fresh substance from the
main protoplastic body, and a consequent thickening of the
individual strings, which at the same time become less tightly
stretched. It is not, however, asserted that lateral coalescence
of strings never occurs; it is only to be expected that in their
rapid vibratory movements strings should come into contact
with one another and remain coherent, but this has not been
directly observed.

It has been stated that before reagents, which are liable to
injure living protoplasm, the strings alter their appearance,
become ropy and slack, and often break away. Similar
changes occur after plasmolysis has been continued for a long
time, and death supervenes in the plasmolysed cells (cf.
de Vries, l.c., p. 66, &c.). In dead cells the contracted proto-
plasm is completely isolated, or only connected with the walls
by a few ropy strings, which differ in general appearance from
those of living cells, and do not show the vibratory movements.

When the strings break away their free ends often execute
irregular movements, while they contract gradually, as described
by Pringsheim and Gardiner, on the one hand to the proto-
plasm, on the other to the cell wall.

It has already been noted by several observers in various
plants that the protoplasm does not always contract asa single
mass. ‘This is sometimes the case in cells of the prothallus,
the protoplasm dividing into two (or more ’) rounded portions ;
when this occurs the masses are usually seen to be connected
by strings of protoplasm of rather coarse texture; these are,
doubtless, of a similar nature to those which connect the con-
tracted protoplasm with the cell wall.

In conclusion, it may be noted that the walls separating
contiguous cells of the prothallus of the above species have
not a perfectly smooth surface, but show, after the protoplasm
has receded, slight inequalities in thickness when observed
with a high power.

Such being the results obtained by the study of plasmolysis
of cells of prothalli of ferns, the next step was to see whether

ON PLASMOLYSIS. 161

these phenomena are of general occurrence in vegetable cells,
and more especially whether they are to be observed in those
of which the plasmolysed condition has already been described
by other writers.

It being already late in the year, young flower stalks of
Cephalaria leucantha (the plant used by de Vries) were
not to be had: experiments were, however, made with a 5 per-
cent. salt solution upon sections of young leafy stems of this
plant, with the result that, though the material was not very
favorable, strings of protoplasm, similar to those seen in the
prothallus, were found connecting the contracted protoplasm
with the cell wall in a large number of cells of the cortical
parenchyma.

Young flower stalks of an allied species (Cephalaria
rigida) were also used: sections were cut through the cortical
parenchyma and treated with 5 per-cent. salt solution. The
same phenomena, as seen in the prothallus, were again repro-
duced here in their chief features: the strings of protoplasm,
at first not well seen, were quite obvious in the cells after the
lapse of one hour (fig. 4).

The observations of Gardiner on the beet were alse verified,
it being found that kere, on plasmolysis with 10 per-cent. salt
solution, strings of protoplasm remain connecting the con-
tracted protoplasm with the cell wall. They have frequent
nodal swellings, but the strings are not so numerous nor so
regular as in the prothallus.

Sections of the flesh of a ripe apple were also subjected to
the same treatment with results similar to those obtained in the
beet.

In leaves of Vallisneria spiralis strings, forming a fine
radiating system, are seen some time after plasmolysis with a
5 per-cent. salt solution.

The diaphragms of the intercellular spaces of water plants
supply very good material for the study of the phenomena of
plasmolysis in parenchymatous cells. Those of the petioles
of Limnocharis, sp. Aponogeton distachyon, Alisma
Plantago, and Pontederia (Eichornia) cerulea,

VOL. XXIII.—NEW SER. L

162 F. O. BOWER.

were used; in all of these the process of plasmolysis was
observed, its main features being the same here as above
described for the prothallus.

Special attention was given to the diaphragms of the petiole
of Pontederia (Eichornia) ccrulea, which consist of
flattened, polygonal, thin-walled cells, in close contact with one
another, except at the angles where three or more cells meet ;
at these points are intercellular spaces, which act as channels
of communication between the cavities above and below the
diaphragm (cf. figs. 5,6). On treating a transverse section,
including a diaphragm, with 1 per-cent. salt solution, a slight
contraction of the protoplasm takes place. In a very large
number of cases it is found that the protoplasm first leaves the
wall at those points where two cells are separated from one
another by a thin septum, while it still remains in contact with
the parts of the wall adjoining the intercellular spaces. This
would not be the case, if the connection of the protoplasm
with the septa were more close than with the walls adjoining
intercellular spaces; hence it may be inferred that it is not so.
In fig. 5, which illustrates this, and which was drawn imme-
diately after plasmolysis, there are no strings to be seen run-
ning to the cell walls; but when plasmolysis is more complete,
and, after the lapse of a short time, numerous strings may here
be seen, as in other cases (fig. 6). It is found that strings
run both to the septa and to the walls adjoining intercellular
spaces, and no distinction in the numbers which run to these
different parts of the wall has been observed. Comparing this
observation with the fact that the strings run in as large num-
bers to the free walls as to the septa in the prothallus, it is
seen that the same inference may be drawn from both cases,
viz. that as far as evidence from plasmolysis goes, the con-
nection of the protoplasm is just as close with the free walls
as with walls separating contiguous cells.

It should be noted that also in Pontederia the septa
dividing contiguous cells have not a perfectly smooth sur-
face, though there are no obvious pits in the usual sense of
the term.

ON PLASMOLYSIS. 163

Observations were also made on the cells of the amphigastria
of Lunularia and Marchantia, a 2 per-cent. solution of salt
being found strong enough to induce plasmolysis. The proto-
plasm in these cells is very meagre; when contracted it was
seen to be connected with the cell wall by a few long, fine
strings of protoplasm.

Filaments of Spirogyra were also treated with salt solu-
tions of various strengths (2, 5, and 10 per cent). The pro-
toplasm of each cell contracts into a rounded mass, usually
leaving the septa entirely, but often remaining in contact with
the lateral walls. Here also fine strings of protoplasm run
from the contracted mass to the walls, more especially to the
septa. They often have nodal thickenings, and execute
obvious vibratory movements. The phenomenon is better seen
on plasmolysis with 10 per cent. than with weaker solutions,
and even then it is seen only with difficulty.

The above observations having been made upon cells with
approximately smooth walls, the question suggests itself, what
will be the relation of these strings of protoplasm to the
pits in walls where these are present? Peculiar interest is
attached to this question since the publication of the obser-
vations of Gardiner on plasmolysis of pitted parenchymatous
cells of the pulvinus of Robinia pseudacacia, and other
plants, in which he had previously demonstrated the con-
tinuity of the protoplasm through the pits.

The leaves of species of Trichomanes serve as excellent
material for the study of this point, since the lateral portions
of the lamina consist of a single layer of cells, of which the
walls separating contiguous cells are thick and have numerous
pits (fig. 7, a, B); the walls in these figures are represented as
rather thicker in proportion than they appear in nature. A
10 per-cent. solution of salt was found to give good results.
Here, as in other cases described, there is usually no very
obvious system of strings to be seen immediately after the
contraction of the protoplasm connecting it with the cell wall;
but, as before, the intervening space soon assumes the silky

164 F. O. BOWER.

striated appearance noted in other objects (fig. 7, a). As time
goes on the striz become more plain, and resolve themselves
into protoplasmic strings. ‘These were observed to run not
only to the lateral walls separating contiguous cells, but also,
and apparently in equal numbers, to the free walls of the cells,
which are not pitted. As in the prothallus, so here fresh
substance is drawn out from the main mass of protoplasm, in
this case as thick conical processes (fig. 7, B), which give a
very striking appearance to the whole protoplasmic body about
two hours after plasmolysis. The observations made in the
prothallus as to apparent branching of strings were confirmed
in the behaviour of the strings of protoplasm in these cells.

It being possible after their thickening to trace the indi-
vidual strings, it could be seen whether they run as a rule or
chiefly to the pits, or whether there is any constant relation
between them and the pits. On examining a large number of
cases I have found that strings of protoplasm often do run to
pits, and that strings from the contracted protoplasm of con-
tiguous cells are often opposite to one another; but that a
much larger proportion of the strings are not opposite to one
another, and run to points on the cell wall where there are no
pits. In other words, I conclude that in Trichomanes
pyxidiferum my observations on plasmolysis give no clue
to there being any special relation of the protoplasm to the
pits. This is, however, no proof that some special relation
does not exist.

Concluding Remarks.

From the above observations it is seen that the connection
between the protoplasm and the cell wall, as shown by plas-
molysis in those cases which have been observed, is closer
than is usually described, or at least implied in current
botanical writings. The objects were selected from very dif-
ferent systematic groups ; it is true their number is small, and
it must be admitted that the effect is not visibly produced in
every cell; nevertheless, though it cannot be asserted that the

ON PLASMOLYSIS. 165

phenomenon described is universal, it must at least be ad-
mitted that it is very general. I may here suggest that some
difference may be found between the relation of the proto-
plasm to the cell wall in young and in old cells ; no such dif-
ference has been uniformly observed by me, though it has
been alluded to by Naegeli.

It has been repeatedly observed in various instances that
according te the evidence of plasmolysis the connection be-
tween the protoplasm and free cell walls is as close as between
the protoplasm and walls separating contiguous cells; also it
has been seen, in the one example investigated in connection
with that point, that there was no evidence to show any
special relation between the protoplasm and pits of an ordinary
parenchymatous tissue [this will of course require confirmation
in other cases]. From these results it may be inferred that
the phenomena observed are due to a close mode of connec-
tion between protoplasm and cell wall, which is uniform
wherever they are in contact with one another in the living
cell. In the light of recent observations on the mode of for-
mation and growth of cell walls by apposition and coalescence
of microsomata, the connection thus demonstrated by plasmo-
lysis acquires a special interest. Unfortunately, the kernel of
the whole matter, viz. the ultimate mode of application of the
protoplasm to the cell wall, cannot be arrived at with cer-
tainty by plasmolysis, owing to the obvious difficulties of
observation of minute details with high powers in uninjured
cells. Still, collateral evidence may be gained, and as such I
regard the observations above described.

I would suggest two possible explanations of the phenomena
observed in plasmolysis, and their bearing upon the ultimate
mode of application of protoplasm to cell wall—(1) that the
main mass of protoplasm on retreating may leave the cell wall
still completely lined with a thin film of protoplasm ; (2) that
the peripheral part of the protoplasm being entangled, as a
network, among the deposited microsomata may, on the con
traction of the main mass, be drawn out at the points of
entanglement, into fine strings like those observed, while the

166 F. O. BOWER.

surface of the wall is for the most part left free, and not
covered by a film of protoplasm.

In the former case the phenomena observed would be
entirely intra-protoplasmic. The process might in fact be
compared with what is seen when two surfaces, having a layer
of a semi-fluid plastic substance, such as canada balsam,
between them, are suddenly separated. Both surfaces remain
covered with a film of the balsam, while between them run
strings of balsam of varying thickness, which are occasionally
branched, and sometimes have nodal thickenings. If the for-
mation of strings in plasmolysis be thus intra-protoplasmic, their
position would, as in the case of the balsam, be mainly deter-
mined by the conformation of the surface of the wall, and by
internal determining causes in the plastic substance itself, and
would not throw light on the present question of the mode of
connection between cell wall and protoplasm. I have repeatedly
examined the cell wails of plasmolysed cells, both in surface
views and when seen edgeways, and have not been able to
observe any continuous film of protoplasm covering their sur-
face. Having, however, learned from the experiments above
detailed that protoplasm may be drawn out into strings so thin
as to remain undefined with very high powers, the failure to
observe such films does not prove their absence, as they might
also be exceedingly thin.

Taking the second possible explanation of the phenomena
into consideration, we have a strong presumption in its favour
from recent observations. In the first place, those of Stras-
burger on the deposition of microsomata on the cell wall, would
suegest that the protoplasm might be, so to speak, entangled
between these microsomata, and thus be continuous into and
held fast by the cell walls. Thus the attachment would not be
equally close over the whole surface of the wall, but would be
most strong at a number of points where the processes of pro-
toplasm are continued into the body of the wall. Secondly,
Fromman asserts that he has seen a continuous network ex-
tending from the protoplasm into the cell wall. Further, we
have evidence that where the protoplasm certainly does pene-

ON PLASMOLYSIS. 167

trate the cell walls (i. e. in sieve plates) it still may retain its
connection with the cell wall after contraction by means of a
number of strings, which run severally to the pores of the
sieve (De Bary, ‘ Vergl. Anat.,’ figs. 72,75). These, if the
second explanation of the above phenomena were true, would
differ in degree but not in kind from the strings of protoplasm
observed on plasmolysis. It is probable, from Gardiner’s
account (l.c.), that the same may be the case in the perforated
pits in the cells of the pulvini on which he has worked.

I have already stated that on careful observation of the
terminal parts of the strings soon after plasmolysis in the pro-
thallus they are often seen to split up close to the cell wall into
fine branches, and that they are thus attached to the cell wall
at a number of points. This observation gives still further
support to the second mode of explanation of these phenomena
of plasmolysis.

Though it is impossible at present to decide with certainty
which of these interpretations of the phenomena is nearer the
truth, the latter seems to me to coincide best with the facts.

It is unfortunately hardly to be anticipated that the phe-
nomena of plasmolysis will yield us any very certain conclusions
as to the ultimate structural relations between cell wall and
protoplasm, since the difficulties are so great in using high
powers on objects at least as thick as one whole cell; and it is
only by the use of high powers that this point can be decided.
We must, therefore, look to the study of fine sections for
further and secure information on this most important
question.

168 ALFRED GIBBS BOURNE.

On Haplobranchus, a New Genus of Capito-
branchiate Annelids.
By

Alfred Gibbs Bourne, B.Sc. Lond.,

University Scholar in Zoology, Assistant in the Zoological Laboratory of
University College, London.

With Plate IX.

OccurRENCE.—This very interesting worm was kindly placed
in my hands for description by Professor Lankester, who
received it last November in quantity from Mr. Thomas
Bolton, F.R.M.S., of Birmingham, accompanied by a sketch of
the animal in its tube, which I have reproduced in a modified
form in fig. 1. Respecting its habitat, I may quote the words
of Mr. W. H. Shrubsoll, who writes as follows :—

‘“‘The worms have been familiar to me for a Jong time, and
occasionally I have met with them in great abundance on the
coast of Sheppey. Not having books of reference at command
I had no means of knowing whether they were described or
not, and I assumed that they were.

** As far as my experience goes, they are always found asso-
ciated with diatoms on the surface of soft mud at the bottom
of gullies, and the presence of the diatoms insures to the ani-
mal a plentiful supply of oxygen.

“There is generally an inch or so of water overlying the
mud, and the diatomaceous film at the bottom is ornamented
with silvery-looking globules of oxygen.

“As it is impossible to collect either the diatoms or the

ON HAPLOBRANCHUS. 169

worms without getting well besmeared, and sometimes walking
knee-deep in the tenaceous mire, it is hardly to be wondered at
that they have hitherto escaped attention.”

Associated with the worm was a species of Nais, which Pro-
fessor Lankester has identified as the Nais littoralis of O.
F. Miller, which has for many years been unrecorded, no
naturalist, in fact, having seen it since Oersted’s description
and figure in 1842. There were also numerous free-living
nematoids and rhabdoceel planarians.

Mr. Bolton tells me that he has seen the worm once before
in a gathering from the mouth of the Liffey.

Anatomy.—The animal is minute, adult specimens not ex-
ceeding 6 mm. in length.

The tube is about twice the length of the animal, and is
composed of particles of mud, with here and there a diatom
(Pleurosigma, sp.).

Segments and appendages.—The “ head” consists of a
prostomium and a peristomium.!

The prostomium is much reduced and hidden by the peris-
tomium, which rises to form a “collar” around it; this collar
is higher upon the ventral than upon the dorsal surface. There
are two prostomial tentacles, which are short, have pigment in
the walls, and are not ciliated; they are much obscured by the
palps and peristomial tentacles (figs. 1, 2, 3, and 5, prost.

1 In the description, I make use of the following nomenclature :—The
Prostomium (Prestomium, Huxley) is the lobe in front of the mouth; it
may bear two kinds of appendages, (1) prostomial tentacles (antennes,
Milne-Edwards and De Quatrefages; cirri, Kinberg), which spring from
its dorsal surface, and generally resemble in character the appendages of
the peristomial somite; and (2) palps (palpi, Kinberg; infero-lateral
prestomialcirri, Huxley; antennes externes, Milne-Edwards and
De Quatrefages), which spring from the lower surface of the prostomium and
differ very considerably in character in different families of annelids. The
Preristomium (Peristomium, Huxley; Mund-segment, Grube), which is
the first somite of the body, and, it may be, the second and third fused with it ;
and although retaining ordinary characters in a few families of annelids, e. g,
Syllide, very generally becomes much modified. Its appendages are peri-
stomial tentacles (tentacules, De Quatrefages).

170 ALFRED GIBBS BOURNE.

tents.) They are united at their base with the palps and
more posteriorly with the two most dorsal of the peristomial
tentacles.

The palps are very long, and, springing ventrally, bend over
at their free ends towards the dorsal region. They are richly
ciliated upon their dorsal surface, and each contains a large
blood-vessel, with the green blood nearly filling up its lumen;
they can thus be instantly recognised, as the peristomial ten-
tacles have no such blood-vessel, but merely prolongations of
the general body cavity.

The prostomium bears at its sides a pair of black pigment
spots (fig. 5, oc.), which can be seen through from the ventral
surface, and appear at first sight to lie upon the collar, but
transverse sections have demonstrated their true position
(fig. 5).

The peristomium, which forms the collar, as stated above,
bears two pairs of appendages, each consisting of a very short
basal piece and two long rami (noto- and neuropodial), of these
the ventral ramus is the longer in each case, but is not so long
asapalp. They are all richly ciliated upon their inner faces,
and contain prolongations of the general body cavity, but no
special blood-vessel.

The mouth, which is hidden by the collar, lies between the
palps and the bases of the prostomial tentacles.

The somites of the body (counting the peristomial somites
as the first) are twelve in number, of these somites 1—9 form
the “thorax,” and differ from somites 1O—12 which form the
“ abdomen.”

The parapodia are very slightly raised from the surface of
the body, and slightly more so in the posterior than in the
anterior somites.

As the peristomum bears no setz the second somite is the
first setigerous somite and bears dorsal capillary sete only
they are of two varieties placed in two bundles with usually
three in each bundle, and resembling respectively figs. 8 and 9,
which represent sete from the following somite. The one
variety has a very long and delicate blade, while in the other

ON HAPLOBSANCHUS. 171
the blade is shorter and wider; in both cases the blade only
occurs on one side of the axis.

The third somite has similar setz but slightly more nume-
rous (the usual number of the various kinds of sete is accu-
rately shown in fig. 2), and bears in addition ventral “ crochet”?
sete (fig. 10). These are not actually forked but evidently
correspond to the “ crochet ” setz of allied worms.

The remaining somites in the “ thorax” bear similar setz,
the crochet setz present slight variations. Figs. 10 and 11
represent the extreme conditions, the capillary sete in the
posterior segments of the thorax gradually approach the con-
dition shown in fig. 12, intermediate in character between figs.
8 and 9.

In the “abdomen,” an inversion in the position of the setz
occurs, the capillary sete becoming ventral, and the crochet
sete dorsal, at the same time their character is changed (fig. I4) ;
the capillary setz become longer, more flexible, are often
bayonet-shaped (not well shown in fig. 13), and the blade
occurs on both sides of the axis; the “crochet ” sete become
very numerous and are closely placed in a transverse row, they
are finely serrated on one side at the free extremity.

Alimentary canal.—The alimentary canal is simple,
there is a constriction between the fifth and sixth somites, and
between the sixth and seventh ; in the sixth somite it is dilated,
from the seventh somite to the anus, in the terminal somite,
it gradually narrows, immediately in front of the anus it is
slightly dilated (fig. 4). The canal is ciliated at any rate in its
posterior half, the anterior portion presents a brown pigment
in the walls.

Blood-vessels.—The complete distribution of the circula-
tory system is not easily determined.

The blood is green.

There is a dorsal vessel which bifurcates in the peristomial
and terminal segments, the two branches in each case turning
round and uniting to form a ventral vessel. The dorsal and
ventral vessels are connected in the tenth and eleventh somites

172 ALFRED GIBBS BOURNE.

by a pair of lateral commissures. Lateral commissures were
also observed in the somites which contained the ova in the
female (somites 4 and 5).

In the central region, the dorsal and ventral vessels pass
towards the sides and appear to form a sinus around the intes-
‘tine, Claparéde thought to have observed this condition in
Fabricia armandi, and De Quatrefages in Amphicorina.

Vessels pass to the head, but it was not possible to make out
their exact origin; these dilate into a sinus at the base of the
peristomial tentacles (Claparéde, ‘Rech. Anat. sur les Anné-
lides, &c., dans les Hebrides,’ 1861, describes. similar sinuses
at the bases of the branchie in Fabricia quadripunctata
and calls them “ branchial hearts’’), but vessels pass from them
into the palps only, a single trunk to each, which alternately
fills and empties, as is the case in Fabricia and in Polydora
and Spio.

Nephridia.—It was impossible to determine definitely the
Nephridia; in somites 10 to 12 (fig. 4, Neph.) paired bodies
are seen at the base of the parapodia, which I take to be
Nephridia.

In the third somite there are two bodies, the structure of
which could not be ascertained, owing to the amount of pig-
ment in the wall, but they are doubtless modified Nephridia
and function as tubiparous glands; they open at the bases of
the parapodia on each side in the same somite.

Gonads.—The sexes are distinct. The spermatozoa are
seen floating in various stages of development in the body
cavity in the thoracic somites 7, 8, and 9, but not in the abdo-
minal somites (fig. 4).

They are confined to the central region of the somite around
the alimentary canal by a membrane. The spermatospheres
are not spherical, but much elongated rope-like bodies.

They cannot be passed from segment to segment, and the
manner in which the spermatozoa are shed is uncertain.

In the females the ova are found in somites 4 and 5 (fig. 2),
and attain a very large size in the body cavity ; their shape is
continually being altered by the movement of the wall of the

ON HAPLOBRANCHUS. ye

intestine. Their large size probably necessitates their passing
to the exterior by rupture of the body wall.

Nervous system.—I have been unable to determine the
structure of the nervous system. The supra-cesophageal gan-
glion nearly fills the prostomium.

There are no caudal eyes.

There are no auditory capsules.

Arrinities.—Haplobranchus comes into the family Ser-
pulide on account of its capitobranchiate nature, but differs
from all hitherto known genera of the family in that the ten-
tacles, while they remain free, are devoid of any secondary fila-
ments and of any trace of cartilaginous support. It agrees with
the sub-family Sabellide in the absence of any thoracic mem-
brane and operculum.

There are certain genera of the Sabellide which present
some approach to its simplicity of structure. These are: Am-
phiglena, Clap.; Fabricia, Blainville; and Amphicorina,
De Quatr. These forms present certain characters in common
which are absent in Haplobranchus. The first pair of Ne-
phridia, belonging to the second somite, which are modified
as tubiparous glands, and which in all true Sabellids open on
each side at the base of the parapodium, are united in these
three genera in the dorsal region, and open by a single median
dorsal pore at the base of the branchie. In Haplobranchus,
although I have not been able to make out their exact rela-
tions, there is no doubt they are not thus specially modified.
These three genera present auditory capsules in the peristomial
segment and caudal eyes, of neither of which is there any trace
in Haplobranchus.

On the other hand, these three genera agree with one another
and with Haplobranchus in the comparatively simple struc-
ture of the head; the prostomium not being completely fused
with the peristomium is still recognisable, and presents prosto-
mial tentacles and palps. The peristomial collar, completely ab-
sent in Amphiglena, is only slightly developed in the other
forms. There is little differentiation of the regions of the body

174 ALFBED GIBBS BOURNE.

—thorax and abdomen; the sete are simpler than in other
sabellids, and the copragogal groove is absent.

A comparison of the heads in these genera seems to throw
considerable light upon the nature of the processes of the head
in the Serpulide.

In Amphiglena! the prostomium remains well developed, bear-

WA YU,
y Sf

Wy Wy eit

\ \) 8
i se.
prost. cA\N

prose. oe.

A. Diagram of head of Amphiglena, dorsal view.

B. Diagram of head of Fabricia,, dorsal view. prost. Prostomium. prost.
tent. Prostomial tentacles. J/. Peristomial somite. co//. Peristomial
collar. oc. Eye-spots. aud. Otocyst. After Claparéde.

ing lateral pigment spots and a pair of tentacles. It is difficult

to make out the nature of the palps from Claparéde’s figures ;

there are two ventral processes which may represent them.

The branchiz which, according to Claparéde, vary in number,

from eight to twelve, have taken up such a position as to form

a circular crown, and bear along their whole length a series of

short, opposite, secondary filaments.

The branchiz appear to spring from the peristomium.

In Fabricia,? also, prostomial tentacles may be definitely
determined.

1 Claparéde, ‘Glanures Zootomiques parmis les Annélides de Port Vendres,’

p. 32, pl. 3.

2 Claparéde, loc. cit., p. 36, pl. 3.

——

ON HAPLOBRANCHUS. 175

The branchiz have become modified owing to the great de-
velopment of the secondary filaments, which arise alternately
and attain the same length as the main stems.

In Haplobranchus, prostomium and tentacles closely
resemble those of Amphiglena, but the palps are well devel-
oped. There is some little uncertainty about the determina-
tion of the organs I have so marked as palps, but their
slightly greater muscular development, their blood supply, and
their close connection with the ventral region of the prostomium,
point to their being different in nature to the other tentacles.

The branchial tentacles it is which are so especially inter-
esting in Haplobranchus, their definite arrangement, united
in pairs at the base, the absence of any secondary filaments,
the rich ciliation upon their inner faces,’ and the absence of
any branch of the closed vascular system in this lumen are all
interesting characters, and point to their being similar to the
peristomial tentacles of other annelids which are just taking
on that branchial function which is the most marked feature
of the whole family. These tentacles on account of their
united bases may really be said to represent two parapodia on
each side, each possessing a notopodial and a neuropodial
ramus, indeed, they remind one very forcibly of the peristomial]
tentacles of such an annelid as Nereis. Their condition
seems to me very strong evidence in favour of their peristomial
nature, and consequently of the peristomial nature of the
branchiz of the Serpulide. This view was entertained by
Milne Edwards,” but De Quatrefages * states that the branchize
of the Serpulide receive their nerve supply from the supra-
cesophageal nerve ganglion, and consequently he considers
them to be prostomial.

Claparéde and Mecnikow* have shown that in Dasychone

1 Since sending my drawings to the press, I have observed a distinct ten-
dency to a grouping in the arrangement of the cilia; upon the surface of the
branchi, groups of cilia springing from a very slightly raised serpentine ridge.

2 *Régne An. ill.,’ pl. 1B, explanation of fig. 2.

3 De Quatrefages, ‘ Hist. Nat. des Ann.,’ tome ij, p. 401.

4 «Zeit. fir wiss. Zoologie,’ Bd. xix, 1869, Taf. xvi.

176 ALFRED GIBBS ROURNE.

lucullana, the first rudiments of the branchiz arise as two
processes which soon bifurcate, and are clearly placed below
the very large prostomium, This is at a stage when three
sete bundles are visible.

The prostomial lobe, which in Dasychone never bears any
tentacles or palps, gradually aborts, and the secondary branchial
filaments appear to have a terminal origin. Thus develop-
mental history, so far as we know it, favours the view of the
peristomial nature of the branchiz.

Systematic DEscRIPTION.
Family—SeErPuLipz. Tribe—Sabellide.
Haplobranchusg, g. n.

Head distinct.

Pro- and peristomium almost fused, two prostomial tentacles,
two palps.

Collar slightly developed.

The paired branchize consist each of four free tentacles
united at the base in pairs, and entirely devoid of secondary
filaments; they are richly ciliated. No blood-vessel in the
branchiz, a single blood vessel in each palp.

Tubiparous glands not united.

Caudal eyes absent.

Auditory capsules absent.

Sexes distinct.

H. aestuarinus, sp.n. Isle of Sheppey, England, W. H.
Shrubsoll ; Mouth of the Liffey, Ireland, Tho. Bolton.

Specific characters where a single species only is known
must always be guardedly put forward, but judging from allied
forms, the following would seem to have such weight :

Length 4—6 mm.

Nine somites (8 setigerous) in the thorax.

Three somites in the abdomen.

The shape of the setz, figs, 8—14.

Blood green.

LATERAL AND CENTRAL EYES OF SCORPIO AND LIMULUS. 177

The Minute Structure of the Lateral and the
Central Eyes of Scorpio and of Limulus.
By
E. Ray Lankester, M.A., F.R.S.,
Jodrell Professor of Zoology,
and
A. G. Bourne, B.Sc.

With Plates X, XI and XII.

Scope.—In the essay entitled “ Limulus an Arachnid,”
published in this Journal in 1881, it has been pointed out by
one of us (Prof. Lankester) that, amongst other very numerous
agreements of structure exhibited by the Scorpions on the one
hand and the King Crab on the other, there is a close super-
ficial coincidence in the disposition and the character of the
eyes; in both we find a single pair of simple central eyes,
and a lateral or marginal pair of “ grouped” or aggre-
gated eyes—the multicorneal lens of the King Crab’s lateral
eye corresponding to the numerous (two to seven) small lenses
placed in groups laterally on the Scorpion’s head.

The object of the investigation, of which the present memoir
records the results, was to ascertain how far there is a real
identity in the minuter structure of the eyes thus compared,
and what precisely is the morphological relationship between
the multicorneal lateral eye of Limulus and the groups of
unicorneal eyes occupying a corresponding position in the
Scorpions.

The evidence adduced in the essay above cited in favour of
a close genetic relationship between the King Crabs and the
Scorpions was so overwhelming and convincing to our minds,
that we entered upon this inquiry with the strong anticipation

VOL, XXIII1,—NEW SER. M

178 PROFESSOR E. RAY LANKESTER AND A. G. BOURNE.

that most important points of agreement would be revealed in
the comparison of the minute structure of the soft parts of the
eyes in question. And we consider that the results we have
obtained confirm the opinion expressed by Prof. Lankester, to
the effect that the minute structure of the eye, when thoroughly
studied, will furnish even more valuable evidence than that
given by any other structural features, in the attempt to trace
out the genetic relationships inter se of the great groups of
Arthropoda. _

Previous. observations.—The classical researches of
Grenacher! have laid the foundations of a new and rational
study of the minute structure of the Arthropod eye. Our
observations have led us to accept, as thoroughly justified, the
main conclusions of that anatomist in reference to the nature
and structure of the morphological factors of the simple and
compound eyes of Arthropoda. To these we shall have to
allude in the course of our descriptions. Here, we have to
point out, that in his large work Grenacher has not given any
account of the eyes of the Scorpions, and only a fragmentary
account of the lateral eye of Limulus.

Von Graber,’ writing subsequently to the publication of
Grenacher’s large work, has endeavoured to “correct” the
conclusions arrived at by Grenacher, and has offered some
original observations on the structure of the lateral and central
eyes of Scorpions. So far as Graber’s “corrections” relate to
the fundamental points, such as the ultimate structure of a
retinal or optic cell, and the relation of optic cells to the ceils
of the vitreous body, we have no hesitation in stating that he
is totally wrong and that Grenacher is right. With especial
reference to the Scorpion’s eye, Graber’s observations and
drawings are very defective and, indeed, altogether misleading
in regard to simple and fundamental features of structure.
Apparently the method of manipulation, the great thickness
of the sections, and similar circumstances, are the cause of
Graber’s errors.

1 ¢Sehorganen der Arthropoden,’ Gottingen, 1879.
2 © Archiv. f. Mikrosk. Anatomie,’ vol. xvii, 1880.

LATERAL AND CENTRAL EYES OF SCORPIO AND LIMULUS. 179

Nevertheless, Graber has the merit of having, as we shall -
mention below, correctly observed some important facts as to
the Scorpion’s retina—for the first time. In reply to Graber’s
article, Grenacher has published a memoir in which, whilst he
very justly rejects the “corrections ” attempted by Graber, he
gives an account of certain observations on the structure of
the Scorpion’s eye—made in order to control the statements of
Graber. Unfortunately these observations of Grenacher were
confined to the central eye, and did not extend to the very
differently-built lateral eye; and moreover, the observations
are not illustrated by any figures. We shall not have occasion
to refer to them again, since they merely furnish the observa-
tional basis which enabled Grenacher categorically to deny
some of Graber’s assertions.

Both this paper of Grenacher’s and that of Graber, to which
it is a reply, deal very largely with the eyes of Myriapods ;
and the structure of these, though not of those of the Scor-
pion’s, is beautifully illustrated in Grenacher’s plates.

Hence it is actually the case that no figures, except the very
erroneous ones of Graber, have been published of the eyes of
Scorpions ; whilst the structure of the lateral eyes of those
animals has not been looked at by the most capable student of
the Arthropod eye.

With regard to Limulus, there is practically nothing else
published relating to the structure of the eyes than the results
given by Grenacher in his large book, of the examination of
the lateral eye of a not too well preserved specimen. No one
has given any account of the minute structure of the central
eye of Limulus.

It is true that Dr. Packard has alluded to this matter in
his memoir on the “Anatomy, Histology, and Embryology
of Limulus,” but it is so abundantly evident that Dr. Packard
has not made use of the ordinary methods of histological in-
quiry in dealing with this and other parts of the King-crab,
that it seems to be the proper course to omit any further refer-
ence to the drawings and descriptions in his memoir which are
supposed to have reference to the histology of the central and

180 PROFESSOR E. RAY LANKESTER AND A. G. BOURNE.

the lateral eyes of Limulus. It is to be regretted that the
sections used by Dr. Packard were not made transparent
before they were put under the microscope and drawn.

Material.—We have studied the central and lateral eyes of
two species of Scorpion, viz. Androctonus funestus, var.
citrinus, Ehr., obtained in the living state by Prof. Lankester
from North Africa, through the kindness of Prof. Carl Vogt,
and of Euscorpius Italicus, Roess. (also of the allied E.
Carpathicus), kindly forwarded to us in the living state by
Mr. Carmichael Gibson, and by Mr. Neville Reid, from the
South of Italy.

We have been able from time to time to purchase living
specimens of the American King-crab, Limulus poly-
phemus, Latr., in London, though we have felt here very
much the want of small specimens not exceeding one or two
inches in the diameter of the prosomatic shield, which would
have been easier to cut and in other respects advantageous.

Methods.—We have been able to ensure, as above shown,
the perfect freshness of the material used.

The soft tissues of the eye were placed, with the piece of
chitinous cuticle adjacent, in absolute alcohol. In the case of
the central eye of Limulus it was found necessary to separate
the soft tissues from the chitin before immersion in alcohol.

Sections were prepared by the method of long soaking, first
in turpentine then in paraffin, and slicing with the improved
Riviére’s microtome. Most of the sections were then, after
removal of the paraffin, carefully depigmented whilst under
observation by the use of dilute nitric acid (about 5 to 10 per
cent.). The process of the destruction of pigment was ar-
rested at various stages. Some of the depigmented sections,
and some of those not treated to remove pigment, were stained
with borax carmine and mounted in Canada balsam. Others
were preserved in glycerine unstained.

We may point out that the excellent method of thorough
impregnation with paraffin enabled us to obtain exceedingly
thin sections of the Scorpion’s eyes, and to preserve complete
series for study.

LATERAL AND CENTRAL EYES OF SCORPIO AND LIMULUS. 181

The lateral eyes of Limulus required special treatment,
on account of the density and extent of the chitinous lens-
area. The vertical sections in this case were cut with the
hand, and were more difficult to obtain in satisfactory condition
than any of the others.

Drawings.—The figures which illustrate this memoir are
not drawn to a constant scale. With a few exceptions, they
are not representations of actual sections, but combination-
drawings, intended to place before the reader results, and not
the crude material from which the results are derived. The
colouring of the drawings is purely conventional. In most
cases the eye-pigment forms, on solution by the nitric acid, a
fine madder-brown tint, which stains the nuclei, and often the
protoplasm, of the cells of an entire section. The pigment
thus becomes diffused as soon as the attempt is made to remove
it. Pigment granules in process of solution appear of a deep
red-brown colour ; when not acted on at all they are absolutely
black, or, in some cases, greenish grey.

Tue LATERAL Eyes oF Scorpions.

Of Androctonus funestus, Ehr.—The lateral eyes of
Scorpions are placed on the margin of the prosomatic shield,
in a group on each side anteriorly. The number of separate
lenses developed differs in various sub-genera, each lens indi-
cating a separate eye. In Androctonus the eyes are more
numerous than in other Scorpions, each lateral group showing
in A. funestus as many as five lenses, three larger and two
smaller. The smaller lenses are equally entitled to count as
eyes with the larger. It is, however, difficult without great
care and minute examination to distinguish mere tubercles of the
chitinous integument from eye-lenses. {This is, of course,
readily done either when sections are cut, or when the sub-
jacent tissues are cleaned away from the ocular area of the
chitinous shield, and the tubercles and lenses are examined by
transmitted light.

182 PROFESSOR E. RAY LANKESTER AND A. G. BOURNE.

It is convenient to speak of the region in which the lateral
eyes of the Scorpion develop as the “ ocular area.”

A section vertical to the chitinous surface of the shield, and
parallel with the optical axis of the lens through one of the
lateral eyes of A. funestus, presents before the pigment is
removed the appearance shown in Plate X, fig. 1, after the
removal yf pigment, the appearance shown in fig. 2.

In the first place we notice in this, as in other Arthropods’
eyes, that the lens is simply a local enlargement of the cuticle,
and that the layer of epidermic cells usually called “ hypo-
dermis,” which produces the general cuticle, and is observable
at the sides of the section (Pl. X, fig. 2c), is continued beneath
the enlarged boss of cuticle, which acts as lens, and is cor-
respondingly increased in dimensions. This eularged portion
of the hypodermis is, in fact, the soft or living tissue of the
eye, and may be distinguished from the lens in front of it by a
special name. We propose to call it the “ommatéum.” The
ommateum and the lens together form the eye.

When series of sections of the ommateum of a lateral eye of
Androct. funestus are carefully studied it is found that the
ommateum is a simple enlargement of the single layer of
cells forming the hypodermis. It consists of a single row
or stratum of cells, which present a distinction among
themselves into two kinds. The two kinds of cells in the omma-
teum of the Scorpion’s lateral eye are—firstly, the retinal or
NERVE-END CELLS (fig. 2, 4), in which the nerve filaments of
the optic nerve terminate (fig. 2, m); and, secondly, INDIF-
FERENT CELLS (fig. 2, fand g), which are narrow and columnar
in form, similar in every respect to the ordinary hypodermis
cells in the neighbourhood of the eye, and like these latter
pigmentiferous,

Both nerve-end cells and indifferent cells of the lateral omma-
teum apparently belong to the epiblastic layer, and are shut
off together with the layer of hypodermis cells from the sub-
jacent connective tissue by a well-marked “ basement mem-
brane,” which in the region of the ommateum may be called
the eye-capsule, or, better, the “‘ ommateal capsule.”

LATERAL AND CENTRAL HYES OF SCORPIO AND LIMULUS. 183

THE NERVE-END CELLS of the ommateum of a lateral eye of
Androctonus funestus are much larger than the neighbouring
indifferent cells. They are elongated, and are disposed somewhat
radially, reaching from the lower surface of the cuticular
lens to close upon the ommateal capsule. The nucleus is placed
near the capsular or filamentary extremity (that which is
connected with the nerve filament) of the nerve-end cell, and
is of large relative size, spherical, and with well-marked
nucleolus. The nerve-end cells appear to possess over their
whole surface a well-marked cuticular substance which encloses
the soft protoplasm. The minuter structure of these cells we
are not prepared to discuss on the present occasion, our object
being morphological rather than histological. A very important
feature in the structure of the nerve-end cells is the existence
of a special rod-like cuticular thickening on the side of each
cell. This thickening is highly refringent, and very possibly
is of a chitinous nature, though we are unable to offer any
evidence as to its chemical nature. In the section drawn in
fig. 2 fragments of these lateral hard-pieces are seen of a
yellow cclour (2). It appears, from further examination of
sections, that the lateral thickening in each nerve-end cell
is so placed as to adjoin and even fuse with the similar lateral
piece of a neighbouring nerve-end cell. The resulting hard-
piece has been called by Grenacher, when observed in the
compound eyes of Insects and Crustaceans, a “rhabdom.”
We may make use of the same term for the composite body
formed by the union of the lateral hard-pieces of the nerve-end
cells of the Scorpion’s eye. At the same time, each hard-
piece in a nerve-end cell may be called a “ rhabdomere.”

The rhabdoms of the lateral eye are not so well developed as
those of the central eye. They appear to be irregular in
shape, and inconstant in the number of cells and rhabdomeres
which take part in their formation. They will be best under-
stood when the structure of the central eye has been de-
scribed.

THE INDIFFERENT CELLS of the ommateum of a lateral
eye of Androctonus funestus are to be distinguished into

184 PROFESSOR E. RAY LANKESTER AND A. G. BOURNE.

two kinds according to their position in the ommateum.
There are, firstly, those which form the periphery of the
ommateum, and are contiguous with the extra-ocular hypo-
dermis cells. These are very long columnar cells, which fill in,
as it were, the optically valueless circumference of the ommateal
capsule. They may be called perineural cells (fig. 2, f).

The second kind of indifferent cells are placed between the
diverging filamentary extremities of the nerve-end cells. They
are very small columnar, closely-fitting cells, quite similar in
character to the general hypodermis cells. ‘They may be called
‘‘interneural cells.” They are by no means easy to observe,
being liable to be destroyed by the action of the acid which
is necessary to remove the abundant pigment with which they
are charged (see fig. 1).

PIGMENT GRANULES appear to be developed in all the cells
of the ommateum as well as in the neighbouring hypodermis
cells. But it is difficult to make out in this and in all Arthro-
pod eyes what precisely is the situation and the limit
of the pigment. Before the pigment is removed observation
is impossible ; when it is dissolved by acid it diffuses and stains
structures previously devoid of pigment. A partial removal
of the pigment by means of solvents seems to be the only
method which can give any indications on this matter, and
even that is unsatisfactory. Pigment granules appear to be
very freely developed in the protoplasm of the ordinary hypo-
dermis cells and of the indifferent cells (both perineural and
interneural) of the ommateum. But in the nerve-end cells
the pigment granules are confined to the surface of the cell,
leaving the axis transparent. It will be seen subsequently
that in the central eyes the nerve-end cells are very nearly if
not quite devoid of intrinsic pigment granules, and one is
led to question whether the pigment which clothes the nerve-
end cells may not in all cases be of external origin. This,
however, cannot, it seems, be maintained. We have to admit
that the nerve-end cells sometimes produce peripheral pig-
ment granules, and sometimes are devoid of pigment.

The relation of pigment to the optical apparatus cannot be

LATERAL AND CENTRAL EYES OF SCORPIO AND LIMULUS. 185

said to be at present properly understood. It is perfectly
certain that in some eyes, and possibly in all, pigment does not
play a primary part in the physiological process set going by
light. Light acts with full effect upon transparent protoplasm,
and no pigment is necessary, converting the energy of light
into the energy of heat, in order that the protoplasm of cells
may constitute an apparatus sensitive to light. The function
of pigment in an eye is a secondary one, as we learn from
the sight of albino varieties. What precisely the significance
of pigment may be in relation to the cells in which the optic
nerve ends, is not yet agreed upon by physiologists.

Of Euscorpius Italicus, Kés.—In figs. 3 and 4 two
sections are drawn of the lateral eye of the little Italian
scorpion, as seen after removal of pigment. They are more
highly magnified than the eyes drawn in figs. 1 and 2, being
in actual size considerably smaller than the corresponding
eye of Androctonus funestus, as shown by a comparison
of the measurements given in the explanation of the plate.
The hypodermis cells are relatively to the nerve-end cells
coarser than in Androctonus funestus. In all essential
respects the eyes of the two species agree. The marginal in-
different cells or perineural cells of the ommateum are larger
relatively than in A. funestus, and so are the interneural
cells, which are correspondingly less numerous than in A.
funestus. The rhabdom is larger and thicker in E. italicus,
whilst further the nerve-end cells present a special structure
which is not in any way indicated in the nerve-end cells of
A.funestus. Each nerve-end cell contains, besides its nucleus,
a globular, highly refringent body (figs. 2, 3, 4), quite uncon-
nected with the rhabdom, though of a substance similar to
that of the rhabdom. These bodies, which it is convenient
to term ‘“ phaospheres,” are usually to be found below the
nucleus of the nerve-end cells, that is to say, near the fila-
mentary extremity of those cells. But there are some nerve-
end cells in every lateral eye of E. italicus which have the
phaosphere placed in front of the nucleus (fig. 4, 7). This

186 PROFESSOR E. RAY LANKESTER AND A. G. BOURNE.

irregularity in the position of the phaosphere is very remarkable.
It is to some extent paralleled by Grenacher’s observation of
the fact that in Epeira diadema all the nerve-end cells in
one eye (posterior dorsal) present a rod-like body i in front of
the nucleus, whilst in a neighbouring eye (Anterior dorsal) all
the nerve-end cells present a rod-like body posterior to the
nucleus. At the same time it is to be observed that the axial
rod of the Spider’s nerve-end cell must be considered as repre-
senting not only the phaosphere but also the laterally-placed
rhabdomere of the nerve-end cell of the Scorpion.

The rhabdoms of the lateral eye of E. italicus are very
nearly as indefinite in their development as in the correspond-
ing eye of A. funestus.

It will be seen below that in the central eye groups of five
nerve-end cells unite by means of their rhabdomeres to form
what Grenacher has called, in the multicorneal eye of Insects
and Crustaceans, a “ retinula,” and in the centre of this retinula
is a five-ridged rhabdom (see Pl. XI, fig. 14).

In the central eye of Scorpions this grouping or segregation
can be made out, though it is by no means so fully expressed
as in the multicorneal (polymeniscous) eye of Insects. In the
lateral eyes of the Scorpions, on the other hand, the grouping
into retinule of the nerve-end cells and the formation of
rhabdoms from rhabdomeres is merely foreshadowed and not
completely carried out.

In fig. 6 a portion of a section transverse to the long axis of
the nerve-end cells of the lateral eye of E. italicus is shown.
Rhabdomeres coloured yellow are seen in section, and it ap-
pears that they have a tendency to unite with one another,
though such a five-sided figure as that to be observed in the
corresponding region of the central eye (Pl. XI, fig. 15) is not
yet attained.

In fig. 5 the appearance of the ends of the nerve-end cells, as
seen when the cuticular lens is removed, isshown. A tendency
of the cells to group in fives can be traced. The pavement-like
appearance of the ends of the long nerve-end cells when thus
viewed has given rise to the erroneous statement on the part of

LATERAL AND CENTRAL EYES OF SCORPIO AND LIMULUS. 187

Von Graber, that a pavement epithelium is disposed on the
deep face of the cuticular lens.

Von Graber’s statements.—Of the numerous points con-
cerning which Von Graber has made erroneous statements in
his writings on Scorpion’s eyes,’ the most important is that
which relates to the fundamental structure of the lateral eye.
Von Graber states (and emphasises his statement by a drawing
professing to be an accurate copy of a preparation of the lateral
eye of Scorpio Europzus, Schr.) that the lateral eyes of the
Scorpions are provided with two rows of cells—a vitreous body
and a retinal body—just as are the central eyes, the two rows
of cells being separated by amembrane. This statement is alto-
gether erroneous. ‘The description and figures which we here
publish show that the ommateum of the Scorpion’s lateral eye
has no “ vitreous body,” and consists of a single layer of cells,
some larger (nerve-end cells), some smaller (interneural and
perineural cells).

Von Graber’s error in this matter has apparently arisen, like
most of the errors to which he commits himself in the same
memoir, from the defective character of his methods of investi-
gation. His sections were too thick and ill directed, and his
macerating and decolorising fluids were allowed to act too
rapidly or for too long a period.

A further error of Von Graber in regard to the lateral eye is
his description of the ommateum as composed of three layers
(besides his non-existent vitreous body)—a layer of nerve-
fibres, a layer of ganglion cells, and a layer of “rod cells”
(nerve-end cells). There are no “ ganglion cells” within the eye-
capsule distinct from the nerve-end cells. Von Graber holds an
altogether erroneous view as to the structure of the nerve-end
cells, which, in opposition to Grenacher (who has studied and
described these structures in other Arthropoda), he declares to
possess three nuclei—an anterior, a middle (that of the rod cell
or rod region), and a posterior (that of the ganglion cell). The
nerve-end cell is thus, according to Von Graber, a compound
body, consisting of three fused cells. He terms it a “ retinal

' * Archiv f. Mikrosk. Anat.’ vol. xvii, 1880, p. 58.

188 PROFESSOR E. RAY LANKESTER AND A. G. BOURNE.

sac” (Schlauch). Grenacher, on the other hand, denies the
existence of Von Graber’s anterior and middle nuclei; for him
the nerve-end cell is a single cell of elongated form, with one
large nucleus—that which Von Graber calls nucleus of the
ganglion cell. Our conclusion as to the nerve-end cells of the
lateral eyes of Euscorpius and Androctonus entirely agree with
those of Grenacher as to these structures in general. Grena-
cher has not examined, it must be remembered, the lateral eyes
of Scorpions. In describing the nerve-end cells of the central
eyes of Scorpions we shall have occasion to point out the
existence of structures adjacent to the nerve-end cells, which
have probably led Von Graber to hold the erroneous views as
to the nerve-end cell which he has advanced. These structures
have escaped the notice of Grenacher.

CrntTRAL EyE oF ANDROCTONUS FUNESTUS.

The central eyes of the Scorpions are considerably larger
(from twice to three times linear) than the lateral eyes. As in
the lateral eye, we distinguish lens and ommateum. The lens
is a simple laminated mass of cuticle, which we must dismiss
on the present occasion without attempting any examination of
its optical properties. It would, no doubt, be important to
compare these with those of the lateral eye, &c.

The ommateum of the central eyes differs essentially from
that of the lateral eyes, in the fact that it is not composed of
one layer of cells, but consists of two layers of cells, one super-
imposed upon the other, and separated from it by a strong
laminated membrane (PI. X, fig. 8%; Pl. XI, fig. 11 2).

The anterior layer (0 in the figures) is known in other simi-
larly constructed eyes as the ‘‘Glask6rper,’” or “vitreous
body,” whilst the hinder layer may be called the “ retina,” or
“retinal body.” As will be seen when we examine the retinal
body more fully, it does not consist of a simple layer of nerve-
end cells, but is complicated by the presence of a large bulk of
nerve-fibres within the eye-capsule, and by the presence of
what is of more importance morphologically, viz. intrusive
connective tissue.

LATERAL AND CENTRAL EYES OF SCORPIO AND LIMULUS. 189

THE EYE-CAPSULE OF OMMATEAL-CAPSULE (fig. 8 d’) is, as
in the case of the lateral eye, a continuation of the well-de-
veloped basement membrane (fig. 8d), which marks off the
hypodermis of the prosomatic shield from the subjacent con-
nective tissue. It is finely laminated, and devoid of nuclei.
The septum (fig. 8 2), which divides the vitreous body from the
rest of the ommateum, is continuous with and part of the cap-
sule. Von Graber has the merit of having discovered this
septal membrane.

The CELLS OF THE VITREOUS Bopy (figs. 8. 9, 110) are closely
similar to those of the general hypodermis, with which they are
in direct continuity ; but they are devoid of pigment.

The long axes of these cells are curiously bent in one portion
of tle vitreous body (see fig. 8). Instead of radiating in lines,
which would meet if continued at some geometrical centre re-
lated to the curved surfaces of the cuticular lens, the vitreous
cells exhibit a bending towards the side marked B in the figure,
which possibly has some relation to the optical axis of the eye.
This seems to lead to the inference that the optical axis differs
considerably from the geometrical axis of the lens; and this
inference is confirmed by the one-sided development of the
hinder part of the ommateum (see figs. 7 and 9).

There is no concretion or formation of refringent substance
in any of the cells of the vitreous body.

The RETINAL BoDy may be divided into the layer of nerve-end
cells and the layer of nerve-fibres—interspaces, and the whole
inner surface of the ommateal capsule being filled in by intru-
sive pigmentary connective tissue.

The NERVE-END CELLS abut upon the septal membrane, which
divides the vitreous body from the retinal body. The opposite
extremity or filamentary extremity of the elongated nerve-end
cells does not come into such close proximity with the ommateal
capsule as in the lateral eyes; a large mass of intracapsular
nerve filaments (fig. 8 7) separates the nerve-end cells from the
capsule. In the lateral eye this mass of intracapsular nerve
filaments does not exist, the nerve filaments perforating the
capsule more immediately than they do in the central eye.

190 PROFESSOR E. RAY LANKESTER AND A. G. BOURNE.

The nerve-end cells are arranged in the central eye in definite
groups of five, more clearly marked than in the lateral eyes,
though not so obviously segregated as in the multicorneal
(polymeniscous) eyes called “compound eyes” in Insects and
Crustacea.

Each group of five cells is entitled, as in the latter case, to
the name proposed by Grenacher of “Retinula.” Each reti-
nula is provided with a five-fluted rhabdom, formed by the
union of the five rhabdomeres, which are produced laterally
each by one nerve-end cell. The nature of these dispositions is
exhibited in the diagrams’ drawn in PI. XI, figs. 14, 16, 17.
A single nerve-end cell with its rhabdomere is shown in fig.
12, where, however, it is drawn of insufficient proportionate
length.

In a view of a horizontal plane (at right angles to the long
axes of the retinulz) it is possible to observe the five-fluted
rhabdoms in optical section when they have the appearance of
five-rayed stars (Pi. XI, fig. 15). This appearance was ob-
served, and described and figured by Von Graber, who appre-
ciated its significance. We thus owe to him, in spite of his
other interpretations which are erroneous, the important dis-
covery that the nerve-end cells of the central eye of Scorpions
are grouped in retinule and possess a compound rhabdom. This
discovery is of great importance, since iu the unicorneal
(monomeniscous) eyes described by Grenacher, whether of
Arachnida or of Insecta Hexapoda, no such segregation of the
nerve-end cells was detected, although the existence of such an
arrangement serves more directly than anything else to con-
nect the structure of so-called simple (monomeniscous) eyes
with that of so-called compound (polymeniscous) eyes.

The INTRA-CAPSULAR NERVE FILAMENTS which are given off
from the filamentary extremities of the nerve-end cells are in
fig. 8 seen to run parallel with the plane of section and issue
from the capsule in groups (nerves) which are placed to the
outer side (8) of the eye. In fig. 9 the section is taken in a
plane which cuts these nerve filaments at right angles to their
long axes, and accordingly they are seen as irregular masses.

LATERAL AND CENTRAL EYES OF SCORPIO AND LIMULUS. 191

The 1NTRUSIVE PIGMENTARY CONNECTIVE TISSUE is a very
important element in the building up of the retinal body,
which has been on the one hand misinterpreted by Von
Graber, and on the other hand overlooked by Grenacher, whose
observations upon the central eye of Scorpions were under-
taken with a view to the controlling of Von Graber’s results.

No intrusive connective tissue (except in the form of blood-
vessels) is described by Grenacher in other monomeniscous
eyes (such as those of Spiders) similar to the central eyes of
Scorpions. It would, perhaps, be worth while searching for it
in those eyes. The structures which we consider as intrusive
connective tissue in the central eyes of the Scorpion may be
compared to the interneural cells of the lateraleyes. Like these
they are pigmentiferous and serve to fill up the spaces between
the several nerve-end cells, and between these and the ommateal
capsule. But whilst we regard the interneural cells as ecto-
dermal in origin, that is, as belonging to the same germinal
layer as the cells of the hypodermis and the great nerve-end
cells, we find reasons for considering the intracapsular pigmentary
tissue of the central eyes of Scorpions as derived from meso-
blast and of the nature of connective tissue.

We have not embryological evidence for this conclusion, and
depend entirely upon the branching, inosculating character of
the pigmentary cells, and upon the analogy of the pigment-
cells surrounding the retinule of the polymeniscous eyes of
Insects and Crustacea, which are very generally held to be of
the nature of connective tissue, as also upon that of the
“‘ packing-tissue”” to be described below in the central eye of
Limulus.

We are by no means anxious to maintain that the more
epithelium-like cells amongst what we are about to describe
as ‘‘intrusive intracapsular connective tissue’ may not be of
distinct origin from other portions of this pigmentiferous frame-
work,and referable to interneural cells of ectodermal nature, but
any such distinctions must be based upon embryological facts,
which we do not possess. In the present state of knowledge
it seems most convenient and justifiable to hold that in the

192 PROFESSOR E. RAY LANKESTER AND A. G. BOURNE.

central eyes of the Scorpions there are no interneural cells of
ectodermal origin, as there are in the lateral eyes, and that
their place is taken by intrusive connective tissue. In any
ease it is by this name that we shall designate a largely de-
veloped pigmentiferous framework, which pervades the hinder
chamber of the ommateal capsule, and has not hitherto been
described in any similar eye.

In fig. 8 (as also in fig. 7), for the sake of clearness in other
details, a large part of the pigmentiferous intracapsular connec-
tive tissue has been omitted. Buta series of epithelium-like
cells (p) and a group of cells resembling adenoid tissue (7) has
been retained in the drawing. All these cells possess before the
de-pigmenting process abundant black pigment granules in
their protoplasm.

The layer of cells (p) closely adheres to the inner wall of the
ommateal capsule, and when the pigment is present gives a
deep black limiting border to the capsule, as shown in fig. 9.
The cells are again seen in the partially de-pigmented prepa-
ration drawn in fig. 10. They may be called the ‘‘ intracapsular
pavement.” At the periphery of the capsule these cells become
continuous with a series of very delicate pigmentiferous
cells, which lie close beneath the capsular septum (7), between
the anterior extremities of the nerve-end cells. These are seen
as flakes of pigment in fig. 9 s, more clearly in fig. 10 s, and
diagrammatically in fig. 11 s and fig. 14s. They may be known
as the “anterior intra-retinular pigment cells.” These cells
are exceedingly thin and delicate, and readily destroyed by the
acid which is used to remove the pigment. It is on this
account that they have escaped the observation of Grenacher,
whilst on the other hand Von Graber has seen their nuclei, and
attributed them not to interstitial cells, but to the nerve-end
cells themselves. These nuclei are undoubtedly the so-called
“anterior nuclei” of Von Graber, which he has seen and
figured with especial clearness in the central eye of Buthus.

A very thin section of the ommateum of the central eye of
Androctonus, which has been but little or not at all acted upon
by acid, shows a second series of pigmentiferous cells similar

LATERAL AND CENTRAL EYES OF SCORPIO AND LIMULUS. 193

to the anterior intra-retinular cells. These lie about the middle
of the length of the retinulz, and are somewhat less compressed
than the anterior cells. They are seen as fusiform patches of
pigment in the section drawn in fig. 9¢. In fig. 10¢ they are
seen more clearly, and in fig. 11 ¢ and fig. 14 they are represented
diagrammatically. These we call the “ median inter-retinular
pigment cells.’ They appear to have given origin to Von
Graber’s “ middle nuclei of the retinal sacs.”

Again, at the base of the retinule, fitting to the rounded
ends of the nerve-end cells is a third series of pigmentiferous
cells, indicated by the letter v in the figures, whilst closely con-
nected with these is a wide-meshed reticulum of pigmentiferous
branching cells (w in the figures), which embraces the bundles
of intra-capsular nerve-fibres, and becomes continuous with
such masses of connective-tissue cells as those marked r in
figs. 8 and 10, and also forms junctions at intervals with the
pigmentiferous pavement cells which line the back of the
ommateal capsule (as shown in fig. 9).

The three series of inter-retinular pigment cells and their
relation to the five nerve-end cells which build up a retinula is
shown diagrammatically in the drawing (fig. 14).

Representations of actual preparations showing these con-
stituents of the retinal body are given in figs. 9 and 10.

CENTRAL EYE OF Evuscorrius ItTaricus, Ros.

The central eye of Euscorpius, the structure of which is
illustrated by the section drawn in fig. 7, requires no special
description. ‘The same elements are present as in Androctonus,
but the cells are relatively of larger size, and consequently
the retinule and rhabdoms are fewer in number.

We have not been able to study the distribution of the pig-
mentiferous tissue of the retinal body so fully in Euscorpius
Italicus as in Androctonus, owing to the treatment to which
the sections were subjected.

It is noteworthy that post-nuclear phaospheres (fig. 7, %)
occur in the nerve-end cells of the central eye of Euscorpius just

VOL, XXIII, —NEW SER, N

194 PROFESSOR E. RAY LANKESTER AND A. G. BOURNE.

as they do in the lateral eye; and occasionally we find pre-
nuclear in place of post-nuclear phaospheres. No phaospheres
occur in the nerve-end cells of Androctonus.

The peculiar shape of the ommateum (as shown in a right
and left vertical section at right angles to the animal’s long
axis) is exhibited in the figure and also the peculiar bending of
the rhabdoms. The drawing is diagrammatic in so far as that
only one plane of nerve-end cells is drawn, whilst the rhabdoms
are represented as uncut. They, of course, really lie each in the
axis of a group of five cells, of which some are omitted from
the drawing for the sake of clearness. Very perfectly pre-
served and well-stained sections, similar to the drawing, were
obtained by the use of nitric acid, followed by borax carmine
(after washing). The inter-retinular pigment cells were, how-
ever, not preserved.

PIGMENT IN THE Optic CELLS OF THE CENTRAL EYEs.

We find it difficult to decide as to whether pigment granules
are ever to be found actually within the nerve-end cells of the
central eyes of Scorpions. Analogy with the nerve-end cells of
the lateral eyes would render it highly probable that the nerve-
end cells in both cases are pigmentiferous, the pigment being
limited to a superficial layer of the cell. At the same time very
thin sections, such as that drawn in fig. 9, seem to show that in
the special instance of the central eye the pigment granules are
not really in the protoplasm of the nerve-end cells, but always in
fine branches and processes of interstitial cells. ‘The question is
one which must be left undecided for the present. It is, however,
important to notice (what will be further described below) that
in Limulus the nerve-end cells of the central eye certainly con-
tain pigment granules within their own proper substance.

CUMPARISON OF THE LATERAL AND CENTRAL EYES OF SCORPIONS
WITH THOSE OF OTHER ARTHROPODS HITHERTO DESCRIBED.

In his great work Grenacher has described the so-called
“ unicorneal ” eyes of several Spiders, the ‘‘ unicorneal”’ eyes
of Insect larve and adult Insects, and the “ multicorneal ” eyes

LATERAL AND CENTRAL EYES OF SCORPIO AND LIMULUS. 195

adult Insects and of Crustacea ; also the “ multicorneal ” lateral
eye of Limulus. In amore recent work Grenacher has described
(‘ Archiv fiir Mikr. Anat.,’ Bd. 18) the eyes of Myriapods.

One of the chief conclusions which appears to us to follow
from Grenacher’s work, when combined with our own observa-
tions on the central eyes of Scorpions, is that the primary dis-
tinction which has to be made amongst the various forms of
Arthropod eyes is not, as has hitherto been maintained, a dis-
tinction into (A) simple or unicorneal eyes, and (B) compound
or multicorneal eyes, but a distinction into (a) eyes with a one-
cell-layered ommateum (i.e. with no vitreous body separated
from the retinal body), and (8) eyes with a two-cell-layered om-
mateum (i. e. with a vitreous body, or layer of cells placed in
front of the retinal body, and usually separated from it by mem-
brane). ‘These may be called respectively Monostichous
and Diplostichous eyes. In both these primary groups it
appears to be possible for the nerve-end cells (see woodcut,
fig. 1) to remain ungrouped—each equal and similar to its
neighbour, as is usual with cells building up cell layers—or,
on the other hand, the nerve-end cells of the ommateum may
segregate and group themselves as Retinule (see fig. 14, Pl. X).
These two conditions we propose to speak of as (1) eyes non-
retinulate, i.e. with the nerve-end cells autonomous, and (2)
eyes retinulate, i.e. with the nerve-end cells segregated.

We have monostichous eyes which are non-retinulate in the
larvee of Insects, according to Grenacher’s descriptions. We
have a monostichous eye which is feebly retinulate in the
lateral eye of the Scorpion, and a highly developed retinulate
monostichous eye in the lateral eye of Limulus. In the dorsal
eyes of Spiders and simple eyes of adult Insects we have ex-
amples of (according to Grenacher’s description) a diplostichous
non-retinulate eye. In the central eyes of Scorpions, on the
other hand, we find for example of the strongly retinulate
diplostichous class.

Further, the so-called compound eyes of Insects and Crus-
tacea are examples of diplostichous retinulate eyes, with cer-
tain additional peculiarities now to be noticed.

196 PROFESSOR E. RAY LANKESTER AND A. G. BOURNE.

' Subsequently, as it would seem, to the segregation of the
layer of nerve-end cells into retinule, optical advantage may be
found in the segregation of the cuticular lens often called cornea.
Thus, from the so-called unicorneal the multicorneal condition
of eye is developed. It seems to be undesirable to speak of the
cuticular lens as a *‘ cornea,” with which it has little analogy,
if by cornea we understand in the first instance the vertebrate
cornea. It will be therefore best to distinguish the simple
one-lensed eyes as ‘‘ monomeniscous,” that with a segregated
lens as “ polymeniscous.” This third alternative of structure
is presented: by Arthropod eyes, which differ in other respects
inter se, but it seems that a non-retinulate eye cannot be
polymeniscous, since the segregation of retinule is the deve-
lopmental antecedent of the segregation of the lens. Hence
we may have and actually can point to monostichous polyme-
niscous eyes (lateral eyes of Limulus) as well as diplostichous
polymeniscous eyes, but all non-retinulate eyes are monome-
niscous.

In this way we are led to correct the conception of the so-
called compound or polymeniscous eye which is current, and
and is thus enunciated by Gegenbaur (‘ Comp. Anatomy,’
English translation, p. 266) : “ A reduction of the retinal ele-
ments of the simple eye gives rise to the retinula, and a com-
pound eye is formed by the gradual concrescence of a number
of simple eyes.” On the contrary, it seems that the compound
eye is formed, not by the gradual concrescence of a number of
simple eyes, but by the segregation of the elements of a simple
eye, which affects first the retina and then the lens. It
appears from a consideration of the structure of the polymenis-
cous lateral eyes of Limulus that even the groups of mono-
meniscous lateral eyes found in Scorpions (and similarly, also,
the closely-set groups of monomeniscous eyes of Myriapods,
which can in some cases be regarded as one large polymeniscous
eye) must be looked upon as resulting from a process of segre-
gation carried further than is necessary for the production of a
compound eye—carried so far, in fact, as to produce from one
original large simple (monomeniscous) eye, not a continuous

LATERAL AND CENTRAL EYES OF SCORPIO AND LIMULUS. 197

polymeniscous eye, but a number of contiguous secondary
simple (monomeniscous) eyes.

The polymeniscous eye presents various elaborations of
structure. In the lateral eye of Limulus we have a monostichous
polymeniscous eye of relatively simple character. On the other
hand, polymeniscous eyes which have developed by the differen-
tiation of diplostichous monomeniscous eyes exhibit the highest
elaboration of structure known in the Arthropod series. In the
first place the retinulee become exceedingly well defined, and
separated from one another by intrusive connective tissue; that
is to say, by connective tissue which, not belonging originally to
the hypodermis layer from which the nerve-end cells and vitreous
body-cells are developed, yet pushes its way in amongst these
elements. This intrusive connective tissue is pigmentiferous.

We have seen it already making its appearance (figs. 9, 10,
11, 14) in the monomeniscous retinulate central eye of Scor-
pions. It is more strongly developed in the typical ‘* compound
eye,’ and serves to isolate very completely each retinula from
its neighbours.

A further speciality of that higher form of polymeniscous eye
known as the compound eye (of Insects and Crustacea) is the
segregation of a third element of the eye in addition to the
segregation of retinule and lens facets; this third element is
the vitreous body. Whilst in the monomeniscous diplostichous
eye, even when retinulate (central eye of Scorpion) the vitreous
body remains homogeneous, consisting of uniformly distributed
columnar cells, in the higher form of polymeniscous diplo-
stichous eye (compound eye of Insects and Crustacea), the
vitreous body, as might be expected, joins in the segregation
which characterises the cuticular lens formed upon it. Not
only do we find the cells of the vitreous body arranging them-
selves in isolated groups similar to, and super-imposed each
upon a retinula, but the intrusive connective tissue advances
into the vitreous layer and cuts off with its pigment cells each
group of vitreous cells from its neighbour, in the same way as it
separates neighbouring retinulz.

Each group of ‘vitreous cells corresponding to a retinula

198 PROFESSOR E. RAY LANKESTER AND A. G. BOURNE.

should be called a “vitrella.” Just as the retinule develop
axial hard structures called “rhabdoms ” by Grenacher, so do
the vitrelle in varying degree develop dense hyaline bodies
within them known as “ crystal-cones.”

The preceding sketch of the degrees of complication of
structure presented by Arthropod eyes which the reader will
find more readily intelligible if reference be made to the plates
of Grenacher’s large work, serves to enable us to estimate cor-
rectly the morphological significance of the two kinds of eye
present in the Scorpions.

In the lateral eyes we have an example of the simplest kind
of Arthropod eye, the monostichous monomeniscous eye. It
may be compared with the simplest eye of this kind studied by
Grenacher, namely, that of the larva of an Insect (Dytiscus).
The woodcut reproduces diagrammatically one of Grenacher’s

Fie. 1.—Eye of larva of Dytiscus, with monostichous, non-retinulate apo-
static ommateum. /. Lens. g. Perineural cells (rudimentary vitreous body).
p. Pigment cells. 7. Nerve-end cells. o. Filaments of the optic nerve.

figures of this eye. As in the Scorpion’s lateral eye, we find a
single row of cell elements continuous with the hypodermis, of
which the lateral members (p. g.) are narrow and columnar,
whilst the more median are connected with nerve filaments and
differentiated as nerve-end cells (7).

As compared with the eye of the larval Dytiscus, the lateral
eye of the Scorpion is in one respect more primitive, in other
respects more elaborate. It is more primitive in this, viz. that
the row of cells forming the ommateum is in continuous con-
tact with the cuticular lens, whereas in the larval Dytiscus eye

LATERAL AND CENTRAL EYES OF SCORPIO AND LIMULUS. 199

the ommateum is deeply cupped leaving a tubular cavity imme-
diately below the cuticular lens. This condition isa step to-
wards the complete pinching in of the perineural cells of the
ommateum, and their separation as an anterior ‘‘ vitreous
layer’? from the deeper lying nerve-end cells.

No Arthropod eye has as yet been described which is so
strictly “‘ monostichous ” as the lateral eye of the Scorpion,
that is to say, which presents so little evidence of any tendency
of the perineural cells to take upa position in front of the nerve-
endcells. The position of the ommateum in relation to the lens
in the lateral eye of the Scorpion is more nearly like that of
the ordinary hypodermis cells in relation to their cuticle, and
may be called ‘ epistatic,” whilst the monostichous eyes de-
scribed by Grenacher (in Myriapods as well as in Insect
arve), are all characterised by a tubular cupping of the om-
mateum, which may be called “ apostatic.”

The relation of the long axes of the cells of the onmateum
to the geometrical (and optical) axis of the lens is widely different
in the two cases. Similar facts as to the direction of the axes
of the cells of the ommateum, which are not unfrequently to
be observed in the eyes of Arthropods, must be taken into
account in any attempt at an explanation of the Arthropod
eye as an optical apparatus; but we are not at present in a
position to make such attempt.

The features in which the lateral eye of Scorpions is more
elaborate than that of the larval Dytiscus are (1) the exist-
ence of interneural cells in the former, and (2) the tendency in
the former of the rhabdomeres of neighbouring nerve-end cells
to unite as rhabdoms.

The two agree in the possession of a well-marked ommateal
capsule continuous with the basement membrane of the hypo-
dermis, and in the ** purity” of the ommateum, that is to
say, its freedom from intrusive connective tissue. Since such
intrusive connective tissue, when it does enter into the omma-
teum, appears to enter there with the function of a pigmentary
investment to the optical elements, we may call an ommateum
which is devoid of such adventitious pigmentiferous tissue

200 PROFESSOR E. RAY LANKESTER AND A. G. BOURNE.

“ autochromic,”’ whilst one which is penetrated by pigmenti-
ferous connective tissue is “‘ exochromic.”’

The central eye of the Scorpion is seen from what has been
said above to take a position among Arthropod eyes very dis-
tinct from that occupied by the lateral eyes. It is diplostichous,
it is definitely retinulate, it is exochromic, and only agrees in
fact with the lateral eye in being monomeniscous. As com-
pared with all other Arthropod eyes hitherto described, the
central eyes of the Scorpion stand alone in being definitely
retinulate, whilst retaining the primitive monomeniscous
character of lens.

It is highly noteworthy that the central eyes of Scorpions,
which in position and naked-eye appearance agree with the
centrally grouped eyes of the Scorpion’s congeners, the Spiders,
should nevertheless present such marked differences from those
eyes as appear from a comparison of the description given in
this memoir of the one and by Grenacher of the other. The
Spiders have apparently simple axial rhabdomeres in their nerve-
end cells in place of laterally developed rhabdomeres, uniting
to form rhabdoms, as in Scorpions. Further, it would appear
from Grenacher’s description that the Spider’s eyes are auto-
chromic, whilst the Scorpion’s central eyes are exochromic—
i.e. the ommateum is penetrated by intrusive connective
tissue.

Having thus noted the peculiarities of the Scorpion’s lateral
and central eyes, we are in a position to compare them with
the lateral and central eyes of Limulus, of which we shall
now give a description. It will be seen that though differing
in some important details, there are, taking all things into
consideration, no Arthropod eyes which so closely agree in
their plan of structure with those of the Scorpions as do those
of Limulus. Whilst this conclusion might be impugned were
we to separate the consideration of the two kinds of eyes, the
lateral and the central, it is indisputable if we compare the
whole set of optical organs in the one animal with the whole
set in the other.

LATERAL AND CENTRAL EYES OF SCORPIO AND LIMULUS. 201

Tue Laterat Eyres or THE AMERICAN Kina Cras,
LIMULUS POLYPHEMUs, Latr.

The two lateral eyes of the King Crab exist on either side
of the prosomatic shield, each as a reniform, smoothly polished
protuberance. The protuberance is the cuticular lens; it is of
great thickness, smooth on its surface, and produced on its
inner surface into a number of conical processes, each of which
is to be regarded as a secondary or segregate lens. The general
features of the structure of this eye have been made known by
Grenacher; we have only to add certain details to the de-
scription and figures given by him.

In Pl. XI, fig. 18, a diagram is given representing a portion
of the polymeniscous lens with its subjacent ommateum.

Making use of the terminology which has been explained in
the course of this paper, we can briefly describe the structure
of the ommateum. Corresponding to each conical facet of the
polymeniscous lens is a retinula (Az). The ommateum is (as
observed by Grenacher) essentially monostichous. The reti-
nulz correspond in position to the apices of the conical
secondary lenses.

That part of the ommateum which clothes the sides of the
lens-cones consists of simple cylindrical cells corresponding to
the perineural cells of the lateral eye of Scorpions, and in the
valleys between neighbouring lens-cones, the ommateal cells are
not to be distinguished from ordinary pigmentiferous hypo-
- dermis cells.

The nerve-end cells, which are combined to form retinule,
are of very large size, as much as ,4,th long. Transverse sec-
tions of the ommateum show that ten cells are united in each
retinula. The difficulties of observation are here even greater
than in the eyes of Scorpions (owing to the thickness and
density of the cuticular lens, which prevents the preparation of
satisfactory sections); but weare inclined to think that tennerve-
end cells, being double the number present in the Scorpion’s
retinula, is the rule, although Grenacher has figured a retinula

202 PROFESSOR E. RAY LANKESTER AND A. G. BOURNE.

of Limulus with as many as fifteen nerve-end cells. Possibly
there isa superior and inferior series of retinula cells interlock-
ing with one another, and amounting to fifteen in number,
which might in certain sections show ten in others, fifteen areas
in section; but we have not obtained definite evidence that such
is the case, and are disposed to consider the retinula as having
the arrangement of cells shown in fig. 19 and in fig. 20. The
nuclei of some of the retinula cells lie nearer to the lens
extremity, in others nearer to the filamentary extremity. Each
retinula cell gives off a coarse nerve filament from its filamentary
extremity.

The rhabdom of the retinule of the lateral eyes of Limulus
is formed by the union of ten rhabdomeres, as shown diagram-
matically in fig. 20. That end of the retinula nearest the lens
touches the conical extremity of the secondary lens, but leaves
an axial space, which is filled neither by the lens nor by the
rhabdom (zy). The rhabdom itself is hollow in its more ante-
rior portion, the constituent rhabdomeres only thoroughly
uniting along the common axis in the deeper region of the
retinula. This is seen in the transverse sections of three reti-
nule at different horizons drawn in fig. 26, where the section
with hollow rhabdom is more anterior (that is, nearer the lens)
than is that to the left with solid rhabdom.

The perineural cells (fig. 19 f) are delicate columnar cells,
much elongated where they adjoin the retinula, and deeply
charged with pigment. They pass over in the lateral regions of
the lens-cones into ordinary pigmentiferous hypodermis cells (¢).

Vertical sections through a lens-cone and subjacent omma-
teum (de-pigmented as a matter of course) show, besides the
retinula cells and the perineural cells, small cells, which are dis-
posed upon and between the adjacent large nerve-end cells of the
retinula. At first sight these might be interpreted as perineural
cells, but their position and distribution do not seem to admit
of this view of their nature. They are apparently intrinsive
connective tissue, which enters the ommateal capsule with
the large group of nerve-fibres attached to the retinula cells.

These cells are seen in fig. 19 and fig. 22, where they are

LATERAL AND CENTRAL EYES OF SCORPIO AND LIMULUS. 203

coloured pink to distinguish them from the perineural cells of
hypedermic origin, and are marked with the letter 7. In trans-
verse sections, owing to the delicacy of the structures and the
destructive action of the acid used to remove the obscuring
pigment, great difficulty is found in tracing the perineural cells
and these intrusive connective-tissue cells. In fig. 26 the
letters f and r indicate what appear to belong to these two sets
of cells respectively.

The whole question of the distribution of pigment granules
in the three sets of cells, viz. nerve-end (retinular), perineural,
and intrusive, is bound up with the proper distinction of the
distribution of the last-named cells.

The nerve-end cells undoubtedly contain some pigment
granules both in the lateral and central eyes of Limulus. The
perineural cells are intensely pigmentiferous. But it is probable
that the chief clothing of pigment to each nerve-end cell—for
example, that which is seen in the right-hand retinula of the
section fig. 26—is furnished by the intrusive connective-tissue
cells disposed between neighbouring retinula-cells.

Connected with the fact that the ommateum of the lateral
eye of Limulus is invaded by intrusive connective tissue, is the
incomplete character of the ommateal capsule. Whilst well
marked in every other region (figs. 19, 25, 26 d), the capsule
is deficient immediately below the retinula where the group of
optic nerve filaments passes out of or into the capsule, and it
is here that the intrusive connective tissue (7) is seen to be
continuous with the extra-capsular connective tissue (e), as
shown in fig. 19.

CoMPARISON OF LATERAL EYE oF LIMULUS WITH THE
LATERAL EyEs OF ScorPION.

The lateral eye of Limulus is shown above to be monosti-
chous, polymeniscous, exochromic (i.e. not purely autochromic),
and seems, therefore, at first sight, to differ largely from a lateral
eye of a Scorpion.

But, as was stated at the commencement of this memoir,

204 PROFESSOR E. RAY LANKESTER AND A. G. BOURNE.

it has been suggested by Prof. Lankester that the comparison
to be made is not between a single lateral eye of the Scorpion
and a whole lateral eye of Limulus, but between the latter and
the complete group of lateral eyes occurring in Scorpions.

When the comparison is thus made, we see that if we sup-
posed a common ancestor of the Scorpion and King Crab to
have exhibited a lateral “ocular area,” which possessed a
single feebly developed cuticular lens, then by two slightly
divergent lines of differentiation we can obtain the grouped
eyes of Scorpio on the one hand, and the polymeniscous eye of
Limulus on the other hand.

The ancestral eye was undoubtedly monostichous, an archaic
character which is retained by both descendants. ‘The archaic
eye was at first non-retinulate, and commenced to exhibit a
tendency to retinulation before the actual divergence of the
Scorpionids and the Limuloids. In the Limuloids the differen-
tiation of retinule and the corresponding differentiation of lens-
cones (facets) became definite and characteristic. It is pro-
bable, if we may judge from the condition of the extinct
Eurypterina, that the lens-cones were at first relatively larger
and shallower and the retinule less concentrated (composed of
more numerous cells) in the Limuloids than they subsequently
became.

In the Scorpionids the segregation of ommateum and lens
took different proportions, The original lens segregrated, not
into a number of contiguous lens-cones, but broke up into a
number of quite separate lenticules, to each of which a portion
of ommateum corresponded. ‘This process was no doubt a very
gradual one, and was essentially determined by the fact that
well-marked retinule did not develop themselves as optical units
in the ommateum of the lateral eye of these ancestral Scor-
pions. Probably the secondary eyes of these Scorpionids were
at first much more numerous and more closely set than in living
Scorpions. Gradually they have become reduced in number
and more widely separated from one another. At the present
day various genera of Scorpions differ in the number of eyelets
present on the lateral ocular area (from two to seven). It is also

LATERAL AND CENTRAL EYES OF SCORPIO AND LIMULUS. 205

important to notice that the number varies even in individuals,
and that supernumerary eyelets of small size are distinguished
as of irregular occurrence by the side of the larger eyelets, thus
seeming to indicate that we have in this region of the Scor-
pion’s prosomatic shield an “ ocular area,’’? which by reversion
may occaionally reproduce the more numerous and closely set
eyelets into which the single ancestral eye of the common
parent of Scorpio and Limulus was divided in the Scorpionid
line of descent.!

THE CEenTRAL Eyes oF THE AMERICAN Kina Cras,
LIMULUS POLYPHEMUs, Latr.

The central eyes of Limulus have not hitherto been ex-
amined. Grenacher appears not to have had an opportunity
of studying them, although he has figured sections of the
lateral eyes. Packard has published some figures relating to
them, which are valueless, because he did not attempt to re-
move the pigment which necessarily obscures their structure.

Hence we have here a perfectly novel subject to deal with.

These eyes are very difficult to study on account of the great
strength and thickness of the lenses and adjacent cuticle. The
ommateum must necessarily be drawn away from the lens in
the fresh state, and then hardened and cut. Our conclusions
are founded on the examination of the central eyes of four fresh
specimens of Limulus.

The anticipation which immediately forces itself on the
mind is, that the central eye of Limulus will prove—if the
assimilation of Limulus and Scorpio be justified—to be, like
that of Scorpio, diplostichous, monomeniscous, and retinulate,
with a more or less abundant intrusive connective tissue in the

' It is important to note the following difference between the lateral eyelet
of a Scorpion and a single element of the King Crab’s lateral eye—in the
former the ommateum contains more than one retinula, it is retinulate, in
the latter it contains but one group of nerve-end cells, truly a retinula when
the whole eye-group is considered but in itself non-retinulate. Thus the
eyelet of the Scorpion is morphologically more (a larger segment of the
original ocular area) than the lens-cone element or eyelet of Limulus.

206 PROFESSOR E. RAY LANKESTER AND A. G. BOURNE.

ommateum (exochromic). This is precisely the character which
is revealed by our sections. In fig. 27 an approximately
median vertical section of one of the two central eyes of
Limulus is represented. The lens is coloured yellow, and the
intrusive connective tissue pink. It is at once clear from the
figure that we have an anterior layer of cells (vitreous body)
(0), separated by firm membrane from a posterior retinal body.
The vitreous body is peculiar in the fact that its cells are not
specially elongated, but are small hke those of the adjacent
hypodermis. The retinal body is highly remarkable. Though
identical in plan of structure with that of Scorpions, it differs
in the exaggerated development of the intrusive connective
tissue, which forms an abundant growth both in front of (s s)
and around (sz) the retinule (2). The individual nerve-end
cells are very large, exactly correspouding to those which form
the well-defined retinule of the lateral eyes. Large groups of
nerve-fibres (mm), surrounded by connective tissue, are seen at
the base of the section passing away from the retinule to form
the optic nerve. We could not define an ommateal capsule,
though there is a perceptible difference between the tissue of
the ommateum and the loose reticulate connective tissue (e),
which is abundantly developed around the eye and in all
other parts of the King Crab’s body.

The retinule are much less definitely constituted in the
central eyes of Limulus than are those of the lateral eyes.
The nerve-end cells are sometimes grouped quite irregularly, but
here and there a definite arrangement of five around a com-
mon axis can be observed (fig. 32, Ret. 2). Those towards the
periphery are less defined, those nearer the centre of the omma-
teum more clearly differentiated. In teasing a fresh ommateum
of the central eye a retinula was isolated, which is diagram-
matically represented in fig. 28. It appeared to be built up
of seven nerve-end cells, and was only partially disengaged from
the adherent connective tissue. In fig. 33 a drawing is given
showing nerve-end cells, and surrounding pigmentiferous con-
nective tissue. The action of nitric acid has dissolved some of the
pigment contained in the nerve-end cells, and stained the proto-

LATERAL AND CENTRAL EYES OF SCORPIO AND LIMULUS. 207

plasm of these cells with the madder-brown colour formed by
the dissolved pigment.

The rhabdoms of the retinule are very irregularly developed.
Those nearer the centre show, however, under favorable cir-
cumstances in transverse section, a five-fluted or seven-fiuted
thabdom, as represented in figs. 29, 30.

It is difficult to estimate the number of the retinule present
in a central eye of Limulus, on account of the want of definite
segregation of the nerve-end cells at the periphery of the omma-
teum. But indications of about twenty are seen when a com-
plete horizontal section is studied.

The most remarkable feature in the central eye of Limulus is
the great development of intrusive connective tissue in
front of and around the retinule. Much of this connective
tissue is pigmentiferous, but we must suppose that certain
tracts of it in front cf the retinul are free from pigment, and
leave a path for the rays of light. Nevertheless we are not
able to say with any certainty which cells are pigmentiferous
and which cells are transparent. The large vesicular cells in
front of the retinule, marked ss, are probably free from
pigment.

The branched and fusiform cells, which are seen everywhere
in sections creeping over and clothing the retinule (sz), are
undoubtedly pigmentiferous.

CoMPARISON OF CENTRAL EYE or LimuLuUs AND ScorPIons.

The great mass of connective tissue present in the ommateum
of the central eye of Limulus is to be regarded as a develop-
ment of the intrusive connective tissue which we have already
seen in the central eye of Scorpions. It is so largely developed in
Limulus as to lead one to regard the retinule as sunk and
buried in it, and suggests the possibility that, at any rate in
the adult, the central eye of the King Crab may have partially
lost its function, At any rate, the irregularly constituted re-
tinule and the abundant connective tissue of the central eye
contrast markedly with the cleanly cut retinule and simple
perineural cells of the lateral eyes.

208 PROFESSOR E. RAY LANKESTER AND A. G. BOURNE.

Although there is this strange exaggeration in parts, the essen-
tial agreement of the central eyes of Limulus with those of
Scorpions is obvious. No other known eye approaches it in
constitution. This agreement is more marked than that of the
lateral grouped eyelets of Scorpions with the lateral polymenis-
cous eye of Limulus, though there is no parallel to the mono-
stichous eyelet of the Scorpion with its “ epistatic” ommateum
to be found so close as that presented by a single element of the
King Crab’s lateral eye, also monostichous and epistatic.

When the two kinds of eyes are compared in the two
animals, each to each, we are abundantly justified in saying
not only that there is no other animal which presents so close
an approach to either of these animals in respect of its eyes as
they do to one another, but even more emphatically that the
agreement is one comprising such a large number of important
details that we are compelled to conclude from it that the
Scorpions and the King Crabs are closely allied representatives
of one class, the Arachnida.

SUMMARY AND TABULAR STATEMENT.

The facts set forth in the preceding memoir are to a large
extent summarised by the plates which accompany it, and the
explanatory description of those plates.

The more general conclusions to which our observations
tend may be gathered from the following tabular arrangement
of some of the chief varieties of Arthropod eyes which are at
present known. The technical terms which have been intro-
duced in the present memoir and recur in the tabular state-
ment are explained in a list appended.

Without assuming to assign their due significance to all
varieties of the Arthropod eye, some of which may very pos-
sibly (e. g. those of certain Crustacea) find no proper place in
the scheme here submitted, we yet think that it is useful to
tabulate the principal facts of structure known as to a large
number of Arthropod eyes in the following way.

We take as primary divisions those eyes with monostichous

209

PROFESSOR E. RAY LANKESTER AND A. G. BOURNE.

‘(lun @ sR palap
-IsU09 UNA} eULMWO
pue sudo apo
-tis B®) sn[ne pue
etpuadoloag

“qr
® St palopisuoo
syataXa yo dnoas 10
Bale IB{NIO BOTTA
ay}) sujne pue
eipuadoloos

‘(tun 8 se pasop

-ISM09 Bare IeyNIO
doy 9173) saAnT
[elojey ‘snp nwiy

eee

* quasoid
* * quasqe

(o1moryooxe)
quasoid
(o1mosyooyne)
quasqe

* orjeysode
* roryeysida

* ST[90 [Ra
-nau1aqut
[eu1apojoq
* * ons
-ST} OAT}
-990U09
DAISNIYUT

} UINIyeTUAWC)

* snoosiuam {jog
SnoosTuaWOUOT

* aqeynuryoy
948] NUI4o1-10 Ny

‘(tun e sR
patepisuos wna,
-RMmO pur 900

sla] apsurs v) sokgy
[etoqe'y ‘snp nwmry

‘(q1an B SB
parapisuos yataoa
a[2uls ®) Sadoy [er
-ayeT ‘uotd100g

' “(TIN B SB
palapisuos syataXo
jo dnois 9joym
ayy) sok [er
-ayeyT ‘awordi00g

*(BUN94S
a[BuIg) wBaiery
yoasuy 191140
pue snosiyig

‘SUAW] SQOHOILSONOPY—'] dnosy

°

NEW SER

XXIII.

VOL,

210 LATERAL AND CENTRAL EYES OF SCORPIO AND LIMULUS.

_*(samo9-[eysf10) vaoeysniy

piv syoasuy qyyoq jo sada punodwoa a4} Jo &ifaxjtA 94} UL SB ][aM su ‘sJoasUT 9UIOS Jo vyeUIUIAys ayy Jo {poq snoasytA

9114 TL SAnvdo doUBYsqus Jo sisoydiowryou yeorsAqd styy,

‘21q"4} 9} UL pavolWaU YOU ST ‘aI[OIJTA WAOJ 04 s][90 944

JO WolRSaisas S41 Woy Javde “(umayemMo jo roXvy dorojue) Lpoq snoa.41A 944 UT doULYSqns 4ueSuTAya4 Jo uOTyIsodep e1L—a' N

“(q1um v sR
porepisuos wma
-emmo Suipuods
-91109 PUR 49IvJ

a] SUIS v) VOORISNID
pue spodexoyy
jo aA payqooniy

: Sie ra * * *
ms Fe x an tee
wee * * * *
“(run
® SB pacapisiod ) mm94g
Baie LeTNIO ATOM ‘saXq yerquag ‘saXor eamag 1 “BUTI 4S
vaoRsnig pue ‘SUIMWUT ‘otda100g yoosuy ‘saapidg

yoosuy podexa
jo aq puno
-W0D FO paqjaoey

‘SHAQ SQOHOILSOTAIQQN—'J] dno

podexay ynpy

* "ee UOTE
WO} 0} payesaises ‘Apoq snoaizt,

- + + + + giduns ‘poq snooayt,

eS (ormoaqooxe)) , , aussry
quasoad :
(otmMoIyoo4yneB) ies i
ate : urd
quasqe a

Moth ee te * snoostuamtsjog
7 5 + = <0) = “Snoosmemenogy

Sot ee ee 8 e+ ggemeayy
24B[NUT}AI-10 NJ

. . . . . . .

LATERAL AND CENTRAL EYES OF SCORPIO AND LIMULUS. 211

ommateum, the more archaic, and those eyes with diplos-
tichous ommateum derived from the monostichous condition.
A few examples clearly transitional between the monostichous
and the diplostichous condition have been described by Gren-
acher (among Myriapods).

It is especially to be noted in reference to the comparison of
monomeniscous and polymeniscous eyes, that the comparison
yields totally different results accordingly as we may choose to
compare with the non-facetted eye of a Spider or of an Insect
larva, on the one hand, a single eyelet of a ‘‘ grouped ” eye or
of a “ compound” eye, or on the other hand, the whole group
or the whole “ compound ”’ eye.

Explanatory List of Terms introduced in this Memoir.

Ommatéum.—aAll the soft tissues of an Arthropod eye, as distinguished from
the cuticular lens.

Nerve-end cells.—The cells of the ommateum, in which the filaments
of the optic nerve terminate.

Perineural cells.—Cells having the same ectodermal (hypodermis) origin as
the nerve-end cells, and surrounding a group of the latter; the rudimentary
vitreous body and marginal cells of monostichous eyes.

Interneural cells.—Cells having the same ectodermal (hypodermis) origin
as the nerve-end cells, but remaining small and unrelated to nerve fila-
ments whilst wedged in between the bases of adjacent nerve-end cells.
Only known as yet in the lateral eyes of Scorpions.

Monostichous.—Of an ommateum which consists of a single layer of cells.

Diplostichous.—Of an ommateum which consists of two layers of cells,
one superimposed on the other.

Vitreous body.—tThe anterior cell-layer of a diplostichous ommateum.

Retinal body.—The posterior, or deep-layer, of a diplostichous ommateum,
often, but not always, separated from the anterior by a septal membrane.

Retinulate.—Of an ommateum in which the nerve-end cells are segregated
to form definite groups, or ‘‘ retinule.”

Rhabdomere.—The axial or lateral hard-piece which is frequently formed
by a nerve-end cell in front of its nucleus.

Rhabdom.—The compound hard-piece formed in the axis of a retinula by
the union of the laterally formed rhabdomeres of its constituent nerve-end
cells.

212 PROFESSOR E. RAY LANKESTER AND A. G. BOURNE.

Phaosphere.—A brilliantly refringent spherical body found in the nerve-end
cells of Euscorpius italicus, Ros, usually behind, but sometimes in
front of the nucleus.

Monomeniscous,—Of the chitinous cuticle in front of an ommateum when
it has the form of a single lens.

Polymeniscous.—Of that condition of the chitinous cuticle in front of an
ommateum when it is segregated or broken up into many lenses, only
found, and then not always, when there is a retinular segregation of the
ommateum itself.

Epistatic.—Of a monostichous ommateum when the anterior ends of its
nerve-end cells abut upon the cuticular lens. Only observed in the lateral
eyes of Scorpio and Limulus.

Apostatic.—Of a monostichous ommateum when it forms a cup or tube-like
depression below the cuticular lens, the lens not entering the lumen of
the cup or tube. —

Autochromic.—Of an ommateum in which all the pigment is developed in
cells of ectodermal (hypodermis) origin, and into which no connective
tissue penetrates.

Exochromic.—Of an ommateum in which some, if not all, the pigment is
developed in intrusive connective-tissue cells, which penetrate deeply be-
tween the proper ectodermal elements of the ommateum.

Vitrella.—A group of cells of a vitreous body which has become segregated
in correspondence with the segregation of the retinal body and of the
lens. A lens-facet, a vitrella and a retinula surrounded by pigmentiferous
cells constitute a single “element” of the compound eye of Insecta
hexapoda and of Crustacea.

The Anatomy and Development of Peripatus
capensis.

By the late

Francis Maitland Baifeur, LL.D., F.R.S.,

Fellow of Trinity College, Professor of Animal Morphology in the University
of Cambridge.

With Plates XIII—XX.

INTRODUCTION.

Tue late Professor Balfour was engaged just before his death
in investigating the structure and embryology of Peripatus
capensis, with the view of publishing a complete monograph
of the genus. He left numerous drawings intended to serve
as illustrations to the monograph, together with a series of
notes and descriptions of a large part of the anatomy of
Peripatus capensis. Of thismanuscript some portions were
ready for publication, others were more or less imperfect ;
while of the figures many were without references, and others
were provided with only a few words of explanation.

It was obviously necessary that Professor Balfour’s work—
‘embodying as it did much important discovery—should be pub-
lished without delay; and the task of preparing his material
for the press was confided to us. We have printed all his
notes and descriptions without alteration.’ Explanations which
appeared to be necessary, and additions to the text in cases in
which he had prepared figures without writing descriptions, to-
gether with full descriptions of all the plates, have been added
by us, and are distinguished by enclosure in square brackets.”

1 Excepting in an unimportant matter of change of nomenclature used
with regard to the buccal cavity.

2 The account of the external characters, generative organs, and develop-
ment, has been written by the editors.

VOL, XX111, —NEW SER, P

214 PROFESSOR F. M. BALFOUR.

We have to thank Miss Balfour, Professor Balfour’s sister,
for the important service which she has rendered by preparing
a large part of the beautiful drawings with which the mono-
graph is illustrated. Many of these had been executed by her
under Professor Balfour’s personal supervision ; and the know-
ledge of his work which she then acquired has been of the
greatest assistance to us in preparing the MSS. and drawings
for publication.

Since his death she has spared no pains in studying the
structure of Peripatus, so as to enable us to bring out the
first part of the monograph in as complete a state as possible.
It is due to her skill that the first really serviceable and accurate
representation of the legs of any species of Peripatus available
for scientific purposes are issued with the present memoir.t

We have purposely refrained from introducing comments on
the general bearing of the new and important results set forth
in this memoir, and have confined ourselves to what was
strictly necessary for the presentation of Mr. Balfour’s dis-
coveries in a form in which they could be fully comprehended.

Mr. Balfour had at his disposal numerous specimens of
Peripatus nove zealandie, collected for him by Professor
Jeffrey Parker, of Christchurch, New Zealand ; also specimens
from the Cape of Good Hope collected by Mr. Lloyd Morgan,
and brought to England by Mr. Roland Trimen in 1881; and
others given to him by Mr. Wood Mason, together with all the
material collected by Mr. Moseley during the Challenger voyage.

A preliminary account of the discoveries as to the embryology
of Peripatus has already been communicated to the Royal
Society.? It is intended that the present memoir shall be fol-
lowed by others, comprising a complete account of all the

species of the genus Peripatus,
H. M. Mosetey.

A. SEDGWICK.

1 The drawings on Pl. XIV, figs. 9 and 10 on Pl. XV, and the drawings of
the embryos (except fig. 37), liave been made by Miss Balfour since Professor
Balfour’s death.

2 ¢ Proc. Royal Soc.,’ 1883.

ANATOMY AND DEVELOPMENT OF PERIPATUS CAPENSIS. 215

PARE .:
DEscRIPTION OF THE SPECIES.
Peripatus capensis (fig. 1).

[The body is elongated, and slightly flattened dorso-
ventrally. The dorsal surface is arched, and darkly pig-
mented; while the ventral surface is nearly flat, and of a
lighter colour.

The mouth is placed at the anterior end of the body, on the
ventral surface.

The anus is posterior and terminal.

The generative opening is single and median, and placed
in both sexes on the ventral surface, immediately in front of
the anus.

There are a pair of ringed antenne projecting from the an-
terior end of the head, and a pair of simple eyes, placed on
the dorsal surface at the roots of the antenne.

The appendages of the body behind the antenne are disposed
in twenty pairs.

1. The single pair of jaws placed within the buccal cavity
in front of the true mouth opening, and consisting each of a
papilla, armed at its termination with two cutting blades.

2. The oral papille placed on each side of the mouth. At
their apices the ducts of the slime glands open.

3. The seventeen pairs of ambulatory appendages, each
provided with a pair of chitinous claws at its extremity.

4, The anal papille placed on each side of the generative
opening.

Colour.—The following statements on this head are derived
from observations of spirit specimens. The colour varies in
different individuals. It always consists of a groundwork of
green and bluish grey, with a greater or less admixture of
brown. The chief variations in the appearance of the animal,
so far as colour is concerned, depend on the shade of the green,

216 =C- PROFESSOR F. M. BALFOUR.

In some it is dark, as in the specimen figured (fig. 1); in
others it is of a lighter shade.

There is present in most specimens a fairly broad light band
on each side of the body, immediately dorsal to the attachment
of the legs. Whis band is more prominent in the lighter
coloured varieties than in the dark, and is especially conspicuous
in large individuals. It is due to a diminution in the green
pigment, and an increase in the brown.

There is a dark line running down the middle of the dorsal
surface, in the middle of which is a fine whitish line.

The ventral surface is almost entirely free from the green
pigment, but possesses a certain amount of light brown. This
brown pigment is more conspicuous and of a darker shade on
the spinous pads of the foot. —

In parts of the body where the pigment is scarce, it is seen
to be confined to the papillae. This is especially evident round
the mouth, where the sparse green pigment is entirely confined
to the papille.

In some specimens a number of white papille, or perhaps
light brown, are scattered over the dorsal surface ; and some-
times there is a scattering of green papillze all over the ventral
surface. These two peculiarities are more especially noticeable
in small specimens.

Ridges and Papille of the Skin——The skin is thrown into a
number of transverse ridges, along which the primary wart-like
papille are placed.

The papillae, which are found everywhere, are specially de-
veloped on the dorsal surface, less so on the ventral. ‘The
papille round the lips differ from the remaining papille of the
ventral surface in containing a green pigment. Each papilla
bears at its extremity a well-marked spine.

The ridges of the skin are not continued across the dorsal
middle line, being interrupted by the whitish line already
mentioned. Those which lie in the same transverse line as
the legs are not continued on to the latter, but stop at the
junction of the latter with the body. All the others pass round
to the ventral surface and are continued across the middle line ;

ANATOMY AND DEVELOPMENT OF PERIPATUS CAPENSIS. 217

they do not, however, become continuous with the ridges of the
other side, but passing between them gradually thin off and
vanish.

The ridges on the legs are directed transversely to their
long axes, i.e. are at right angles to the ridges of the rest of the
body.

The Antenne are ringed and taper slightly till near their
termination, where they present a slight enlargement in spirit
specimens, which in its turn tapers to its termination.

The rings consist essentially of anumber of coalesced primary
papillz, and are, therefore, beset by a number of spines like
those of the primary papille (described below). They are
more deeply pigmented than the rest of the antenna.

The free end of the antenna is covered bya cap of tissue like
that of the rings. It is followed by four or more rings placed
close together on the terminal enlargement. ‘There appears to
be about thirty rings on the antenne of all adults of this
species. But they are difficult to count, and a number of
small rings occur between them, which are not included in the
thirty.

The antenne are prolongations of the dorso-lateral parts of
the anterior end of the body.

The Eyes are paired and are situated at the roots of the
antenne on the dorso-lateral parts of the head. Each is placed
on the side of a protuberance which is continued as the an-
tenna, and presents the appearance of a small circular crystal-
line ball inserted on the skin in this region.

The rings of papille on that part of the head from which
the antenne arise lose their transverse arrangement. They
are arranged concentrically to the antennal rings, and have a
straight course forwards between the antenne.

The Oral Papille are placed at the side of the head. They
are attached ventro-laterally on each side of the lips. The
duet of the slime gland opens through their free end. They
possess two main rings of projecting tissue, which are especially
pigmented on the dorsal side; and their extremities are covered
by papille irregularly arranged,

218 PROFESSOR F. M. BALFOUR.

The Buccal Cavity, Jaws, and Lips are described below.

The Ambulatory Appendages—— The claw-bearing legs are
usually seventeen in number; but in two cases of small females
we have observed that the anal papillz bear claws, and pre-
sent all the essential features of the ambulatory appendages.
In one small female specimen there were twenty pairs of claw-
bearing appendages, the last being like the claw-bearing
anal papille last mentioned, and the generative opening being
placed between them.

The ambulatory appendages, with the exception of the fourth
and fifth pairs in both sexes, and the last pair (seventeenth) in the
male, all resemble each other fairly closely. A typical appen-
dage (figs. 2 and 3) will first be described, and the small
variations found in the appendages just mentioned will then
be pointed out. Each consists of two main divisions, a larger
proximal portion, the leg, and a narrow distal claw-bearing
portion, the foot.

The leg has the form of a truncated cone, the broad end of
which is attached to the ventro-lateral body-wall, of which it
appears to be, and is, a prolongation. Itis marked by anumber
of rings of primary papille, placed transversely to the long
axis of the leg, the dorsal of which contain a green and
the ventral a brown pigment. These rings of papille, at
the attachment of the leg, gradually change their direction
and merge into the body rings. At the narrow end of the cone
there are three ventrally placed pads, in which the brown pig-
ment is dark, and which are covered by a number of spines
precisely resembling the spines of the primary papilla. These
spinous pads are continued dorsally, each into a ring of
papille.

The papille of the ventral row next the proximal of these
spinous pads are intermediate in character between the primary
papille and the spinous pads. Each of these papillee is larger
than a normal papilla, and bears several spines (fig, 2). This
character of the papilla of this row is even more marked in
some of the anterior legs than in the one figured; it seems
probable that the pads have been formed by the coalescence of

ANATOMY AND DEVELOPMENT OF PERIPATUS CAPENSIS. 219

several rows of papillz on the ventral surface of the legs. On
the outer and inner sides of these pads the spines are absent,
and secondary papille only are present.

In the centre of the basal part of the ventral surface of the foot
there are present a group of larger papille, which are of a slightly
paler colour than the others. They are arranged so as to form
a groove, directed transversely to the long axis of the body, and
separated at its internal extremity by a median papilla from a
deep pit which is placed at the point of junction of the body and
leg. The whole structure has the appearance, when viewed
with the naked eye, of a transverse slit placed at the base of
the leg. The segmental organs open by the deep pit placed
at the internal end of this structure. The exact arrangement
of the papillz round the outer part of the slit does not appear
to be constant.

The foot is attached to the distal end of the leg. It is
slightly narrower at its attached extremity than at its free end,
which bears the two claws. The integument of the foot is
covered with secondary papill, but spines and primary papille
are absent, except at the points now to be described.

On each side of the middie ventral line of the proximal
end of the foot is placed an elliptical elevation of the integu-
ment covered with spines. Attached to the proximal and
lateral end of this is a primary papille. At the distal end
of the ventral side of the foot on each side of the middle
line is a group of inconspicuous pale elevations, bearing
spines.

On the front side of the distal end of the foot, close to the
socket in which the claws are placed, are two primary papilla,
one dorsal and the other ventral.

On the posterior side of the foot the dorsal of these only
is present. The claws are sickle-shaped, and placed on papilla
on the terminal portion of the foot. The part of the foot on
which they are placed is especially retractile, and is generally
found more or less telescoped into the proximal part (as in the
figure).

The fourth and fifth pairs of legs exactly resemble

22" PROFESSOR F. M. BALFOUR.

the others, except in the fact that the proximal pad is broken
up into three, a small central and two larger lateral. The
enlarged segmental organs of these legs open on the small
central division.

The last (17) leg of the male (Pl. XIV, fig. 4) is
characterised by possessing a well-marked white papilla on the
ventral surface. This papilla, which presents a slit-like open-
ing at its apex, is placed on the second row of papille count-
ing from the innermost pad, and slightly posterior to the axial
line of the leg.

The Anal Papille, or as they should be called, genera-
tive papille, are placed one on each side of the generative
aperture. They are most marked in small and least so in
large specimens. That they are rudimentary ambulatory
appendages is shown by the fact that they are sometimes pro-
vided with claws, and resemble closely the anterior appendages.

PAR Tedd:
ALIMENTARY CANAL.

The alimentary canal of Peripatus capensis forms, in
the extended condition of the animal, a nearly straight tube,
slightly longer than the body, the general characters of which
are shown in figs. 6 and 7.

. For the purposes of description, it may conveniently be
divided into five regions, viz. (1) the buccal cavity with the
tongue, jaws, and salivary glands, (2) pharynx, (3) the ceso-
phagus, (4) the stomach, (5) the rectum.

The Buccal Cavity—The buccal cavity has the form of a
fairly deep pit, of a longitudinal oval form, placed on the
ventral surface of the head, and surrounded by a tumid lip.

[The buccal cavity has been shown by Moseley to be formed
in the embryo by the fusion of a series of processes surround-

ANATOMY AND DEVELOPMENT OF PERIPATUS CAPENSIS. 221

ing the true mouth-opening, and enclosing in their fusion
the jaws.

The lip is covered by a soft skin, in which are numerous
organs of touch, similar to those in other parts of the skin
having their projecting portions enclosed in delicate spines
formed by the cuticle. The skin of the lips differs, however,
from the remainder of the skin, in the absence of tubercles, and
in the great reduction of the thickness of the dermis. It is
raised into a series of papilliform ridges, whose general form is
shown in fig. 5; of these there is one unpaired and median
behind, and a pair, differing somewhat in character from the
remainder, in front, and there are, in addition, seven on each
side.

The structures, within the buccal cavity are shown as they
appear in surface views in figs. 5 and 7, but their real nature
is best seen in sections, and is illustrated by Pl. XVI, figs. 11
and 12, representing the oral cavity in transverse section, and
by Pl. XVI, figs. 17 and 18, representing it in horizontal
longitudinal sections. In the median line of the buccal cavity
in front is placed a thick muscular protuberance, which may
perhaps conveniently be called the tongue, though attached to
the dorsal instead of the ventral wall of the mouth. It has
the form of an elongated ridge, which ends rather abruptly
behind, becoming continuous with the dorsal wall of the pha-
rynx. Its projecting edge is armed by a series of small teeth,
which are thickenings of the chitinous covering, prolonged
from the surface of the body over the buccal cavity. Where the
ridge becomes flatter behind, the row of teeth divides into two,
with a shallow groove between them (Pl. XV, fig. 7).

The surface of the tongue is covered by the oral epithelium,
in parts of which are organs of special sense, similar to those in
the skin; but its interior is wholly formed of powerful muscles.
The muscles form two groups, intermingled amongst each other.
There are a series of fibres inserted in the free edge of the
tongue, which diverge, more or less obliquely, towards the skin
at the front of the head anteriorly, and towards the pharynx
behind. The latter set of fibres are directly continuous with

222 PROFESSOR F. M. BALFOUR.

the radial fibres of the pharynx. The muscular fibres just
described are clearly adapted to give a sawing motion to the
tongue, whose movements may thus, to a certain extent, be
compared to those of the odontophor of a mollusc.

In addition to the above set of muscles, there are also trans-
verse muscles, forming lamine between the fibres just described.
They pass from side to side across the tongue, and their action
is clearly to narrow it, and so cause it to project outwards from
the buccal cavity.

On each side of the tongue are placed the jaws, which are,
no doubt, a pair of appendages, modified in the characteristic
arthropodan manner, to subserve mastication. Their structure
has never been satisfactorily described, and is very complicated.
They are essentially short papille, moved by an elaborate and
powerful system of muscles, which are armed at their free ex-
tremities by a pair of cutting blades or claws. The latter struc-
tures are, in all essential points, similar to the claws borne by
the feet, and, like these, are formed as thickenings of the
cuticle. They have therefore essentially the characters of the
claws and jaws of the Arthropoda, and are wholly dissimilar
to the sets of Chetopoda. The claws are sickle-shaped and,
as shown in Pl. XIV, fig. 5, have their convex edge directed
nearly straight forwards, and their concave or cutting edge
pointed backwards. Their form differs somewhat in the different
species, and, as will be shown in the systematic part of this
memoir,! forms a good specific character. In Peripatus ca-
pensis (Pl. XV, fig. 10) the cutting surface of the outer blade
is smooth and without teeth, while that of the inner blade
(fig. 9), which is the larger of the two, is provided with five or
six small teeth, in addition to the main point. A more important
difference between the two blades than that in the character
of the cutting edge just spoken of, is to be found in their relation
to the muscles which move them. The anterior parts of both
blades are placed on two epithelial ridges, which are moved by
muscles common to both blades (Pl. XVI, fig. 11). Posteriorly,

1 Some material for this memoir was left by Prof. Balfour, which will be
published separately.

ANATOMY AND DEVELOPMENT OF PERIPATUS CAPENSIS. 223

however, the behaviour of the two blades is very different.
The epithelial ridge bearing the outer blade is continued back
for a short distance behind the blade, but the cuticle covering
it becomes very thin, and it forms a simple epithelial ridge
placed parallel to the inner blade. The cuticle covering the
epithelial ridge of the inner blade is, on the contrary, prolonged
behind the blade itself as a thick rod, which, penetrating back-
wards along a deep pocket of the buccal epithelium, behind the
main part of the buccal cavity for the whole length of the
pharynx, forms a very powerful lever, on which a great part
of the muscles connected with the jaws find their insertion.
The relations of the epithelial pocket bearing this lever are
somewhat peculiar.

The part of the epithelial ridge bearing the proximal part
of this lever is bounded on both its outer and inner aspect by
a deep groove. The wall of the outer groove is formed by
the epithelial ridge of the outer blade, and that of the inner
by a special epithelial ridge at the side of the tongue. Close to
the hinder border of the buccal cavity (as shown in Pl. XVI,
fig. 12, on the right hind side), the outer walls of these two
grooves meet over the lever, so as completely to enclose it in
an epithelial tube, and almost immediately behind this point
the epithelial tube is detached from the oral epithelium, and
appears in section as a tube with a chitinous rod in its interior,
lying freely in the body cavity (shown in Pi. XVI, figs. 13—
16 le). This apparent tube is the section of the deep pit
already spoken of. It may be traced back even beyond the end
of the pharynx, and serves along its whole length for the
attachment of muscles.

The greater part of the buccal cavity is filled with the tongue
and jaws just described. It opens dorsally and behind by the
mouth into the pharynx, there being no sharp line of demarca-
tion between the bnecul cavity and the pharynx. Behind the
opening into the pharynx there is a continuation of the buccal
cavity shown in transverse section in fig. 13, and in longitudi-
nal and horizontal section in fig. 17, into which there opens
the common junction of the two salivary glands. This diver-

224, PROFESSOR F. M. BALFOUR.

ticulum is wide at first. and opens by a somewhat constricted
mouth into the pharynx above (Pl. XVI, fig. 13, also shown
in longitudinal and horizontal section in fig. 17). Behind it
narrows, passing insensibly into what may most conveniently
be regarded as a common duct for the two salivary glands
CPISKN I; digs 17):

The Salivary Glands.—These ‘two bodies were originally
described by Grube, by whom their nature was not made out,
and subsequently by Moseley, who regarded them as fat bodies.
They are placed in the lateral compartments of the body cavity
immediately dorsal to the ventral nerve cords, and extend for
a very variable distance, sometimes not more than half the
length of the body, and in other instances extending for nearly
its whole length. Their average length is perhaps about two
thirds that of the body. ‘Their middle portion is thickest,
and they thin off very much behind and to a slight extent in
front. Immediately behind the mouth and in front of the
first pair of legs, they bend inwards and downwards, and fall
(fig. 7) one on each side into the hind end of the narrow
section of the oral diverticulum just spoken of as the common
duct for the two salivary glands. The glandular part of these
organs is that extending back from the point where they bend
inwards. This part (fig. 16) is formed of very elongated
cells supported by a delicate membrana propria. The section
of this part is somewhat triangular, and the cells are so
long as to leave a comparatively small lumen. The nuclei
of the cells are placed close to the supporting membrane,
and the remainder of the cells are filled with very closely
packed secretory globules, which have a high index of refrac-
tion. It was the presence of these globules which probably
led Moseley to regard the salivary glands as fat bodies. The
part of each gland which bends inwards must be regarded as

the duct.
The cells lining the ducts are considerably less columnar

than those of the gland proper. Their nuclei (fig. 14) are
situated at the free extremities instead of at the base of
the cells, and they are without secretory globules. The

ANATOMY AND DEVELOPMENT OF PERIPATUS CAPENSIS. 225

cells lining the ducts of the salivary glands pass, without
any sharp line of demarcation, into those of the oral epi-
thelium, which are flatter and have their nuclei placed in the
middle.

The Pharynx.—The pharynx is a highly muscular tube (fig.
7) with a triangular lumen (figs. 14, 15), which extends from
the mouth to about half way between the first and second pair
of legs. It is lined by a flattish epithelium bounded by a cuticle
continuous with that of the mouth. On the dorsal side is a
ridge projecting into the lumen of the pharynx. This ridge
may be traced forwards (Pl. XVI, figs. 11—14) into the
tongue, and the two grooves at the side of this ridge, forming
the two upper angles of the triangular lumen, may be followed
into those at the sides of the tongue. The muscles of the
pharynx are very highly developed, consisting of an intrinsic
and an extrinsic set. ‘The former consists, as is best seen in
longitudinal sections, of (Pl. XVIII, fig. 23) radial fibres,
arranged in somewhat wedge-shaped lamine, between which
are rings of circular fibres. The latter are thicker externally
than internally, and so also appear wedge-shaped in longitu-
dinal sections. Very characteristic of the pharynx are the two
sympathetic nerves placed close to the two dorsal angles of
the triangular lumen (fig. 14, sy).

The pharynx of Peripatus is interesting in that it is unlike,
so far as I know, the pharynx of any true Arthropod, in all of
which the region corresponding with the pharynx of Peripatus
is provided with relatively very thin walls.

The pharynx of Peripatus has, on the other hand, a very
close and obvious resemblance to that of many of the Cheetopoda,
a resemblance which is greatly increased by the characteristic
course of the sympathetic nerves.

The form of the lumen, as already pointed out by Grube,
resembles that of the Nematoda.

The sophagus.—Behind the pharynx there follows a narrow
cesophagus (fig. 7,0 ¢) shown in section in fig. 16. It has
‘somewhat folded and fairly thick walls, and lies freely in the
central division of the body cavity without any mesenteric

226 PROFESSOR F. M. BALFOUR.

support. Its walls are formed of five layers, viz. from without
inwards.

(1) A peritoneal investment.

(2) A layer of longitudinal fibres.

(3) A layer of circular fibres, amongst which are numerous
nuclei.

(4) A connective-tissue layer supporting (5) a layer of fairly
columnar hyaline epithelium, bounded on its inner aspect by
a cuticle continued from that of the pharynx. In front it
passes insensibly into the pharynx, and beyond the region
where the dorsal walls of the pharynx have clearly com-
menced, the ventral walls still retain the characters of the
cesophageal walls. The cesophagus is vertically oval in front,
but more nearly circular behind. Characteristic of the ceso-
phagus is the junction of the two sympathetic nerves on its
dorsal wall (fig. 16). ‘These nerves cannot be traced far beyond
their point of junction.

The Stomach—tThe next section of the alimentary tract is
the stomach or mesenteron (fig. 6). It is by far the largest part
of the alimentary tract, commencing at about the second pair
of legs and extending nearly to the hind end of the body. It
tapers both in front and behind, and is narrowest in the middle,
and is marked off sharply both from the cesophagus in front and
the rectum behind, and is distinguished from both of these hy
its somewhat pinker hue. In the retracted condition of the
animal it is, as pointed out by Moseley, folded in a single
short dorsal loop, at about the junction of its first with its
second third, and also, according to my observations, at its
junction with the rectum; but in the extended condition it is
nearly straight, though usually the posterior fold at the junction
of the rectum is not completely removed. Its walls are always
marked by plications which, as both Moseley and Grube have
stated, do not in any way correspond with the segmentation
of the body. In its interior I have frequently found the
chitinous remains of the skins of insects, so that we are
not justified in considering that the diet is purely vegetable.
It lies free, and is, like the remainder of the alimentary

ANATOMY AND DEVELOPMENT OF PERIPATUS GAPENSIS. 227

tract, without a mesentery. The structure of the walls of
the stomach has not hitherto been very satisfactorily de-
scribed.

The connective tissue and muscular coats are extremely
thin. There is present everywhere a peritoneal covering, and
in front a fairly well-marked though very thin layer of muscles
formed of an external circular and an internal longitudinal
layer. In the middle and posterior parts, however, I was
unable to recognise these two layers in section; although in
surface view Grube found an inner layer of circular fibres
and an outer layer formed of bands of longitudinal fibres,
which he regards as muscular.

The layer supporting the epithelium is reduced to a base-
ment membrane. The epithelial part of the wall of the
stomach is by far the thickest (fig. 20), and is mainly com-
posed of enormously elongated, fibre-like cells, which in the
middle part of the stomach, where they are longest, are
nearly half a millimétre in length, and only about :006 mm.
in breadth. Their nuclei, as seen in fig. 20, are very elon-
gated, and are placed about a quarter of the length from the
base.

The cells are mainly filled with an immense number
of highly refracting spherules, probably secretory globules,
but held by Grube, from the fact of their dissolving in
ether, to be fat. The epithelial cells are raised into nu-
merous blunt processes projecting into the lumen of the
stomach.

In addition to the cells just described there are present in
the anterior part of the stomach a fair sprinkling of mucous
cells. There are also everywhere present around the bases of
the columnar cells short cells with spherical nuclei, which are
somewhat irregularly scattered in the middle and posterior
parts of the stomach, but form in the front part a definite
layer. I have not been able to isolate these cells, and can
give no account of their function.

The rectum extends from the end of the stomach to the
anus. The region of junction between the stomach and the

228 PROFESSOR F. M. BALFOUR.

rectum is somewhat folded. The usual arrangement of the
parts is that shown in fig. 6, where the hind end of the stomach
is seen to be bent upon itself in a U-shaped fashion, and the
rectum extending forwards under this bent portion and joining
the front end of the dorsal limb of the U. The structure of
the walls of the rectum is entirely different to that of the
stomach, and the transition between the two is perfectly sudden.
Within the peritoneal investment comes a well-developed mus-
cular layer with a somewhat unusual arrangement of its layers,
there being an external circular layer and an internal] layer
formed of isolated longitudinal bands. The epithelium is fairly
columnar, formed of granular cells with large nuclei, and is
lined by a prolongation of the external cuticle. It is raised
into numerous longitudinal folds, which are visible from the
surface, and give a very characteristic appearance to this part
of the alimentary tract. The muscular layers do not penetrate
into the epithelial folds, which are supported by a connective
tissue layer.

Nervous SystTEM.

The central nervous system consists of a pair of supra-
cesophageal ganglia united in the middle line, and of a pair of
widely divaricated ventral cords, continuous in front with the
supra-cesophageal ganglia.

It will be convenient in the first instance to deal with the
general anatomy of the nervous system and then with the
histology.

Ventral Cords—The ventral cords at first sight appear to
be without ganglionic thickenings, but on more careful exami-
nation they are found to be enlarged at each pair of legs (Pl.
XV, fig. 8). These enlargements may be regarded as imper-
fect ganglia. There are, therefore, seventeen such pairs of
ganglia corresponding to the seventeen pairs of legs. There
is in addition a ganglionic enlargement at the commencement
of the cesophageal commissures, where the nerves to the oral
papille are given off (P]. XVIII, fig. 22 or. g.), and the region
of junction between the esophageal commissures with the supra-

ANATOMY AND DEVELOPMENT OF PERIPATUS CAPENSIS. 229

cesophageal ganglia, where another pair of nerves are given
off to the jaws, (Pl. XVIII, fig. 22 7 ») may be regarded as the
anterior ganglion of the ventral cords. There are, ‘therefore,
-according to the above reckoning, nineteen pairs of ganglia
connected with the ventral cords.

The ventral cords are placed each in the lateral compart-
ments of the body cavity, immediately within the longitudinal
layer of muscles.

They are connected with each other, rather like the pedal
nerves of Chiton and the lower Prosobranchiata, by a number
of commissures. These commissures exhibit a fairly regular
arrangement from the region included between the first and
the last pair of true feet. There are nine or ten of them be-
tween each pair of feet (Pl. XIX, fig. 26). They pass along
the ventral wall of the body, perforating the ventral mass of
longitudinal muscles. On their way they give off nerves
which innervate the skin.

In Peripatus nove zealandiz, and probably also in P.
capensis, two of these nerves, coming off from each pair of
ganglia, are distinguished from the remainder by the fact that
they are provided with numerous nerve-cells, instead of being
composed of nerve-fibres only, like the remaining commissures
(Pl. XIX, fig. 26 gco). In correlation with the nerves given
off from them to the skin the commissures are smaller in the
middle than at the two ends.

Posteriorly the two nerve-cords nearly meet immediately in
front of the generative aperture, and between this aperture
and the last pair of feet there are about six commissures passing
between them (Pl. XV, fig. 8). Behind the generative aper-
ture the two cords bend upwards, and, as is shown in fig. 8,
fall into each other dorsally to the rectum. ‘The section of
the two cords placed dorsally to the rectum is solely formed ‘of
nerve-fibres ; the nerve-cells, present elsewhere, being here
absent.

In front of the ganglion of the first foot the commissures
have a more dorsal situation than in the remainder of the body.
The median longitudinal ventral muscle here gradually thins

VOL. XXIIIL—NEW SER, Q

230 PROFESSOR F. M. BALFOUR.

out and comes to an end, while the commissures pass imme-
diately below the wall of the pharynx (Pl. XVI, figs. 14, 15).
The ventral cords themselves at first approach very close to
each other in this region, separating again, however, to en-
velope between them the pharynx (Pl. XVIII, fig. 22).

There are eleven commissures in front of the first pair of legs
(Pl. XVIII, fig. 22). The three foremost of these are very close
together, the middle one arising in a more ventral position
than the other two, and joining in the median ventral line a
peculiar mass of cells placed in contact with the oral epithelium
(fig. 14). It is probably an organ of special sense.

The ventral cords give off a series of nerves from their outer
borders, which present throughout the trunk a fairly regular
arrangement. From each ganglion two large nerves (figs. 8,
22, 26) are given off, which, diverging somewhat from each
other, pass into the feet, and, giving off branches on their way,
may be traced for a considerable distance within the feet along
their anterior and posterior borders.

In front of each of the pair of pedal nerves a fairly large
nerve may be seen passing outwards towards the side of the
body (fig. 22). In addition to this nerve there are a number
of smaller nerves passing off from the main trunk, which do
not appear to be quite constant in number, but which are
usually about seven or eight. Similar nerves to those behind
are given off from the region in front of the first pair of legs,
while at the point where the two ventral cords pass into the
cesophageal commissures two large nerves (fig. 22), similar to
the pairs of pedal nerves, take their origin. These nerves may
be traced forwards into the oral papille, and are therefore to
be regarded as the nerves of these appendages. On the ventral
side of the cords, where they approach most closely, between
the oral papille and the first pair of legs, a number of small
nerves are given off to the skin, whose distribution appears
to be to the same region of the skin as that of the branches
from the commissures behind the first pair of legs.

From the esophageal commissures, close to their junction
with the supra-cesophageal ganglia, a nerve arises on each side

ANATOMY AND DEVELOPMENT OF PERIPATUS CAPENSIS. 231

which passes to the jaws, and a little in front of this, apparently
from the supra-cesophageal ganglion itself, a second nerve to the
jaws also takes its origin (Pl. XVIII, fig. 2272). These two
nerves I take to be homologous with a pair of pedal
nerves.

Between the nerves to the jaws and those to the oral
papillee a number of small nerves take their origin. Three of
these on each side pass in a dorsal direction and one or two in
a ventral one.

The Supra-csophageal Ganglia—The supra-cesophageal gan-
glia (figs. 8 and 22) are large, somewhat oval masses, broader
in front than behind, completely fused in the middle, but
free at their extremities. Each of them is prolonged

anteriorly into an antennary nerve, and is continuous behind
“with one of the csophageal commissures. On the ventral
- surface of each, rather behind the level of the eye, is placed a
very peculiar protuberance (fig. 22 d), of which I shall say
more in dealing with the histology of the nervous system.

A number of nerves arise from the supra-cesophageal ganglia,
mainly from their dorsal surface.

In front are the immense antennary nerves extending along
the whole length of each antenna, and giving off numerous
lateral twigs to the sense organs. Near the origin of the
antennary nerves, and rather on the dorsal surface, there spring
a few small twigs, which pass to the skin, and are presumably
sensory. The largest of them is shown in Pl. XVII, fig. 19.
About one third of the way back the two large optic nerves
take their origin, also arising laterally, but rather from the
dorsal surface (Pl. XVII, fig. 19 p and £). Each of them
joins a large ganglionic mass placed immediately behind the
retina. Nearly on a level with the optic nerves and slightly
nearer the middle dorsal line a pair of small nerves (fig. 19 p)
spring from the brain and pass upwards, while nearly in the
same line with the optic nerves and a little behind them a
larger pair of nerves take their origin.

Behind all these nerves there arises from the line of suture
between the two supra-cesophageal ganglia a large median nerve

232 PROFESSOR F. M. BALFOUR.

which appears to supply the integument of the dorsal part of
the head (PI. XV, fig. 8; Pl. XVI, figs. 11—14 dn).

Sympathetic System—In addition to the nerves just de-
scribed there are two very important nerves which arise near
the median ventral line, close to the hind end of the supra-
cesophageal ganglia. ‘The origin of these two nerves is shown
in the surface view (fig. 22 s y, and in section in fig. 11). They
at first tend somewhat forwards and pass into the muscles near
the epithelium lining the groove on each side of the tongue.
Here they suddenly bend backwards again and follow the
grooves into the pharynx.

The two grooves are continuous with the two dorsal angles
of the pharynx; and embedded in the muscles of the pharynx,
in juxtaposition with the epithelium, these two nerves may
easily be traced in sections. They pass backwards the whole
length of the pharynx till the latter joins the esophagus.
Here they at once approach and shortly meet in the median
dorsal line (fig. 16). They can only be traced for a very
short distance beyond their meeting point. ‘These nerves are,
without doubt, the homologues of the sympathetic system of
Cheetopods, occupying as they do the exact position which
Semper has shown to be characteristic of the sympathetic
nerves in that group, and arising from an almost identical part
‘of the brain.

Histology of the Nervous System.

Ventral Cords—The histology of the ventral cords and
cesophageal commissures is very simple and uniform. They
consist of a cord almost wholly formed of nerve-fibres, placed
dorsally, and a ventral layer of ganglion cells (figs. 16 and 20).

The fibrous portion of the cord has the usual structure, being
formed mainly of longitudinal fibres, each probably being a
bundle of fibres of various sizes, enveloped in a sponge-work
of connective tissue. The larger bundles of fibres are placed

1 Vide Spengel, ‘Oligognathus Bonelli, Naples Mittheilungen,’ Bad. iii,
pl. iv, fig. 52.

ANATOMY AND DEVELOPMENT OF PERIPATUS CAPENSIS. 293

near the inner borders of the cords. In this part of the cord
there are placed a very small number of ganglion cells.

The layer of ganglion cells is somewhat crescent-shaped in
section, and, as shown in figs. 16 and 20, envelopes the whole
ventral aspect of the fibrous parts of the cord, and even creeps
up slightly on to the dorsal side. It is thicker on the inner
than on the outer side, and increases considerably in bulk at
each ganglionic enlargement. ‘The cells of which it is com-
posed are for the most part of a nearly uniform size, but at the
border of the fibrous matter a fair sprinkling of larger cells is
found. : Ha

The tracheal vessels supplying the nervous system are placed
amongst the larger cells, at the boundary between the gan-
glionic and fibrous regions of the cords.

With reference to the peripheral nerve-stems there is not
much to be said. They have for the most part a similar struc-
ture to the fibrous parts of the main cord, but are provided with
a somewhat larger number of cells.

Sheath of the Ventral Cords—The ventral ‘cords are enve-
loped by a double sheath, the two layers of which are often in
contact, while in other cases they may be somewhat widely
separated from each other. The inner layer is extremely thin
and always very closely envelopes the nerve-cords. The outer
layer is thick and fibrous, and contains a fair sprinkling of
nuclei.

Supra-cesophageal Ganglia.—In the present state of our
knowledge a very detailed description of the histology of the
supra-cesophageal ganglia would be quite superfluous, and
I shall confine myself to a description of the more obvious
features in the arrangement of the ganglionic and fibrous por-
tions (Pl. XVII, fig. 19 a—e).

The ganglion cells are in the first place confined, for the
most part, to the surface. Along the under side of each gan-
glion there is a very thick layer of cells, continuous behind,
with the layer of ganglion cells which is placed on the under
surface of the cesophageal commissures. ‘These cells have,
moreover, an arrangement very similar to that in the ventral

234 PROFESSOR F. M. BALFOUR.

cords, so that asection through the supra-cesophageal ganglia
has an obvious resemblance to what would be the appearance
of a section through the united ventral cords. On the outer
borders of the ganglia the cells extend upwards, but they end
on about the level of the optic nerve (fig. 19D). Immediately
dorsal to this point the fibrous matter of the brain is exposed
freely on the surface (fig. 19 a, B, &c., a). I shall call the region
of fibrous matter so exposed the dorso-lateral horn of white
matter.

Where the two ganglia separate in front the ganglion cells
spread up the inner side, and arch over so as to cover part
of the dorsal side. Thus, in the anterior part, where the two
ganglia are separate, there is a complete covering of ganglionic
substance, except for a narrow strip, where the dorso-lateral
lobe of white matter is exposed on the surface (fig. 19 a). From
the point where the two ganglia meet in front the nerve-cells
extend backwards as a median strip on the dorsal surface (fig.
19p and). This strip, becoming gradually smaller behind,
reaches nearly, though not quite, the posterior limit of the
junction of the ganglia. Behind it there is, however, a region
where the whole dorsal surface of the ganglia is without any
covering of nerve-cells.

This tongue of ganglion cells sends in, slightly behind the
level of the eyes, a transverse vertical prolongation inwards
into the white matter of the brain, which is shown in the series
of transverse sections in fig. 19 5, and also in the vertical longi-
tudinal section (PJ. XVIII, fig. 21), and in horizontal section
in Pl. XVIII, fig. 22.

On the ventral aspect of each lobe of the brain there is pre-
sent a very peculiar, bluntly conical protuberance of ganglion
cells (Pl. XVIII, fig. 22), which was first detected by Grube
(No. 10), and described by him as “a white thick body of a
regular tetrahedral form, and exhibiting an oval dark spot in
the middle of two of the faces.’ He further states that it is
united by a delicate nerve to the supra-cesophageal ganglion,
and regards it as an organ of hearing.

In Peripatus capensis the organ in question can hardly

ANATOMY AND DEVELOPMENT OF PERIPATUS CAPENSIS. 2385

be described as tetrahedral. It is rather of a flattened oval
form, and consists, as shown in sections (P]. XVII, fig. 19 cand
D, D), mainly of ganglion cells. In its interior is a cavity with
a distinct bounding membrane: the cells of which it is com-
posed vary somewhat in size, being smallest near the point of
attachment. At its free end is placed a highly refractive,
somewhat oval body, probably forming what Grube describes
as a dark spot, half embedded in its substance, and kept in
place by the sheath of nervous matter surrounding it. This
body appears to have fallen out in my sections. The whole
structure is attached to the under surface of the brain bya
very short stalk formed of a bundle of cells and nervous fibres.

It is difficult to offer any interpretation of the nature of this
body. It is removed considerably from the surface of the animal,
and is not, therefore, so far as I can see, adapted to serve as an
organ of hearing.

The distribution of the white or fibrous matter of the ganglia
is not very easy to describe.

There is a central lobe of white matter (fig. 19 5), which
is continuous from ganglion to ganglion, where the two are
united. It is smaller behind than in front. On its ventral side
it exhibits fairly well-marked transverse commissural fibres,
connecting the two halves of the ganglion. Laterally and
somewhat ventrally it is prolonged into a horn (fig. 19 p, , 3),
which I propose calling the ventro-lateral horn. In front it is
placed in a distinct protuberance of the brain, which is placed
ventrally to and nearly in the same vertical plane as the optic
nerve. This protuberance is best shown in the view of the
brain from below given in Pl, XVIII, fig. 22. This part of the
horn is characterised by the presence of large vertically-directed
bundles of nerve-fibres, shown in transverse section in fig. 19 p.
Posteriorly the diameter of this horn is larger than in front
(fig. 19 E, F, G), but does not give rise to a protuberance on
the surface of the brain owing to the smaller development of
the median lobe behind.

The median lobe of the brain is also prolonged into a dorso-
lateral lobe (fig. 19, a), which, as already mentioned, is freely

236 PROFESSOR F. M. BALFOUR.

exposed on the surface. On its ventral border there springs the
optic nerve and several pairs of sensory nerves already described
(fig. 19 p, &) while from its dorsal border a pair of sensory
nerves also spring, nearly in the same vertical plane as the optic
nerves.

Posteriorly where the dorsal surface of the brain is not covered
in with ganglion cells the dorso-lateral horn and median lobe
of the brain become indistinguishable.

In the front part of the brain the median lobe of white matter
extends dorsalwards to the dorsal strip of ganglion cells, but
behind the region of the transverse prolongation of these cells,
into the white matter already described (p. 234), there is a more
or less distinctly defined lobe of white matter on the dorsal
surface, which I propose calling the postero-dorsal lobe of white
matter. It is shown in the transverse sections (fig. 19 Fr and
G,c). It gradually thins away and disappears behind. It is
mainly characterised by the presence on the ventral border of
definite transverse commissural fibres.

THE SKIN.

The skin is formed of three layers.

1. The cuticle.

2. The epidermis or hypodermis.

3. The dermis.

The cuticle is a layer of about 0:002 mm. in thickness. Its
surface is not, however, smooth, but is everywhere, with the
exception of the perioral region, raised into minute secondary
papillz, the base of which varies somewhat in diameter, but is
usually not far from 0:02 mm. On the ventral surface of the
body these papille are for the most part somewhat blunt, but
on the dorsal surface they are more or less sharply pointed. In
most instances they bear at their free extremity a somewhat
prominent spine. ‘The whole surface of each of the secondary
papillae just described is in its turn covered by numerous
minute spinous tubercles. In the perioral region, where the
cuticle is smooth, it is obviously formed of two layers which
easily separate from each other, and there is I believe a similar

ANATOMY AND DEVELOPMENT OF PERIPATUS CAPENSIS. By

division elsewhere though it is not so easy to see. It is to be
presumed that the cuticle is regularly shed.

The epidermis, placed immediately within the cuticle, is
composed of a single row of cells, which vary, however, a good
deal in size in different regions of the body. The cells excrete
the cuticle, and, as shown in fig. 32, they stand in a very
remarkable relation to the secondary papille of the cuticle just
described. Each epidermis cell is in fact placed within one of
these secondary papille, so that the cuticle of each secondary
papilla is the product of a single epidermis cell. This relation
is easily seen in section, while it may also be beautifully shown
by taking a part of the skin which is not too much pigmented,
and, after staining it, examining from the surface.

In fig. 82 a region of the epidermis is figured, in which the
cells are exceptionally columnar. ‘The cuticle has, moreover,
in the process of cutting the section, been somewhat raised
and carried away from the subjacent cells. The cells of the
epidermis are provided with large oval nuclei, which contain a
well developed reticulum, giving with low powers a very gra-
nular appearance to the nuclei. The protoplasm of the cells
is also somewhat granular, and the granules are frequently so
disposed as to produce a very well-marked appearance of stria-
tion on the inner end of the cells. The pigment which gives
the characteristic colour to the skin is deposited in the proto-
plasm of the outer ends of the cells in the form of small
granules. An attempt is made to show this in fig. 32.

At the apex of most, if not all, the primary wart-like papillz
there are present oval aggregations, or masses of epidermis
cells, each such mass being enclosed in a thickish capsule (fig.
31). The cells of these masses appear to form the wall of a
cavity which leads into the hollow interior of a long spine.
These spines when carefully examined with high objectives
present a rather peculiar structure. The base of the spine is
enveloped by the normal cuticle, but the spine itself, which
terminates in a very fine point, appears, as shown in fig. 31, to
be continuous with the inner layer of the cuticle. In the
perioral region the outer layer of the cuticle, as well as the

238 PROFESSOR F. M. BALFOUR.

inner, appear to be continued to the end of the spines. Within
the base of the spine there is visible a finely striated substance
which may often be traced into the cavity enclosed by the cells,
and appears to be continuous with the cells. Attached to the
inner ends of most of the capsules of these organs a delicate
fibrillated cord may be observed, and although I have not in
any instance succeeded in tracing this cord into one of the nerve-
stems, yet in the antenne, where the nerve-stems are of an
enormous size, I have satisfied myself that the minute nerves
leaving the main nerve-stems and passing out towards the skin
are histologically not to be distinguished from these fibrillated
cords. I have therefore but little hesitation in regarding these
cords as nerves.

In certain regions of the body the oval aggregations of cells
are extremely numerous; more especially is this the case in the
antenne, lips, and oral papille. On the ventral surface of the
peripheral rings of the thicker sections of the feet they are
also very thick set (fig. 20 Pp). They here form a kind of pad,
and have a more elongated form than in other regions. In the
antennee they are thickly set side by side on the rings of skin
which give such an Arthropod appearance to these organs in
Peripatus.

The arrangement of the cells in the bodies just described led
me at first to look upon them as glands, but a further inves-
tigation induced me to regard them as a form of tactile organ.
The arguments for this view are both of a positive and a nega-
tive kind.

The positive arguments are the following:

(1) The organs are supplied with large nerves, which is dis-
tinctly in favour of their being sense organs rather than
glands.

(2) The peculiar striz at the base of the spines appear to me
like the imperfectly preserved remains of sense hairs. ~

(3) The distribution of these organs favours: the view that
they are tactile organs. They are most numerous on the an-
tennze, where such organs would naturally be present, especially
-inacase like that of Peripatus, where the nerve passing to

ANATOMY AND DEVELOPMENT OF PERIPATUS CAPENSIS. 239

the antenne is simply gigantic. On the other hand, the an-
tenn would not be a natural place to look for an enormous
development of dermal glands.

The lips, oral papillae, and under surface of the legs, where
these bodies are also very numerous, are situations where tactile
organs would be of great use.

Under the head of negative arguments must be classed those
which tell against these organs being glandular. The most
important of these is the fact that they have no obvious
orifice. Their cavities open no doubt into the spines, but
the spines terminate in such extremely fine points that the
existence of an orifice at their apex is hardly credible.

Another argument, from the distribution of these organs
over the body is practically the converse of that already used.
The distribution being as unfavorable to the view that they
are glands, as it is favorable to that of their being sense
organs.

THE TRACHEAL SYSTEM.

The apertures of the tracheal system are placed in the de-
pressions between the papillee or ridges of the skin. Each of
them leads into a tube, which I shall call the tracheal pit (fig.
30), the walls of which are formed of epithelial cells bounded
towards the lumen of the pit by a very delicate cuticular
membrane continuous with the cuticle covering the surface of
the body. The pits vary somewhat in depth; the pit figured
was about 0°09 mm. It perforates the dermis and terminates
in the subjacent muscular layer. The investigation of the
inner end of the pit gave me some little trouble.

Transverse sections (fig. 30) through the trunk containing a
tracheal opening show that the walls of the pit expanded
internally in a mushroom-like fashion, the narrow part being,
however, often excentric in relation to the centre of the expanded
part.

Although it was clear that the trachez started from the
expanded region of the walls of the pit, I could not find that
the lumen of the pit dilated into a large vesicle in this part,

24.0 PROFESSOR F. M. BALFOUR.

and further investigation proved that the trachez actually
started from the slightly swollen inner extremity of the narrow
part of the pit, the expanded walls of the pit forming an
umbrella-like covering for the diverging bundles of trachee.

I have, in fig. 30, attempted to make clear this relation
between the expanded walls of the tracheal pits and the
tracheee. In longitudinal sections of the trunk the trachea.
pits do not exhibit the lateral expansion which I have just
described, which proves that the divergence of the bundles of
trachez only takes place laterally and not in an antero-posterior
direction. Cells similar in general character to those of the
walls of the tracheal pits are placed between the branches of
tracheze and somewhat similar cells, though generally with
more elongated nuclei, accompany the bundles of trachee as far
as they can be followed in my sections. The structure of
these parts in the adult would, in fact, lead one to suppose
that the trachez had originated at the expense of the cells of
pits of the epidermis, and that the cells accompanying the
bundles of tracheze were the remains of cords of cells which
sprouted out from the blind ends of the epidermis pits and
gave rise in the first instance to the tracheze.

The trachez themselves are extremely minute, unbranched
(so far as I could follow them) tubes. Each opening by a
separate aperture into the base of the tracheal pit, and measuring
about 0:002 mm. in diameter. They exhibit a faint transverse
striation which I take to be the indication of a spiral fibre
[Moseley (‘ Phil. Trans.,’ 1874, Pl. 73, fig. 1) states that the
trachez branch, but only exceptionally.]

Situation of the tracheal apertures.—Moseley states
(No. 13) that the trachez arise from the skin all over the surface
of the body, but are especially developed in certain regions. He
finds “a row of minute oval openings on the ventral surface of
the body,” the openings being “situate with tolerable regu-
larity in the centres of the interspaces between the pairs of
members, but additional ones occurring at irregular intervals.
Other similar openings occur in depressions on the inner side
of the conical foot protuberance.” It is difficult in preserved

ANATOMY AND DEVELOPMENT OF PERIPATUS CAPENSIS. 241

specimens to make out the exact distributions of the tracheal
apertures, but I have been able to make out certain points
about them.

There is a double row of apertures on each side of the
median dorsal line, forming two sub-dorsal rows of apertures.
The apertures are considerably more numerous than the legs.
There is also a double row of openings, again more numerous
than the legs, on each side of the median ventral line between
the insertions of the legs. Moseley speaks of a median row in
this position. I think this must be a mistake.

Posteriorly the two inner rows approach very close to each
other in the median ventral line, but I have never seen
them in my section opening quite in the middle line. Both
the dorsal and ventral rows are very irregular.

I have not found openings on the ventral or dorsal side of
the feet but there are openings at the anterior and posterior
aspects of the feet. There are, moreover, a considerable
number of openings around the base of the feet.

The dorsal rows of tracheal apertures are continued into the
head and give rise in this situation to enormous bundles of
trachee.

In front of the mouth there is a very large median ventral
tracheal pit, which gives off trachez to the ventral part of the
nervous system, and still more in front a large number of
such pits close together. The trachee to the central nervous
system in many instances enter the nervous system bound up
in the same sheath as the nerves.

Tue Muscunar System.

The general muscular system consists of—(1) the general
wall of the body ; (2) the muscles connected with the mouth,
pharynx, and jaws; (3) the muscles of the feet; (4) the
muscles of the alimentary tract.

The muscular wall of the body is formed of—(1) an external
layer of circular fibres ; (2) an internal layer of longitudinal
muscles ; (3) a layer of transverse fibres.

242 PROFESSOR F. M. BALFOUR.

The layer which I have spoken of as formed of circulat
fibres is formed of two strata of fibres which girth the body
somewhat obliquely (P]. XVIII, fig. 25). In the outer stra-
tum the rings are arranged so that their ventral parts are
behind, while the ventral parts of the rings of the inner stratum
are most forward. Both in the median dorsal and ventral lines
the layer of circular fibres become somewhat thinner, and
where the legs are attached the regularity of both strata is
somewhat interfered with, and they become continuous with a
set of fibres inserted in the wall of the foot.

The longitudinal muscles are arranged as five bands (vide
fig. 16), viz. two dorsal, two lateral, and three ventral. The
three ventral may be spoken of as the latero-ventral and
medio-ventral bands.

The transverse fibres consist of (1) a continuous sheet on
each side inserted dorsally in the cutis, along a line opposite
the space between the dorsal bands of longitudinal fibres, and
ventrally between the ventro-median and ventro-lateral bands.
Each sheet at its insertion slightly breaks up into separate
bands. They divide the body cavity into three regions—a
median, containing the alimentary tract, slime glands, &c.,
and two lateral, which are less well developed, and contain the
nervous system, salivary glands, segmental organs, ec.

(2) Inserted a little dorsal to the transverse band just de-
scribed is a second band which immediately crosses the first,
and then passes on the outer side of the nervous cord and
salivary gland, where such is present, and is inserted ventrally
in the space between the ventro-lateral and lateral longitudinal
band.

Where the feet are given off the second transverse band
becomes continuous with the main retractcr muscular fibres in
the foot, which are inserted both on to the dorsal side and
ventral side.

Muscular system of the feet.—This consists of the
retractors of the feet connected with the outer transverse
muscle and the circular layer of muscles. In addition to these
muscles there are intrinsic transverse muscles which cross the

ANATOMY AND DEVELOPMENT OF PERIPATUS CAPENSIS. 243

cavity of the feet in various directions (Pl. XVIII, fig. 20).
There is no special circular layer of fibres.

Histology of the muscle.—The main muscles of the
pody are unstriated and divided into fibres, each invested by a
delicate membrane. Between the membrane and muscle are
scattered nuclei, which are never found inside the muscle
fibres. The muscles attached to the jaws form an exception
in that they are distinctly transversely striated.

Tue Bopy Caviry AND VASCULAR SYSTEM.

The Body Cavity, as already indicated, is formed of three
compartments—one central and two lateral. The former is by
far the largest, and contains the alimentary tract, the genera-
tive organs, and the mucous glands. It is lined by a delicate
endothelial layer, and is not divided into compartments nor
traversed by muscular fibres.

The lateral divisions are much smaller than the central, and
are shut off from it by the inner transverse band of muscles.
They are almost entirely filled with the nerve-cord and salivary
gland in front and with the nerve-cord alone behind, and their
lumen is broken up by muscular bands. They further contain
the segmental organs which open into them. They are pro-
longed into the feet, as is the embryonic body cavity of most
Arthropoda.

The Vascular System is usually stated to consist of a dorsal
heart. I find between the dorsal bands of longitudinal
fibres a vessel ina space shut off from the body cavity bya con-
tinuation of the endothelial lining of the latter (fig. 16). The
vessel has definite walls and an endothelial lining, but I could
not make out whether the walls were muscular. The ventral
part of it is surrounded by a peculiar cellular tissue, probably, as
suggested by Moseley, equivalent to the fat bodies of insects.
It is continued from close to the hind end of the body to the
head, and is at its maximum behind. In addition to this vessel
there is present a very delicate ventral vessel, by no means
easy to see, situated between the cutis and the outer layer of
circular muscles.

244, PROFESSOR F. M. BALFOUR.

SEGMENTAL ORGANS.

A series of glandular organs are found in Peripatus which
have their external openings situated on the ventral surface of a
certain number of the legs, and which, to the best of my belief,
end internally by opening into the lateral compartments of the
body cavity. These organs are probably of an excretory nature,
and I consider them homologous with the nephridia or seg-
mental organs of the Cheetopoda.

In Peripatus capensis they are present in all the legs.
In all of them (except the first three) the following parts may
be recognised :

1. A vesicular portion opening to the exterior by a narrow
passage.

2, A coiled portion, which is again subdivided into several
sections.

3. A terminal section ending by a somewhat enlarged opening
into the lateral compartment of the body cavity.

The last twelve pairs of these organs are all constructed in a
very similar manner, while the two pairs situated in the fourth
and fifth pairs of legs are considerably larger than those behind,
and are in some respects very differently constituted.

It will be convenient to commence with one of the hinder
nephridia. Such a nephridium from the ninth pair of legs is
represented in fig. 28. The external opening is placed at the
outer end of a transverse groove placed at the base of one of the
feet, while the main portion of the organ lies in the body cavity
in the base of the leg, and extends into the trunk to about the
level of the outer edge of the nerve-cord of its side. The ex-
ternal opening (0s) leads into a narrow tube (s d), which
gradually dilates into a large sac (s).

The narrow part is lined by small epithelial cells, which are
directly continuous with and perfectly similar to those of the
epidermis (fig. 20). It is provided with a superficial coating
of longitudinal muscular fibres, which thins out where it passes
over the sac, along which it only extends for a short distance,

The sac itself, which forms a kind of bladder or collecting

ANATOMY AND DEVELOPMENT OF PERIPATUS OAPENSIS. 245

vesicle for the organ, is provided with an extremely thin wall,
lined with very large flattened cells. These cells are formed of
granular protoplasm, and each of them is provided with a large
nucleus, which causes a considerable projection into the lumen
of the sac (figs. 20,29s). The epithelial wall of the sac is sup-
ported by a membrana propria, over which a delicate layer of
the peritoneal epithelium is reflected.

The coiled tube forming the second section of the nephridium
varies in length, and by the character of the-epithelium lining
it may be divided into four regions. It commences with a region
lined by a fairly columnar epithelium with smallish nuclei (fig.
28 scl). The boundaries of the cells of this epithelium are
usually very indistinct, and the protoplasm contains numerous
minute granules, which are usually arranged in such a manner
as to give to optical or real sections of the wall of this part of
the tube a transversely striated appearance. These granules
are very probably minute balls of excretory matter.

The nuclei of the cells are placed near their free extremities,
contrary to what might have been anticipated, and the inner
ends of the cells project for very different lengths into the in-
terior, so causing the inner boundary of the epithelium of this
part of the tube to have a very ragged appearance. This
portion of the coiled tube is continuous at its outer end with
the thin-walled vesicle. At its inner end it is continuous with
region No. 2 of the coiled tube (fig. 28 s ¢ 2), which is lined by
small closely-packed columnar cells. This portion is followed
by region No. 3, which has a very characteristic structure (fig,
28 sc). The cells lining this part are very large and flat, and
contain large disc-shaped nuclei, which are usually provided
with large nucleoli, and often exhibit a beautiful reticulum.
They may frequently be observed in a state of division. The
protoplasm of this region is provided with similar granules to
that in the first region, and the boundaries of the cells are usually
very indistinct. The fourth region is very short (fig. 28s ¢ 4),
and is formed of small columnar cells. It gradually narrows
till it opens suddenly into the terminal section (so ¢), which
ends by openirg into the body cavity, and constitutes the most

VOL. XXILI,—NEW SER, R

246 PROFESSOR F. M. BALFOUR.

distinct portion of the whole organ. Its walls are formed of
columnar cells almost filled by oval nuclei, which absorbs
colouring matters with very great avidity, and thus renders
this part extremely conspicuous. The nuclei are arranged in
several rows.

The study of the internal opening of this part gave me some
trouble. No specimens ever show it as rounded off in the
characteristic fashion of tubes ending in a cul-de-sac. It is
usually somewhat ragged and apparently open. In the best
preserved specimens it expands into a short funnel-shaped
mouth, the free edge of which is turned back. Sections
confirm the results of dissections. Those passing longitudinally
through the opening prove its edges are turned back, forming
a kind of rudimentary funnel. This is represented in fig. 29,
from the last leg of a female. I have observed remains of
what I consider to be cilia in this section of the organ. The
fourth region of the organ is always placed close to the thin-
walled collecting vesicle (figs. 28 and 29). In the whole of the
coiled tube just described the epithelium is supported by a
membrana propria, which in its turn is invested by a delicate
layer of peritoneal epithelium.

The fourth and fifth pairs are very considerably larger than
those behind, and are in other respects peculiar. The great
mass of each organ is placed behind the leg, on which the
external opening is placed, immediately outside one of the
lateral nerve-cords. Its position is shown in fig. 8.

The external opening, instead of being placed near the base
of the leg, is placed on the ventral side of the third ring
(counting from the outer end) of the thicker portion of the leg.
It leads (fig. 27) into a portion which clearly corresponds with
the collecting vesicle of the hinder nephridia. This part is
not, however, dilated into a vesicle in the same sort of way,
and the cells which form the lining epithelium have not
the same characteristic structure, but are much smaller. Close
to the point where the vesicle joins the coiled section of the
nephridium the former has a peculiar nick or bend in it. At
this nick it is firmly attached to the ventral side of the foot by

ANATOMY AND DEVELOPMENT OF PERIPATUS OAPENSIS. 247

muscles and trachez, and when cut away from its attachment
the muscles and trachez cannot easily be detached from it.
The main part of the coils are formed by region No. 1, and
the epithelial cells lining this part present very characteristically
the striated appearance which has already been spoken of. The
large-celled region of the coiled tube (fig. 27) is also of con-
siderable dimensions, and the terminal portion is wedged in
between this and the commencing part of the coiled tube.
The terminal portion with its internal opening is in its histo-
logical characters exactly similar to the homologous region in
the hinder nephridia.

The three pairs of nephridia in the three foremost pairs of
legs are very rudimentary, consisting, so far as I have been
able to make out, solely of the collecting vesicle and the duct
leading from them to the exterior. The external opening is
placed on the ventral side of the base of the feet, in the same
situation as that of the posterior nephridia, but the histological
characters of the vesicle are similar to those of the fourth and
fifth pairs.

GENERATIVE ORGANS.

[The sexes are distinct, and the average size of the females
appears to be greater than that of the males.

The only outward characteristic by which the males can be
distinguished from the females is the presence in the former of
a small white papilla on the ventral side of the 17th pair
of legs (Pl. II, fig. 4). At the extremity of this papilla the
modified crural gland of the last leg opens by a slit-like
aperture.

The generative orifice in both sexes is placed on the ventral
surface of the body, close to the anus, and between the two
anal papille, which are much more marked in small specimens
than in large ones, and in two cases (of females) were observed
to bear rudimentary claws.

1. The Male Organs. Pl. XX, fig. 43.
The male organs consist of a pair of testes (¢e), a pair of

248 PROFESSOR F. M. BALFOUR.

prostates (pr) and vasa deferentia (vd) and accessory glandu-
lar tubules (f).

All the above parts lie in the central compartment of the
body cavity. In addition, the accessory glandular bodies or
crural glands of the last (17th) pair of legs are enlarged and
prolonged into an elongated tube placed in the lateral com-
partment of the body cavity (a@ q).

The arrangement of these parts represented in the figure
appears essentially that which Moseley has already described
for this species. The dilatations on the vasa deferentia,
which he calls vesicule seminales, is not so marked; nor can
the peculiar spiral twisting of this part of the vas deferens
which he figures (No. 13) be made out in this specimen.
The testes are placed at different levels in the median com-
partment of the body cavity, and both lie on the same side of
the intestine (right side).

The arrangement of the terminal portions of the vas de-
ferens is precisely that described by Moseley. The right vas
deferens passes under both nerve-cords to join the left, and
form the enlarged tube (p), which, passing beneath the
nerve-cord of its side, runs to the external orifice.
The enlarged terminal portion possesses thick muscular
walls, and possibly constitutes a spermatophore maker, as
has been shown to be the case in P. N. Zealandiz, by
Moseley.

In some specimens a different arrangement obtains, in that
the left vas deferens passes under both nerve-cords to join the
right.

In addition to the above structures, which are all described
by Moseley, there are a pair of small glandular tubes (/), which
open with the unpaired terminal portion of the vas deferens at
the generative orifice.

2. Female Organs. Pl XIX, fig. 33.

The female organs consist of a median unpaired ovary and
a pair of oviducts, which are dilated for a great part of their
course to perform a uterine function, and which open behind

ANATOMY AND DEVELOPMENT OF PERIPATUS CAPENSIS. 249

into a common vestibule communicating directly with the
exterior.

Ovary.—In the specimen figured the following is the
arrangement :

The ovary lies rather to the dorsal side in the central com-
partment of the body cavity, and is attached to one of the
longitudinal septa separating this from the lateral compart-
ment. It lies between the penultimate and antepenultimate
pair of legs.

The oviducts cross before opening to the exterior. The
right oviduct passes under the rectum, and the left over the
rectum. ‘They meet by opening into a common vestibule,
which in its turn opens to the exterior immediately ventral to
the anus. It has not been ascertained how far this arrange-
ment, which differs from that observed by Moseley, is a normal
one. The young undergo nearly the whole of their develop-
ment within the uterus. They possess at birth the full num-
ber of appendages, and differ from the parent only in size and
colour.

NoTES ON ADDITIONAL GLANDULAR Bopizs IN THE LEGs
[CruRAL Guanps].

1. They are present in all except the first.

2. They open externally to the nephridia (Pl. XVIII,
fig. 20), except in the fourth and fifth pairs of legs, in which
they are internal.

3. A muscular layer covers the whole gland, consisting, I
believe, of an oblique circular layer.

4. The accessory gland in the male (fig. 43, ag) is proba-
bly a modification of one of these organs.

[The structure and relations of these glands may be best
understood by reference to Pl. XVIII, fig, 20. Each consists of
a dilated vesicular portion (f gl) placed in the lateral com-
partment of the body cavity in the foot, and of a narrow duct
leading to the exterior, and opening on the ventral surface
amongst the papille of the second row (counting from the
internal of the three foot pads (fig. 20 P).

250 PROFESSOR F. M. BALFOUR.

The vesicular portion is lined by columnar cells, with very
large oval nuclei, while the duct is lined by cells similar to
the epidermic cells, with which they are continuous at the
opening.

In the last (17th leg) of the males of this species, this gland
(vide above, note 4) possesses a slit-like opening placed at
the apex of well-developed white papilla (Pl. XIV, fig. 4).
It is enormously enlarged, and is prolonged forward as a long
tubular gland, the structure of which resembles that of the
vesicles of the crural glands in the other legs. This gland lies
in the lateral compartment of the body cavity, and extends
forward to the level of the 9th leg (Pl. XV, fig. 8, and
PI. XX, fig. 43). Itis described by Professor Balfour as the
accessory gland of the male, and is seen in section lying
immediately dorsal to the nerve-cord in fig. 20, @g.]

PART II
Tuer DEVELOPMENT OF PERIPATUS CAPENSIS.

[The remarkable discoveries about the early development of
Peripatus, which Balfour made in June last, shortly before
starting for Switzerland, have already been the subject of a
short communication to the Royal Society (‘ Proc. Roy. Soc.’
No. 222, 1882.) They relate (1) to the blastopore, (2) to the
origin of the mesoblast.

Balfour left no manuscript account or notes of his discovery
in connection with the drawings which he prepared in order
to illustrate it, but he spoke about it to Professor Ray Lan-
kester and also to us, and he further gave a short account of
the matter in a private letter to Professor Kleinenberg.

In this letter, which by the courtesy of Professor Kleinenberg
we have been permitted to see, he describes the blastopore as an
elongated slit-like structure extending along nearly the whole
ventral surface; and further states, as the result of his exami-
nation of the few and ill-preserved embryos in his possession,

ANATOMY AND DEVELOPMENT OF PERIPATUS CAPENSIS. 251

that the mesoblast appears to originate as paired outgrowths
from the lips of the blastopore.

The drawings left by Balfour in connection with the dis-
coveries are four in number: one of the entire embryo, show-
ing the slit-like blastopore and the mesoblastic somites, the
other three depicting the transverse sections of the same em-
bryo.

The first drawing (fig. 37), viz. that of the whole embryo,
shows an embryo of an oval shape, possessing six somites,
whilst along the middle of its ventral surface there are two
slit-like openings, lying parallel to the long axis of the body,
and placed one behind the other. The mesoblastic somites
are arranged bilaterally in pairs, six on either side of these
slits. The following note in his handwriting is attached to this
drawing :

“Young larva of Peripatus capensis.—I could not make
out for certain which was the anterior end. Length 1:34
millimetres.”

Balfour’s three remaining drawings (figs. 40—42) are, as
already stated, representations of transverse sections of the
embryo figured by him as a whole. They tend to show, as he
stated in the letter referred to above, that the mesoblast
originates as paired outgrowths from the hypoblast, and that
these outgrowths are formed near the junction of the hypoblast
with the epiblast at the lips of the blastopore.

In fig. 40 the walls of the mesoblastic somites appear con-
tinuous with those of the mesenteron near the blastopore.

In fig. 41, which is from a section a little in front of fig. 40,
the walls of the mesoblastic somites are independent of those
of the mesenteron.

Fig. 42 is from a section made in front of the region of the
blastopore.

In all the sections the epiblast lying over the somites is
thickened, while elsewhere it is formed of only one layer of
cells; and this thickening subsequently appears to give rise
to the nervous system. Balfour in his earlier investigations on
the present subject found in more advanced stages of the

252 PROFESSOR F. M. BALFOUR.

embryo the nerve-cords still scarcely separated from the
epiblast.1

We have since found, in Balfour’s material, embryos of a
slightly different age to that just described. Of these, three
(figs. 34, 35, 36) are younger, while one (fig. 38) is older than
Balfour’s embryo.

Stage A.—The youngest (fig. 34) is of a slightly oval form,
and its greatest length is ‘48 mm. It possesses a blastopore,
which is elongated in the direction of the long axis of the em-
bryo, and is slightly narrower in its middle than at either end.
From one end of the blastopore there is continued an opaque
band. This we consider to be the posterior end of the
blastopore of the embryo. The blastopore leads into the
archenteron.

Stage B.—In the next stage (fig. 35) the embryo is elon-
gate-oval in form. Its length is ‘7 mm. The blastopore is
elongated and slightly narrowed in the middle. At the pos-
terior end of the embryo there is a mass of opaque tissue. On
each side of the blastopore are three mesoblastic somites. The
length of the blastopore is ‘45 mm.

Stage C.—In the next stage (fig. 36) the features are much
the same as in the preceding. The length of the whole
embryo is °9 mm.

The following were the measurements of an embryo of this
stage with five somites, but slightly younger than that from
which fig. 36 was drawn.

Length of embryo . . . . . - ‘4mm,
33 blastopore : : : 5 F "461.5

Distance between hind end of blastopore and hind end of body y | eae
fe » front end of body and front end of blastopore “06 75

The somites have increased to five, and there are indications
of a sixth being budded off from the posterior mass of opaque
tissue. The median parts of the lips of the blastopore have
come together preparatory to the complete fusion by which the
blastopore becomes divided into two parts.

1 ©Comparative Embryology,’ vol. i. p. 318.

ANATOMY AND DEVELOPMENT OF PERIPATUS CAPENSIS. 253

Stage D.—The next stage is Balfour’s stage, and has been
already described.

The length is 1°34,

It will be observed, on comparing it with the preceding em-
bryos, that while the anterior pair of somites in figs. 35 and
36 lie at a considerable distance from what we have called the
anterior end of the embryo (a), in the embryo now under con-
sideration they are placed at the anterior end of the body, one
on each side of the middle line. We cannot speak positively as
to how they come there, whether by a pushing forward of the
anterior somites of the previous stage, or by the formation of
new somites anteriorly to those of. the previous stage.

In the next stage it is obvious that this anterior pair of so-
mites has been converted into the preoral lobes.

The anterior of the two openings to which the blastopore
gives rise is placed between the second pair of somites; we
shall call it the embryonic mouth. The posterior opening
formed from the blastopore is elongated, being dilated in front
and continued back us a narrow slit (?) to very near the hind
end of the embryo, where it presents a second slight dilatation.
The anterior dilatation of the posterior open region of the
blastopore we shall call the embryonic anus.

Lately, but too late to be figured with this memoir, we have
been fortunate enough to find an embryo of apparently precisely
the same stage as fig. 37. We are able, therefore, to give a
few more details about the stage.

The measurements of this embryo were :

Length of whele embryo . : : - 1:32 mm.
Distance from front end of body to frat end of mouth as Hisehess
si embryonic mouth to hindend of embryonicanus 52 ,,
3 from hind end of embryonic anus to hind end
of body : ‘. ; é AE PO) ae
Length of embryonic anus. - Skea
bs part of blastopore behind “Eiauaie anus . i jp ot ese
Greatest width of embryo . ; : : : ‘64. 5

Stage E.—In the next stage (figs. 838 and 39) the flexure
of the hind end of the body has considerably increased. The

254 PROFESSOR F. M. BALFOUR.

anterior opening of the blastopore, the embryonic mouth, has
increased remarkably in size. It is circular, and is placed
between the second pair of mesoblastic somites. The anterior
dilatation of the posterior opening of the blastopore, the em-
bryonic anus, has, like the anterior opening, become much
enlarged. It is circular, and is placed on the coneavity of the
ventral flexure. From its hind end there is continued to the
hind end of the body a groove (shown in fig. 39 as a dotted
line), which we take to be the remains of the posterior slit-like
part of the posterior opening of the blastopore of the preceding
stage. The posterior dilatation has disappeared. The embryo.
has apparently about thirteen somites, which are still quite
distinct from one another, and apparently do not communicate
at this stage with the mesenteron.

The epiblast lying immediately over the somites is, as in the
earlier stages, thickened, and the thickenings of the two sides
join each other in front of the embryonic mouth, where the
anterior pair of mesoblastic somites (the preoral lobes) are
almost in contact.

The median ventral epiblast, i.e. the epiblast in the area,
bounded by the embryonic mouth and anus before and behind
and by the developing nerve-cords laterally, is extremely thin,
and consists of one layer of very flat cells. Over the dorsal
surface of the body the epiblast cells are cubical, and arranged
in one layer.

Measurements of Embryo of Stage E.

Length of embryo. ; : . . - 112mm.
Greatest width é : ‘ 2. “Oda
Distance from front end of anibeyanie oats to hind end of

embryonic anus . . - 5 >. “48 55
Greatest length of embryonic rapa 5 : “16° 25
Length between hind end of embryonic mouth and front aud

of embryonic anus : - ‘ : .* “00 "aa

These measurements were made with a micrometer eyepiece,
with the embryo lying on its back in the position of fig. 38, so

ANATOMY AND DEVELOPMENT OF PERIPATUS GAPENSIS. 255

that they simply indicate the length of the straight line con-
necting the respective points.

This is the last embryo of our series of young stages. The
next oldest embryo was 3°2 mm. in length. It had ringed
antenne, seventeen (?) pairs of legs, and was completely dou-
bled upon itself, as in Moseley’s figure.

The pits into the cerebral ganglia and a mouth and anus
were present. There can be no doubt that the mouth and
anus of this embryo become the mouth and anus of the
adult.

The important question as to the connection between the
adult mouth and anus, and the embryonic mouth and anus of
the Stage EK, must, considering the great gap between Stage E
and the next oldest embryo, be left open. Meanwhile, we may
point out that the embryonic mouth of Stage E has exactly
the same position as that of the adult; but that the anus is
considerably in front of the hind end of the body in Stage E,
while it is terminal in the adult.

If the embryonic mouth and anus do become the adult
mouth and anus, there would appear to be an entire
absence of stomodeeum and proctodeum in Peripatus, unless
the buccal cavity represents the stomodeum. The latter
is formed, as has been shown by Moseley, by a series of
outgrowths round the simple mouth-opening of the embryo,
which enclosing the jaws give rise to the tumid lips of the
adult.

For our determination of the posterior and anterior ends of
each of these embryos, Stage A to E, we depend upon the
Opaque tissue seen in each case at one end of the blastopore.

In Stage A it has the form of a band, extending backwards
from the blastopore.

In Stages B—D, it has the form ofan opaque mass of tissue
occupying the whole hind end of the embryo, and extending a
short distance on either side of the posterior end of the blas-
topore.

This opacity is due in each case to a proliferation of cells
of the hypoblast, and, perhaps, from the epiblast (?).

256 PROFESSOR F. M. BALFOUR.

There can be no doubt that the mesoblast so formed gives
rise to the great majority of the mesoblastic somites.

This posterior opacity is marked in Stage C by a slight
longitudinal groove extending backwards from the hind end
of the blastopore. This is difficult to see in surface views,
and has not been represented in the figure, but is easily seen
in sections.

But in Stage D this groove has become very strongly marked
in surface views, and looks like a part of the original blasto-
pore of Stage C.

Sections show that it does not lead into the archenteron,
but only into the mass of mesoblast which forms the posterior
opacity. It presents an extraordinary resemblance to the
primitive streak of vertebrates, and the ventral groove of
insect embryos.

We think that there can be but little doubt that it is a part
of the original blastopore, which, on account of its late appear-
ance (this being due to the late development of the posterior
part of the body to which it belongs), does not acquire the
normal relations of a blastopore, but presents only those rudi-
mentary features (deep groove connected with origin of meso-
blast) which the whole blastopore of other tracheates presents.

We think it probable that the larval anus eventually shifts
to the hind end of the body, and gives rise to the adult anus.
We reserve the account of the internal structure of these embryos
(Stages A—E) and of the later stages for a subsequent memoir.

We may briefly summarise the more important facts of the
early development of Peripatus capensis, detailed in the
preceding account.

1. The greater part of the mesoblast.is developed from the
walls of the archenteron.

2. The embryonic mouth and anus are derived from the re-
spective ends of the original blastopore, the middle part of the
blastopore closing up.

3. The embryonic mouth almost certainly becomes the adult
mouth, i.e. the aperture leading from the buccal cavity into
the pharynx, the two being in the same position. The em-

ANATOMY AND DEVELOPMENT OF PERIPATUS CAPENSIS. 257

bryonic anus is in front of the position of the adult anus, but
in all probability shifts back, and persists as the adult anus.

4, The anterior pair of mesoblastic somites gives rise to the
swellings of the preoral lobes, and to the mesoblast of the
head.!

There is no need for us to enlarge upon the eestees of
these facts. Their close bearing upon some of the most im-
portant problems of morphology will be apparent to all, and
we may with advantage quote here some passages from Bal-
four’s ‘ Comparative Embryology,’ which show that he himself
long ago had anticipated and in a sense predicted their dis-
covery.

** Although the mesoblastic groove of insects is not a gastrula,
it is quite possible that it is the rudiment of a blastopore,
the gastrula corresponding to which has now vanished from
development.” (‘Comparative Embryology,’ vol. i, p. 378.)

“ TRACHEATA.—Insecta. It (the mesoblast) grows inwards
from the lips of the germinal groove, which probably Fepreconts
the remains of a blastopore.” (‘Comparative Embryology,’
vol. ii, p. 291.)

“It is, therefore, highly probable that the paired ingrowths
of the mesoblast from the lips of the blastopore may have
been, in the first instance, derived from a a of archenteric
diverticula.” (‘Comparative Embryology,’ vol. ii, p. 294.)

The facts now recorded were discovered in June last, only a
short time before Balfour started for Switzerland; we know but
little of the new ideas which they called up in his mind. We
can only point to passages in his published works which seem to
indicate the direction which his speculations would have taken.

After speculating as to the probability of a genetic connec-
tion between the circumoral nervous system of the Ccelen-
terata,and the nervous system of Echinodermata, Platyelminthes,
Cheetopoda, Mollusca, &c., he goes on to say:

“ A circumoral nerve-ring, if longitudinally extended, might

' We have seen nothing in any of our sections which we can identify as of
so-called mesenchymatous origin.

258 PROFESSOR F. M. BALFOUR.

give rise to a pair of nerve-cords united in front and behind
—exactly such a nervous system, in fact, as is present in many
Nemertines (the Enopla and Pelagonemertes), in Peripatus
and in primitive molluscan types (Chiton, Fissurella, &c.).
From the lateral parts of this ring it would be easy to derive
the ventral cord of the Chetopoda and Arthropoda. It is
especially deserving of notice, in connection with the nervous
system of the above-mentioned Nemertines and Peripatus,
that the commissure connecting the two nerve-cords behind
is placed on the dorsal side of the intestines. As is at once
obvious, by referring to the diagram (fig. 231 B), this is the
position this commissure ought, undoubtedly, to occupy if de-
rived from part of a nerve-ring which originally followed more
or less closely the ciliated edge of the body of the supposed
radiate ancestor.” (‘Comparative Embryology,’ vol. 11, pp.
311, 312.)

The facts of development here recorded give a strong addi-
tional support to this latter view, and seem to render possible
a considerable extension of it along the same lines. ]

[The editors of the present memoir intend to prepare for
publication a complete monograph of all the species of Peri-
patus known, with figures in extension of the materials for
that purpose collected by the late Prof. Balfour. They would,
therefore, feel extremely obliged for the loan of any specimens,
especially of species from the West Indies. Such specimens
would be most carefully preserved from injury, and returned
after inspection and comparison. Any such should be sent to
Mr. A. Sedgwick, Trinity College, Cambridge.]

List or Memoirs on PERIPATUS.

(1) M. Lanspown GuiLpine.— An Account of a New Genus of Mollusca,”
‘ Zoological Journal,’ vol ii, p. 443, 1826.

(2) M. Anpourn anp Mitnr-Epwarps.—“ Classific des Annélides et descrip-
tion le celles qui habitent les cdtes de France,” p. 411, ‘Ann. Scien.
Nat.,’ ser. i, vol. xxx, 1833,

ANATOMY AND DEVELOPMENT OF PERIPATUS CAPENSIS. 259

(3) M. Gervats.— Etudes p. servir a Vhistoire naturelle des Myriapodes,’’
‘Ann. Scien. Nat.,’ ser. ii, vol. vii, 1837, p. 38.

(4) Wimemann.—‘ Wiegmann’s Archiv,’ 1837.

(5) H. Mrtyz-Epwarps.—“ Note sur le Peripate juluforme,” ‘Ann.
Scien. Nat.,’ ser. ii, vol. xviii, 1842.

(6) Brancuarp.—* Sur lorganization des Vers,” chap. iv, p. 137—141
‘Ann. Scien. Nat.,’ vol. viii, 1847.

(7) QuarreracEs.—“ Anat. des Hermelles, note on,” p. 57, ‘Ann. Scien-
Nat.,’ ser. ili, vol. x, 1848.

(8) Quatreraces.— Hist. Nat. des Annelés,’ 1865, Appendix, pp. 675-6.

(9) Dr Buainvitte.—‘ Suppl. au Dict. des Se. Nat.,’ vol. i.

(10) Ep. Gruse.— Untersuchungen tib. d. Bau von Peripatus Hd-
wardsii,’ ‘Archiv fiir Anat. und Physiol.,’? 1853.

(11) Sazrncer.—* Moskauer Naturforscher Sammlung,” ‘ Abth. Zool.,’ 1869.

(12) H. N. Moserey.—‘ On the Structure and Development of Peripatus
capensis,” ‘ Proc. Roy. Soc.,’ No. 153, 1874.

(13) H. N. Moserey.—“ On the Structure and Development of Peripatus
capensis,” ‘ Phil. Trans.,’ vol. clxiv, 1874.

(14) H. N. Moserey.—‘ Remarks on Observations by Captain Hutton,
Director of the Otago Museum, on Peripatus nove zealandiz,”’
‘Ann. and Mag. of Nat. History,’ Jan., 1877.

(15) Carrain Hurron.—“ Observations on Peripatus nove zealandie,”
‘Ann. and Mag. of Nat. History,’ Nov., 1876.

(16) F. M. Batrour.—* On Certain Points in the Anatomy of Peripatus
capensis,” ‘Quart. Journ. of Micr. Science,’ vol. xix, 1879.

(17) A. Exnst.—‘ Nature,’ March 10th, 1881.

2

260 DR. E. KLEIN.

On a Morphological Variety of Bacillus
Anthracis.

By

E. Klein, M.D., F.R.S.,

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

With Plate XXI.

In my Report of 1881 to the Medical Officer of the Local
Government Board, reprinted in this Journal, January, 1883, I
have mentioned the occurrence, in artificial cultivations of
Bacillus anthracis, of peculiar torula-like cells connected
with the Bacillus anthracis. I have stated there (this
Journal, January, p. 37) that observing the growth of the
threads of Bacillus anthracis in “ cell specimens,” it is found
that the growing ends of the threads are occasionally seen in con-
nection with a row of torula-like cells, i.e. spherical or elliptical
cells closely placed so as to form a complete chain. The torula-
like cells are larger than the cubical cells, which I there
described as constituting the elements of the Bacillus
anthracis.

I have since followed this change more in its details, and
propose in the following pages to describe the results of my
observations.

Every cultivation made of Bacillus anthracis in the pork
broth or the mixture of pork broth and gelatine in test-tubes
or flasks, as described in the above-named paper, when ex-
posed to a temperature of 20°—25° C., shows some, although
not many, of the torula-like variety. But I have succeeded in

MORPHOLOGICAL VARIETY OF BACILLUS ANTHRACIS. 261

obtaining at will a copious crop of this variety, by sowing the
ordinary typical Bacillus anthracis of an artificial culti-
vation on to the surface of the solid gelatine pork broth,
contained in test-tubes or flasks plugged with sterile cotton
wool (see my former paper), and keeping these cultures at
ordinary temperature of the room, i.e. about 15—20° C. The
growth under these conditions proceeds slowly, and taking out
a sample of the growth after two or three days, we find the
astonishing fact that almost the entire growth, or the greater
majority of it, consists of the torula variety. What one finds is
this—spherical or slightly elliptical cells, of a diameter varying
between 0:0013 and 0:0026 mm., or more, isolated, or in small
groups, or more commonly in longer or shorter chains (see figs.
1, 3, and 6). Some chains are composed of cells which are
twice the diameter of the cells of a neighbouring chain. In
the fresh state these cells possess a distinct membrane, a clear
contents, and one or two granules. In dried and stained
specimens the cells are stained deeply as a whole.

Some of the cells show constrictions dividing them into two,
three, or even four small cells, or they are in a state of GEMMA-
TION, like a true torula,a smaller or larger knob protruding either
in the long axis of the chain, or, what is not at all uncommon, to
the side of it. In consequence of this, and also in consequence
of the division of the largest cells into three or four daughter
cells, we find often attached to the side of a chain one or more
cells (see figs. 1, 3, and 4). When dividing into four, we
obtain a form similar to a sarcina.

Besides these chains we meet with long rows of cells, of
which the majority or minority are spherical and large, the re-
mainder elliptical, and approaching more or less the ordinary
thin, rod-shaped elements of the typical bacillus. _When stain-
ing such a specimen, we notice this fact, that the staining does
not extend uniformly over the whole chain, but that the deeply
stained spherical or elliptical cells are separated from one another
—or rather connected with one another—by a very faintly
stained thin, longer or shorter bridge. Fig. 2 illustrates this
point.

VOL, XXIII, —NEW SER. 8

262 DR. kK. KLEIN.

It is noticed at the same time that the torula cells of the
same chain are of various sizes, some being twice and three
times as big as others (see fig. 2). Whether small or large
they may show division or gemmation into two or three, or
four. The most interesting forms of chains are those which
contain here and there one or two big torula cells amongst
a long series of small cells, the former resembling a sort of
sporangium. And, indeed, such big cells may contain two,
three, or four granules.

As the cultivation proceeds many of the torula chains
are gradually converted into the thin typical threads of
Bacillus anthracis; here and there in the thread a
more or less elliptical cell is still recognisable, or a series
of them, so as to denote the way the thread has origi-
nated. In fig. 4 the transition is very well shown. Some
of the threads that have become typical threads of Anthrax
bacillus, when followed up to their growing end, still show
in a very marked manner the torula nature of their cells.
Although when one examines these preparations, one is at
first inclined to assume that there exist in these cultivations
two distinct kinds of organisms, viz. (a) typical threads of
Anthrax bacillus, and (0) a kind of torula—so great is the
contrast between the two; still, when examining carefully the
specimen, all doubt soon disappears, since one finds not only all
intermediate forms, but the two kinds along the very same
thread. I have shown such specimens to many of my friends
well able to judge on these points, and I had no difficulty
whatever in demonstrating to them the correctness of my views,
although at first sight they were hardly inclined to admit it.

In the accompanying illustrations I have selected some of the
more pronounced forms, and the reader will have no difficulty
in recognising the transitions between the two forms. When the
torula-like cells of the chain become elliptical, and when these
still more elongate, a thin typical thread of Bacillus an-
thracis will be the result.

As the cultivation proceeds, say after several weeks, the
number of the chains and cells of the torula variety gradually

MORPHOLOGICAL VARIETY OF BACILLUS ANTHRACIS, 263

diminish, and the greater amount of the growth is com-
posed of the typical smooth, thin threads, with here and there
an indication in them of a_ spherical or elliptical torula-

like cell.
Iam indebted to my friend Dr. George Maddox for some

photographs which he made of some of the specimens; the
figures 6, 7, and 8 are exact copies of them. This gentleman
has kindly furnished me with an abstract of a description by
Dr. Antoine Magnin (translated by Dr. Sternberg, of Boston)
of certain unpublished observations by M. Toussaint, according
to which this experienced observer has seen the Bacillus
anthracis growing in a “ Ranvier’s’”’ moist chamber at 37—
40° C., undergo changes, by which the protoplasm in some
becomes collected into larger or smaller spherical or elliptical
sporangia, each of them yielding several “spores.” The mem-
brane finally bursts, and the “ spores” are freed. Toussaint
has seen these spores undergo germination and elongation into
typical Anthrax bacilli. There can be no doubt that some
of my torula-like cells correspond to Toussaint’s sporangia. I
have repeated Toussaint’s experiments, but could not see the
freeing of the “spores.” 1 do not consider the granules
present in the cells to be spores; I have only been able to see
that the protoplasm of the cells divides into two, three, and
four ; but each of these is capable of growing up into a large
spherical or oval cell. These granules are not spores in
any sense, at least not in that sense in which are the bright,
highly-refractive oval corpuscles that appear in the threads of
Bacillus anthracis, and in other kinds of bacilli when
they are supplied with sufficient amount of oxygen. The
granules stain well in anilin dyes, the typical spores do not.
The small cells derived by gemmation and division from our
larger spherical cells, are identical in structure and in value with
their parent cells, but the typical bright oval spores of Anthrax
bacillus are, as is well known, altogether different structures
from the parent cells.

In our case the oval or spherical cells, no matter whether
large or small, are capable of division and gemmation, and

264 DR. E. KLEIN.

hereby are capable of producing daughter cells; further, the
spherical or oval cells in our case become ultimately elongated,
so as to contribute to form the typical rods and threads of
Bacillus anthracis. These rods give origin, by transverse
division, to cubical cells, such as are noticed in the ordinary
Bacillus anthracis after staining with anilin dyes.

We see, then, that the Bacillus anthracis shows two dis-
tinct morphological varieties, viz. one: the typical bacillus, the
other: a torula-like form; the cells are spherical or elliptical,
and capable of gemmation like a real torula cell, or of di-
viding into two, three, or four new cells like a true schyzomy-
cetes. The spherical and elliptical torula-like cells elongate,
and are transformed into ordinary typical bacillus. The torula-
like cells are not spores, nor do they form sporangia.

An interesting fact that I observed is this, that as time goes
on the torula variety, in chains as well as in threads, undergoes
the same degeneration which I observed in ordinary cultures
of Bacillus anthracis in fluid media, and have described
in my paper printed in the January number of this Journal,
1883, viz. the protoplasm gradually disintegrates into granules
and these gradually dissolving, leave the empty sheath behind.
Of a formation of real spores there is nothing to be seen in these
cases; such a culture has lost all power of infection. This is
another proof that no spores are formed by the torula cells.

The torula-like cells, while intact, are physiologically just as
poisonous as the ordinary Bacillus anthracis, since guinea-
pigs and rabbits invariably die of typical anthrax when in-
oculated with them.

The bacillus found in such animals is always the ordinary
Bacillus anthracis.

MORPHOLOGICAL VARIETY OF BACILLUS ANTHRACIS. 265

Note on the Foregoing.
By

E. Ray Lankester, M.A., F.R.S.

My friend, Dr. Klein, kindly suggests that I should add to
his paper a note calling the reader’s attention to the series of
varieties of Bacterium rubescens which were described and
figured by me in vol. xiii, p. 408 (1873), and in vol. xvi, p. 27
(1876) of this Journal, since they offer an interesting parallel
to the important fact discovered by Dr. Klein.

Dr. Klein’s observation establishes that Bacillus anthracis
is, so far as form is concerned, a “ Protean species,’ and once
for all demonstrates the error of those who, like Koch, have
arrived at the conclusion that the forms of Bacteriacez are
fixed and breed true.

I have enumerated in the papers above referred to the pos-
sible variations of form known at that date as occurring amongst
Bacteriacee. ‘The particular variety of Bacillus anthracis
now described by Dr. Klein adds a new form to the list there
given. ,

Using the term “ plastid’ to denominate the unit of struc-
ture of the Bacteriacew, I have pointed out four categories
under which their form-characters may be grouped, viz.:

A. Shape of the plastids.

.B. Substance of the plastid; relation of protoplasm and cell
wall in each plastid. ,

c. Distribution of colour (when present) in each plastid.

pv. Mode of aggregation of the plastids.

Under heading A we have the following possibilities:
1. Spherical; 2. Biscuit-shaped; 3. Bacillar; 4. Filamen-
tous; 5. Acicular; 6. Serpentine; 7. Spiral; 8. Helicoid.

Under heading B we have as possibilities—

1. Plastids clean; or, on the other hand,

2. Plastids gloeogenous (producing a jelly-like cell wall)

266 NOTE BY PROFESSOR E. RAY LANKESTER,

And again as alternatives—

8. Plastids homogeneous (no cell wall distinguishable from
contents) ; or,

4, Plastids loculate (unilocular or moltilocular).

Under heading c we have the colouring matter—

1. Diffuse ; or,

2. Locular (confined to the substance within the loculi).

Lastly, under heading p we have as varieties in the mode of
ageregation of the plastids the following, which undoubtedly
is not an exhaustive list :—1. Linear (thread-like) ; 2. Stellar ;
3. Globose; 4. Massive; 5. Arborescent; 6. Catenular; 7.
Reticular; 8. Tesselate.

Dr. Klein’s new form of Bacillus anthracis consists of
plastids of spherical shape, and each plastid is unilocular
and clean; the plastids are in catenular aggregation. But
a special kind of catenular aggregation is exhibited by Dr.
Klein’s growth, inasmuch as there is an irregularity in the
size of the aggregated spherical plastids. At approximately
regular recurrent intervals large spherical plastids are inter-
posed between series of smaller ones. Such an arrangement
of spherical plastids aggregated in chains is familiar to botanists
in the case of the Alga Nostoc, and hence this variety of ca-
tenular aggregation in the Bacteriacezee may be conveniently
registered as “ nostocoid catenular aggregation.”’

The commonly known twisted filaments obtained in culti-
vations of Bacillus anthracis, and figured in fig. 4 @ of
Dr. Klein’s plate, are linear aggregates of very short
bacillar plastids. It will be found most convenient to sepa-
rate these very short cylindrical plastids from the longer
forms (twice or many times as long as they are broad) under
the name “micro-bacilli”—the longer cylindrical plastids,
such as are commonly observed in early stages of a growth of
Bacillus subtilis, being called “ bacilli’? as heretofore.
Short rods built up by linear aggregation of micro-bacilli are
not to be confused with homogeneous macro-bacilli consisting
of one elongated plastid. The twisting of the linear aggre-
gates into a rope, shown in fig. 4 a, introduces to us the

MORPHOLOGICAL VARIETY OF BACILLUS ANTHRACIS. 267

necessity of an additional series of terms descriptive of the
grouping of the aggregates; that is to say, we require terms
describing the combinations of the second order formed by
the union of those combinations of the first order, which we
have called “ aggregates of plastids.” Of such “‘ combinations
of aggregates ” the present may be described as “ funicular.”

Thus, in a systematic way we should give the following
description of fig. 4 a:

** Plastids—Shape, microbacillar; substance, homoge-
neous, clean; colour, wanting.

“ Agoregates of plastids.—Linear.

*‘Combinations of aggregates. — Funicular, the filaments
twisted and recurrent.”

268 DE. ES. “KLIN:

Note on a Pink Torula.
By

E. Klein, ME.D., F.R.S.,

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

Some time ago I examined for my colleague at St. Bartholo-
mew’s Hospital, Dr. W. J. Russell, F.R.S., a sample of dis-
tilled water contained in a water bottle, through which 25 cubic
feet of London fog air had been passing every hour for twenty-
four hours. In this water were present, besides numbers of
soot and dirt particles, large numbers of mycelia, or what
appeared to be the mycelium of penicillium and mucor.

There were also present bacilli in the shape of longer or
shorter, apparently smooth threads, and also a few ordinary
torula cells, Saccharomyces cerevisie. With this water I
inoculated a few test-tubes plugged with sterilised cotton
wool containing neutral, or slightly acid pork broth, such as
I used for other cultivation purposes (see my paper in this
Journal, January, 1883), and placed them in ‘the incubator at
32° C. After several days there was present in the test-tubes
a fair amount of a whitish, or rather colourless nebulous sedi-
ment, which, when examined under the microscope, was com-
posed of the most exquisite threads singly and in spiral bundles of
the above bacillus. There were also present some short bacilli
of the above kind; they were all non-moving. The bundles
of spirally convoluted threads were identical with the typical
cable-like bundles of Bacillus anthracis, and it would have
been very difficult to recognise a difference at first sight; but
they were not, of course, anthrax bacilli, as was soon ascer-
tained by experiment. Besides these bacilli there were pre-

NOTE ON A PINK TORULA. 269

sent in the culture numerous cells of the yeast Saccharomyces
cerevisie. The cells are oval, consisting of a limiting mem-
brane and a homogeneous, highly refractive protoplasm, and in
it at one place one or two vacuoles. In some of them there
was one large bright corpuscle present at one side of the pro-
toplasm. In some the protoplasm appears slightly granular.
The cells are of different sizes, some twice as big as others.
Their sizes are as follows: the big cells 0°009 mm. by 0:01 mm.,
the small ones 0:005 mm. by 0:008 mm. The small ones are
evidently young forms, since they could be seen to sprout out,
and to become constricted off from bigger ones.

As regards the process of reproduction, it appeared to me to
be that of gemmation only. Hereby large groups of cells,
some chain-like, were formed, which groups by enlargement
become soon confluent into larger masses.

The pork broth, kept at 52° C. for several weeks, became so
concentrated, that when taken out of the incubator and allowed
to cool, almost solidified. Keeping it in this state at the
ordinary temperature of the room, it was noticed, after some
days, that the growth appeared on the surface of the nourish-
ing material in the shape of minute whitish spots or flat
droplets, which, as they gradually enlarged, assumed a distinct
pinkish colour. The enlargement in breadth and thickness
proceeded in a few days so far that the whole surface of the
almost solid nourishing material became covered with a
pinkish film, in which, however, the individuality of the drop-
lets could still be recognised. Under the microscope these
pink droplets are composed entirely of torula cells of exactly
the same nature and size as those above described. They are,
no doubt, the same organisms, as will appear also from other
facts presently to be mentioned.

The cells themselves do not possess any colour when looked
at under the microscope, singly or in a thin layer, but they
appear of a pinkish tint when viewed as a group, or in a thick
layer.

I have sown out from this layer of pink torula cells on to
boiled white of egg, solid gelatine, and mixture of gelatine and

270 DR. E. KLEIN.

pork broth, used in my experiments on Anthrax bacillus.
With the egg I have not obtained any satisfactory results, but
with the gelatine and the mixture of gelatine and pork broth I
have obtained beautiful crops. The sowing was done with
the point of a capillary glass tube on to the free surface of the
nutritive material (contained in flasks or test-tubes, plugged with
sterilised cotton wool) ; and after an incubation of about four
days, the vessel being kept at ordinary temperature of the room,
there appeared the first signs of the growth having taken root,
in the shape of a minute pinkish droplet; this gradually spread
in breadth and thickness. The very interesting fact observed
with this increase was this: the masses growing downward
into the nutritive material remained colourless, whereas those
spreading on the free surface were pink, both being composed
of exactly the same torula cells.

The thicker the layer became, the deeper the pink tint. The
gelatine does not become liquified by the growth, and in this re-
spect it differs from a growth of micrococci, bacteria, or bacilli.

Sowing the pink torula into the depth of fluid nutritive
material, such as pork broth, and keeping it at the bottom of
the fluid, it is noticed that no matter whether growing at ordi-
nary temperature of the room, or in the incubator at 30—35° C.,
it remains colourless, and when of considerable amount,
appears like a whitish precipitate at the bottom of the fluid.

Sowing this colourless torula on to a free surface, it again
gives origin to pink growth. But also in the same tube the
at first colourless torula, i.e. while growing at the bottom of
the fluid, may, when reaching the free surface, give origin to
the pink growth.

Another interesting fact I have observed is this, that when a
copious growth of pink torula has made its appearance on the
surface of the solid nourishing material (gelatine), and this
nourishing material is made fluid, so that the pink growth
sinks to the bottom, and the material is again allowed to
solidify: it will be observed that the pink mass retains its
colour, that is to say, that the torula, once pink does not loose
its colour when removed from the free access of air. But the

NOTE ON A PINK TORULA. a

new increment of the mass at the bottom of the now solid
nourishing material is not pink, but colourless.

Schroter (Cohn’s ‘ Beitrage zur Biologie d. Pflanzen,’ ii
Heft. p. 112) mentions in a footnote that he observed occasion-
ally on discs cut from a potato, mucous droplets of a pinkish
colour, which, when examined under the microscope, were seen
to consist entirely of torula cerevisiw. The cells were not
coloured.

Myfriend Professor Lankester informs me, that when carrying
on his researches on Bacterium rubescens (see this Journal,
New Series, vol. xiii) in the laboratory of the Botanic Garden,
Oxford, he observed a pink torula which spontaneously made
its appearance in a test-tube containing Pasteur’s solution.

pa hed H. MARSHALL WARD.

Observations on Saprolegnie.
By

HH. Marshall Ward, B.A.,
Berkeley Fellow of Owens College, Victoria University.

With Plate XXII.

Tue Saprolegniz are a family of Fungi characterised
especially by their aquatic habits and correlated delicate struc-
ture ; they are for the most part Saprophytes, flourishing on
the decaying bodies of animals or plants in water, though
several are now known as parasites on living members of both
kingdoms, in which they cause profound destructive changes,
sometimes ending in the death of the host.

To the latter parasitic forms belong Saprolegnia de
Baryi and 8. Schachtii, found in the cells of Algz and
Hepatic respectively, according to Walz! and Frank,? and
especially the Saprolegnia of the “ salmon disease,” according
to Professor Huxley.°

Generally described, the Saprolegnie consist of a thallus,*
bearing reproductive organs of two kinds—zoosporangia
and oogonia, with or without accompanying antheridia.
The thallus consists of long, branched, tubular hyphe, of
which the main portions are free or “ extramatrical ;” shorter,

1 “Bot. Zeitg.,’ 1870, p. 537.

2 « Krankheiten der Pfl.,’ erste Halfte, p. 383.

3 This Journal, 1882, p. 311.

4 It seems almost necessary to preserve the gencral name Thallus here, as
De Bary has done in his recent memoirs, although the Saprolegnie are

accepted as Fungi, chiefly on physiological grounds.
5. De Bary, ‘ Beitrige zur Morph. u. Phys. der Pilze,’ 4th ser., 1881. This

ON SAPROLEGNIA. te

rhizoid-like branches penetrating the tissues from which the
fungus radiates.

De Bary has lately pointed out that the “intramatrical’
portion of the thallus does not spread far in the tissues, and no
extension of the fungus occurs by outward developments from
its internal branches; the hyphz outside, however, continually
send downward prolongations, which take root and spread
slightly in the attacked tissues, and thus the area of ‘ intra-
matrical’’ hyphe becomes extended.!' Hence, if zoospores, &c.,
are prevented from again attacking the matrix from without,
the Saprolegnia thallus does not extend far within the tissues
—an important distinction between these fungi and those which,
like Pythium, &c., attack a host at one point and send rami-
fications in all directions inside the tissues. The difference is,
roughly speaking, analogous to that between a banyan tree and
a bamboo, in so far that the former extends its area of feeding
ground by sending down prolongations from its outer branches
to root afresh in the matrix or earth, while the latter extends
itself under the surface by means of underground shoots, which
protrude here and there further from the parent stock.

The tubular hyphe forming the thallus vary much in
diameter, partly according to position, main stems and branches
being thicker than secondary and tertiary ones; their cellulose .
walls are very thin and transparent, and enclose protoplasmic
and oily contents. ‘These latter are, as a rule, very coarsely
granular, causing the thallus branches to appear dense and
opaque, usually with a yellowish hue.

Septa occur very rarely in the tubular branches, but are
always found separating the zoosporangia and the sexual re-
productive organs from the purely vegetative portions of the
thallus.

The Saprolegnie have usually been stated to be a-nucleate
in common with other fungi.” This, however, is not the case ;
publication is now the most important authority for the morphology of the
group.

' Loe. cit., pp. 95 to 97.

2 Luerssen, ‘Med, Pharm, Bot.’ 1879, B. 1, p. 72, and Sach’s ‘ Lehr-
buch,’ iv ed.

274 H. MARSHALL WARD.

Schmitz has described nuclei in these and allied fungi, as well
as in many others of the lower cryptogams.! I have lately
also found nuclei of a very definite character in the mycelium
of an allied fungus, and shall show that a perfectly definite
division of nuclear masses occurs in the zoosporangia of
Achlya and Saprolegnia. It is at least certain that the
Saprolegniz can no longer be regarded as devoid of nuclei,
and the same probably holds good for all but the very lowest
cryptogams; according to Schmitz the Phycochromacee and
Schizomycetes.

The reproduction of the Saprolegnie takes place by
means of asexual zoospores, produced in long zoosporangia ;
and sexual (at least morphologically they must be considered
so) oospores, produced in oogonia, with or without accom-
panying antheridia. The details concerning both these
kinds of structures may be deferred for the moment.

With this introduction, the immediate object of the present
essay may be entered upon; that is, to describe some observa-
tions made during the past summer and autumn on species of
Achlya and Saprolegnia, the two most important genera of
the group. ‘These observations are not all equally valuable or
new (although some of the facts were observed before I was
aware that others had discovered them),* but they have been
made quite independently of the literature,*and are thus of some
service as confirmatory evidence to those who wish to study the
subject further.

1 ©Sitzber. d. niederrhein Gesel. in Bonn,’ 1880. Quoted also in the
appendix to the recent Engl. trans. of Sach’s ‘ Textbook.’

2 T must take this opportunity of thanking Prof. De Bary, not only for
giving me material, but also kind advice and references in connection with

this work,

3 The now copious literature consists chiefly of the following, among
others :—Pringsheim, several papers in ‘ Jahrb. fiir wiss. Bot.,’ i, ii, and ix.
De Bary, Jahrb, f. wiss. Bot.,’ ii. Walz, ‘Bot. Zeit.,’ 1870. Cornu, ‘Ann.
de. Sc. Nat., 5. v.,’ vol. xvi,

The most important memoir, from a general morphological point of view, is
De Bary, ‘ Beitrage zur Morph. u. Phy. d. Pilze,’ 1881,

*_ See also Huxley on ‘‘ Saprolegnia in Relation to Salmon Disease,” this Jour-
nal, July, 1882.

ON SAPROLEGNIA. PW KS

I shall adopt the simple plan of describing what I have seen
and drawn, together with methods employed, leaving more
general conciusions until afterwards.

1, Achlya polyandra.—Masses of débris of “meal-worms”
on which this species had been grown some months previously,
and which had been kept in a cool cellar during the interval,
were placed, together with a fresh meal-worm in a large
deep glass beaker perfectly clean, with a considerable quantity
of boiled and filtered water, and a glass plate over the top; the
whole stood in a well-lighted room at the ordinary (summer)
temperature. In the course of two or three days, during
which the water was several times replaced, the floating grub
was seen to be developing pale, cottony filaments in all direc-
tions, on and in the water around. These filaments proved to
be slender, straight tubes, filled with hyaline protoplasm in
which numerous large granules were scattered, especially in
the larger specimens; the walls became distinctly coloured
blue in Schultz’s solution, and in H,SO,, after treatment with
iodine, the protoplasmic contents becoming yellow in the former
reagent.

After a considerable mass of these radiating tubules had _ be-
come developed, certain of them were found to bear zoospo-
rangia. The development of a zoosporange was observed
many times in the following manner :—A broad glass slip being
placed in the water under the whole growth of Achlya, the
attacked meal-worm was lifted up bodily, and transferred to
the stage of the microscope; plenty of water being carefully
added to the specimen, the upper, more or less floating
branches, could be easily observed with a Zeiss D with a little
care. The great advantage of this or a similar method is that
the Achlya goes on growing almost undisturbed, and fresh
water can be continually added as evaporation goes on. If a
higher power is needed, it is very easy to place a small piece
of very thin, perfectly dry and clean glass, so as to float on
the flooded object, and remove it dexterously afterwards.
These very delicate cryptogams will not grow in a normal man-
ner under the pressure of an ordinary cover-slip, if continued,

276 H. MARSHALL WARD.

and the above method (or that of suspending a small speci-
men grown on a fly’s leg in a drop of water under a cover-slip)
is advantageous in many ways. I may add, however, that with
care observations may be made with Zeiss E without any
cover-slip at all on favorably situated portions.

The Zoosporangium is simply the terminal portion of
a branch spreading freely in the water; this becomes slightly
dilated into a club-shaped body, into which very granular pro-
toplasm collects, giving the young zoosporange a dull grey
appearance, easily detected with a good handlens. The apex of
this body remains blunt, and the walls are not thickened
During the course of about an hour after the protoplasm has
slowly accumulated, the following changes occur :—(1) A thin
septum becomes formed at the base of the dilated portion,
separating its dark grey, non-vacuolated, coarsely granular
protoplasm from the more sparsely granular, vacuolated contents
of the rest of the branch; and (2) an aggregation of the con-
tents of the zoosporange around numerous centres takes place.
To take an instance actually observed (Pl. XXII, fig. 1). The
gradual swelling and filling of the zoosporange (a) was completed
by about lla.m.; at 11.15 (2) the septum had become formed,
and the protoplasmic contents were already denser and showing
signs of aggregation at many centres; at 11.45 this had pro-
ceeded so far that no doubt of the existence of a multitude of
small semi-detached masses could be entertained (fig. 1 ¢).

And now followed a most remarkable phenomenon. At
about twelve o’clock the appearance shown at ¢ was replaced
by one similar to that shown at 6; this had occurred in many
previous examples, and puzzled me exceedingly, and I was
accordingly prepared to watch the exact sequence of events in
this instance. The protoplasm at 11.50 was distinctly divided
up into a large number of nearly globular independent masses,
slightly compressing one another. No membrane could be de-
tected around any of these, though, as subsequent investiga-
tions showed, a sort of watery-looking, clear boundary stood
between the masses. At one to two minutes after 12 the
masses and their clear boundaries became indistinguishable ;

ON SAPROLEGNIA. Pa)

this fading away of the granular appearance took place so
quickly that it might almost be termed sudden, and the proto-
plasmic mass was now in a condition apparently like that he-
fore the division, except that its granules were smaller and
probably more numerous, and the mass seemed more trans-
lucent than before. During the next three minutes a large
number of small, clear, equidistant areas (&) made their
appearance in the now almost hyaline, fine-grained protoplasm,
anda curious, pale, watery look had replaced the dark grey ap-
pearance of the earlier stages. These bright vacuole-like spots
seemed approximately equal in number to the masses which
preceded them (D), and one could well believe that the clear
spaces form at points corresponding to the centres of the former
bodies ; this, however, could not be decided. In three or four
minutes after the last condition was figured, the almost sudden
reappearance, so to speak, of the rounded or polygonal solid
masses occurred (F), as if the contents had gone back to the
state figured at p. This time, however, the separation of the
masses became more evident, and at ten minutes past 12 the
apex of the sporangium gave way suddenly, and the whole
mass of separated blocks of protoplasm suddenly flowed out
into the surrounding water, and remained at the mouth of the
zoosporangium as a spherical clump of protoplasmic globules
(figs. 3 and 4).

The following peculiarities concerning these bodies and their
exit were noticed. At the moment before the apex of the
zoosporangium bursts the isolated, though closely packed,
masses of protoplasm showed slight amceboid movements, and
during the rapid expulsion were actively changing their form ;
the instant they reached the extericr, however, they all became
strictly spherical, if free, slightly polygonal if compressed by
neighbouring ones (fig. 4). The whole process of exit and
rounding off only occupies a few seconds, and in a well-grown
mass of Achlya dozens of zoosporangia may be emptying their
amoeboid contents at the same time; in such cases, also, the
presence of this pheuomenon can be detected with a hand-glass,
the heads of globules being quite distinct in a good light.

VOL, XXIII.—NEW SER, 7

278 H. MARSHALL WARD.

The future behaviour of these zoospores—for such they
must be considered—may be described, as before, from what
was observed in a given example. Careful examination of one
of the more loosely attached specimens on the outer portions
of the spherical group (figs. 3 and 4) convinces the observer
that an extremely delicate envelope becomes developed at the
periphery of the resting globular zoospoore (fig. 5a), and the
_ body remains in this condition for some hours.

The specimen referred to was drawn at 1 p.m., and remained
in this condition until about 4 p.m. Soon after this there were
signs of change going on in the neighbouring specimens, and
this particular globe was carefully watched. At 4.14 a slight
protuberance made its appearance at one side, rapidly increased
in size during the next minute or two (fig. 5 d—d), and a clear
space (e) was then seen separating the delicate envelope from
the granular, slightly amceboid protoplasmic contents, which
were, in fact, becoming withdrawn to pass through to the outside.
A minute afterwards the whole of the protoplasm was outside,
except a minute papilla, which slipped out forthwith (fig. 5 f, 9),
and. the mass commenced to writhe slowly in an ameboid
manner outside the very delicate empty envelope, in the side
of which could be seen the minute pore through which the zoo-
spore had slipped out. At 4.18, the moment of complete exit,
a clear spherical vacuole was seen at one side of the zoospore
(g); the latter then quickly acquired a reniform shape, and
from the sinus (corresponding to the hilus of the kidney), two
minute cilia with knobbed tips were observed to spring forth,
quickly grow in length, apparently at the expense of the knobs
at their ends, and begin to wave slowly about. The zoospore
now (4.25) commenced to swing perceptibly as the lashing of
the cilia became more vigorous; nevertheless it did not move
away for some time; at 4.35, however, the zoospore was freely
and rapidly moving about, and at once disappeared from the
field.

This process, described as faithfully as possible in one case,
was repeated by the contents of the rest of the spheres com-
posing the globular mass at the mouth of the zoosporangium

ON SAPROLEGNIA. 279

(fig. 4), and all were nearly in the same stage of development
at the same instant; consequently the exit of the zoospores
from the spherical envelopes can be readily observed when the
critical time is carefully watched.

The mass of empty envelopes remain behind, appearing as
an exceedingly delicate network (fig. 6), and even simulating
parenchyma of great tenuity, mutual pressure causing the
spheres to become polygonal. Here and there the minute pore
can be observed as a dark spot in the side of the envelope, and
sometimes a zoospore escapes much later than its companions.

There is little more to be said of the zoospores. Some time
after their exit they come to rest, round off, and each at once
commences to put forth a simple tube (fig. 7), having first lost
its cilia and vacuole, and acquired several brilliant granules,
which become arranged around the periphery. If the germina-
tion occurs on a proper matrix, such asa meal-worm, fly, &c.,
the tube enters and commences to grow into a rhizoid-like por-
tion, a new thallus becoming developed from outside. On glass,
&c., the tube soon reaches the end of a limited growth, its con-
tents fade, and the whole dies.

Among other abnormalities in the course of phenomena such
as the above, mention should be made of one which is not un-
commonly met with, and which I have drawn at fig. 8. In
certain of the zoosporangia the completely separated zoospores
remain behind, rounded off, and form their delicate membranous
envelopes while still in the cavity; not only so, they germi-
nate in this position, each pushing a short tube through the
sporangium wall (fig 8, 8) before emptying its contents on
the exterior. In the example figured, the apex of the sporan-
gium became open in the usual manner at length. The figure
B represents part of A under a higher power and twenty-
three hours later. These ‘* Dictyuchus” forms were obtained
from specimens of Achlya polyandra which had remained
about six hours in the same water on a slip of glass; towards
the end of the period the fungus was obviously passing into
a state of inanition. As fig. 8 B shows, the sporangium be-
comes filled with an apparent tissue of extreme delicacy—the

280 H. MARSHALL WARD.

empty membranes of the zoospores—and the name “ Dictyu-
chus” was given to express the net-like structure thus pro-
duced. Whether the genus “ Dictyuchus” exists on a firmer
basis than this I do not know.'

With respect to the sexual reproductive organs of this
Achlya, my observations cover a considerable field ; as before,
the description applies strictly to what I have seen. The
oogonia and antheridium branches become produced in
large quantities when the cultivated Achlya is allowed to re-
main quite still, floating on the surface of abundance of water ;
their presence is soon detected with a good hand lens, and
further examination gives the following information concerning
them. yi

The oogonia arise as globular or nearly pear-shaped
swellings of the ends of very short branchlets, developed at
nearly equal intervals along the course of a vigorous branch
(fig. 9); the short branchlet is usually much smaller in dia-
meter than the parent twig, but resembles it in possessing thin
walls and coarsely granular protoplasmic contents, and in its
cylindrical shape. The balloon-like terminal swelling receives
a large supply of protoplasm, which accumulates in it as a yel-
lowish-grey dense mass, and then becomes shut off from the
pedicle by a thin, sharply-marked septum (fig. 10). In the
pedicle, which is about as long as the longer diameter of the
oogonium, the remaining protoplasm is much more watery
and poorer in granules; the latter is inserted sharply, as it
were, into the parent branch, and there is no septum at the
base—the cavity of the two remains continuous throughout.
In vigorous specimens (fig. 10) the granules often seem to be
arranged in rows, embedded in the layer of transparent proto-
plasm lining the cylindrical cell walls. The groups of
Oogonia, marked by their yellowish-grey contents while
young, present a striking object (fig. 9), like groups of berries
developed in racemose order; the pedicles are usually slightly
curved in various directions. As well shown in fig. 9, the
oogonium-bearing branches may be of various orders, very

1 De Bary, loc. cit., p. 94, says this abnormality occurs in other species.

ON SAPROLEGNIA. 281
commonly secondary and tertiary. Since the thallus has accu-
mulated much material, and the asexual reproductive organs
have been for the most part emptied when the oogonia arise,
it is usual to find empty zoosporangia terminating the main
twigs. The production of lateral branches from beneath the
sporangia is characteristic of Achlya, and that such may bear
oogonia is sufficiently demonstrated by fig.9 8. Such is the
typical mode of development of the oogonium. Before pro-
ceeding to describe the changes which its contents undergo,
we may examine the mode of origin and growth of the so-called
“antheridial” branches.

These are longslender tubes,springing from the main branches
from points either close to the oogonia (fig. 10) or at greater
distances apart, or even from separate branches. The diameter
of the tube is commonly less than that of the pedicel, but may
equal it: within its thin walls are finely granular, watery pro-
toplasmic contents, not always easily distinguished. As seen
in fig. 12, the “antheridial branch” arises as a simple
tube ; it often begins to form branches soon after its origin,
and these spread in all directions, curling and waving as they
do so. In this manner they become wrapped or coiled around
objects, such as neighbouring branches or oogonia, with
which they come in contact (fig. 98). It is in this coiling
of the antheridium branch about an oogonium that the
first stage of a proper sexual process has been recognised by
earlier observers.

During the development of. the coiling antheridial tubes
above described, the granular, yellowish-grey contents of the
oogonium, undergo certain changes, which result in their
complete transformation into the egg-cells or oospheres. A
clear, almost watery spot appears in the centre of the mass
(fig. 11 @), and slowly increase in bulk as the dense grey
granular protoplasm recedes to the walls; in this latter are
large, fatty-looking granules, which seen from the surface
(fig. 11 4) are in slow but evident motion. This retirement
of the protoplasm to the sides is followed by another process ;
a collection of the whole mass into two or more clumps, which

282 Hl. MARSHALL WARD.

then slowly round off as naked oospheres (fig. 12), consist-
ing of the fatty protoplasm only, suspended in the oogonium
cavity, which appears otherwise empty. This collecting of the
protoplasm to form the eggs, or oospheres, is a remarkable
process in more respects than one; it takes place slowly, and
occupies several hours altogether. I willconfine my description
to one case observed.

An oogonium was favorably situated for observation
from above, and at about 12 noon had attained the stage
figured at fig. 11 a. The coarsely granular protoplasm
aggregating on the walls, was in a state of continuous slow-
flowing motion, quite distinct to one observing a given granule
from the upper side (fig. 11 64); this specimen was watched
carefully from this time forward till nearly 5 p.m., and under-
went changes which were figured as follows (fig. 14):

For a long time the mass on the walls slowly heaved and
flowed, without its lateral continuity becoming broken. At
about 1.30 to 2 o’clock, however, the surface view showed that
the dense layer was breaking up into more distinct masses ;
and at 2.35 oblique, broad bands of fatty granules represented
the connection between two large masses aggregated at the
sides (fig. 14 f). On watching the uppermost of these bands,
the slow breaking up and passage over to either side of the
granular protoplasm was distinctly observed (fig. 14 g, h).
About 5 or 10 minutes before 4 the whole of the protoplasm
was thus collected into two equal lumps, still somewhat flat-
tened on one surface to the walls of the oogonium, and stand-
ing on opposite sides (fig. 14 2,4). The next five minutes were
occupied in the collection of a few scattered granules, the rais-
ing up of the centre of each lump from the wall, and its
ultimate withdrawal altogether towards the centre of the
oogonium (fig. 14 7). During the latter process, the egg-
masses were distinctly amceboid; each had its surface alter-
nately raised into lumps and smoothed off again, and in some
cases small particles of the protoplasm became detached and
taken up again.! There is not the slightest doubt as to the

1 This detachment of protoplasmic masses occurs still more decidedly, ac-

ON SAPROLEGNIA. 283

accuracy of these observations ; the amceboid motion continued
for some time, then slowly. ceased, and at 10 minutes past 4,
the two perfectly spherical oospheres lay obliquely in the
oogonium, mutually in contact, as shown in fig. 14. The
oospheres in this condition are apparently ready for “ fertili-
zation,” and the following phenomena occur.

One or more of the antheridial tubes, coiled closely about
the oogonium (figs. 10, 11, &c.), while the above described
processes have been going on, begins to send a tubular process
through the oogonium wall, at or about the time when the
oospheres are smooth and rounded off; the tubular process
thus sent into the cavity of the oogonium (fig. 12) has been
termed the “ fertilising-tube.” It is a direct prolongation of
the “‘antheridial branch,” and contains finely granular proto-
plasm ; it grows for some time in the cavity of the oogonium,
coming in contact with the oospheres—even running on their
surfaces. I have never seen it enter an oosphere, nor have I
seen it open at the end or emptied of contents. Whether any-
thing passes from it to the oospheres cannot be decided; but
De Bary gives such strong reasons for doubting that any fer-
tilising process whatever occurs, and supports his conclusions
by so many examples and so much observation that it would
be presumptuous to attempt to decide the question without de-
voting at least equal energies and time to the task. So far as
my observations go, they decidedly fail to supply evidence
for the view that anything is emptied from the tube into the
oogonium or oosphere. Before offering any further remarks
on this subject, it will be convenient to describe the remaining
observations made on other species.

Achlya apiculata is the name by which Professor De Bary
designates a species not yet (I believe) described ; my observa-
tions on this form are not yet sufficient to enable me to do more
than depict the formation of the zoospores. I have never seen
the Oogonia or Antheridia.
cording to De Bary, in Saprolegnia ferax ; he thinks it due to the throwing

off of water. May not the bodies, however, be of the nature of the “Polar
cells ” thrown off from the animal ovum preparing for fertilization ?

284 H. MARSHALL WARD.

The zoosporangium (fig. 15) differs somewhat in shape from
that of A. polyandra, the apex especially being more pointed.
In a specimen carefully watched for some hours, the sporan-
gium was at first filled with very finely granular grey proto-
plasm, and two or three large vacuoles remained below, abutting
on the somewhat swollen-looking septum; the tube below the
septum contained many and large vacuoles, the nets or bridles
between which were slowly streaming. Such being the condi-
tion of affairs at 9.30, the only observed difference at 9.50 was
that the vacuoles had disappeared from the zoosporangium, and
the fine-grained protoplasm reached close up to the now more
sharply-marked septum. About 10 o’clock the tip of the spo-
range appeared brighter and marked by faint longitudinal striz
(fig. 15 c), and a slight tendency to the formation of brighter
areole: seemed evident in the protoplasm. At 19.10 this was
distinctly marked (fig. 15 p); the protoplasm arranged itself
slowly into polygonal masses, each with a brighter central part.
This stage lasted for nearly ten minutes, the division lines be-
coming brighter and sharper, until the blocks stood nearly iso-
lated, and then, quite suddenly, at 10.20, the separation lines
disappeared and the blocks fused together, and a uniform grey,
granular mass (g£) resulted as before. This particular sporange
was not observed further, but in fig. 16 are drawings of what
was seen in another specimen from the same cultivation. At
10.35 the breaking up into the preliminary blocks was nearly
complete (a, fig. 16) and very distinct ; the hard, sharp division
lines disappeared quite suddenly about two to three minutes
later, and then the evenly granular protoplasm became marked
out into bright areas (fig. 16 4) by small vacuole-like points.
These increased slowly in size, and at 10.45 the sporange pre-
sented a peculiar lustrous aspect, the granules appearing re-
markably sharp and black in the bright, watery-looking matrix.
It seemed also that there were relatively more vacuoles than
preliminary divisions: this difficult point could not be decided.
At 10.50 the second series of division planes were established
(fig. 16 c). I could not satisfy myself that each vacuole
occupied the centre of one of the blocks, though such was un-

ON SAPROLEGNIA. 285

doubtedly true sometimes ; it seemed that in some cases a large
block became further cut up into smaller ones. This process
proceeded very rapidly, and by 10.53 the zoospore masses (fig.
16 d) were finally isolated, and slipped out in the next two or
three minutes. ‘The further fate, &c., of the zoospores need
not be described in detail ; they behave essentially as before.

SAPROLEGNIA.

The observations on this genus will be confined to the forms
of Saprolegnia ferax (Pringsheim), and the following de-
scriptions, &c., will refer particularly to that called S. mon-
oica in the sense of the above author.

Through the kindness of Prof. De Bary, I was enabled to
infect “ meal-worms” and house-flies with S. monoica, and
in two or three days had excellent cultivations floating in
abundant water as before. The methods of observation, &c.,
need not be detailed; they are practically the same as those
described for Achlya.

Fig. 17 shows the various stages of development of the zoo-
sporangium and zoospores; the segmentation of the proto-
plasm takes place as before, and need not be further described.
At the completion of the second segmentation, the masses of
protoplasm behave in a manner quite different from those of
Achlya, however, since, instead ef simply slipping out of the
apex of the zoosporangium and then rounding off, they ac-
quire two terminal cilia at once, and pass off as actively
moving zoospores (fig. 17 f). Each zoospore is a top-shaped
mass of finely granular protoplasm, with two very long cilia
actively waving at its pointed (forward) end, and with a sort of
zone of three small vacuoles around its broader part (figs. 17 g,
19 a). In this condition it moves rapidly from the point of
exit, coming to rest (2) after some minutes. With care it is
quite possible to watch a zoospore through all its changes.
Fig. 19 shows the phases actually seen in the case of a zoospore
emitted from the zoosporangium in fig. 17. It became free
about 9 a.m., and moved actively for ten minutes, rounding off
and losing its cilia and vacuoles in an instant (c f. fig. 19 a,b).

286 H. MARSHALL WARD.

In this quiescent condition it remained for some hours un-
changed, excepting that an envelope was gradually formed on
its surface. At 2.30 p.m. the contents of the little sphere came
out (¢ and d) as an ameeboid, naked mass, which gradually
acquired’ a kidney-shape and a large vacuole, and developed
two lateral cilia; these latter, as before, arose as two minute
knobbed processes, which slowly increased in length and began
to wave, causing the body of the zoospore to swing more and
more, and at length (about 3 p.m.) to move away. This par-
ticular specimen was then lost. But I observed another (f to
¢) for nearly an hour and a half; it was just coming to rest (9)
about 4 o’clock, and had commenced to germinate before 5
p-m., growing very rapidly (¢) and then dying. In another
case (fig. 18) I followed the second more closely. It escaped
from the envelope (a, 0) about 2.10 p.m., and swarmed as a
kidney-shaped spore (c) for nearly half an hour; it then lost its
cilia, writhed two or three times in an ameeboid manner (d),
and suddenly became rounded off (e) as a naked spore. This
was at 2.55 p.m. At 3.15(f)it began to germinate, by throw-
ing out a slender tube, which had reached a considerable
length by 4 o’clock (7), when the whole was dying.

In the normal condition of affairs such a germinal tube
enters the body of the insect, and continues the life-cycle. In
some cultivations one often finds bright white clumps of ger-
minating zoospores (fig. 20) lying at the bottom of the water;
these result from numerous zoospores coming to rest about the
same time, falling quietly through the still water, and, again,
germinating almost simultaneously.

There is little more to be said concerning these processes.
The zoosporangia of this Saprolegnia vary in shape
within wonderfully wide limits; some are almost as broad as
long, others nearly tubular, while pyriform, top-shaped, and
irregular specimens of all kinds occur. In cultivations, allowed
to starve from want of renewed water, &c., imperfect and dis-
torted sporangia reach a certain stage of development, and
then, acquiring very thick walls, remain in a resting con-
dition, springing into activity again when the conditions of the

ON SAPROLEGNIA. 287

environment improve. I have drawn one or two specimens of
such dormant branches of the thallus of Sapr. monoica at
fig. 21. Ihave not yet obtained oogonia of Saprolegnia
monoica, and must refer to the literature ;! drawings of the
ripe oospores are given at fig. 22, but no attempt to produce
them by cultivation on my part have yet succeeded.

The immediate object of this paper, to describe accurately
a few careful observations in the hope that they may help to
stimulate others to pursue the subject, is now ended; but it
may be well, before concluding, to call attention to some general
conclusions which have been drawn lately, and on which such
observations as the above throw light.

Apart from the question as to whether the Saprolegnia be
regarded as true Fungi or not, they may certainly be con-
sidered as forming a distinct group of parasitic and saprophy-
tic organisms inhabiting water, and multiplying by means of
zoospores and oospores as above described.

As respects the asexual mode of reproduction, by means of
zoospores, all observers are now fairly in accord as to the main
facts. With respect to the processes of incomplete segmenta-
tion preceding the formation and escape of the zoospores from
the sporangium, it appears to be best explained as a pheno-
menon of nuclear division, in which the cell plate first
formed becomes used up again. Biisgen,? who observed a
similar process in several other cases, draws attention to Stras-
burger’s discovery,’ that in the development of pollen grains
and spores it sometimes happens that a “ primary cell plate ”
is first formed, and then disappears, as if its materials were used
up again. This certainly appears to explain the phenomena
of the division, &c., inside the sporangium of Saprolegnia;
but why should the protoplasm make so many tentative efforts,
so to speak, before once more growing out as a thallus? Why

1 The development of oogonia, &., in this form is very fully given by De
Bary, ‘ Beitr. z. Morph.,’ &c., iv.

? « Die Entwicklung der Phycomyceten-sporangien,” Jahrb. f. wiss. Bot., B.
xiii, 1882.

3 *Zell-bildung und Zell-theilung,’ ed. iii.

288 H. MARSHALL WARD.

should the protoplasmic masses, once having become zoospores,
still hesitate (if the word may be permittted) before growing
on, and, having rested awhile, again become zoospores, but of
a different kind?

It might be suggested that the entire series of phenomena
should be connected and looked at in some such way as the
following :

Ist. The zoospore masses are formed, excreting a clear inter-
calary substance (primary cell wall of Strasburger), which they
then take up again.

2nd. A more energetic separation follows, resulting in com-
plete isolation, passage out, and removal to a distance. This
active phase, though more energetic and lasting than the pre-
ceding, is in its turn superseded by a resting state, and the’
protoplasm excretes the substances for a membrane.

3rd. After the period of rest the protoplasm once more moves
actively (having left its membrane behind) as a still more ener-
getic zoospore—at least it moves for a longer period—which in
its turn comes to rest, but only for a short time prior to
germination.

4th. It then, having formed certain brilliant granules and a
cell wall, throws out a germinal tube at the expense of its con-
tents; this soon dies if no proper matrix be at hand.

Unfortunately this restatement of the matter does not seem
to help us. One can dimly see that the little protoplasmic
zoospore undergoes processes of activity and rest —possibly
partial exhaustion—and it is not absurd to conceive that some-
thing is gained by an active vacuolated stage.

In Achlya—the above applies to Saprolegnia'—the first
and second stages occur as before, only the second stage seems
to be less energetic, and the ameeboid bodies only succeed in
reaching the mouth ofthe sporangium. The third and fourth
stages are much the same.

In the “ Dictyuchus” form the second stage is still more

1 De Bary, however, says that both zoospore stages may become abnormally
suppressed, and the germinal tube be formed at once on leaving the sporan-
gium; this increases the difficulty. Loe. cit., p. 94.

ON SAPROLEGNIA. 289

abbreviated; it consists merely in complete isolation. The
resting globules germinate in sittin the zoosporangium.

That we are brought face to face here with a profound
problem in its simpler forms is obvious. Perhaps the only
light it affords us as yet is the suggestion once more of the
exceedingly complex nature of the changes proceeding in the
simplest piece of protoplasm. It appears somewhat significant
that the second form of zoospore—the reniform one with lateral
cilia—is that most constant. This is the only form in the
nearest fungoid allies of the Saprolegine, and must probably
be regarded as the most ancestral form; nevertheless it is
not easy to suggest how or why the other zoospore was ac-
quired.

With respect to the “sexual reproductive organs” of this
group, much has been written and many theories advanced since
Alexander Brown and Pringsheim first described them and
their relations. The antheridia were first believed to pour
granular matter into the oogonium amongst the ova (00-
spheres). Then Pringsheim discovered that the oospheres in
certain cases become normal oospores without the appearance
of antheridia. Certain small antherozoid-like bodies were
then believed to be set free and find their way into the
oogonia amongst the oospheres. Meanwhile other observers
denied that the “antheridia” either formed antherozoids,
or that the tubes sent into the oogonium emptied anything
into its cavity.

The discussion seems to have been somewhat in this state
when Cornu,' in 1872, described the process of fertilisation,
&c., as consisting neither in the formation and entry ot
antherozoids, nor the emptying of granules between the
oospheres, &c., but in the passage of protoplasmic contents
from the antheridium through the fertilising tube and
into the substance of the oospheres. This view has been
accepted somewhat widely.

De Bary seems to have maintained for some years that, in
some cases at least, no passage of material takes place through

1 «Ann. d. Se. Nat.,’ 5th ser., t. xv.

290 H. MARSHALL WARD.

the tube—at any rate, not as protoplasm; but that the tube
remains closed at the end, and never enters the subsiance of
the oosphere; on the contrary, the antheridium either
remains coiled round the oogonium, or the tube which it
sends into the cavity simply touches or pushes the oo-
spheres. Hethus thought that the fertilising influence must
pass through the closed walls of the tube which remains
closed.

Pringsheim, in 1874,! again examined the question, and
came to the conclusion that Cornu was wrong, and that where
the tube comes in contact with the oosphere it remains quite
distinct from it, however closely applied. ‘Thus no slow pass-
age over of protoplasm into the substance of the oosphere
occurs, and hardly any, if any, contents of the antheridium
disappear. Pringsheim further came to the conclusion that in
some cases, since the oospheres become ripe oospores
without any antheridial branches coming near them, the phe-
nomenon must be considered one of parthenogenesis.

De Bary’s lately published views have been already referred
to. He finds, after prolonged and exact researches, that not
only does no observable passage of anything take place through
the fertilisation tubes; not only does the naked oosphere
clothe itself with a membrane (thus indicating that it no longer
requires fertilisation) without the contact of the tube, but that
normal, ripe oospores are produced habitually in some forms
without an antheridium branch ever being formed at all.
Such cases De Bary considers not ‘‘ parthenogenetic,” in
Pringsheim’s sense, but apogamous.

One more point may be shortly adverted to. It appears
as said to be a constant phenomenon in certain forms, perhaps
in all, that the masses of protoplasm forming the oospheres
throw off smaller or larger portions of their substance during
their ameeboid movements preceding their final rounding off
as smooth oospheres; if these detached masses of protoplasm
are to be regarded as of the nature of the “ polar cell”

1 * Jahrb. f. wiss. Bot.,’ B. ix.

ON SAPROLEGNI A. 291

observed to be thrown off by animal ova prior to fertilisation,
may not the hypothesis thrown out by Balfour apply also to the
cases observed by De Bary ?

De Bary shows that these protoplasmic bodies are taken
up again, and that such oospheres as have again absorbed
the thrown-off bodies, become ripe oospores, capable of ger-
mination after a period of rest without being fertilised. Bal-
four suggested that the “ polar cells” are thrown off to prevent
parthenogenesis, i.e. to prevent the egg dividing up and
developing an embryo which has not benenied (in Darwin’s
sense) by receiving protoplasm from a distance; the further
development of the non-fertilised oospores (ova) of Sapro-
legniz may be possible because the ‘‘ polar cells” are again
absorbed? Here, however, the proper limits of the present
essay have been passed.

1 C f. Balfour’s ‘ Comparative Embryology,’ p. 58.

292 DR. VINCENT HARRIS.

On Double Staining Nucieated Blood-Corpuscles
with Anilin Dyes.

By

Vincent Harris, M.D.,
Demonstrator of Physiology at St. Bartholomew’s Hospital.

THE usefulness of the process of staining tissues with several
dyes has been abundantly proved. The general effect aimed
at is the staining of each separate part in a different colour, so
that for the purposes of histological demonstration each shall
be distinct and clear. In the use of certain of the most com-
monly employed and easily manipulated dyes, e.g. hema-
toxylin and picrocarmine, it is believed that a definite effect
may be always calculated upon when they are used in combi-
nation. With anilin stains, however, the results arrived at
appear to differ very materially if the methods of employment
are made to vary in even a very slight degree, and this has
been one of the causes of the restricted use of very beautiful
staining colours. It has been shown by several experimenters
that with combinations of anilin colours, there is a tendency
at any rate for certain dyes to pick out and stain different
parts of a tissue; but J think I am right in believing that no
certain result has hitherto been expected, except in the case of
a very few combinations.

There is no doubt that unless more definite results be
obtained with combinations of these dyes, the hope that with
them it will be possible to pick out each element of a tissue in
a different colour, and each kind of morbid growth in a similar
manner, will be long deferred.

The following notes were made during a series of experiments,
the object of which was to find out the best combination of anilin

DOUBLE STAINING NUCLEATED BLOOD-CORPUSCLES. 293

dyes for double staining, as some of the methods recommended
had, in my hands, turned out very unsatisfactorily. During
the year 1881 I had, at the request of Dr. Vandyke Carter,
devoted a considerable time to the preparation and investiga-
tion of the organs of patients who had died of “ spirillum
fever,’ and also of those of animals which had been experi-
mentally inoculated with the fever virus. We wished to de-
monstrate in these tissues and organs the presence of spirilla
as had been done with ease in the blood. We used, according
to the advice of Professor Koch, several anilin dyes in turn
with anything but certain results. The indefiniteness, there-
fore, of the staining struck me much, and induced me to think
that something might be done in the way of obtaining more
satisfactory combinations. ‘The demonstration of the bacillus
in tubercle has brought into prominence the manipulation of
combinations of these dyes, and the fact that since the first
process of Koch was introduced, large numbers of modifica-
tions of it have been brought forward, notably by Ehrlich,
Baumgarten, Ermengen, Gibbes, and others, point to the con-
clusion that a considerable amount of work remains to be done
before the subject is fully develoded. These experiments in
staining the nucleated corpuscles were undertaken, as I have
said, with a view of finding out whether, with such definite
elements of tissue, certain results could be obtained by staining
with certain anilin dyes in solutions of definite strength in
regular sequence.

Method of experiment.—After trying several methods,
the one I adopted was as follows :—Blood from various animals
—frogs, newts, &c.—was spread in thin layers upon #-inch
cover-glasses and allowed to dry in the air. Certain of the anilin
colours were then chosen as the first or primary colours; these
were fuchsin, eosin, rosein, and rosanilin, in aqueous
or dilute spirit solution, as red dyes; iodine or malachite
green, as green dyes; and methyl violet, Hoffman’s vio-
let, or gentian violet, as violet dyes. All of these were in
aqueous or dilute spirit solution. With each primary colour
was included a separate series of experiments. Several drops

VOL. XXI1I,—NEW SER, U

294, DR. VINCENT HARRIS.

of ‘the first solution were allowed to remain upon the dried
blood for about five minutes, and were then washed off with a
gentle stream of water from a wash-bottle. The cover-glass
was then dried in the flame of a spirit-lamp and allowed to
cool. When ready for the second dye, a small quantity was
dropped upon the cover-glass and allowed to remain the same
length of time. <A second washing with a stream of distilled
water followed until the washings were all but colourless ; dry-
ing and mounting in Canada balsam concluded the process.
Treatment with alcohol and clove oil previous to mounting in
several instances quite vitiated my result, and so I gave up
that method of dehydration and clearing. In cases where I
thought it possible to employ three dyes, the third was used in
a manner almost exactly similar to the one described above.
The method of fixing the corpuscles with osmic acid did not in
the least improve the staining of the corpuscles. The chief
precaution which appeared to be necessary was not to allow the
blood to coagulate, and to place it under conditions of rapid,
but natural, drying, e.g. in the sun’s rays.

ARRANGEMENT OF THE COMBINATIONS OF CoLouURs:

Series A.—Primary colour, red; tried with orange, yellow,
green, blue, violet, and brown.

lst Dye. 2np Dyess.
Substances used. Substances used.
Red— Yellow—Anilin Primrose.
Fuchsin—Lake. Orange—Tropeolin, or Aurin.
Rosein—Crimson. Green—lodine Green.
Kosin—Pink. Blue—Methylen Blue.

Violet—Hoffman’s.
Brown—Bismarck.

SERIES A.

Redand orange. (Kosin and aurin.)
This combination was unsuccessful, as the solution of aurin
had to be made with absolute alcohol, it being such a very in-

DOUBLE STAINING NUCLEATED BLOOD-CORPUSCLES. 295

soluble substance, and entirely replaced the eosin, which was a
saturated watery solution. The whole of the corpuscles were
stained a deep orange. So far I have been able to do little
with aurin as a dye, its great insolubility causes every solution
speedily to deposit crystals.

Redand yellow. (fuchsin and anilin primrose.)

Fuchsin, a salt of rosanilin, is a fine lake dye, partly soluble
in water, freely soluble in dilute spirit. Anilin primrose, a
penetrating yellow dye of the colour of picric acid, almost in-
soluble in water, and only partly in methylated spirit, from
which it quickly deposits crystals. After some difficulty in
obtaining a good specimen the corpuscles were found to have
stained thus. The nucleus, a yellowish red, not unmixed
crimson, the remainder of the coloured corpuscles a light
yellow, and the colourless corpuscles a light red. The com-
bination was not a good one, as the yellow proved a very diffi-
cult dye to turn out, but, judging by results, it had the greater
affinity for the protopiasm of the corpuscles, and less for the
nucleus.

Red and green. (Rosein and iodine green.)

Rosein, similar to fuchsin, but of a more deep crimson, is
partly soluble in water, very soluble in dilute spirit. Iodine
green is freely soluble in water; an excellent combination.
The coloured corpuscles were stained a bright red, with bluish-
green nuclei. The colourless corpuscles were easily made out
to be of three varieties in the blood (of newt): 1, entirely
stained in green ; 2, partly stained in green and partly yellowish
red, the nucleus green, and the surrounding protoplasm of the
other colour; 3, the large masses of nuclei-like bodies, said to
be developing colourless corpuscles, were deeply stained in
green.

The relation in size of the nuclei of the colourless corpuscles
to the stroma is a very variable one. Sometimes the nucleus
appears to occupy nearly the whole cell, at others, perhaps, not
more than a fourth or fifth of it. It would be possible to de-
scribe a larger number of different kinds of colourless cor-
puscles than the above if their reaction to the dyes were alone

296 DR. VINCENT HARRIS.

considered, but no doubt the effects of the staining fluids were
not constant.

Redand blue. (Fuchsin and methylen blue.)

Certainly one of the most successful combinations. The
methylin blue was used as a saturated solution in absolute
alcohol. In this case, as in most of the others, the blue was
used second to the fuchsin, and vice versa, with similar
results. In these specimens the nuclei of the coloured cor-
puscles were deeply stained blue. Of the remainder of the
corpuscle, a light greenish hue with the edge a-bright pink, or
where the staining had been less deep the whole of the cor-
puscle, except the nucleus, was stained pink with an edge of
a deeper but similar colour. The staining of the colourless
corpuscles was peculiar ; some were alight bluish green. These
were irregular and branched; others had a deeper purple
colour with unstained spots (vacuoles?) ; and a third variety
appeared to be stained in two colours; in these a large central
mass looked like an immense nucleus. Other varieties might
be mentioned, as in the last combination, but as it is possible
that the size, amount of granules, and the staining are, except
in the three varieties, mere differences in amount not in kind,
it is scarcely necessary to mention them. For purposes of
demonstrating the divisions, irregularities in shape and in
varieties of colourless corpuscles, I am strongly inclined to
recommend staining with methylen blue and fuchsin as the
best combination possible.

Red and blue, 2nd combination. (Fuchsin and soluble
anilin blue.)

Nuclei stained red, as well as the colourless corpuscles ;
stroma a light blue. It is as good a combination as the last.

Red and violet. (Kosin and methyl violet.)

With these colours there appeared to be a mixture of the
dyes in the corpuscle, and the nuclei were not distinct.

Red and brown. (Fuchsin and Bismarck Brown.)

Bismarck brown is an anilin dye of considerable utility ; it
is partially soluble in water, more so in water with a few
drops of glycerin added to it, and easily soluble in dilute

DOUBLE STAINING NUCLEATED BLOOD-CORPUSCLES. 297

methylated spirit. The solution used was 2 per cent. in
dilute spirit. The corpuscles stained easily and in a fairly
regular manner, the nucleus a deep red, the stroma a fine
brown. The colourless corpuscles pinkish red. In a few
cases I noticed a mixture of the colours in the corpuscles.
In using Bismarck brown, a dye much employed on the Con-
tinent, I find it best to immerse the specimens in it for twenty
to thirty hours, and then they will retain their colour even
if passed slowly through the dehydrating and clearing fluids.

Red and brown, 2nd combination. (Kosin and ve-
suvin.) Vesuvin and Bismarck Brown are said to have the same
chemical formula, and are probably identical.

The vesuvin was used in a strong aqueous solution. The
corpuscles easily stained with eosin, nuclei and _ colourless
corpuscles a deep pinkish colour, and the stroma of the
coloured corpuscles a light pink. After double staining with
vesuvin the stroma stained a light yellowish brown, leaving
the nuclei and the colourless corpuscles stained as before. A
very successful combination.

Series B.—Primary colour green; tried with brown, red,
orange, yellow, blue, and violet.

Ist Dye. Qnp Dyess.
Substances used. Substances used.
Green—lIodine or Malachite Brown—Bismarck.
Green. Red—Flamingo or Ponceau.

Orange—Aurin and Anilin Orange.
Yellow—Anilin Primrose.
Blue—Bleu de Lyon.
Violet—Methyl Violet.

SERIES B.

Green and brown. (Iodine green and Bismarck brown.)

The colourless corpuscles and the nuclei of the coloured
corpuscles distinctly green; the stroma brown.

Green andred. (Iodine green and flamingo.)

Flamingo, a deep brownish red, soluble in water partly, but
freely in dilute spirit. The latter solution was the one em-
ployed. The nuclei of the coloured and the whole of the

298 DR. VINCENT HARRIS.

colourless corpuscle showed a deep bluish green, and the stroma
was coloured pink.

Green and red, 2nd combination. (Malachite green and
ponceau.)

Malachite green is freely soluble in water, and was used in
aqueous solution. The nuclei of the coloured corpuscles
stained green and the stroma a light pink. The colourless
corpuscles of the same shade as the nuclei. On keeping the
specimens it was found that the green dye almost entirely
disappeared.

Green and orange. (Malachite green and fluorescin.)

The nuclei of the coloured corpuscles stained a light yel-
lowish green, as did also the colourless corpuscles. The
stroma stained a yellow colour. The green stain was very
temporary.

Green and orange. (Malachite green and aurin.)

Entirely unsuccessful.

Green and yellow. (Iodine green and anilin primrose.)

The colourless corpuscles and the nuclei of coloured cor-
puscles stained green ; the stroma of the latter yellow.

Greenand blue. (Iodine green and Bleu de Lyon.)

Double staining quite unsuccessful. The Bleu de Lyon was
employed in a dilute spirit solution.

Green and violet. (Malachite green and methyl violet.)

Combination not good. The nuclei stained a very light
purple, as did also the colourless corpuscles, whilst the stroma
was a pinkish yellow. The green and violet apparently
mingled in staining the nuclei.

Series C.—Primary colour, violet; tried with brown, red,
orange, yellow, green, and blue.

Ist Dyz. Qnp Dyess.
Substances used. Substances used,
Violet— Brown—Bismarck.
Methyl Violet. Red—Flamingo, Eosin, Anilin Scarlet.
Gentian Violet. Orange—Tropeolin.

Hoffman’s Violet. Yellow—dAnilin Primrose.
Green—lodine Green.
Blue—Methylen Blue.

DOUBLE STAINING NUCLEATED BLOOD-CORPUSCLES., 299

Violet and brown. (Hoffman’s violet and Bismarck
brown.)

An excellent combination. An aqueous solution of the
violet was used, and a dilute spirit solution of the Bismarck,
brown. The result showed excellent double staining. The
nuclei of the coloured corpuscles and the colourless corpuscles
were stained a reddish brown, and the stroma a light brown.
The brown evidently stained the whole of the corpuscle
stroma, and nucleus, but met the violet in the nucleus, and
together stained it the colour it presented.

Violet and red. (Hoffman’s violet and flamingo.) Fla-
mingo is a mixture of rosanilin and Bismarck brown.

The nuclei and stroma stained two shades of the same colour,
probably a mixture of the two used, mauve.

Violet and red, 2nd combination. (Gentian violet and
anilin scarlet.)

Unsuccessful,

Violet and red, 3rd combination. (Gentian violet and
eosin. )

Nuclei of coloured corpuscles, deeply stained red, with the
colourless corpuscles a similar colour with their nuclei. Stroma
a light pink. |

Violet and orange. (Hoffman’s violet and tropeolin.)

Tropzolin used in 1 % watery solution. Double staining
entirely failed.

Violet and yellow. (Gentian violet and anilin primrose),

Strangely enough the corpuscles were stained two shades of
green.

Violet and blue. (Methyl violet and methylen blue.)

Certainly one of the best combinations tried. The methyl
violet is a very pink dye, and the blue a very deep blue. The
latter stained the nuclei, the former the stroma.

Resulis—From my experiments I draw the following conclu-
sions as to double anilin staining of the nucleated corpuscles.
It seems reasonable to look upon such corpuscles as made up of
only a few varieties of tissue, and as such I have spoken of
them.

300 DR. VINCENT HARRIS.

1. The only entirely successful combinations were the
following :
Rosein and anilin green.
Fuchsin and methylen blue.
Fuchsin and Bismarck brown.
Fosin and vesuvin.
Iodine green and Bismarck brown.
Hoffman’s violet and Bismarck brown.
Anilin violet and methylen blue.

2. The green dyes were not at all permanent. I have proved
this with both malachite and iodine greens.

3. Even with the above successful combinations the results
varied in a most extraordinary manner, whilst the circum-
stances of the staining operation and the solutions appeared to
be unvaried, the very greatest care being required to produce
a constant result. One thing necessary for success was cer-
tainly that the solutions should be quite fresh. This is likely
to prove a great objection to the general introduction of anilin
dyes into use. The simple method of dehydration employed,
of course, could not be employed in the preparation of tissues,
although it does for blood, sputa, &c.

4, The result was materially affected by the time each dye

was allowed to remain in contact with the blood.
It is worthy of note that according to the evidence of com-

petent authorities, various chemically diffused anilin dyes have
been sold under the same commercial name; and so, both in
the preceding notes and also in the annexed table, it should
be said that the anilin dyes used were obtained from Messrs.
Hopkins and Williams, Hatton Garden, W.C. The follow-
ing table (drawn up August, 1882) includes the dyes used in
above experiments :

I am much indebted to Mr. Meldola, of Messrs. Brooks,
Spiller & Co., for valuable information as to the chemical com-
position and relations of many of the above anilines, and of
the commercial names, &c., of others. This information he
kindly furnished at the request of Dr. Russell.

301

DOUBLE STAINING NUCLEATED BLOOD-CORPUSCLES.

‘urqdsoyd puv urplorsor jo ornqxtut & st neoouOg ,

tr ee

“gtaids
ura[qnjos—ayd
-Ing S,a9[]1dg

*19]8M UT “TOS
—2]0I A 0}1e0u
‘antq uviaXyz,

"10 YOM
ur ‘jos A[aaa1j—
Sutyeuruopaid
ang ey} 49]
-OIA UeI4QUaDH

“gtids ut *[Os
Ayaaay § ATperused
JOyeM WI ‘OS—

suryeurmopard
poy oy “497
“OrK TAGIOT"

“qruids
oyOyIp ur pues
Joye Ul ‘OS
S[901J—4 9 01 A.

SLE OTE

*LUIOI A

“OYE M
UI ‘jos syoo.7
SSCL a el

sOUITP
—oan|[q Is8199g

“TOYO M
at ‘jos Ayaary
SUSE, ASO)

“qraids
Ul pue J9yeA
ur ‘jos Ayaary
—onjgq daaq
Aion @ Sontg

ua Ay4oW
‘quids Suo034s ut
os yaaa} £104
“BM UL *[OSTI—
uokryT op nolg

"OYE M
‘jos Ayaoay
Pe NT
SVs ea TOR

Ut

.

‘TaTg

“quatds

UI pur Jaye
ur ‘jos Ayoory
—us0Iy onig
ssoy @ “WadIy
Se ad

“q1I1ds IO 1048A\
ut ‘fos Ayaexy
—usaIy ong
‘Ud01H) DUIPOT

Brictcares)

“grids “qyour
‘jos Ajjaed
Ajuo—aso1
SCO ie ELE ELOY,

Ul

"40908910
“ng Atingrynvoq
Suloq uoryny
“OS at} “qratds
UI ‘JOS ‘1a}eM
UI *[OSUI—Ao]
-[2X Yystuse.

‘alosai ony

*MOTTSX

“qiids
Ul pue 10qeA

UL *[OS—ULUBIJJES

“qtuds
ur ‘{jao1y you 4nq
‘os o10ur £ 104eM
ur ‘jos Ayyernaed
—AISURIC) SIMO]
PX ‘urydsoyg

“qtatds “13000

UI OS a1OU £104
“BM Ur "Jos ATyaed
—Sa]vog SuIua4

“SID aor[aX daoq

“quuids 94nq
“Ip UL*TOS f dayem UT
‘jos A}jaed—uts yong

“‘yatds oqnjrp ut “fos
Ajaaqy 1048 UT “JOS
Ayyaed — wipIuesoy

“quids oynqip
ut Ajaory faaqeM Ut
‘jos A]2ed—uosut19
pat daap ,‘neaou0g

‘quuids ‘yyout ut os fT
-901J $1948 UI ‘TOS AT

ur ‘urpoedody, |-yaed— par ystumorq

“0741p ‘0941p

—dISURIO UI[IUY

"JOyooye o4ny
-OSqe UI OS aL0UU
‘glaids Suoiys ut
"jos Ajqaed § 197em

daop ‘osurmey iy,
“grads

payeAyyour ut os AT
-90JJ £199BA\ UI "JOST
sr LACS Pe Ne

“1OJVM UL *[OS

Ur ‘josul—utiny | AJao1j—yurg ‘aso

‘TONVUGO,

"Cay

"19JBM UI “TOS

—urprosfayo

"19qeM UL
‘[OS—uUrIANSO A

“qtatds a4nqIp
UL *[OS £10]8A\
ur ‘jos Aqjeryaed

—yolvwsig

‘NMOUT

a a ee ee |

“LIWIdg NI GNV UALV AA NI SHILITIGNTIOG UIAHL HLIM ‘STA NITINY doIHOD AHL AO LIT aaldIssvIp

302 WALTER GARDINER.

Some recent Researches on the Continuity of
the Protoplasm through the Walls of Vege-
table Cells.

By

Walter Gardiner, B.A.,
Late Scholar of Clare College, Cambridge.

(From the Jodrell Laboratory, Royal Gardens, Kew).

Havine been for some time engaged in investigating the
subject of the continuity of protoplasm through the walls of
vegetable cells, it was with no small degree of interest that I
read Dr. Elsberg’s paper! in the hope of finding something
that would be of value for my research. In this, however, I
was disappointed, and having carefully gone over his paper,
and worked through his methods, I resolved to publish my
results, believing it to be of extreme importance in a subject
such as he treats of, that his statements should, if correct, re-
ceive every confirmation and support, or if any mistake had
arisen, that such mistakes should as quickly as possible be
rectified.

There are several points in his paper that I should like to
touch upon before giving my own conclusions.

As Dr. Elsberg at the outset admits that he is not a botanist,
it is perhaps not surprising to find a want of accuracy in
his botanical terminology. Thus, he talks of leaf paren-
chyma cells as ‘‘ presenting blunt polygons separated from
one another by a shining rim of cellulose,” meaning, I
suppose, polygonal cells with thin cell walls. He then pro-
ceeds to rechristen protoplasm, and proposes to substitute for

1 «Quart. Journ. Mic. Sci.,’ No. lxxxix, Jan. 1883.

RESEARCHES ON PROTOPLASMIC CONTINUITY. 303

it the certainly not euphonious name bioplasson. To this it
may be answered that, although the word protoplasm conveys
in some ways an unsatisfactory idea, yet its persistent and wide
use, would warrant its being retained, rather than that scientific
terminology should be burdened with another new word, when
the value of the original word has been perfectly well defined,
and has for the biological student a perfectly clear meaning.

As far as I can judge Dr. Elsberg appears to confuse reticu-
late arrangement with reticulate structure, for he uses the
same expression, “reticulated living matter” for both. As
examples of such reticulated living matter, he gives Zygnema
cruciatum; the description of which he quotes from Sachs—
only altering the words “ primordial utricle”’ for parietal
sac.”! Other examples are: young cells of Zea mais,
Fritillaria imperialis, and Vicia faba; hairs of Tra-
descantia virginica, and Cucurbita. All these are of
course examples of reticulate arrangement of protoplasm, and
have nothing to do with the structure of the protoplasm
itself.

Passing on to where he treats of the analogy between ani-
mals and plants, his terminology again becomes somewhat
confused, in his endeavours to carry the analogy too far; for
comparing the fact, that just as the animal cell is limited by
its layer of cement substance, so is the plant-cell limited by
its layer of cellulose, he proposes to commemorate “ Schleiden
and his cell doctrine,” by making the word cellulose subserve
for the limiting membrane of both the animal and the plant-
cell. It is quite obvious, however, that this is impossible, for
the term cellulose is a name applied to a definite chemical
substance with definite properties, and does not necessarily
carry with it the idea of a limiting membrane at all. Cement
substance, so far as I am aware, does not, for example, give a
blue colour with iodine and sulphuric acid, nor furnish gun-
cotton when acted upon with nitric acid.

We now come to the most important part of the paper, where
Dr. Elsberg treats of the perforation of the cell wall.

' Sach’s ‘Text-book of Botany,’ 1882, p. 46.

304 WALTER GARDINER.

The first experiments were made upon Nierembergia
(printed Norembergia) gracilis.! He took pieces of the
flower—whether calyx, corolla, stamens, or pistil does not
transpire—treated them with a 2 per-cent. solution of silver
nitrate for half an hour, or with a *d per-cent. solution of gold
chloride for forty minutes, washed, exposed to daylight, and
examined. ‘The silver nitrate preparations when seen from the
surface, showed the cell walls stained dark brown, and demon-
strated that every here and there were interruptions in their
continuity. I have unfortunately been unable to obtain flowers
of this plant, and have in consequence not had any opportunity
of making this observation for myself, but I should like to
point out that such pronounced and frequent interruption is
quite opposed to our present knowledge, and certainly to the
results I myself have obtained. It is much more probable that
the walls were pitted, and that the pit membrane being thin
escaped observation. The same may be said of the figure of
the hair of Nierembergia, the transverse walls of which are
probably pitted in a similar manner to those of the walls of
Athzea hairs, a figure of which occurs in Sach’s Text-book.?
If Dr. Elsberg’s figure is drawn to scale it can scarcely be
wondered at,if he has made a mistake, for his magnifying
powers have not been sufficiently high. Very frequently the
pit membrane is so thin that without very careful preparation,
it cannot be recognised under the highest powers, and in many
cases the only way to bring out such a membrane is to stain
the protoplasm and leave the membrane unstained, or to stain
and swell the membrane itself with Schultz solution (Chlor.
Zinc Iod.).

But it was from the study of sections of the petiole of Ficus
elastica, when treated with silver nitrate, that Dr. Elsberg
has obtained his most conclusive results.

He gives a drawing of one of his preparations, and it is its
appearance, and the appended description of it, which, perhaps,
forms the most startling part of the whole paper; for we are

1 « Nierembergia gracilis,” Hook, ‘ Bot. Mag.,’ 58, 3108.
2 Loe. cit., p. 43.

RESEAROHES ON PROTOPLASMIC CONTINUITY. 305

told that “what has been sometimes described by authors,
especially in growing tissues, as ‘intercellular spaces,’ and
“middle lamelle ” in the cellulose were revealed to be, in a
number of instances, accumulations and filaments of living
matter wedged in between the plant cells.”’ Since it is impossible
to understand how an intercellular space (if Dr. Elsberg really
means space) can be an accumulation of anything, one must
proceed to deal with the question of the middle lamella being
an accumulation of living matter, &c. It is an undoubted
fact that the substance of the middle lamella resembles proto-
plasm in many of its properties. Thus, like protoplasm, it
resists the action of strong sulphuric acid in cases where it has
attained to any pronounced degree of development, and, like it
again, dissolves in strong potash or in Schultz’s mixture ; and
it is very noticeable that many reagents which are used as
special stains for the protoplasm will also stain the middle
lamella. But whatever view be taken as to the nature of the
middle lamella and the thickened cell wall, and no matter
whether one accepts the intussusception theory of Nageli,! or
the apposition theory of Schmitz? and Strasburger,? it is quite
certain that at the time of its first formation the cell wall is
essentially cellulose, and is thickened by deposits of cellulose
substance. If protoplasm in any way enters into the constitu-
tion of, or forms an integral part of, that structure which we
recognise as cell wall, it is, to say the least of it, hard to ima-
gine, even on the well-nigh exploded intussusception theory,
that such large quantities of protoplasm should be present, not
only to replace the structure which we are accustomed to regard
as middle lamella, and as such consisting of altered cell wall,
but even large areas on either side of it, in such proportion as

' Nageli, “ Die Starkekorner.”

Nageli and Schwendener, ‘“‘ Das Microscop,” &c.

2 Schmitz, “ Sitzber. d. niederrhein Ges. in Bonn,’’ 1879 and 1880.

3 Strasburger, “ Bau und Wachstum,” Leipzig, 1882. In connection with
this subject, cf. also Schimper, ‘“‘ Ueber das Wachstum der Starkekorner,”
‘Bot. Zeit.,’ 1881, 186; and Mayer, ‘“ Ueber die Structur der Starkekorner,”
‘Bot. Zeit.,’ 1881, 844.

306 WALTER GARDINER.

to cause the wall to consist as much of protoplasm as of
cellulose.

It would rather be expected that if protoplasm does perforate
the substance of the cell wall, such perforations would assume
the form of fine threads, of such a degree of tenuity that they
could only be recognised with great difficulty, involving very
careful preparation and the use of very high powers. This
subject has already been ably dealt with by Strasburger,'
whose ideas have received confirmation from Tangl’s? researches
published in his work, on the Structure of the Endosperm Cells
of Strychnos, Phenix, and Areca, and from the results I
myself have obtained in the pulvini of Mimosa, Robinia, and
Amicia. In my later work, which will shortly be published
by the Royal Society, I shall be in a position to show that, as
far as my investigations have as yet progressed, there has not
only not been the least suggestion of the presence of large
quantities of protoplasm in the cell wall, but also that no
examples of reticulate arrangement have been met with in
those cases where perforation actually takes place.

I now propose to give the experiments which were made
with a view of testing Dr. Elsberg’s results.

Unfortunately the name of the grass he investigated is not
given. I examined in detail two grasses, viz. Poa nemoralis
and Bromus maritensis. In each case it was apparent that
when mounted in dilute glycerine a distinct network structure
could be made out in the chlorophyll grains. The boundary
line of each grain was badly defined, and it was very hard to
recognise with any certainty whether the reticulate appearance
was confined to the immediate substance of the grain, or
whether it extended beyond these limits. Indeed, in some

1 € Ueber den Bau und das Wachstum der Zellhaute,’ p. 246.

2 Pringsheim, ‘Jahrb.,’ vol. xii, p. 170.

3 © Quart. Journ. Mier. Sci.,’ Oct., 1882; ‘Roy. Soc. Proe.,’ Nov. 11th,
1882. See also the conclusions arrived at by Russow from the callus reaction
given by the closing membrane of the nits of Phloem parenchyma. ‘ Sitzber,
der Dorpat Naturfor.,’ 1882, p. 350, and ‘Bot. Central,’ viii, 1888, p. 271,
also ‘ Strasburger Sitzber. d Niederrh Ges.’ 4 Dec. 1882, p. 12.

RESEARCHES ON PROTOPLASMIC CONTINUITY. 307

instances, it certainly appeared to extend from the chlorophyll
grain into the general cell protoplasm.

In order to see whether any abnormal appearances had been
brought about by the action of the dilute glycerine, some
pieces of the blade were mounted in expressed cell sap and
examined. As the thickness of the blade was, however, too
great to allow of satisfactory observations being made with
high powers, a small piece was teased out before mounting ;
and since it was found, that during the teasing process and
rupture of the tissues, no observable alterations had taken
place in the cells, when such a preparation was compared with
an uninjured one, in subsequent experiments teased out pre-
parations were made use of.

When such a preparation is mounted in SenTe cell sap
and examined, it becomes apparent that the outline of each
chlorophyll corpuscle is quite defined and distinct, and that
little or nothing can be made out of the reticulate structure.
I was also quite unable to observe any network in the general
cell protoplasm. If, however, dilute glycerine, or simply
water, be run under the cover glass, the corpuscles will be
seen to gradually swell up, and, in so doing, to display more
and more distinctly a reticulate structure. The outline also
becomes more and more diffuse, and one almost begins to
make out that the network appears to extend beyond the
grain into the protoplasm. I am of opinion, however, that this
is not the case.

The structure of the unaltered chlorophyll grain, and the
action of reagents upon it, can be much better followed in
thin leaves with large grains. I found, for instance, Selagi-
nella uncinata very good material. And just as the chloro-
phyll grains in uninjured cells of zrial plants will swell in
the way I have described when treated with water, so will
those of water plants when the cell becomes broken into or
otherwise injured, e.g. the chlorophyll grains of Chara,
Vallisneria, and Elodia. In each case a reticulate appear-
ance is first produced, which, upon prolonged treatment, gives
way to a granulation, and is followed at length by complete

308 WALTER GARDINER.

disorganisation. The fact, however, appears worthy of notice,
that whatever light the action of reagents may throw upon the
constitution of chlorophyll corpuscles, yet that pronounced
reticulation of structure and diffuseness of outline are not
observable in grains which are normal and unaltered.

My results then agree with those of Dr. Elsberg in so far as
the reticulate structure of chlorophyll grains is concerned,
but I have been unable to trace any reticulation in the proto-
plasm itself. Since I was unfortunately unable to obtain any
Nierembergia materialit only remains for me to deal with
Ficus elastica. When transverse sections of the petiole
were examined in dilute glycerine I could make out the reticu-
lation in the chlorophyll corpuscles, but, as in the case of the
grass, was quite unable to see anything of the kind in the pro-
toplasm. ‘Thin transverse sections, treated for half an hour
with a 2 per-cent. solution of silver nitrate, washed, exposed to
daylight, and mounted in glycerine, exhibited a structure
somewhat similar to the figure drawn by Dr. Elsberg, viz. that
on the cut surfaces of the cell walls were a number of ex-
ceedingly small, darkly-stained patches, separated from one
another by light and unstained narrow areas. The reduction
appeared not to have taken place uniformly all over the section,
being, for instance, specially pronounced just under the epi-
dermis.

The most obvious questions that arose were: Are these dark
patches confined to the surface of the section, or are they
present, as one would naturally suppose, in the entire thick-
ness of the wall, and thus admit of being seen at any focus?
Secondly, what is their nature? Do they consist of stained
cellulose.

Now, the epidermal and cortical parenchyma cells being
freely pitted, it is easy to focus to any determinate depth by
fixing upon any given pit. I examined carefully in this
manner several thin and well-prepared sections, but was unable
to see any staining whatsoever below the free surface. On
the contrary, my observations led me distinctly to the conclu-
sion that the black patches were granules resting upon the

RESEARCHES ON PROTOPLASMIC CONTINUITY. 509
cut surface, and that the substance of the wall itself was quite
free from them.

I proceeded to make several experiments to test the truth
of this conclusion. When sections of the petiole are cut in
water a considerable amount of latex escapes from the injured
surface and runs over the section, and it seemed not impossible.
that the latex mechanically deposited on the cut surfaces of the
cell walls had reduced the silver. In order to expel the latex
as much as possible before cutting the sections, a short piece
of the petiole was taken and fitted into a bored india-rubber
cork, which was then tightly fastened into the shorter limb of
a manometer tube. The manometer was filled with mercury,
with the exception of a short length next the cork, which con-
tained water. ‘Then, under a pressure of about 50 inches of mer-
cury,a current of water was rapidly driven through the petiole
tissue, and the latex was almost entirely expelled. Sections of
the petiole however, showed the same granular appearance as
in the first instance, thus proving that the latex had not been
the cause of the reduction. Indeed, direct observation of
sections of fresh petiole showed that little or no reduction had
taken place in the laticiferous cells.

It was still possible, however, that some of the contents of
the other cells might have been smeared over the walls either
in the act of cutting or of escaping when cut. ‘Thin sections
of thé petiole were cut, some of which were vigorously shaken
with water in a test-tube. They were both treated together
with silver nitrate. In the shaken-up sections there was much
less reduction than in the others. I then resolved to try alcohol
material where coagulation and hardening of the protoplasm
would occur, and there would be a greater propability of
getting a clean surface. In order to make out whether the
action of the alcohol would interfere with the reduction of the
silver, sections of fresh petiole were cut in water, and then
treated with absolute alcohol. Having been washed with
water, and treated with silver nitrate in the usual manner,
they were examined, and it was seen that they exhibited the
granulation quite as well as fresh sections. ‘Thus having

VOL, XX1I1.—NEW SER, x

310 WALTER GARDINER.

established that alcohol did not interfere with the reduction,
I cut sections of petiole which had been for twenty-four hours
in absolute alcohol. The sections were washed and manipu-
lated as before. However, hardly any reduction was found to
have taken place on the cell walls, although the cell-
contents themselves exhibited well-marked reduction. This,
again, suggested that escaped cell-contents were the cause of
the granulation.

To make this quite plain I took sections which were well
reduced, and were in every way satisfactory preparations, and
with a camel-hair brush freely brushed their surfaces. Nearly
all the granules disappeared from the surface of the cell wall.
In the cell lumen they were still numerous where the contents
had not fallen away, but the cut surfaces of the walls them-
selves were quite clean and bright, and so was the entire thick-
ness of the cell wall.

Again, having taken sections of fresh material, i brushed
some well before the treatment with silver nitrate. Others I
did not brush. The former showed no reduction. The latter
exhibited well-marked granulation in the usual manner.

If the reduction be allowed to take place in diffused daylight
the granules are small. If in sunlight they are large. Con-
sequently, one may vary at pleasure the size of the granules,
and therefore the size of the meshes of the reticulum.

I think these experiments have sufficiently established that
the appearance described by Dr. Elsberg is simply due to the
fact that granules of reduced silver are deposited on the cut
surfaces of the cell walls, and that no staining occurs in the
substance of the wall itself. The whole appearance of proto-
plasmic continuity can be brushed away by mere mechanical
means, and the size of the granules can be varied at will. The
reduction of the silver in the cell wall is caused by some
of the cell contents which have escaped from the cell lu-
men.

The gold chloride preparations are not nearly so successful
as the silver nitrate, for much less reduction occurs. I was
unable to detect any network in the protoplasm, nor could I

RESEARCHES ON PROTOPLASMIC CONTINUITY. dll

trace any perforation of the cell wall by protoplasmic fila-
ments.

The identification of the substance causing the reduction
now simply resolves itself into a micro-chemical investigation.
That it was a very powerfully reducing agent was evident from
the fact that gold chloride, silver nitrate, osmic acid, and
chromic acid were all reduced by the cell contents of a very
great number of the parenchymatous cells. When tested with
alcannin it was shown that the presence of resin was confined
to the laticiferous cells and to the cuticle, and there being no
oil globules in any of the cell contents, with the exception of
the latex, it was probable that the reaction with osmic acid
had not been caused by oil or fat. The chromic acid reaction
pointed to tannin. Sections were therefore treated with ferric
chloride, when the cells occupying the same position as those
which had especially reduced the osmic, chromic acid, &c.,
were turned a brown-green colour, thus proving conclusively the
presence of tannin.!

In order to examine the distribution of the tannin cells,
transverse and longitudinal sections were treated with chromic
acid. They are shown to be present in the tissue in very great
numbers, and are especially abundant just under the epidermis
(at the very place where the greatest reduction of silver nitrate
occurs), and are arranged around the vascular bundle, being
also dotted about irregularly in the tissue. In longitudinal
section they are shown to be arranged in rows, end to end,
and their cell contents exhibit a fine reticulation.

A longitudinal section of alcohol material treated with
chromic acid shows that tannin has escaped over the cut
surface, and thus gives confirmation to the other results. In
the reduction experiments I placed along with the Ficus sec-
tions, sections of material such as the endosperm of Phoenix
dactylifera, where I knew that perforation of the cell wall

1 The tannin occurring in these cells is evidently not of the same character
as gallotannic acid which gives a black colour with ferric chloride. It is pro-
bably related to catechu-tannic acid, which also gives a brown-green when
treated with the above-mentioned reagent.

312 WALTER GARDINER.

did occur, and obtained no satisfactory results whatever.
Two years ago I made a number of experiments with gold
chloride and silver nitrate, and was forced to conclude that
they were unsatisfactory for botanical research. In my later
work also I have tried many modifications which have met
with no success as far as their use for studying the perfo-
ration of the cell wall by protoplasm is concerned, and I can
only add that the experiments made in connection with Dr.
Elsberg’s paper have fully confirmed my previous conclusions.

Before leaving the subject I should like to make a few
remarks upon a more important paper, viz. Professor Froim-
mann’s “ Beobachtungen uber Structur und Bewegungserschei-
nungen des Protoplasma der Pflanzenzellen.”’ !

The only part of his paper that I shall venture to comment
upon is that which deals with the perforation of the cell wall
and the subject of protoplasmic continuity. His results in
this direction may be summed up in his own words, in which
he claims to have established “ that protoplasmic nets pass
from one cell to anether, and connect neighbouring cells with
one another by means of either smaller or larger gaps and
crevices in the membrane.” ‘This structure was especially
clearly seen in the epidermal and hypodermal cells of the leaves
of Rhododendron ponticum and Dracena Draco, but
the leaves of Aloé arborescens, Crocus, Hyacinthus,
and Mentha were also investigated.

I propose to give several quotations from Professor From-
mann’s paper which will serve to illustrate the exact nature of
his statements, and at the same time make it quite clear, which
of these I wish to deal with, and to criticise.

On page 9, when treating of the epidermal and hypodermal
cells of Rhododendron ponticum, we find: “ The inter-
cellular spaces contain nets and granules ;” and further on—
“ The partition walls, however, do not always completely shut
off neighbouring cells from one another, but are pretty fre-
quently interrupted by gaps and crevices which are generally

' © Beob. uber Structur und Beweg. d. Protoplasma der Pflanzenzellen,’
Jena, 1880,

RESEARCHES ON PROTOPLASMIC CONTINUITY. 313

very narrow, so that only reticular threads or one or two series
of meshes find room in them, but they sometimes attain greater
breadth.”

On page 10, when speaking of the mode of formation of the
cell wall from the protoplasm—‘‘ And this view is also sup-
ported by the occurrence here and there observed of chloro-
phyll grains and coloured portions of nets, not only in the
crevices, but also in the substance of the partition walls into -
which they appear as it were forced.”

On page 11 the following statement occurs :—“ In the latter
(i.e. the cuticle) chlorophyll corpuscles are deposited here
and there.” :

The next quotation, on page 17, refers to Dracena Draco:
‘“‘ The partition walls separating the epidermal cells from one
another and from the subjacent cells may, for a short distance,
lose their brilliancy, but interruptions of continuity are more
frequent which either appear isolated, or three or five of them
on one partition wall; are partly very narrow, partly wider ;
may reach the diameter of a chlorophyll” grain, and are tra-
versed by isolated threads, or by narrow reticular bands, fre-
quently showing isolated thickened threads and nodes, by
means of which neighbouring cells are in connection.”

Again, page 21: “ Narrow cracks and wider crevices,
reaching the diameter of a chlorophyll granule, or of a
nucleus, may occur with varying frequency.”

On page 29, referring to the epidermis of Crocus and
Hyacinthus: “The threads of the neighbouring reticular
layers are sunk into the membrane, while in places where
cracks and crevices appear on the partition walls, the threads
extend through them, and connect the nets of neighbouring
cells with one another.”

Lastly, on page 38: “ But continuous roundish or band-
shaped lamellie are also deposited in those layers of membrane
which shut off the epidermis cells towards the outside, and
may appear on their surface, and be enclosed in their texture,
and as further in them as well as in the thicker partition walls
chlorophyll grains also occur, it can admit of no doubt that

314 WALTER GARDINER.

reticular protoplasm may enter into the structure of cell mem-
rane to a greater or less extent.”

Briefly stated, the principal facts involved in these state-
ments are: that open passages of a very appreciable size are of
very frequent occurrence in the common cell wall. That chlo-
rophyl corpuscles and protoplasmic reticula occur embedded in
its substance. That the intercellular spaces may contain gra-
nules and nets. That these nets and reticula of protoplasm
may be traced into the cell wall, and are particularly clearly
defined in the case of epidermal cells, running from the cell
lumen out into the cuticle.

With all deference to Professor Frommann, I cannot but
think that every one of these statements would be received
with some surprise by almost any botanist who is at all
acquainted with the histology of tissues.

I have investigated in as careful a manner as_ possible
the leaves of Rhododendron ponticum and Dracena
Draco, in order to give Professor Frommann’s results a fair
test. Transverse and longitudinal sections, as well as sections
parallel to the leaf surface, were examined in water, in cell sap,
and in dilute glycerine. Both fresh material and that preserved
in picric acid and absolute alcohol, were made use of. Iodine,
Chlor. Zinc. Iod., and hematoxylin, which latter Schmitz! so
successfully employed in his researches on the structure of
protoplasm, and the nucleus, were used as staining reagents.
Professor Frommann used expressed cell sap, sugar solution,
and dilute glycerine as fluids for mounting his preparations,
and employed methyl green as a stain; but since he expressly
states that the staining due to this reagent was confined to the
nucleus, and did not affect the nets and reticula, one must con-
clude that most of his observations were made upon pre-
parations which were simply mounted in the fluids before
mentioned.

1.do not intend to enter into detail with regard to the subject
of the intimate structure of the protoplasm. Suffice it to say
that in a great measure my results agree with those of Pro-

1 Schmitz, ‘Sitzber. d. niederrhein Ges. in Bonn,’ 1879.

RESEARCHES: ON PROTOPLASMIC CONTINUITY. 315

fessor Frommann, although I am unable to see the nets and
reticula with anything like the clearness with which he de-
scribes them in his paper, and figures them in his drawings.
There is, however, a distinct reticulation in the cells of both
Rhododendron and Draceua, especially in the former.
This reticulation is shown much more clearly by staining with
hzmatoxylin, tissue that has been preserved in picric acid, and
afterwards washed with alcohol. In Rhododendron very
many of the cells, and especially the pallisade- parenchyma cells,
contain tannin, and in these the reticulation is especially evi-
dent. I have frequently found this to be the case with tannin
cells in general. Treatment with chromic acid, osmic acid,
dilute potash, or dilute nitric acid will generally bring out a
reticulate structure, although at present I am unable to give
any explanation of the phenomenon. My attention was first
drawn to the fact when investigating the tannin cells which
occur in the pulvinus of Robinia pseudacacia.

As far as regards the reticulate structure, I can, therefore, in
the main bear out the statements of Professor Frommann, whose
results certainly accord with those of Schmitz! and Strasburger ;”
both these investigators having established that a reticulation
can be observed in the protoplasm and the nucleus. The
chlorophyll grains in the same way exhibit reticulation, but I
should like to point out, as I did in the case of Dr. Elsberg’s
research, that they are much swollen and somewhat disorganised
by the action of dilute glycerine. Nevertheless, after the most
careful preparation with picric acid and absolute alcohol and
subsequent staining, they still exhibited a distinct reticulate
structure, and agree fully with the description given by
Pringsheim® of the structure of chlorophyll grains in
general.

As to the occurrence of chlorophyll grains in the cell wall,
it is scarcely necessary to state that after very careful examina-
tion no such case was observed. Were I to attempt to explain

' Schmitz, loc. cit.

2 Strasburger, loc. cit.
3 Pringsheim, ‘ Lichtwirkung und Chlorophyll function,’ 1881, p. 28.

.

316 WALTER GARDINER.

Professor Frommann’s mistakes in this direction I might sug-
gest with regard to the occurrence of chlorophyll grains in the
cell walls that he was looking down through the thickness of
a cell wall upon a chlorophyll grain that had got into a pit;
and in the case of the grain in the cuticle, I can only put
forward the explanation that he was viewing a chlorophyll
grain in the guard cell of a stoma through the cuticle of an
epidermal cell.

There is the simpler view, that during swelling, and also by
mechanical means, some of the protoplasm was carried on to
the cut surface of the cell wall; and fig. 4, Plate I, which re-
presents a subepidermal cell of Dracena, certainly gives some
colour to this idea, although I put it forward with some diffi-
dence. Numerous preparations, treated and stained in various
ways, showed no sign of there being either granules or nets, or,
finally, any protoplasmic structure whatsoever in the intercel-
lular spaces.

Now, as to the subject of holes, gaps, and crevices in the cell
wall. At the outset I cannot but feel that Professor Frommann
was somewhat unfortunate in taking for his investigation such
small-celled tissue as occurs in the leaves of Dracena and
Rhododendron, and especially so as regards the epidermal
cells. In both these leaves, and particularly in Rhododen-
dron, the epidermal and parenchyma cells are very freely
pitted, and it is quite evident that what Professor Frommann
has taken for open passages between the cells are in reality
pits, each of which is closed by its own pit membrane. In
Plate II, figs. 4 and 5, he gives a drawing of the so-called
holes, but even a cursory examination of carefully prepared and
thin sections treated with Chlor. Zine Iod., or otherwise appro-
priately stained, will at once convince one that in every case a
closing membrane is present, and that the pits are not open.
As I have mentioned in the earlier part of this paper, the
occurrence of open pits in living cells would be quite opposed
to our accepted ideas of cell structure and cell mechanism. In
his descriptions, however, Professor Frommann gives passages
which suggest that he has noticed, but not identified, the pit-

RESEARCHES ON PROTOPLASMIC CONTINUITY. 317

closing membrane. For instance, on page 9 he observes “short
threads crossing from one side of the membrane (i.e. the mem-
brane of the cell wall bordering a gap) to the other;” and
again, page 17, ‘a somewhat stouter and more strongly refrac-
tive thread not unfrequently unites the portions of membrane
bordering the gap in a bridge-like manner, and crosses the
threads which pass through it.” On page 11 I may quote a
passage which shows the want of accuracy in his botanical
terminology, for he describes as middle lamella “the layers
immediately below the cuticle, which are double or three
times as thick as the partition walls between the epidermal
cells.”

The last point I have to deal with is the question of the
perforation of the cell wall, and the possibility of following
protoplasmic structures into its substance. Tang]! has shown
that it is impossible to see anything of the protoplasmic threads
in the cell walls of the endosperm cells of Strychnos, .
Phenix, and Areca, by direct observation in such fluids as
dilute glycerine, and even by ordinary staining. In each
case a special mode of preparation must be employed. I can
fully bear out his statements, and, indeed, in some of my
most striking examples of the occurrence of protoplasmic
filaments in the cell wall, it was quite impossible to see any-
thing of them when examined in the usual manner. I was
unable to observe anything of the kind in the epidermal cells
of Dracena and Rhododendron, but I think that some
satisfactory explanation can be given of the appearances which
Professor Frommann describes. If sections of Draczna be
examined in dilute glycerine, what appears to be a reticulate
structure can be distinctly observed, both on the upper, and the
side walls of the epidermal cells. ‘The outlines of the walls
bounding the cell lumen are not well defined, and, as Professor
Frommann says, the protoplasm seems to gradually merge, as
it were, into the cell wall, as he endeavours to represent in
fig. 25, plate I. Indeed, the whole appearance is most striking.
If, however, excessively thin and exactly transverse sections

! Pringsheim, ‘ Jahir,,’ |. ¢,

318 WALTER GARDINER.

be treated with Chlor. Zinc Iod. the walls swell; the boundary
bordering on the cell lumen becomes more distinct ; and definite
granules can be recognised deposited in the substance of the
swollen wall. I made several experiments to test the nature
of these granules. When the sections are warmed in dilute
potash solution, the granules appear to become somewhat
aggregated together, and signs of commencing solution can be
recognised. When boiled with a 5 per-cent. solution of potash
they are totally dissolved, the substance of the cell wall is left
quite clear, and its limits sharply defined. When sections are
treated with ether, and afterwards with boiling alcohol, consi-
derable solution takes place, attended with a clearing up of the
structure. These reactions appear to indicate that the granules
consist of wax, and, as De Bary! has shown, the presence of
wax is of frequent occurrence, not only on the surface of the
cuticle, but even embedded in the substance of the cuticularised
layers of the cell wall. The presence of these granules appears
to explain, in a satisfactory manner, both the diffuseness of
outline and the appearance of reticulation.

In Rhododendron, in the same way, an appearance of stria-
tion approaching to reticulation occurs. When examined in
Chlor. Zine Iod. it becomes apparent that very great cuticu-
larisation of the epidermis has taken place, the cuticularisation
extending even to the transverse walls. It is also apparent
that the striation is confined to the cuticularised layers, and
is separated from the cell lumen by a thin layer of cell wall,
which still gives the cellulose reaction. If treated with a
solution of potash, the cuticularised portion still shows stria-
tion, while the rest of the cell wall becomes clear and tran-
sparent. Examination of thin sections mounted in glycerine
supports these observations in every way. I need only refer
to Sach’s ‘'Text-book’? to show that the occurrence of striation
in the cuticularised layers of the cell wall of epidermal cells is
a perfectly well known phenomenon, His figure of the epi-

1 * Vegl. Anat.,’ Leipziz, 1877, p. 87.
2 Sach’s, loc, cit., p. 35.

RESEARCHES ON PROTOPLASMIC CONTINUITY. 319

dermal cell of Ilex aquifolium presents a case in point. I
used in this investigation one of Zeiss No. 1 microscopes,
Oculars 2 and 4, Objectives D and F, and a microscope of
Hartnack’s with two very excellent Objectives, No. 10, and an
Immersion, known as “ Foyer 1 m/m” (No. 132), Oculars
2 and 4.

320 J. EK. BLOMFIELD.

Review of Recent Researches on Spermatoge-
nesis.

By
J. E. Blomfield, M.A.,

Srnce the publication of my two former papers on the sub-
ject of Spermatogenesis (vide this Journal January, 1880, and
July, 1881) some accounts of the researches of other observers
have been published in foreign periodicals, which go far to
support the attempt made in the papers referred to, to establish
a general plan of Spermatogenesis which should be applicable
to the animal kingdom, or, at any rate, to a large part of
it. While our knowledge of the formation of the ovum was
fairly perfect, observations on the development of the cor-
responding sexual element, though by no means deficient in
number, yet were wanting in completeness and wideness of
scope. For instance, in such an animal as the frog, many had
observed the process for a short time and drawn conclusions as
to how this was conducted from data which were neeessarily
incomplete ; since the observations to be of any value in such
animals which procreate once in the year, must be continued
during the whole period from the time of procreation till the
next crop of spermatozoa is ready.

The publications now referred to, tend strongly to prove the
truth of the views expressed in my former papers, and confirm
the conclusion that the process as found in Mammals is closely
similar to that of Mollusca. This conclusion was, at the time,
somewhat speculative since it was founded on interpretations of
the drawings of other observers different to those which they
themselves had imposed upon their observations.

The first papers of which it is proposed here to give an ac-

RECENT RESEARCHES ON SPERMATOGENESIS. O21

count are by M. Duval, who has chosen for the investigation of
the general phenomena of Spermatogenesis the snail, with other
Pulmonate Gastropods, and the frog. Both of these animals
were also chosen by myself, though I took my earliest type
from the lower divisions of the animal kingdom, in the shape
of the earth-worm, an animal which exhibits the phenomenon
of Spermatogenesis, as found in the majority of the Vermes,
with singular clearness and facility for study.

M. Duval’s first communication on the subject is published
in the ‘ Rev. de Science Nat.,? tome vii, June, 1878, in which
he treats of Spermatogenesis in Pulmonate Gastropods, and
takes Helix as the typical representative of the class. Had I
been aware of the existence of this paper when I published
my researches over the same ground I should have referred to
it then, but I did not discover its existence till too late.

M. Duval commences by showing the importance of ex-
tending observations which apply to animals whose time of
procreation is limited to a definite period of the year, over
sufficient time to embrace the whole process, not as many have
done, making observations only at one time of the year, and
then drawing conclusions from observations which are, of
course, imperfect. Taking the condition of the ovotestis as
it exists in the winter, it is found to contain a few bundles of
spermatozoa and some free spermatozoa, while its wall is lined
with indifferent cells, to which he gives no particular name,
but which obviously correspond to the testicular and ovarian
epithelia. In the spring certain of these cells are seen to
enlarge, some rapidly, so that it soon becomes obvious that
they are destined to form the ova, while others grow more
slowly, and never attain the same size. These he calls male
ovules, and they form the starting-point from which the sper-
matozoa are evolved. ‘These male ovules consist of granular
cells containing one well-marked nucleus. The first change
they undergo consists in the development of other nuclei by
endogenous formation, which, by their further multiplication
by division, form the immature spermatozoa or spermatoblasts.
And here is the first point where his observations differ from

322 J. E. BLOMFIELD.

mine ; I did not observe any endogenous formation of nuclei,
and thought that the large multinucleated mass arose by
division of the nucleus. ‘These large multinucleated masses
are found during the late spring and early summer months,
when the male ovules have given place to mulberry-like masses
(grappes) adherent to the walls of the gland. These bodies
consist of several pyriform cells, each one containing one nu-
cleus, or a nucleus in the act of dividing into two, which are
united by strands of clear substance to a central mass containing
a large, well-marked nucleus. ‘This central mass he calls the
“‘ mother cell,” and its nucleus ‘ the principal nucleus.”

He gives no name to these cell groups, but in my paper 1
proposed to call them sperm-polyplasts. It is best to have some
name for such cell masses, as they are of very frequent if not
universal occurrence in the development of the spermatozoa of
various members of the animal kingdom. The term _blas-
tophor or blastophoral cell was applied by me to the body, which
takes no part in the formation of the spermatozoa, but plays the
part rather of a support for the others, which M. Duval calls
“ mother cell,” a term objectionable on account of its having
been applied to quite different factors of the process by other
authors. The term “spermatoblast”’ is used by me for the
young condition of the cells which actually and individually
develop each into a spermatozoon. ‘These terms may seem to
many to be unnecessary, but any one who has studied the subject
will find that the variety of names in use, and the manner
in which one author applies one term to one thing and another
applies the same term to a totally different, do not conduce to
clearness, and if the process is to be generalised to one or
more plans, definite names are necessary to point definite stages
in the course of events.

The next stage concerns the formation of the spermatozoa from
the spermatoblast, which is (according to Duval) brought about
in a more complicated manner than that described by me. In
each spermatoblast, which has much the shape of a racquet,
appears in the region of the handle, a minute spot exhibiting
great affinity for staining reagents, which he calls the “ cepha-

RECENT RESEARCHES ON SPERMATOGENESIS. 320

lic corpuscle.” This gradually enlarges until it has assumed
the shape and size of the head of the mature spermatozoon,
that is, more or less pear-shaped. At the same time, or directly
after, the tail of the spermatozoon appears in an unexplained
manner in the substance of the spermatoblast, while the sper-
matoblastic nucleus disappears, so that the group of spermato-
blasts have given place to a bundle of spermatozoa united in a
bundle and supported on the ‘“ mother cell.” The further
history concerns only the ‘* mother cell” and principal nucleus
(my blastophor) which, according to M. Duval’s observations
and my own, undergoes fatty degeneration and disappears.

M. Duval’s next communication is contained in the ‘ Rev.
de Science Nat.,’ Sept., 1879, in which he gives an account of
the two kinds of spermatozoa which since the time of Von
Siebold have been known to be present in Paludina. This
paper need not detain us longer than to mention that he finds
that the two kinds of spermatozoa develop on the same plan
as those of the snail, and that they are veritably two different
kinds of spermatozoa, a fact which some naturalists have
disputed.

The next publication from the pen of this writer is in the
‘Rev. de Sci. Nat.,’ Sept., 1880, on the “ Spermatogenesis of
the Common Frog,” which also formed part of the subject of
my second paper; and on comparing the two the similarity of
the figures are striking, though the interpretations drawn from
them are not quite in agreement.

He commences his description by showing the condition
of a testicular crypt during the winter, when the spermatozoa
are seen to be united into bundles, which bundles are arranged
in a radial manner, with remarkable regularity around the
walls of the sac. The wall of the sac itself is formed of
fibrous tissue and cells, and of the latter two kinds can be dis-
tinguished, one the “male ovules,” the other smaller and
more granular, embedded firmly among the male ovules, which
constitute the source from which the male ovules are derived.
This condition of things obtains throughout the winter till
the spring, when the spermatozoa are discharged and changes

324 J. E. BLOMFIELD,

begin in the “‘ male ovules ” for the preparation of next year’s
crop of spermatozoa. ‘These changes consist in the multipli-
cation of the nucleus till large multinucleated masses are pro-
duced, comparable to those found in the snail. By the end of the
spring these have reached a considerable size, and the next
step consists in an arrangement of the nuclei around the peri-
phery of the mass, while the centre part is free from them,
and undergoes a kind of liquefaction, thus giving rise to a
vesicular body. “As this body has grown from the original
“male ovule,”’ it has carried the smaller granular cells which
surrounded it on its surface towards the lumen of the sac,
where they may be seen resting on the protoplasm of the vesicle.
Each vesicle is now composed of a mass of uninucleated bodies,
which are the spermatoblasts, and the next stage is the trans-
formation of the spermatoblasts into spermatozoa. 'The exact
mode in which this is brought about, M. Duval is unable to
state; but reasoning from what he found in the snail, he
thinks it probable that the process is the same, and that a
“cephalic corpuscle’’ is produced while the spermatoblastic
nucleus is dissolved and the filament is formed from the plasma:
but the existence of the ‘‘ cephalic corpuscles ” he is unable
to determine.

The immature spermatozoa are so arranged that the tails
point to the interior of the vesicle, while the heads are arranged
radially around the concave inner face.

When the spermatoblasts have nearly changed into mature
spermatozoa dehiscence of the vesicle takes place, and the
spermatozoa are carried back to the wall of the sac, bringing .
the granular cells with them, where they become arranged in
bundles, and the same condition of things is found as that
which formed the starting point.

The granular cells, which have been carried back, undergo an
increase in size, and become ‘ male ovules ;”’ destined to repro-
duce the crop for the next year; but in teased preparations M.
Duval finds that often a bundle of spermatozoa is attached to
one of these granular cells, an attachment which he regards as
quite accidental, though, at one time, he was tempted to see a

RECENT RESEARCHES ON SPERMATOGENESIS. 325

connection between this cell and the “ mother cell” with its
‘principal nucleus” of the Mollusca. Further examination
caused him to abandon this view, as it was possible to find
distinct gradations between this cell and the “ male ovules ;”
and here is the chief point where his observations and mine
disagree. Undoubtedly these granular cells do exist and give
rise to the “male ovules” (spermatospores) forming part of
the testicular epithelium, but, according to my view of the
matter, the cells which are carried back with the young sperma-
tozoa are distinct from these and are, as it were, left behind
during the formation of the vesicle, being homologous and
analogous with the blastophoral cell of the Mollusca, and, like
that, undergoing a process of fatty degeneration. For, if a
testis be taken in the spring, at the right time, after the dis-
charge of the spermatozoa there will be found in the testicular
sacs a number of cells whose nuclei are disintegrated into
coloured granules, whose surface is marked with stric, repre-
senting the points of attachment of the spermatozoa, and whose
substance contains many fat granules, obviously representing
the degeneration of the supporting cell. This view of the
existence of the blastophoral cell and explanation of its origin
and function will be seen to receive considerable support from
the study of the spermatogenesis of a group in many ways
allied to the Amphibia, viz. the Selachians, to which I will
now proceed.

The most recent account of the development of the sper-
matozoa in the Selachians is contained in the ‘ Journal de l’Anat.
et de la Physiol., "No. 4, 1882, in a paper by M. G. Hermann.

The testis of the Selachians is remarkable for its simplicity
and points of resemblance to the ovary, the primitive germinal
cells, forming the germinal epithelium, sink into the subjacent
tissue in chains or groups (cordons de Pfliiger) in the same
way as they do in the ovary; but instead of forming Graafian
follicles they make up the ampulle in which the germinal
cells, male ovules, or spermatospores undergo changes which
result in the production of a crop of spermatozoa, and when
this process is over, the whole follicle or ampulla atrophies

VOL, XXIII.—NEW SER. ¥

326 J. KE. BLOMFIELD.

Fre. 1.—Spermatogenesis of Selachians. A. Transverse section of a very
young testicular ampulla of Scyllium. 0. male ovules of Hermann (sper-
matospores, Blomfield). 7. Prismatic cells with nuclei in course of
division. m, 2. Small cells resulting from division of the prismatic cells.
B. Portion of a similar transverse section at a later stage. o. Nuclei of
the male ovules now constituting multi-nucleate polyplasts and increased
in number by division of the prismatic cells. m. Terminal nuclei bounding
the lumen of the ampulla. c. Similar piece at a later stage. o. The
nuclei no longer arranged in rows. 2. ‘lhe basilar nucleus of Semper
(blastophoral cell of Blomfield). pb. Section showing two sperm-polyplasts
or mother-cells longitudinally cut at a stage immediately preceding the
conversion of the spermatoblasts (/) into spermatozoa. p. Axial proto-
plasm of the polyplast. c¢. Basilar portion of the polyplast. 2. Basilar
nucleus (¢c and ~ form the blastophoral cell). 2. Transverse section of
similar cells. m’, m. Intercalary processes of the protoplasm of the poly-
plast. to L. Conversion of a spermatoblast (final stage of division of
the nucleus and protoplasm of an original mother cell) into a spermatozoon.

n. Nucleus. zw. ‘ Corpuscle precurseur,” 0. Cephalic nodule, m. Middle-
piece. /. Flagellum.

Fic. 2.—Spermatogenesis of the rat. a. Seminiferous cell of Renson (sper-

matospore, Blomfield). 3. Multinuclear cyst of Renson (sperm-polyplast,
Blomfield), resulting from the growth and division of the nucleus of a.
c. Fragments of the “cellules de soutien” of Renson (blastophoral cell of
Blomfield) as seen in a teased preparation. D: First phase of the sup-
posed penetration (Renson) of the nematoblasts (young spermatozoids)
into the “ cellule de soutien ” (blastophoral cell), ©. Later phase of the
same. F. tod. Development of a spermatoblast (nematoblast, Renson)
into young spermatozooid, occurring, according to Renson, before its
connection with the “cellule de soutien” or blastophoral cell,

328 J. E. BLOMFIELD.

forming a degenerated tissue much in the same way as the
Graafian follicle forms the corpus luteum.

In a young ampulla formed in this way two kinds of cells
are found, the first are large, round, and granular, possessing a
large, well-defined nucleus. These are the male ovules; the
second kind are smaller and more irregular and fill up as it
were the interstices left between the male ovules. The whole
ampulla is continuous with the upper part of the cordon de
Pfliiger, which becomes its excretory duct. The second kind
of cell above mentioned is separable into two categories ; some
are prismatic in shape, situated on the basement membrane
and squeezed in between the male ovules and obviously
form male ovules, thus increasing the number of these bodies
and the size of the ampulla, while the others are placed more
on the internal surface of the male ovules and are often seen
in the act of division, in fact it is by their division that the
first few cells of the mulberry-like body soon to be described
are produced.

In a section of an ampulla (fig. 1, A), rather older than the
one we have just described, the male ovules, now more
properly called sperm-polyplasts, are arranged radially round
the periphery of the ampulla; and but few of the first
kind of small cells alluded to above are left, having been
converted into male ovules, while the second kind are seen
more distinctly conical in shape and placed at the end of the
polyplast next to the lumen of the ampulla. The polyplast
itself now (fig. 1, B) consists of four, five, or six, the number
varying according to age, nuclei arranged radially in a series
and terminated by one of these conical cells. When the activity
of this terminal conical cell is as it were exhausted in giving
rise to these several cells, the cells themselves commence to
multiply by division, and the outlines between each poly-
plast become indistinct from the mutual pressure exerted, but
each group is well marked by an elongated nucleus which now
comes into prominence at the base of each conical mass (fig. 1,
c,”) or polyplast, which is called by Semper the basilar nucleus,
and is regarded by him as the first formed nucleus, an obserya-

RECENT RESEARCHES ON SPERMATOGENESIS. 329

tion confirmed by M. Hermann. It is needless to point out
the identity of this basilar nucleus and its plasma with my
blastophoral cell of Helix and the Frog.

The ampulla is now filled with polyplasts (called mother cells
by M. Hermann) each consisting of 30 or 40 more small cells
and a basilar nucleus towards its peripheral end; and now
commences the process to which M. Hermann limits the term
Sperinatogenesis, viz. the conversion of the spermatoblasts into
spermatozoa ; but before describing this change we must lodge
a protest against the restriction of the term to this portion of
the change undergone by a cell in giving rise to spermatozoa.
Surely it should be applied to the whole of the changes which
are undergone when a germinal cell gives rise to its crop of
spermatozoa, as oogenesis is understood to refer to the pheno-
mena presented by an ovarian epithelial cell in becoming a
mature ovum.

The next change that the spermatoblasts undergo is one of
position (fig. 1, D), something similar to what occurs in the frog,
only, instead of forming a vesicle, they form a sort of tube
filled with granular protoplasm. When this is completed, the
change into spermatozoa commences.

The first event (fig. 1, F to L) is the appearance in the
spermatoblast of what appears to be a local consolidation of the
protoplasm composing the spermatoblast which M. Hermann
calls “corpuscle precurseur.’’ It has but a short existence,
and, as far as could be made out, no connection with any
part of the mature spermatozoon (wz). Before this has vanished
a small projection becomes visible at one point of the nu-
cleus (0), bearing, however, no relation to the position of the
“corpuscle precurseur.” ‘This is the (1) ‘nodule cepha-
lique,” which is destined to form the head of the sperma-
tozoon. It seems to be a thickening of the membrane of the
nucleus. While the cephalic nodule is being formed, (2) a
rod or bar (m) appears in the substance of the spermato-
blast, which is in connection, by its central end, with the
nucleus, and by its distal end with a fine leaf-like strand of
plasma, which is situated at the end of the spermatoblast,

Qs

330 J. E. BLOMFIFLD.

uext the hollow of the tube and extends into (3) the flagellum
or caudal filament (/).

The bar forms the middle piece (Mittel Stick, of Schweigger
Seidel), while the caudal filament helps to complete the tail of
the spermatozoon. The caudal filament would appear to be
formed out of the plasma which fills up the hollow of the tubu-
lar polyplast. The further changes consist in a growth of the
cephalic nodule over the rest of the nucleus in the form of a
cap, and the elongation of the middle piece, which becomes
freed from the cell body while the caudal filament remains
attached to its distal extremity.

The ampulla now consists of a mass of tube-like polyplasts,
each with its basilar nucleus, and each built up by a bundle
of spermatozoa, whose tails project towards the lumen of the
ampulla. As the next step, these are discharged, and now
nothing remains except the blastophoral cell. Each blasto-
phoral cell consists of plasma holding a basilar nucleus towards
its base, and a body, which arose during the later development
of the spermatozoa, placed about the middle of cell, and called
by Semper ‘‘ the problematic body.”

Nothing is known about this body. It does not appear to
consist of fat, but possibly it is connected with the atrophy of
the blastophoral cells and ampulla which now takes place. (See
Semper’s ‘ Arbeiten,’ 1875.)

We now pass to the description of Spermatogenesis in
Mammals, contained in a paper by Dr. George Renson in the
‘ Archiv de Biologie,’ tome iui, fascicule i1, 1882, according to
which in its main outlines the process is strikingly similar to
what we have before described in other classes.

The paper in question commences with a resumé of the
work of previous observers, especially V. Ebner, Sertoli, and
V.la Valette St. George. A few pages are devoted to descrip-
tions of methods which are those in ordinary use, and then the
author commences the process as seen in the Rat, an animal
which all observers acknowledge to be the best for obser-
vation.

He first describes the elements which are seen in a prepara-

RECENT RESEARCHES ON SPERMATOGENESIS. 301

tion of the teased testis treated with osmic acid, (1) large,
rounded cells, with one, two, or three nuclei (fig. 2, a). These
are the seminiferous cells of Sertoli. (2) Multinucleated
masses (fig. 2, B), to which he applies the word Kysts. The pro-
toplasm is collected round each nucleus, so that they are rather
to be regarded as collections of separate cells. (3) Nemato-
cysts (fig. 2, F to J), or cells in which the nucleus has undergone
changes to form the head of the spermatozoon. Each is pro-
vided with a filament. (4) Large oval nuclei (fig. 2, c), sur-
rounded by masses of hyaline protoplasm, which are really the
basal portion of the “cellules de soutien,” to be mentioned
immediately.

By means of 4} % of alcohol of Ranvier, he found another
element, which he regards as extremely worthy of attention.
These are large cylindrical cells with expanded bases, contain-
ing in the expanded end a large smooth nucleus, and support-
ing at the other the nematocysts or immature spermatozoa.
He identifies these cells as the cellule de soutien of Meckel,
and the cellule fixe of Sertoli, as well as part of the spermato-
blast of V. Ebner, who applied this latter name to the cell
under consideration, taken together with the nematocysts sup-
ported on its central end (fig. 2, p and B£).

As regards the origin of these cellules de soutien, and their
connection with the nematocysts, our author thinks that they
arise independently from cells placed next the lining membrane
of the seminal tubule, and that they come into connection with
the nematocysts as they grow towards the centre of the tubules
supporting them, and by their growth pushing them as mature
spermatozoa into the lumen. He gives as a reason for the dis-
tinctness of the two that very young nematocysts are not found
in connection with the cellules de soutien, but, viewing the
subject in the light which we have gained from the study of
Spermatogenesis in other classes, 1 think that any one will
allow that it is extremely probable that there is a closer con-
nection between them than our author has allowed, and that
the ‘‘cellules de soutien” are homologous and analogous with
the body which has been described in the foregoing papers as

3302 J. E. BLOMFIELD.

mother cell with basilar nucleus, or mother cell with principal
nucleus or sperm-blastophor of my nomenclature.

The author then describes a series of sections showing the
relations which the above elements bear to one another.

In a typical section next to the membrana propria are found
two kinds of cells, the germinative cells, whose origin is doubt-
ful, and irregular cells, containing a smooth large nucleus,
which is the basal portion of the cellule de soutien. ‘The
germinative cells (my spermatospores) multiply by division to
form the multinucleated masses, which he calls Kysts (my sperm-
polyplasts) ; the nuclei, with a small portion of protoplasm, then
become free, constituting the seminiferous cells (my spermato-
blasts), and then the nucleus of each seminiferous cell begins
to undergo changes to form the spermatozoon, when the term
nematocyst is applied to it. When this state is reached, they
become attached to and supported on the “ cellule de soutien,”
where they remain till they become mature spermatozoa. After
throwing off the spermatozoa, the “ cellule de soutien” appears
to undergo fatty degeneration in its central part, while its peri-
pheral part remains and forms the “ stellate cells,” interspersed
among the germinative cells, which were described by Sertoli
and Meckel.

While these changes have been going on, the germinative
cells have given rise to seminiferous cells, which, embedded
between the radial projections of the cellules de soutien, are
ready for the next crop of spermatozoa, and thus the process is
coutinued.

The origin of the “ cellules de soutien” is not given; but if
we assume that the germinative cell and the cellule de soutien
both arise from the division of a primordial testicular ceil (a
spermatospore), we can bring these observations into agreement
with the generalised process as described by me.

It remains to describe the exact changes which a nematocyst
undergoes in becoming a spermatozoon. First, a caudal fila-
ment is protruded, and at the pole of the nucleus opposite to
this point a thickening of the membrane takes place, near the
nucleus in the accessory corpuscle. The nuclear thickening

RECENT RESEARCHES ON SPERMATOGENESIS. 333

spreads till it occupies a hemisphere of the nucleus; then a
swelling is seen at the pole of the hemisphere, which form the
‘baton terminal” (Spitzenknopf of Meckel), a formation quite
independent of the accessory corpuscle which some observers
have believed to have formed this structure. ‘The hemisphere
continues to thicken till it forms a hood (capuchin, Kopfkappe),
while the caudal filament becomes united to the nucleus by a
granule at the opposite pole. The protoplasm of the nemato-
cyst forms a hyaline tube, in which can be seen the caudal fila-
ment. ‘The nucleus becomes flattened, the nuclear thickening
with the baton terminal disappears, but its lower limit is
marked by a circular line, and the mature spermatozoon is
formed.

The formation of the spermatozoon, as here briefly described,
has many points of similarity with that above mentioned in
Selachians. In both there is a body formed (corpuscle pre-
curseur, Selachians, accessory corpuscle, Mammals), which takes
no share in the process. In both a caudal filament is protruded,
in both a thickening of the nucleus forms the head, and in both
there is a consolidation of the plasma, which may be regarded
as forming the middle piece; and here let me make a few
remarks on this portion of the spermatogenetic process.

It is now well determined that the nucleus forms the head
of the spermatozoon, and the plasma of the spermatoblast gives
rise to the tail, but does the whole of each take part in the
process? It seemed to me in the cases which I selected for
study that the head was formed by simple elongation or
change of shape of the nucleus, while drawing out of the
plasma gave origin to the tail; but it is more than probable
that in many cases the process is more complicated than this,
and possible that it differs in different animals. The similarity
of the process in Selachians and Mammals would seem to
suggest that it is conducted on one plan, which plan would
seem to be that a portion of the nucleus—possibly the mem-
brane and the chromatin—form the head, while a caudal fila-
ment forms the extreme portion of the tail, the intermediate
part having origin in solidification or modification of the sub-

334 J. E. BLOMFIELD.

stance of the spermatoblast. Further researches on this point
are needed, especially with reference to the constitution of a
nucleus disclosed by the researches of Flemming.

These remarks would be incomplete without reference to the
observations of M. Sabatier, who in ‘ Comptes rendus,’ 94,
1882, pp. 172-3, has recorded his observations on Spermato-
genesis in Salmacina (one of the Serpulide). In this worm he
finds that a spermatospore or mother cell becomes covered by
multiplication and budding of nuclei with a mass of clavate
cells (protospermatoblasts) which do not themselves give rise
to spermatozoa, but become detached, and then, by nuclear
multiplication and budding, produce a crop of spermatoblasts
(deutospermatoblasts), from which the spermatozoa originate.
He uses these observations to explain the process of the forma-
tion of testicular ampulle in the lower Vertebrates, and to com-.
plete its comparison with the formation of a Graaffian follicle,
as suggested by Balbiani.

According to this latter observer an ampulla is formed by
a central cell surrounded by smaller ones. The central cell
represents the female part, and takes no share in the formation
of spermatozoa, while the smaller ones represent the male part
ef the primordial indifferent cells. This state of things is
reversed in the Graaffian follicle of the ovary, in which the
central cell (the female portion) undergoes development at the
expense of the surrounding epithelial cells of the follicle
representing the male element. This central cell of the am-
pulla could, on M. Sabatier’s view, be represented by the sper-
matospore, the cells lining the ampulla would be the proto-
spermatoblasts, and the cells which we have called spermato-
blasts would be deutospermatoblasts.

While this paper was being written a notice appeared in
the ‘ Zoologische Anzeiger’ for the 19th Feb., 1883, referring to
researches made by Max v. Brunn on the double form of sper-
matozoon found in Paludina, to which, in conclusion, I must
make a brief reference. He finds that the formation of the
head does not take place endogenously and without the nucleus,
as stated by M. Duval, but that it is produced as described by

RECENT RESEARCHES ON SPERMA'TOGENESIS. 335

myself in Helix by division of the nucleus. He accounts for
the discrepancy in the observations by supposing that the former
observer obtained deceptive appearances bya too prolonged action
of osmic acid. The object with which his observations were
made was to assign some explanation to the striking fact of the
presence of two kinds of spermatozoa—the hair-like form and the
worm: like form—in Paludina. He found that only the small
hair-like form was concerned in the fertilisation of the ovum, in
which process the other larger form took no part, and, as far as
he was able to determine, this latter had no function at all.
Having failed to find any physiological reason for the presence
of this form, he was obliged to fall back on morphological
explanations, and the hypothesis he offers is that this large
form represents a disturbed or arrested development of a sper-
matozoon ; that the cell from which it arises is, as it were, a
female cell, and that the testis, as seen in Paludina, is a transi-
tion form to the hermaphrodite glands of Pulmonates and
Opisthobranchs, in which the distinction between male and
female cells is at an early stage impossible.

336 G. F. DOWDESWELL.

Note on a Minute Point in the Structure of the
Spermatozoon of the Newt.

By
G. F. Dowdeswell, MW.A., Kc.

THE general structure of the spermatozoon of the water
Newt (Triton cristalus) has recently been well and accur-
ately described by Dr. H. Gibbes in this Journal.! It is toa
point therein, which has hitherto escaped notice, that I wish
here to call attention.

The spermatozoon consists, as described (loc. cit.) of the
“‘ body,” to which is attached a fine, narrow, translucent mem-
brane, bordered by what is usually termed “the filament,”
which takes its origin from what may be called the neck, the
upper or thickest extremity of the body. This membrane and
“ filament” evidently consist of protoplasm, being highly con-
tractile ; in the fresh state rhythmical waves of contraction may
be seen passing up them, and producing that remarkable ap-
pearance of spiral rotation, which in similar cases was often a
source of perplexity to microscopists. The “ body” also
appears to be protoplasmic, both behaving in the same manner
towards reagents, but the upper and thickest part, ‘‘ the neck”
—the “ elliptical body ” mentioned by Gibbes—appears to be of
somewhat different constitution, as in some cases it stains
much more deeply and readily than the rest of the body. Sur-
mounting this, forming a cap as it were, is a long, finely taper-
ing conical head, which, as already shown (loc. cit.), is of
materially different constitution to the other parts, being
apparently less stable, swelling up readily when treated with
water, and being easily altered and destroyed by other re-

' Vol. xix (1879), p. 487.

MINUTE STRUCTURE OF THE SPERMATOZOON OF NEWT. 3837

agents. It stains more readily than the body and membrane,
but not so deeply as what I have termed the neck. Towards
the extreme end, from tapering very regularly, the head
becomes somewhat abruptly more constricted for the last few

fy Aes —== SSS

——____

“1 Ties, I.

micromillimeters of its length, and is here, in unstained pre-
parations, more highly refracting than the rest, its substance
appears more dense; probably this end portion is solid and
the remainder hollow (of which preparations stained with car-
mine present very much the appearance), and shows a double —
contour. At the extreme point of this head there is a minute
barb (see woodcut, fig. 1). In successful preparations it may be
very distinctly seen and readily measured, and this even when
unstained. I have already! referred to its existence, and on
further examination of better preparations do not find it so ultra
minute as I at first thought it. It is indeed of very appreciable
magnitude, being in breadth about 1°5 mic. m. (0:0015 mm.), and
in length 2:0 mic, m. (0°002 mm.), though obviously accurate
measurements of such objects are difficult, for to detect the
actual termination of an impalpably fine point is not always
possible.”

Such a determinate and remarkable structure.as that here

1 See this Journal ante, vol. Ixxxii, N. S., 1882.

2 In such measurements I have found great advantage in the use of a cob-
web micrometer, admirably constructed by Messrs. Ross, which has the second
web, which is usually fixed, movable; this both saves time and promotes
accuracy, as in the usual form (with only one web movable) it is almost
impossible, by means of the mechanical stage, to bring an object into exact
contact with the fixed web, which is done at once with ease and certainty by the
second movable one. Having now used this a good deal, I certainly prefer
it to any other plan; whatever arrangement is adopted, however, it is neces-
sary to determine the value of the seale of the eye-piece with a stage micro-
meter, as the least variation in the conditions of the instrument, as e.g.
slightly turning the screw collar of the objective, appreciably alters their
relations,

308 G. F. DOWDESWELL.

described cannot be supposed to exist without some purpose
recent researches at once suggest that this is to attach the
spermatozoon to, and enable it to penetrate into the ovum in
the early stages of fertilisation, as has been shown to occur by
Fol and others; and we should expect to find a similiar forma-
tion in other spermatozoa. In those, however, which I have
hitherto examined I have not detected it, and the structure of
many, as that of the toad, and of most mammalia, does not
appear to admit of its existence.

To prepare the spermatozoa for the examination of this
object the first essential is to get them as nearly as possible in
contact with the cover-glass and flat upon it; this requires
some care, to avoid their drying, by which they are materially
altered. They may be preserved by several methods, either
by treating for twelve to twenty-four hours with a concen-
trated solution of picric acid, a dilute solution of chromic
acid, by Dr. Klein’s method with a 9 per-cent. solution of
ammonium chromate, by iodine, by silver nitrate, or by osmic
acid or gold chloride; the latter are convenient as being
quicker. I have myself most usually employed picric acid.
For staining I have found glycerine magenta’ the best method,
as it stains all parts as strongly as desired. ‘To show the
general structure alcoholic carminate of aminonia is the most
satisfactory, but it does not stain the barb deeply. Other
aniline dyes I have not found answer so well. If it be in-
tended to examine the preparation with a homogeneous, or
“ oil”-immersion objective, it should be mounted in Canada
balsam, the objective having, as is generally known, no advan-
tage, and, indeed, being inferior to dry glasses for objects be-
tween which and the cover-glass there is either a film of air,
or of any fluid the refractive index of which is much different
from that of glass.

The use of glycerine as a mounting fluid for preparations
stained with any of the aniline dyes is at best trouble-

1 & Magenta cryst., 1 part; glycerine, 200 parts; alcohol, 150 parts; aq.,
150 parts; immerse the preparation in the solution for from two to four
minutes according to the depth of colouring required, and then wash,

MINU'TE STRUCTURE OF THE SPERMATOZOON OF NEWT. 339

some,' and sooner or later, to my experience, the staining runs,
and the preparation is spoiled. Solutions of acetate of potash or
chloride of calcium I have not found satisfactory; the form,
even of such resistant objects as bacteria, in some cases becoming
materially altered by these reagents. With Canada balsam,
even when dissolved in chloroform or turpentine, I have not
found the preparations fade, as has sometimes been said to be
the case, and as we should have expected ; nor, if they are suffi-
ciently washed in alcohol and passed through oil of cloves,
will they run; the risk, however, both of fading and running
may be entirely obviated by using benzine as a solvent for the
balsam, or by employing it undiluted and liquified by
warmth.

In examining this structure I have employed the ,'-th
homogeneous immersion of Messrs. Powell and Lealand, which
having the very high numerical aperture of 1:38° gives, with
admirable light and definition, an amplification of about 3400
diameters, with an eye-piece of 3 m. focal length; the barb,
however, in a suitable preparation, may be readily seen and
examined with a good }th objective. Even with much lower
powers, as e.g. the 74th P.and L. I have recognised it,
dependent however much upon the method of illumination
employed: for, as is generally recognised, good illumination
will show an object with a much lower power than is requisite
in the ordinary way. The best means of this as yet available
is the direct light of the flame of a paraffin lamp turned edge-
wise to the observer, whether with or without a substage con-
denser. This was recommended by Dr. L. Beale thirty years
ago, and is now again frequently adopted. Light reflected
from any mirror is in some way inferior to direct light, and
this not owing to the double reflecting surface of ordinary
mirrors, for I have tried them silvered on the upper surface
without any material advantage.

' The method is, add an equal bulk of glycerine to the aqueous solution of
the aniline dye used, stain somewhat more deeply than requisite, mount on
slide with cover glass in the staining fluid, which is to be gradually replaced
as the water evaporates by plain glycerine,

340 PROFESSOR RAY LANKESTER AND A. G. BOURNE.

On the Existence of Spengel’s Olfactory Organ
and of Paired Genital Ducts in the Pearly
Nautilus.

By

E. Ray Lankester, M.A., F.R.S.;
Jodrell Professor of Zoology,

and

A. G. Bourne, B.Sc.

A MALE and a female specimen of Nautilus pompilius
were purchased a few years since by one of us for the Zoolo-
gical Museum of University College, London. Leisure and
opportunity for the study of these specimens have recently
been afforded, and we propose to briefly report here on two
interesting additions to knowledge which our observations have
yielded.

The specimens were in an excellent state of preservation for
the purpose of dissection, though not fit for histological study.
The male was not purchased as such, and its sex was not re-
cognised until it was withdrawn from its shell and the circum-
oral tentacular apparatus examined.

With the exception of the specimens reported on by Van der
Hoeven no adult males of Nautilus have been carefully ex-
amined, that referred to by Keferstein (in Bronn’s ‘ Classen
und Ordnungen des Thierreichs,’ Weichthiere) being an imma-
ture specimen. A description and figure of the circum-oral
tentacular apparatus of the male Nautilus and a comparison of
this with the corresponding region in the female will form the
subject of « memoir by Mr. Bourne in another publication.
Here we shall confine ourselves to two points.

PEARLY NAUTILUS. 341

Spengel’s Olfactory Organ.

The extremely important observations of Spengel on the
olfactory organ of Mollusca (‘ Zeitsch. wiss. Zoologie,’ vol.
35) lead to the conclusion that there is very generally if not
universally present in the Mollusca an olfactory organ placed
near to or in relation with each gill, and that this organ
receives its nerve from the ‘visceral loop’? or commissure,
which, sometimes short, sometimes long, in some molluscs
twisted, in others straight, joins to one another the pair of
so-called “ visceral” ganglia.

We adopt the name “ osphradium” for this molluscan organ
of smell, proposed by Professor Lankester in his article ‘ Mol-
lusca,” in the ‘ Encyclopedia Britannica.” The osphradium is
thus distinguished by its name from all other organs presumed
to have the olfactory function, whether placed on the head;
lips, tentacles, or elsewhere. The osphradium of molluscs is
Spengel’s olfactory organ, it lies near the gill and tests the
respiratory medium.

Spengel was unable to find an osphradium in Cephalopoda.
He appears not to have had the opportunity of examining
Nautilus, where a well-developed pair of osphradia exist,
although we have not been able to detect their representatives
in Sepia or Octopus.

The osphradia of Nautilus are in the form of a pair of teat-
like papillz placed upon the body-wall of the subpallial chamber,
a little to the outer side of the muscular attachments of the an-
terior pair of gills, one corresponding to the right gill and the
other to the left gill (woodcut, figs. 1, 2, o/f.).

These papille have been seen and figured by previous
writers (Van der Hoeven, Keferstein), but no suggestion as to
their significance has ever been made. We were led to infer
from their position that they represented Spengel’s olfactory
organ, and proceeded to test that hypothesis by an examination,
firstly, of their microscopic structure, and secondly, of their
nerve-supply.

The inquiry into their microscopic structure was entirely
negative. Our specimens were not sufficiently well preserved

VOL, XXIII.—NEW SER. Z

342 PROFESSOR RAY LANKESTER AND A, G. BOURNE.

to enable us to say whether the epithelium of these papillz is
specially modified or not.

Vise per.

inet

Fic. 1.—View of the postero-ventral surface of the body of a female
Nautilus pompilius, as seen when the mantle-skirt is reflected.
Drawn from the object by Mr. A. G. Bourne, and reduced to one half
the natural diameter. a. Muscular band from the foot (siphon) to the

_ body-wall. 4. Valvular ridge of the siphon. ¢. The reflected border of
the mantle-skirt. az. Anus. 2. Peculiar median group of post-anal
papillae of unknown significance. g. 2. Nidamental gland. *. ov. Right
oviduct’s aperture. /. ov. Left oviduct’s aperture. zeph. a. Aperture of
the left anterior nephridial sac (in front of the left anterior gill-plume).
neph. p. Aperture of the left posterior nephridial sac (in front of the left
posterior gill-plume). vise. per. Left aperture of the viscero-pericardial
sac. olf. The left osphradium (Spengel’s olfactory organ).

With regard to the second test we obtained very satisfactory
evidence. The basi-branchial papille of Nautilus possess pre-
cisely that nerve-supply which is characteristic of the molluscan

PEARLY NAUTILUS. 343

osphradium, viz. they are innervated by nerves arising from the
visceral commissure.

AY

ge an

_ &n. x VISCPEL

Fie. 2.—View of the postero-ventral surface of the body of a male Nau-
tilus pompilius, as seen when the mantle-skirt is reflected. The gills
and the foot are cut short. Drawn from the object by Mr. A. G. Bourne,
and reduced to one half the natural diameter. a. Muscular band from
the foot (siphon) to the body-wall. c. The reflected border of the mantle.
skirt. az. Anus. . Peculiar median group of post-anal papille of
unknown significance. pe. Penis-like opening of the right sperm-duct.
1. sp. Aperture of the left sperm-duct. xeph. a., neph. p., vise. per., and
o/f., as in Fig. 1.

The nervous system of Nautilus is represented in fig. 3. In
the Cephalopoda, as in some other forms (most Pteropoda, some
Gastropoda), the visceral and pleural ganglia are not separated
from one another, but form a continuous nervous band (pi. vise.).
Nerves to the mantle (m.) proceed from the pleural portion of
this transverse band, whilst a large visceral nerve (x. vise.)
proceeds from each of the contiguous visceral portions of the
same band, which represent the visceral ganglia. These large
visceral nerves give off each a genital ganglion (gen.) in the
neighbourhood of the genital ducts, and then, taking a super-
ficial course, divide each into two branches, one supplying the
anterior the other the posterior branchia of its side,

344 PROFESSOR RAY LANKESTER AND A. G. BOURNE.

From the portion of the nerve between the anterior and pos-
terior branchial nerve is given off the nerve to the olfactory

[f-p4 \ } he Ne bry
i
xe Ly SF * Ye ye olf

Jn. tf br.

“a

Fic. 3.—Diagram of the nervous system of Nautilus pompilius. Drawn
from the object by Mr. A. G. Bourne. cer. Cerebral ganglion. ped.
Pedal ganglion. opt. Optic ganglion (resting on the cerebral). pi.
Pleural ganglion. vise. Visceral ganglion. and y. Ganglion-like en-
largements on nerves passing from the pedal ganglion to the infero-
median lobe of the inner circlet of circum-oral tentacular lobes. m.
Nerves from the pleural ganglion to the mantle. 2. vise. The genito-
branchial nerve, or chief visceral nerve (of the left side). x. v. Nerve
accompanying the vena cava which lies between this and the similar nerve
of the right side. ge. The left genital ganglion. x. sup. br. Anterior
branchial nerve. 2. izf. br. Posterior branchial nerve. x. o/f. Olfactory
nerve entering the left osphradium. ov. The oviduct (right side).
olf. p. p. The right osphradium.

papilla or osphradium. ‘This nerve supply is closely paralleled
in such Gastropoda as Haliotis.

An examination of Octopus showed that in that Cephalopod a
similar distribution of visceral nerves obtains, but the branch
on each side corresponding to that which supplies the olfactory
papilla in Nautilus simply ramifies beneath the skin, there

PEARLY NAUTILUS. 345

being no papilla or prominence of any kind in Octopus cor-
responding to the osphradium of Nautilus.

The paired Oviducts and Sperm-ducts.

By all previous writers (Owen, Valenciennes, Van der Hoe-
ven, Huxley, Keferstein, Woodward, Macdonald) the genital
ducts of Nautilus have been described as unpaired. A single
oviduct is said to be present in the female, and a single
sperm-duct ending in a penis-like process in the male.
Jhering, in an article in the ‘ Zeitschrift fiir wiss. Zoologie,’ vol.
XX1x, p. 989, in which an attempt is made to discuss the homo-
logies of the excretory organs and genital ducts of Mollusca,
says: “We are prevented from assuming that the relations (of
these two sets of organs) is the same in the Cephalopoda as in
the Lamellibranchia, were we otherwise inclined so to do, by
the fact that in Nautilus only a single. oviduct exists. Until
we have the developmental history of Nautilus before us, there
is no prospect of any progress in this direction.”

Without endorsing Jhering’s statements in any way upon
other points, we may point out that, without any reference to
the developmental history of Nautilus, the supposed fact that
in Nautilus only a single oviduct exists is demonstrated to be
no fact at all.

In the female Nautilus (fig. 1) we find on the postero-
ventral wall of the body overhung by the mantle-skirt no less
than nine apertures. The central aperture is the anus (qam.).
The two pairs right and Jeft of this, just in front of the hinder
gill-plumes, are the two openings of the great viscero-pericardial
chamber (vise. per.) and of the two posterior nephridial sacs
(neph. p.) respectively. Right and left, in front of the an-
terior gill-plumes, we find the pair of apertures appropriate to
the anterior nephridia (xeph. a.). Nearer the middle line is
placed on the right side the great oviducal aperture, with
plaited lips (7. ov.), and at the corresponding point left of the
middle line is placed the aperture, which we have marked
1. ov,

346 PROFESSOR RAY LANKESTER AND A. G. BOURNE.

It is curious that this aperture has been overlooked by every
student of the Nautilus excepting Keferstein. At the same time
Keferstein failed to apprehend its true significance.

Keferstein showed that the aperture (/. ov.) leads by a duct
into the “ pyriform appendage,” originally described by Owen
as lying in close connection with the ventricle of the heart.
The nature of this pyriform appendage neither Owen nor any
subsequent observer was able to divine. Keferstein first
showed that it communicates with the exterior by means of the
aperture (/. ov.), discovered by him.

But this aperture is the exact left-hand representative of the
large oviducal aperture (7. ov.). ‘This suggested to us the inquiry
as to whether the relations of the “ pyriform appendage ”’
are such as to favour the supposition that it is a rudimentary
left oviduct. We find that they are; and we conclude that
in Nautilus the left oviduct is reduced to a rudimentary condi-
tion, becomes constricted at the point where it joins the ovary,
and so ends blindly as “ the pyriform appendage,” whilst still
opening to the exterior by the left genital pore (. ov.). The re-
lation of the right and left oviducts to the ovary, to the ven-
tricle of the heart, and to one another, is shown in the diagram
(fig. 4.) Whether the ovary is a medium structure, or whether,
on the other hand, the pyriform appendage represents an
aborted ovary, as well as a rudimentary blindly ending oviduct,
is a matter for further inquiry.

Our conclusion as to the nature of the aperture (/. ov.), and
the pyriform appendage of the female, was greatly strengthened
by our discovery of a precisely similar disposition of parts in the
male.

The same number of apertures is present in the male
Nautilus (fig. 2) as in the female. Instead of the right ovi-
duct with plaited mouth, we have a right sperm-duct produced
into a large penis-like structure (pe.). The aperture marked
I. sp. in fig. 2 has not hitherto been noticed, nor has the pyri-
form appendage been observed hitherto in a male Nautilus.
We find that, just as in the female, the left aperture (/. sp.)
leads into a “‘ pyriform appendage,” which ends blindly. As

PEARLY NAUTILUS. 34.7

Fic. 4.—Diagrams of the male and female generative organs of the Pearly
Nautilus, to show the relation of the rudimentary duct of the left side to
the testis and ovary respectively, and of the cardiac ventricle to the
organs of both sides. Drawn by Mr. A. G. Bourne. 7. Testis.
O. Ovary. Ac. Accessory gland of the male apparatus. 4//. Albumini-
parous gland of the female apparatus. WV. Needham’s sac in the male,
in which the spermatophores are formed. P. Penis (in the male),
R.G.O. Right genital orifice. Z. G.O. Left genital orifice, Pyr. Owen’s
pyriform appendage, attached by membrane to /., the cardiac ventricle
with its four branchial veins, and also to the testis or the ovary. Fo.
Foramen in the membrane, which attaches the pyriform appendage to
the ventricle and to the testis or ovary. The foramen places two portions
of the viscero-pericardial sac in free communication with one another,

348 PROFESSOR RAY LANKESTER AND A. G. BOURNE.

shown in the diagram (fig. 4), the relations of these parts in the
male is (as in the female) such as to leave little doubt that we
have in them a rudimentary condition of the left spermatic
duct and its external opening. Whether the pyriform append-
age in any way represents also a rudimentary testis, must
remain for the present uncertain.

The significance of the occurrence of paired genital ducts in
Nautilus is not so great as it would be were it not the fact that
both in the Octopoda and in Ommastrephes among Decapoda,
the female genital ducts are paired, and both equally well de-
veloped, although the sperm-duct of the males is (with the
solitary exception of Eledone moschata, recorded by Kefer-
stein) single.

The single oviduct of those Cephalopoda which have but
one, and the single sperm-duct also, is, in all cases, that of the
left side; whilst in Nautilus the right oviduct and the right
spermduct is large and functional, the left being that which is
rudimentary.

Nevertheless, on account of its archaic character, and of the
great significance of the primitive Cephalopod structure in re-
lation to the morphology of Mollusca generally, any divergence
in Nautilus from the condition obtaining in other forms has
possibly and even probably a special significance. Such
divergence may be the remnant of an ancestral condition, and
in so archaic a form as Nautilus, is not readily to be dis-
missed as “an adaptation”’ peculiar to that form.

Hence the doubt which might have arisen through Nautilus
as to the primitive arrangement of the molluscan genital ducts
is removed. ‘They are now shown to be paired ducts, as in
Chiton, and as in Lamellibranchs.

On the Ancestral Form of the Chordata.
By

A. A. W. Hubrecht,
Professor of Zoology at the University of Utrecht.

CS eee

With Plate XXIII.

Aw all important question was raised in biology when the
Law of Development came to be recognised as the only true
explanation of the facts as they lie before us. This was the
question: ‘‘ From what invertebrate stock are the Vertebrata
evolved, and which amongst the Invertebrates at present living
approaches most closely in its organisation to this primitive
parent stock.” In 1868 the solution appeared to have been
found when Kowalevsky’s splendid researches concerning the
development both of Amphioxus and of the Ascidians could be
compared side by side. The Tunicate larva was for the time
being proclaimed to be the missing link, to be of all Inverte-
brates the closest approach to the much-looked-for parent form.

Since then the aspect of things has changed and later inves-
tigations, more especially those of Dohrn and of Ray Lankester,
have rendered it nearly certain that the Tunicata must, on
the contrary, be looked upon as degenerate Vertebrates which
can be hardly of much use in helping us to the failing clue.

Dohrn, Semper, Hatschek, Leydig, Kleinenberg, and Eisig
are amongst the foremost who have suggested, and most
brilliantly expounded and argued, that the Annelids offer the
greatest number of points of resemblance with the Vertebrates ;
that they and the Arthropods have descended together with
the Vertebrata from a primitive type, distantly agreeing in

VOL, XXIII,—NEW SER, AA

300 A. A. W. HUBRECAT

shape with Polygordius, and that the only postulate which
this assumption necessarily implies, is the old idea of
Geoffroy St. Lhlaire that the ventral side of Annelids and Arthro-
pods is homologous with the dorsal side of the Vertebrata.
These naturalists explain the difference in situation of the
mouth and cesophagus, with respect to the cerebral ganglion,
by divers subtle hypotheses, which, however, generally dis-
agree with each other. Their views are nevertheless rapidly
gaining ground, although the school of Gegenbaur and
Haeckel has never been reconciled to them. Gegenbaur looks
upon two lateral cords such as are present in Nemertines
as a very primitive arrangement, from whence might at
any rate be derived the ventral nerve-cord of Annelids and
Arthropods. Harting (‘ Leerboek der Dierkunde,’ 1874) was
inclined to accept the possibility of a similar dorsal coales-
cence out of which a spinal cord might take origin. Balfour
and myself felt strongly inclined to choose this side in the
contest—he in once more tracing the outlines of a similar
explanation (‘ Development of Elasmobranch Fishes’), I
in recapitulating the facts, such as they had made them-
selves known to me in the organisation of certain Nemer-
tines, in which, indeed, a tendency towards approxima-
tion of the lateral cords on the dorsal side was unmis-
takable (‘ Verh. der Kon. Akad. van Wetenschappen,’ Am-
sterdam, 1880). When Balfour, in the second volume of his
‘Comparative Embryology,’ made himself a definite advocate
of this view in opposition to the combatants for the Annelidan
affinities, it may be safely inferred that many of the younger
naturalists paused to reconsider the claims of both hypotheses.
The great difficulty which is encountered in any attempt to
point out a definite group amongst invertebrates most closely
related to the primitive Vertebrata is the total absence of
anything resembling so important and so early-formed an
organ as the Vertebrate Chorda dorsalis. Attempts to find
anything like it amongst the Annelids, even amongst Polygor-
dius and its archaic allies have proved either futile or barren.
I will at present attempt to point out in what group of

ON THE ANCESTRAL FORM OF THE CHORDATA. 35L

invertebrate animals we do indeed find an organ which, in my
estimation, ranks equal to the vertebrate notochord, and thus
supplies the much-desired transitional form by which the Chor-
data are allied to the lower Metazoa, and in fact to such forms
as have neither the much specialised organisation of the seg-
mented animals (Arthropods and Annelids), nor require to be
turned upside down before their homology with the lowest
Vertebrata is admissible.

That I venture to state this hypothesis before I am able
to bring into the field a well-arranged host of facts in its sup-
port, must be ascribed to my desire to induce such fellow-
workers in biology as have more leisure and better occasion for
investigating the numerous problems it suggests than I have,
into taking up a question which cannot be but looked upon as
of the highest significance for modern morphology.

According to my opinion the proboscis of the Ne-
merteans, which arises as an invaginable structure
(entirely derived, both phylo- and ontogenetically,
from the epiblast), and which passes througha part
of the cerebral ganglion, is homologous with the rudi-
mentary organ which is found in the whole series of
Vertebrates without exception—the hypophysis cere-
bri. The proboscidian sheath of the Nemerteans is
comparable in situation (and development?) with the
chorda dorsalis of Vertebrates.

After this brief statement of my hypothesis I will enter into
a short discussion of its different details.

It is not my intention to consider the numerous modifications
of structure which the hypophysis cerebri presents in different
adult Vertebrates, nor its glandular appearance, the con-
nection into which it enters with blood-vessels, &c.; but I
wish to restrict myself to the comparison of its very first onto-
genetic stages, in which it may be presumed most purely to
reproduce its ancestral character.

We find the hypophysis originating as an invagination from
the epiblast, arising either independently on the outer surface

852 A. A. W. HUBRECHT

(this, according to Dohrn’s .interesting researches, is the case
in one of the lowest of the vertebrate scale—Petromyzon),
this invagination at the same time being directed
towards the anterior termination of the notochord,
and lying in its direct continuation (figs. 1 and 2), or
(as is the case in the higher Vertebrates) not directly on the
outer surface, but on that portion of the epiblast which has
become the stomodeum (fig. 6). In the latter case it arises
as a median dorsal outgrowth from the mouth-cavity, directed
towards that portion of the under surface of the brain where,
between Prosencephalon and Metencephalon, the infundi-
bulum travels downwards, this being at the same
time the limit up to which the notochord extends
forwards under the brain. The fact that an outgrowth
from the brain thus grows downwards to meet this epi-
blastic invagination sufficiently indicates that in ancestral
generations, where the hypophysis was a less rudimentary
organ, some sort of connection existed between it and the
cerebral thickening of the central nervous system.

The constant presence in all Vertebrates of an organ so rudi-
mentary as the hypophysis, and about the significance of which
no plausible explanation has as yet been offered, has already
been insisted upon above.

Both facts are in favour of regarding it as a very ancient
structure, which was once of great importance, and had a
different and at the same time a more definite physiological
value.

In tracing this ancestral significance, the relation to the brain
and the somewhat less direct but, nevertheless, unmistakable
relation to the notochord must not be lost sight of.

We will now consider the ontogenetic and phylogenetic his-
tory of the Nemertine proboscis. In the lower Platyelminths the
researches of v. Graff, lately crowned by his brilliant mono-
graph, have brought to light the different stages through
which retractility of a portion of the tactile anterior extremity
of the body, in which urticating elements are present, leads
to the appearance of a definite proboscidian structure, which

ON THE ANCESTRAL FORM OF THE CHORDATA. 3503

obtains a special musculature, and which finally (in Graff’s
Rhabdocoel family of the Proboscida) has definitely
become a proboscis, which is directly comparable to that of
Nemertines, situated like it above the intestine, internally (ex-
ternally when everted) clothed by the direct continuation of
the outer layer of epiblast, and serving tactile and at the same
time—through its nematocysts—aggressive purposes.

The proboscis of Nemertines is thus directly related to this
important structure of the lower flat-worms, as was already
noticed in Gegenbaur’s ‘Grundziige’ (1870). We find urti-
cating elements largely developed in the proboscidian coating
of Paleo- and Schizonemertini, whereas in the Hoplonemer-
tini the tactile significance may perhaps have come to predomi-
nate, if we judge from the extremely complicate arrangement
and massive development of nervous tissue in the proboscis of
these forms, which, moreover, is here provided with a central
stylet-shaped armature.

As to the ontogenetic development of the Nemertean pro-
boscis, the great majority of authorities are in accordance that
it develops as an invagination from the epiblast,
which commences at the anterior extremity and gradually
pushes backwards. Extensive details as to its successive
developmental stages are, however, not yet to hand, only the
principal fact above mentioned being generally accepted.

It is highly important to notice that in this backward course
the proboscis takes its way between the two anterior thicken-
ings of the lateral nerve-cords, which in Carinella constitute the
simplest Nemertean brain, and in other genera become more
or less subdivided, the right and left halves being united by a
thick commissure (fig. 3), ventral in relation to the proboscis,
and by a thinner one dorsal to it. In all cases the pro-
boscis passes through the ring of nerve tissue thus
formed; in all cases the proboscidian sheath reaches
forwards to the level of this nervous commissure,
through which the proboscis passes (fig. 4).

If we may look upon the spinal cord and brain of vertebrates
as a dorsal coalescence of lateral trunks similar to those of the

304. A. A. W. HUBRECHT

Nemertines (as was already advocated in my paper “ zur Ana-
tomie und Physiologie des Nervensystems der Nemertinen,”
Amsterdam, 1880), then the double proposition above enunci-
ated necessarily leads to the conclusion that the spot just
mentioned corresponds to that part of the verte-
brate brain where the hypophysis (proboscis) bends
upwards towards the central nervous apparatus and
where the notochord (proboscidian sheath) termi-
nates, i.e. the region of the primitive fore-brain.
This proposition at the same time implies the homology be-
tween the vertebrate fore-brain and part of the nervous lobes
of Platyelminth ancestors. i

It remains to be further inquired into—and the facts as they
lie before us are very suggestive in this direction—whether
perhaps the distinction between the two pair of lobes as they
are present in most Nemertines may not have been perpetuated
in the vertebrates, these superior lobes (after dorsal coalescence
of the two halves of the nervous system) becoming the fore-
brain, the inferior ones the equivalents of mid- and hind-brain.
The following two points are in favour of such an interpre-
tation: (1) the nerves for the higher sense organs, eyes,! and
olfactory (?) pits start from the superior brain lobes;
(2) the strong nerve which on both sides supplies the an-
terior (respiratory, M‘Intosh!) region of the «esophagus,
and for which in a former paper I have proposed the
name of N. vagus, takes its origin in the inferior lobes
(figs. 3 and 5).

Upon the dorso-median coalescence of these inferior lobes
and of the lateral stems, above the intestine and the probos-
cidian sheath, the latter must have become severed anteriorly
from its connection with both nerve-system and _ proboscis.
Might not the fact of the anterior end of the notochord being
bent upwards in several of the lewer Elasmobranchs (ef.

1 It is of course understood that the ectodermal eyes of Nemertines are
not directly comparable to the myelonic vertebrate eye. However, it is im-

portant that Graff has already succeeded in demonstrating true cerebral eyes
in other Platyelminths (‘ Monogr. der Turbellarien ’) !

ON THE ANCESTRAL FORM OF THE CHORDATA. 305

Gegenbaur, ‘ Das Kopfskelet der Selachier,’ pl. ix, figs. 1 and
2) be interpreted as a reminiscence of this connection ?

A further character which is common to the two epiblastic in-
vaginations, known as hypophysis and as proboscis respectively,
is the shifting of their external opening. Amongst Nemertines
examples are found which form a parallel to the larger bulk of
Vertebrates (fig. 6) where the hypophysis does not arise (as in
Petromyzon) independently on the outer surface, but where it
is an invagination directed upwards from the roof of the mouth
cavity. Both in Malacobdella and in Akrostomum (a genus of
Hoplonemertini, instituted by Grube, in which I place, for ex-
ample, M*‘Intosh’s Amphiporus bioculatus and Amph. has-
tatus, and of which I have myself examined several specimens)
the opening for the proboscis is not independently
situated at the anterior extremity, but is found on
the dorsal wall of the intestinal tract, just inside
the mouth (figs. 7—10). I have the strongest reason to
believe, upon which I will not here further enter, that this
is a secondary modification, and that the separate opening
is the original state of things, phylogenetically related to the
separate proboscis of certain Rhabdocoels.

The facts here advanced may justify usin looking
upon the Platyelminth (s. str. Nemertean) proboscis
as the homologue of the vertebrate hypophysis, as
was implied in the first part of our proposition.

The proboscidian sheath in Nemertines is a cavity closed on
all sides and lined by an epithelium. It is situated in the
median dorsal line, above the intestine, just inside the muscu-
lar body wall, to which it is more or less firmly attached. Mus-
cular fibres serve to a large extent towards the thickening of
the tube we are considering.

It terminates in the immediate vicinity of the anus, and
reaches forwards to just in front of the cerebral ganglia, which
in Schizo- and Paleonemertini are situated at a short distance in
front of the ventrally situated mouth. In the Hoplonemertini

306 A. A. W. HUBRECHT

the mouth has travelled forwards till close to the tip of the head,
the intestinal tract thus passing beyond the proboscidian sheath
anteriorly. In certain other Nemertines the proboscidian
cavity does not extend through the whole length of the body
posteriorly. This is, for example, the case in that genus which
must be regarded for several reasons as the least differentiated,
primitive type—the genus Carinella. It isonly in the anterior
region of the body that the proboscis and the cavity surrounding
it are found, the latter situated as usual above the intestine.
Here, too, the mouth is found ventrally, the opening for the
proboscis terminally. One other genus—Drepanophorus—
deserves special mention, in so far as the proboscidian sheath
has the bulk of its cavity increased by lateral thin-walled
sacs, metamerically placed, one above each lateral cecum
of the intestine, and communicating with the cavity of the
sheath by narrow perforations of the muscular tissue of its
wall. Nemertes carcinophila is said both by M(‘Intosh
and Barrois to be without a special proboscidian sheath.
Barrois found the proboscis much reduced (according to him as
an effect of parasitism) and floating in the general body-cavity.
Not having examined this species myself, and not having
either ever met with a general body-cavity in other Nemer-
tines, I would venture to suggest the necessity of a careful re-
examination of this species, which might prove to be not
without importance for the préblem we are considering.

The type according to which the proboscidian sheath is built
up is very similar throughout the whole group, although the
muscular elements in its wall may increase in number (fig, 16)
and become more complicately arranged, or its size may be consi-
derably reduced. It is capable of a very considerable increase
in width corresponding to the movements, the rapid retraction,
or the mode of coiling up of the proboscis it encloses. It is,
moreover, filled with a fluid, containing corpuscles characteristic
in shape, and in one case—Cerebratulus urticans—charac-
teristic in chemical properties, viz. by the presence of hamo-
globin. This fluid is in no way connected with the fluid
circulating in the longitudinal and transverse blood-vessels.

ON THE ANCESTRAL FORM OF THE CHORDATA. 357

The dorsal blood-vessel takes its course beneath the probos-

cidian sheath, between it and the intestine; in many instances
it is enclosed in the muscular wall of the sheath in the fore-
most part of the body, above the cesophagus. The possibility
of a comparison with the subnotochordal rod of Vertebrates
ought to be considered.

The inner epithelium lining the cavity of the sheath is very
marked and everywhere present; it is least conspicuous in
Carinella, owing perhaps to the considerable extension which
the sheath had undergone in all specimens that have hitherto
been examined in view to this point.

This being the general arrangement of the proboscidian
sheath, it now concerns us to examine what is known about its
development in the embryo. The data available are very scanty,
and in some respects contradictory. Barrois describes it in
certain species of Lineus as developing from the reticulum,
the mesoblastic tissue between the epi- and hypoblast, and
gradually extending backwards at the same rate as the develop-
ing proboscis pushes it in that direction. In Amphiporus the
development of the proboscidian sheath was studied by the
same observer, and according to his description there is a
remarkable divergence from the development in Lineus.

In Amphiporus the proboscidian sheath is not formed gra-
dually, travelling slowly backwards along the median dorsal
line, but the sheath suddenly appears all round the
“whole length of the proboscis. It is here formed out
of the fatty mass, which also gives rise to the diges-
tive tube.

Tetrastemma, another Hoplonemertine, corresponds closely,
according to the same observer, with Amphiporus just de-
scribed.

Salensky, who has lately given a very short account (‘ Biolo-
gisches Centralblatt,’ 1883) of the development of Nemertes
(Borlasia) vivipara, ascribes a mesoblastic origin to the probos-
cidian sheath. Nevertheless, he noticed what appeared to him
to be a connection between the first origin (Anlage)
of the esophagus and that of the proboscis. As he

358 A. A. W. HUBRECHT

has postponed giving the details of this connection toa later
publication we cannot at present judge of its significance.

Hoffman is the only other author who gives any details
about the formation of the proboscidian sheath. According to
his account of sections made of Tetrastemma, a portion of the
proboscis is split off from the dorsal surface of the alli-
mentary canal. The muscular proboscidian sheath is meso-
blastic in origin. This observation, which can hardly be
brought to agree with the epiblastic origin of the proboscis,
noticed above, might perhaps allow of a different interpreta-
tion. As a simple suggestion I would advance, that perhaps
Hoffmann may have mistaken the formation of the inner por-
tion of the proboscidian sheath (so often confused with the
proboscis !) for that of a part of the proboscis itself.

Hypoblastic formative elements internally would then coa-
lesce with mesoblastic derivates, more especially muscular
elements, exteriorly applied to the former and constituting
together the proboscidian sheath, i.e. the wall of the probos-
cidian chamber. .

Such an interpretation would appear to be more acceptable
than the coalescence of a tubiform derivate of the hypoblast,
with an invagination of the epiblast travelling backwards, the
fusion of these two giving rise to the definite, cylindrical, evers-
ible proboscis. Balfour, in his ‘ Comparative Embryology,’ is
not inclined to accept Hoffmann’s statements without further,
confirmation.

Still this observation, if it may be interpreted as proposed,
would be of importance and its repetition much to be desired.
This and Barrois’ description above cited appear to open the
prospect that embryology may eventually succeed in
demonstrating for the proboscidian sheath, or for one
of its constituent parts, a hypoblastic origin.

If such be proved to be the case, not only its situa-
tions but also its development would correspond to
that of the notochord of the lower Vertebrates. Still,
considering in how many cases the origin of the notochord in
Vertebrates is apparently mesoblastic (this phenomenon being

ON THE ANCESTRAL FORM OF THE CHORDATA. 309

considered as secondary, the hypoblastic origin as the primary,
or ancestral arrangement), it cannot be considered as absolutely
necessary, that in the other offshoot, the Nemertines, the hypo-
blastic origin of the proboscidian sheath should first be demon-
strated before any homology between notochord and proboscidian
sheath may be accepted. In Nemertines as well as in Verte-
brata the mesoblastic origin of the proboscidian sheath might
be a secondary condition. In that case much value could not
be attributed to special cases of coincidence in the embryolo-
gical data, and it would be the most primitive representatives
of both groups which would more especially be likely to fur-
nish evidence of a conclusive character. However, we must
here wait for more circumstantial evidence before being justi-
fied in further advancing in the domain of speculation. Yet,
with respect to this question, it must not be overlooked that
the fact of the cavity of the proboscidian sheath carrying pecu-
liar corpuscles, and of its being a closed sac lined by an
epithelium, tends far to give it the general character of a
body-cavity, a coelomic diverticulum, which is indeed situated
dorsally and longitudinally, but which by those general cha-
racters would lead us to expect a derivation from the archen-
teron, rather than a schizoceelic origin in mesoblastic tissues.

If we were inclined to accept the mesoblastic derivation as
primary, and wished to picture to ourselves a possible common
origin of notochord and proboscidian sheath in the common an-
cestor of Vertebrates and Nemertines, we should have to postu-
late as a-still more primitive arrangement an axial thickening
of the mesoblastic tissues, which became more solid in the one
and hollowed out by the proboscis in the other. This would
clash both with the hypoblastic origin of the vertebrate noto-
chord and with the phylogenetic significance of the hypo-
physis.

We have now to consider certain aspects of the suggested
homology between notochord and proboscidian sheath.

There is no doubt that the fully developed notochord of a
Vertebrate is a structure of an entirely different character from
the proboscidian sheath of a Nemertean. The one is a solid,

360 A. A. W. HUBRECHT

rod-like organ, the other a hollow tube. However, at the very
earliest stages of its formation the notochord of the primitive ver-
tebrates (cf. Hatchek’s ‘ Development of Amphioxus’) possesses a
central groove, which is a derivate of the archenteron, and which
only secondarily, in accordance with the further differentiation
of the tissues of the notochord, obliterates.

The ulterior difference in histological structure of the one, a
cellular tissue eminently vacuolar, and of the other: the cellular
lining and fluid contents of a tube, the cavity of which does, as
a rule, not obliterate, is, however, no serious objection to their
eventual homology. In more than one instance modern mor-
phology recognises solid cellular strings to be homologous with
others containing a cavity inside them.

The different degree in which muscular tissue takes part in
the constitution of the proboscidian sheath (figs. 16—18) must
of course not be overlooked, the more so as it is entirely absent
in the notochord and its envelopes. However, in Nemertines
this muscular tissue can be shown to be most closely related to
the function of the proboscis, and in fact to be sometimes ex-
ceedingly reduced. For this reason its importance as a point
of comparison must not be over-rated.

All these differences arein the last instance due to the different
significance in the animal economy which has in the two groups
been attained by this organ. In the Vertebrates this central
rod-like structure, sustaining the mesoblastic somites in their
progressive development, has a significance as a temporary axis,
round which these processes take place. Its important cha-
racter as a primitive, i.e. as an ancestral organ, is recognised
notwithstanding, or rather just because of, its gradual disap-
pearance in the adult forms of the higher groups, where its

1 T must not omit to call attention to certain papers in the volume for 1882
of the ‘Archiv fiir Anatomie und Physiologie” ‘They came into my hands
after the completion of my MS. The one is by Lieberkiihn, “ Ueber die
Chorda der Siugethieren ;” the other by Braun, “ Entwickelungsvorginge am
Schwanzende der Saugethieren.” Both naturalists have succeeded in demon-
strating that in different regions of the body the notochord is at first a
hollow, tubiform structure (fig. 1]). Braun found the same in
birds. Kolliker, Strahl, and others have lately come to similar results !

ON THE ANCESTRAL FORM OF THE CHORDATA. 361

significance as a sustaining axis has been replaced by that
of the vertebral column.

That in a far distant ancestor of the vertebrates it may sim1-
larly have been subservient to the lodging of a retractile pro-
boscis, tactile in function, appears to me to follow from
a careful consideration of.the relations between the
hypophysis and the notochord, and between the
first-named rudimentary organ and the brain.

An important fact, which in conclusion I must make men-
tion of, is a phenomenon which I have repeatedly observed in
the posterior portion of the proboscidian sheath of different
species of Cerebratulus, very long Nemertines, where the
sheath reaches down to the posterior extremity of the body.
Whereas in young specimens of this species the sheath was a
hollow tube down to the very end, in older and larger speci-
mens the aspect of things had changed. In the posterior ex-
tremity of the body the cavity was here nearly filled up by a
continuous cellular tissue with distinct nuclei (fig. 18),
sometimes even entirely obliterated. This cellular tissue is
sometimes apparently glandular, the arrangement in some cases
even such that it must be interpreted as a set of radial acini, by
which the surface is considerably enlarged. Future investiga-
tions will have to decide whether the cases of evident oblite-
ration may be interpreted as a step towards real solidification
of that part of the tube which is comparatively of the smallest
value for the general function in connection with the expul-
sion of the proboscis. This change of function and histolo-
gical appearance is only present in the more primitively organ-
ised groups, which rarely make use of their proboscis; in the
more highly specialised Hoplonemertini, where the proboscis
is in constant play, and the development of the muscular ele-
ments in the proboscidian sheath vary considerably, it was
nowhere found.

Apart from the argument which can be derived from the
nature of this cellular coating the significance of this pheno-
menon will have to be carefully inquired into. Along such a
line of development we might picture to ourselves the eventual

362 A. A. W. HUBRECHT

conversion of a hollow proboscidian sheath into a solid noto-
chord, the more so as functionally the proboscidian sheath in
Nemertines may already be looked upon as an axis, around
which the other organs symmetrically arrange themselves as
they do around the notochord in Vertebrates. It must at the
same time be borne in mind that the muscular coating in this pos-
terior portion is found to be considerably reduced and replaced
by a more or less homogeneous and comparatively thin sheath.

Having thus far considered the arguments which are at pre-
sent available for insisting upon the homology between proboscis
and hypophysis, on the one hand, and between proboscidian
sheath and notochord on the other, it now remains to inquire
whether there are points in the anatomy of Nemertines, beyond
those just now exposed, which either corroborate or weaken the
evidence hitherto advanced in favour of the suggestion that the
Nemertines, more than any other known group of Invertebrates,
resemble the ancestors of the Protochordata.

I need hardly insist upon the fact that I do not advocate any
direct relation between existing Nemertines and existing Verte-
brates; my argument goes no further than the attempt to show
that the general plan of structure of a Nemertine is more in
accordance with that of a vertebrate animal than is, for ex-
ample, that of the Archiannelida, and that the link connecting
Ceelenterate ancestors with Vertebrate descendants has most
probably comprised forms in which two lateral nerve-cords
were present, ultimately coalescing dorsally, and in which an
epiblastic proboscis served for purposes which have either been
given up or have been replaced by others when the animals
gradually exchanged the Platyelminth for the Chordate type.

Simultaneously with this passage from the Cclenterate
type to the Chordate the highly important processes must have
been gone through which lead to the formation of a body-cavity,
separate from the archenteron with which, as embryology teaches
us, certain diverticula are originally in open communication,
ultimately becoming constricted off and developing into the
splanchnic and somatic layers, which have the body-cavity
between them,

ON THE ANCESTRAL FORM OF THE CHORDATA. 363°

The brilliant researches of Lang on Gunda segmentata,
and of Hatschek on the ‘ Development of the Amphioxus,’ must
here in the first instance guide us; and anybody having care-
fully perused those important contributions, and having com-
pared them with each other, must have been struck by the
great probability of the view advocated by Lang, that the ali-
mentary diverticula of these Platyelminths are the fore-runners
of the arrangement of the cclom in the higher enteroccelous
worms, and that through this link a glimpse is gained at the
road along which Annelids may have developed out of an an-
cestral Platyelminth stock.t

On the other hand, the stages in the development of Amphi-
oxus, where a double set of lateral diverticula of the archenteron
is present (fig. 12), which ultimately become converted into the
mesoblastic somites, appear to be of very great importance,
in so far as they render it highly probable that in the an-
cestry of vertebrates certain forms with metamerically placed
alimentary ceca must have obtained, of which the larval stage
of Amphioxus is the reminiscence. In the remaining verte-
brates the primitive alimentary diverticula giving rise to the
celom are reduced to two. This appears to be an ulterior
simplification. An attempt to explain this simplification, and
to bring the process of the formation of the coelom in Amphi-
oxus under the same head with that in the other vertebrates,
has at the present moment not yet been made. It is, however,
sure to be made some day by the leading authorities on the
subject. For the present it may suffice to point out that the
ulterior development of the mesoblastic somites in the bulk
of the vertebrates re-establishes the homology with the more
primitive processes in Amphioxus.

For us this larval stage of Amphioxus is all the more inter-
esting, because it must lead up to Platyelminths, correspond-
ing with Gunda in the presence of alimentary cxca, metame-

’ It must here be noticed that Lang has only very lately (‘ Biologisches
Centralblatt,’ May, 1883) emitted serious doubts concerning his own proposi-
tions. It remains to be seen whether future investigations will not tend to
confirm his original suggestive hypothesis rather than these doubts.

364 A. A. W. HUBRECHT

rically placed and of a general internal metamerisation,
but differing from Gunda in such important respects as the
presence of the forerunners, both of the hypophysis and of the
notochord, two structures no trace of which is found in the
salt-water Triclades. Such Platyelminths must needs
have resembled the present Nemertines more than
anything else.

Here the important question at once thrusts itself upon us:
Has the formation of a cceelom already been arrived at in the Ne-
mertines or not? i.e. have these animals a body-cayity developed
out of and separated from the primitive digestive tract or not ?
Although I have formerly, when attempts were made to bring
the Nemertines under the so-called Parenchymatous Flat-worms,
combated those attempts, and endeavoured to show that the
regular arrangement of digestive and generative czca, the deve-
lopment of muscular septa between them, &c., went contrary to
it, yet now that our ideas about the significance of a true body-
cavity as an ultimate derivate of the archenteron have of late
years gained so considerably in clearness and definition, I
should hesitate to affirm that any such body-cavity is present
in Nemertines, and would be inclined to answer the question
proposed above negatively.

Both in the more highly differentiated Hoplonemertini and
in the more primitive Schizo- and Palzonemertini, I have
met with numerous instances in which all the space which
remained free between the muscular body-wall on the one
hand, and the intestinal, generative, proboscidian, and circu-
latory cavities on the other, was one unbroken mass of con-
nective tissue.

Sometimes, more especially around the cesophagus, occurred
what were apparently fissures and cavities in this tissue. They
were not lined by an epithelium (are perhaps in communica-
tion with the vascular system ?), and could best be compared
to a true Schizocelom (Huxley), i.e. fissures in a mesoblastic
tissue.

All this makes me very much inclined to look upon the
alimentary diverticula of the Nemertines in the same light as

ON THE ANCESTRAL FORM OF THE CHORDATA. 365

Lang does upon those of Gunda; incipient celomic sacs,
comparable to those of the larval stage of Amphioxus.

A question very difficult to answer is this: How do these
alimentary diverticula eventually come to exchange their
function and significance to such an extent? If they were
originally acquired with a view to an enlargement of the diges«
tive surface, they must in the course of time, as they became
constricted off, have lost this function,. and in its stead have
developed powerful layers of epithelial muscular tissue in their
walls, which then represent the successive myomeres,
and which finally supplant the original muscular
body-wall (Hautmuskelschlauch), itself never divided
into myomeres, and originally derived from the
epiblast.

Traces of this epiblastic muscular sheath, primitively en-
veloping the myomeres, which secondarily spring from the
alimentary diverticula, appear to be found in certain Verte-
brates, externally to their general musculature.

It remains for the present unsolved what were the leading
factors in this important transformation, the general outlines
of which we have here only touched upon.

We have now to compare Nemertines and primitive Verte-
brates under another important head: foremost cesophageal
diverticula and their relation to respiratory functions and sen-
sorial (?) apparatus. Here, too, I do not wish tu enter into a
thorough discussion of the subject ; an enumeration of the chief
points may suffice for the present.

A special respiratory apparatus in the form of external
branchize has never been met with in Nemertines. In a very
early stage of embryological development, however, two
lateral diverticula, situated inthe very foremost
portion of the wsophagus in front of the mouth,
bud out from its wall (Bitschli, Barrois, and others),
and are inthis stage directly comparable to similar
diverticula which arise in the same region in the
Balanoglossus larva, and there give rise to the first
pair of branchial slits (figs. 14 and 15),

VOL, XXI1IL—NEW SER. BB

366 A. A. W. HUBRECHT.

In Nemertines these diverticula become constricted off from
their point of origin—the cesophagus—and entering into con-
nection with invaginations from the epiblast, which bring
about a free access of the external sea-water, they become
converted—at least in the large section of the Schizonemertines
—into an apparatus which I have proved to be subservient
(‘Zur Anatomie u. Physiologie der Nemertinen,’ p. 28) toa
process of cerebral respiration, in which oxygen is carried to
the nervous system itself, the cellular elements of which are
in this subdivision profusely provided with hemoglobin.

I am not prepared to say that in the great subdivision of
the Hoplonemertini, where the central nervous apparatus is no
longer provided with hemoglobin, but where, on the contrary,
the circulating fluid is, these diverticula, which continue to de-
velop in the same way out of the esophagus, are also—and in the
first place—subservient to a respiratory process. JI am rather in-
clined to believe that in this group the cephalic grooves—as the
epiblastic invaginations travelling inwards to meet the hypo-
blastic diverticula in question are called—remain more especially
adapted for sensiferous purposes, probably of olfactory nature.
The way in which the complicated organs in the adult, the so-
called side organs, develop, remains quite the same: an out-
growth from the esophagus coalesces with an ingrowth from the
epiblast, the principal difference being that the connection with
the brain-lobes is no more so intimate, and that the apparatus is
connected with the brain by a special set of nerves. In some
species it continues to be situated behind the brain, in others
it becomes placed in front of the central nervous apparatus.

It appears to me that these facts are not without significance.
However, I refrain for the present from a further discussion,
and would merely wish to point to an interesting detail in the
development of Amphioxus lately brought to light by Hatschek’s
researches, It is the presencein the anterior region of the
esophagus, in front of the mouth, of two lateral hypo-
blastie diverticula, differing in their nature and in their
further development, both from the archenteric diverticula
(mesoblastic somites), and from the branchial outgrowths of

ON THE ANCESTRAL FORM OF THE CHORDATA. 367

the esophagus. These two diverticula, originally symmetrical,
become constricted off from the hypoblast, and in their further
development they have a different fate, the left one com-
municating with the exterior by a ciliated opening,
which appears in the epiblast, the right one forming an epi-
thelial lining in the preoral body-region. ‘The left one was
looked upon by Kowalevsky as a special sense-organ
of the larva.

Although I am not prepared for the present to furnish any
evidence in this direction, I would call attention to the simi-
larity in development between these structures and the cephalic
diverticulum of Nemertines. Considering the amount of dege-
neration which in several respects Amphioxus appears to have
undergone, it does not appear impossible that the left lobe is
really a temporary olfactory organ, the right one having entered
upon other functions, and having lost its original significance.

These cesophageal diverticula of Amphioxus would stand
about in the same relation to the posterior paired outgrowths
of the esophagus which ulteriorly give rise to the branchial
slits of this animal, as would the two primary larval diver-
ticula of Balanoglossus, giving rise to the first pair of gill-slits,
to the following ones successively appearing behind them. In
Nemertines only one corresponding pair of respiratory diverti-
cula is encountered, and they may remain in connection with
‘those portions of epiblastic ingrowths which form the primary
constituents of a sensorial (olfactory ?) apparatus in certain of
the higher differentiated genera in the way we have above
traced.

The far reaching significance of our starting point has
obliged us to throw a rapid glance at the principal points-in
which Nemertines allow of a certain degree of comparison with
Vertebrates, and it would lead us too far offif we were to follow
this up for the secondary, less important points, or for those
which are at present not fully enough known to allow of any
fruitful comparison. Amoug the latter I count the excretory
and the generative apparatus. Do the closed generative sacs

368 A. A. W. HUBRECHT.

of Nemertines arise as part of the ccelom (cf. Lang, ‘ Gunda
segmentata)? What is the morphological significance of the
generative ducts which establish a direct communication be-
tween these sacs and the exterior, and which are recognisable
on the outer surface as a double set of symmetrical pores? Is
the paired nephridium provided with internal openings or is
it not? These and other questions will have to be diligently
studied and solved before comparison can extend itself in the
domain of these organs.

With respect to the vascular system, it is not unimportant
that in Nemertines it is on the whole a closed system of vessels,
sometimes carrying corpuscles charged with hemoglobin,
sometimes colourless, and giving off a system of transverse
connecting vessels, which link together the three longitudinal
stems. These transverse vessels do not give off any capillaries,
and are metamerically placed with unbroken regularity, one
for each internal metamere (intestinal diverticulum). Ifindeed
the suggested homology might prove to hold good between these
diverticula and the -mesoblastic somites of Amphioxus, the
significance of this regular disposition, one for each of the
transverse subdivisions of the body, corresponding in a general
way to the arrangement of the aortic arches in vertebrate em-
bryos could not be overlooked.

In conclusion I would point out that the speculations and
suggestions contained in the last pages ought to be distin-
guished from the contents of the first part of this article.
They have not in any way contributed to the formulating of
the hypothesis there brought forward; they are merely the
sequel in a train of thoughts which, starting from a comparison
of such important and primitive organs in both Vertebrates
aud Nemertines as are the nervous system, the hypophysis
and the notochord necessarily extended itself to other struc-
tures and organs occurring in both groups.

With respect to these, we must await more thorough inves-
tigations before pushing our speculations further.

THE RENAL ORGANS (NEPHRIDIA) OF PATELLA. 369

The Renal Organs (Nephridia) of Patella.
By

J. TT. Cunningham,
Fellow of University College, Oxford.

THE existence of two renal organs in Patella was first
pointed out by Professor Lankester in 1867,! and he stated at
the same time that he believed he had discovered, by careful
dissection, a minute orifice leading from the pericardium into
the left and smaller of the two. In 1877 Von Jhering, in a
paper on the ‘ Morphology of Molluscan Kidneys,’” con-
firmed the account given by Professor Lankester, at the same
time emphasising the fact that the two renal sacs are dis-
tinct, as had already been stated by Lankester, who says,
after describing the two renal papille with their openings to
the exterior, “ These two orifices represent two renal
organs, as in Lamellibranchs.” Von Jhering was unable to
find any pericardial openings. In April, 1881, Professor Lan-
kester and Mr. A. G. Bourne® examined fresh limpets as to the
pericardial orifice of the kidneys. They found that injections
from the pericardium passed sometimes into the right and
sometimes into the left renal sac, but that there was only one
orifice leading into a narrow subanal tract belonging to the
organ of the right side. Professor Lankester suggested to me
some time ago that I should endeavour to settle definitely this
question of the reno-pericardial pore in Patella by cutting a
series of sections through the parts, and so tracing their relations

1 ¢Ann. and Mag. Nat. Hist.,’ 3rd series, vol. 20, 1867.
2 «Zur Morph. der Niere der Mollusken Zeitschr. f. w. Zool.,’ Bd xxix.

3 «On the originally bilateral character of Renal Organ of Prosobranchia,”
‘Ann. and Mag. Nat. Hist.,” vol. vii, 1881.

370 J. T. CUNNINGHAM.

to one another. My investigations were commenced in the
Zoological Laboratory of University College, London, under
the supervision of Professor Lankester, and were subsequently
completed at Naples in the winter 1882-83.

I injected several specimens of Patella vulgata from the
south coast of England with gelatine solution coloured with
Berlin blue by means of a fine-pointed pipette inserted into the
pericardium of the animal in the fresh state, and after harden-
ing I cut out the piece to be examined, stained it whole, and
cut it embedded in paraffin. I frequently found the injection
in the cavities of both kidneys, and soon became convinced
that each kidney had a separate communication with the peri-
cardium ; but the difficulty of getting complete series of sections
with the tissues well preserved, in which the channels of com-

.unication could be satisfactorily traced by means of the injec-
tion, was-so great that I was unable to discover the exact
character and relation of the canals leading from the pericardium
to the cavity of the kidneys. At the Naples Zoological Station
I have used for injection simply the cold solution of Berlin blue
in water, which has many advantages over the gelatine solution ;
it penetrates more easily, and is seen in sections as a blue line
along the sides of the cavities which it has reached, while the
gelatine has the two disadvantages of solidifying before the in-
jection is complete, and contracting during the process of har-
dening. After many trials I have succeeded in obtaining
complete series of sections from injected and uninjected speci-
mens, in which the two canals of communication can be traced
through their whole length.

The species of Patella which I have used at Naples, and
which is most common there, is P. coerulea; its shell is much
flatter than the P. vulgata of the English coast, and it does
not attain such a large size; but I have found no differences
between the organs of the two species.

When a Patella is removed from its shell in the fresh state
the dorsal surface appears of a deep black colour, which is due
to the presence of pigment in the superficial epithelial cells.
This pigment can easily be washed off. When this is done, and

THE RENAL ORGANS (NEPHRIDIA) OF PATELLA. 371

the roof of the mantle cavity cut away, the superficial parts of
several organs are seen.

The pericardium occupies the left half of the posterior
border of the mantle cavity. Projecting from this border on the
right are three papillze with apertures at the ends; the central
one is the anal papilla, and on each side of it is the orifice of a
renal organ. The left organ is small, extending from the
right border of the pericardium as far as the rectum, while its
extent from before backwards is the same as that of the peri-
cardium. The right renal organ is distinguished by its
darker colour, and extends round behind the left organ and
part of the pericardium, and over the greater part of the dorsal
surface, except a small portion where the liver is visible. On
dissection’ the cavity of the right kidney is found to extend
under the visceral mass from the right side of the animal as far
as the median line, forming here a flat sac between the muscle
of the foot below and the genital gland above; it also sends a
prolongation under the rectum and left kidney, between these
and the liver, which extends to the wall of the pericardium.

g | ) = : ! o<
> AS US SSOeAe hy
KE 5 : vey.
4 VEO Wife WEES
\
h ag af

Diagram of a transverse section of Patella passing through the two renal organs,
the rectum and the pericardium, so as to show the two reno-pericardial
canals and their pores. @. Main cavity of the larger nephridium. ad. Sub-
anal tract of the larger nephridium. 4. Cavity of the smaller nephridium.
c. Pericardium. d. Rectum. e. Liver. “£ Stomach. yg. Integument,
i. Black line representing the renal epithelium. 7. Reno-pericardial
pore of the smaller nephridium. 4%. Ditto of the larger nephridium.

The canals which lead from the pericardium in passing to
the right to reach the renal cavities ascend slightly, so that in

Bie J. T. CUNNINGHAM.

a series of sections the end opening into the latter is first met
with when the series begins from the anterior end of the pre-
pared piece of tissue. The actual opening of the canal into the
main cavity of each kidney is seen in such sections with the in-
jection in the canal and in the aperture. A triangular piece of
tissue is seen before each opening, which lower down joins
with the tissue of the projection, which contains the canal,
and therefore forms a sort of valve open towards the kidney
and closed towards the pericardium. The part of the canal in
the immediate neighbourhood of the aperture into the kidney
is lined by cellular epithelium, which consists, apparently, in
some places of more than one layer of cells. There are in
many sections indications that this epithelium is ciliated, but
whether it is so or not I could not definitely determine.
Towards the part of the canal communicating with the peri-
cardium the epithelium disappears; it may pass into a flat
epithelium lining the whole interior of the pericardium, but my
sections do not show this clearly.

I have been able to discover and study these canals in series
of sections prepared withcut injection, and in these the epi-
thelium is seen much more clearly.

With regard to the position of these canals, it is to be ob-
served that the projections towards the cavities of the kidneys,
through which they run, lie close to the external surface of the
body, that surface which forms part of the floor of the mantle
cavity ; this is especially true of the canal belonging to the left
renal organ. In the second place, they are not far distant
from the external apertures of the kidneys, as is shown by the
fact of the sections in which they are seen including a portion
of the mantle cavity.

It is difficult to understand how it can have come about
that the right kidney of Patella communicates with the peri-
cardium by passing under the intestine instead of over it, unless
the intestine at one time passed through the pericardium; in
this case it could in passing out again during ancestral history,
and becoming separated from the pericardium, have passed
between the two openings of the kidneys ; but if the pericar-

THE RENAL ORGANS (NEPHRIDIA) OF PATELLA, 378

dium had always been above the intestine as in Chiton, and
had then passed to the left to take up the position it has in
Patella, the channels of communication which attach the
kidneys to the pericardium would have both been above the
intestine. It may be suggested that the original condition in
the ancestors of Patella was similar to that still obtaining in
Halictis and Fissurella where the intestine is surrounded by
the ventricle and pericardium,

The structure of the kidneys in Patella resembles that of
other Molluscs. The external orifice leads into a central flat-
tened cavity lined throughout by the secreting epithelium
which is characteristic of renal organs throughout the Mol-
lusca. This central cavity sends off numerous diverticula
which branch and ramify and form a spongy tissue beneath
the integument on the dorsal surface, and to a less extent in
the other regions bordering the kidneys. These glandular
diverticula alternate with large irregular lacunz containing
blood ; the wall of the diverticulum consists of the secretory
epithelium supported by a thin layer of nucleated, fibrous,
connective tissue, and from the latter trabecule of the same
tissue pass off here and there and stretch across the blood-
lacune to reach the wall of a neighbouring diverticulum.

These trabecule occasionally contain BANDS OF MUSCULAR
FIBRES. This is true of both the small kidney of the left side,
and that of the right, so that both can be contracted to a
certain extent.

The minute structure of the secreting epithelium is some-
what difficult to make out; when a portion of either kidney is
teased on a slide in the fresh state and examined in sea water,
a number of loose spherical vesicles are seen scattered over the
field ; these are of various sizes and contain a varying number
of small dark coloured concretions soluble in potash; they are
almost all ciliated. When the edge of a piece of renal tissue is
looked at similar cells are seen forming a regular ciliated epi-
thelium ; when the surface of the same epithelium is observed,
the outline of the cells is seen to be polygonal. The great
quantity of concretions present makes it impossible to decide

OVA J. T. CUNNINGHAM.

whether there is but one layer of cells or more than one. A
nucleus is often visible in the cells. In sections the epithelium
is often badly preserved ; its cells seem to be very delicate and
easily destroyed, and the presence of the concretions here also
obscures the cell-outlines. When the epithelium is well pre-
served and stained it has the following appearance. Near its
base the nuclei of the cells are seen surrounded by the concre-
tions, and the vertical boundaries of the cells are also visible.
Towards the cavity of the gland the epithelium terminates in a
number of clear rounded projections ; in sections I have never
been able to discover the cilia. It is evident that the rounded
projections are the vesicular cells still attached; it is these
which drop off so easily and are seen free in teased prepara-
tions. Their remains can also be often seen in sections in the
lumen of the gland, and the process of secretion is evidently
effected by these cells becoming detached in the renal cavities,
and then breaking up and allowing the concretions to escape.
But there is evidently another layer of cells in the epithelium
beneath the vesicular projections, cells which contain concre-
tions but are not yet vacuolated, and which show a nucleus;
these take the place of the mature secretory cells when they
fall into the lumen, and go through the same course of deve-
lopment as these. Won Jhering has observed the stages of
this development in the kidney cells of Tethys,’ in prepara-
tions of the fresh kidney. I have not been able to trace the
process in this way in Patella. Very often the renal epithe-
lium in sections does not show any vesicular projections, these
having disappeared in the course of preparation ; but the pro-
jections present variations in size and prominence which prob-
ably correspond to variations in secretory activity.

The structure of the two renal organs in Patella is exactly
the same; the difference in colour which they present to the
naked eye is due to the fact that the urinary concretions are
more numerous in the right kidney than in the left. The cavities
of the two kidneys do not communicate. ;

I have also cut some series of sections through the nuchal

1 Von Jhering, ‘ Morph. Jahrbuch,’ Bd. 2.

THE RENAL ORGANS (NEPHRIDIA) OF PATELLA. 375

papille of Patella, in order to test the accuracy of the figure
which Spengel has given of their structure.!| There is little
doubt that part of each of these papillz consists of a sensory organ
homologous with that which exists on the attachment of the
gillin Haliotis. A nerve from the visceral commissure can be
traced up to each papilla, as Spengel describes it, and in sec-
tions a nervous ganglion is seen beneath the epithelium, as in
his figure. Besides this ganglion there is to be seen in the
sections a cellular structure, oval in shape, surrounded by con-
nective tissue, and divided by trabecule into compartments of
various size. This may be the branchial rudiment, but it lies
beneath the epithelium, and is not a special development
of it.

1 « Die Geruchsorgane, etc., der Mollusken,” ‘Z. f. w. Z.,’ Bd. xxxv.

376 Cc. O. WHITMAN.

A Rare Form of the Blastoderm of the Chick,
and its Bearing on the Question of the For-
mation of the Vertebrate Embryo.

By

C. ©. Whitman, Ph.D.

With Plates XXIV and XXV.

In the summer of 1878, while in Leipsic, I obtained a blas-
toderm of the chick, which presented at least one unusual and
very remarkable feature. The egg had been kept in an incu-
bator, at a temperature of 37° to 38° C., for eighteen hours.
After cutting away a portion of the shell and removing as
much of the white of the egg as could be done without injury
to the embryonic disc, the remainder of the egg, while still in
the shell, was carefully dropped into a bowel of nitric acid
(10%). The embryonic portion was then freed from the
coagulated white by the aid of a feather, and, after fifteen
minutes’ immersion, was carefully cut around by sharp scissors
and floated into a watch-glass. The vitelline membrane was
then removed by the aid of pincers, and the yolk detached by
gently shaking the blastoderm in the watch-glass. The acid
was next turned off and replaced with distilled water, several
times renewed. After being thus thoroughly washed from the
acid it was stained in an aqueous solution of carmine, then
passed through several grades of alcohol, and finally mounted
in balsam. The whole process was accomplished without
causing any distortion or wrinkle, and without the loss of any
portion of the blastoderm, as the preparation still shows.

RARE FORM OF THE BLASTODERM OF THE CHICK. 377

The following topographical measurements were taken from
the mounted preparation by the aid of the camera lucida:

Blastoderm 9:12 mm. long, 8:04 mm. wide.

Marginal notch (“ Randkerbe,” Rauber) ‘46 mm. deep.

Area pellucida 2-4 mm. long, 1°76 mm. wide.

Distance from the anterior end of the area pel. to the corresponding
margin of the blastoderm 3°56mm.

§ posterior a ss ae 316 ,,
rH right side ” ” ” 2°8 ”
ay ' left side . z $9 3°48 ,,

Primitive streak 1:04 mm. long; the distance between its
fore end and the anterior border of the area pellucida 1:2 mm.
Distance of the marginal notch from the posterior end of the
primitive streak 2°2 mm.

The blastoderm is subcircular, and the marginal notch is
situated on the right side instead of at the hind end of the
main axis.!_ The area pellucida has an irregular pyriform out-
line, with the smaller end directed backward. It is placed a
little excentrically, being a little nearer the right side than the
left.

In the pellucid area two regions are easily distinguished ;
one, which is sometimes called the embryonic shield (s), is
pyriform, occupies the central and rear portion of the field,
and bears the primitive streak ; the other is less opaque, peri-
pheral in position, and crescentic in shape. The horns of the
crescent are not quite symmetrical, the left being the longer, so
that the hind end of the shield lies nearer the right than the
left side of the pellucid area. The shield shades off very
gradually into the clearer crescentic region.

The primitive streak is very well marked, and shows a well-
defined primitive groove. In front of the primitive groove a
linear opaque area is seen to vanish a little before the front
margin of the shield, and appears to be a continuation of
the primitive streak itself. This linear area is the “ kopfort-
satz” of Kolliker (“Axenplatte,’ Kupffer), and represents,

' The main axis passes to the left of the primitive streak, but parallel
with it.

378 C. O. WHITMAN.

according to Dursy and Balfour, the commencement of the
notochord.

The primitive groove does not end at the hindmost point of
the area pellucida, but curves to the right near where it leaves
the shield, and PASSES ON THROUGH THE AREA OPACA, TER-
MINATING IN THE MARGINAL NotcH. ‘The primitive streak
cannot be followed beyond the inner edge of the area opaca.
Not the slightest trace of a falciform expansion! of its posterior
end could be detected. No head-fold has yet appeared to mark
the limit of the prolongation of the primitive streak.

The only exceptional feature of this blastoderm is the con-
tinuation of the primitive groove to the marginal notch; and I
believe this is the first time that such a feature has been re-
corded. Balfour, His, Rauber, and others have expressly
stated that they have never been able to trace this groove to
the margin of the blastoderm. Lauber, His, Balfour, and
Pander have observed cases in which a marginal notch was
present, but not a single instance where the groove extended
to the notch. Although this is probably the first time that
such a peculiarity has been described, I cannot say that it is
the first time it has been observed. Indeed, I happen to know
that one other similar case was found only a few days before I
obtained the one here described. It was my friend, Dr. A.
Bohm, the present assistant of Professor Kupffer, who found
the first case. What use he has made of his preparation |
cannot say, but I am quite certain that he has never published
any account of it. I have long deferred the publication of this
paper in the expectation that Dr. Bohm would make known
his discovery, and in the hope of obtaining another specimen
myself for sections. As neither of these events have been ful-
filled, I have decided to describe the blastoderm in my pos-
session.

As to the fact that the primitive groove extends to the
notch, there is no room for a shadow of doubt. That arm of
the groove which passes through the area opaca could be seen
with the naked eye even before the blastoderm was cut from

1 See Kupffer, 27, 28, 29, 30.

RARE FORM OF THE BLASTODERM OF THE CHICK. 379

the yolk, and the same is still true in the mounted prepara-
tion. By transmitted light I can see with the unaided eye the
primitive streak and the external arm of the primitive groove
as a clear line leading from the notch to the base of the streak.
The inner arm of the groove is just barely recognisable under
the same conditions. The external arm is a little more than
twice the length of the inner arm, and forms with the latter
an angle of about 110°. In the blastoderm found by Dr.
Bohm, the posterior portion of the groove formed a straight
line with the anterior, and thus the marginal notch fell
directly behind the primitive streak.

What now is the meaning of this continuity between the
primitive groove and the marginal notch? It is certainly a
deviation from the avian type of development, of such rare
occurrence that it must be admitted to be an anomaly ; still it
is, if I am not mistaken, an anomaly for which comparative
embryology furnishes a very satisfactory explanation. It is a
well-known fact that the embryos of all Amniota are formed
near the centre of the blastoderm, while those of all anamni-
otic vertebrates have their origin at the margin. It is now
generally believed that the central position has been derived
secondarily from the marginal, and this belief has given rise to
various speculations in regard to the meaning of the primitive
streak and its relation to the embryo. Intimately connected
with these questions is another relating to the mode of forma-
tion and the growth in length of the embryo.

That these questions are not very easy of solution is evident
from the fact that, after half a century of patient research, a
decade or more of plodding “ microtomization,” and the indus-
trious accumulation of embryometrical tables as the fruit of
sections made in every nameable plane, the most eminent
authorities in embryology cannot agree in their interpretation
of the primitive streak, nor even in an opinion as to the part it
plays in the formation of the embryo. When Kupffer and
Rauber differ as to what constitutes the blastopore of the
chick ; when Goette, Kolliker, His, and Waldeyer affirm that
the primitive streak is directly concerned in the formation of

380 Cc. O. WHITMAN.

the trunk of the embryo, while Dursy and Balfour assert that
it entirely atrophies, or at most (Balfour) forms only the “ tail
swelling” and a part of the ventral wall of the post-anal gut;
and when His and Rauber maintain that the embryo is formed
by a retrogressive concrescence of the two symmetrically placed
halves of the germinal ring, while Balfour and others contend
that there is no such concrescence, and that the lengthening of
the embryo is entirely by intussusceptional growth, it might
seem like presumption on my part to offer any suggestions on
these disputed points. It is far from my intention, however,
to enter into an elaborate consideration of all these questions,
and I have no conclusions to present which will take me very
far from the plain highway of observed fact.

As to the meaning of the primitive groove, I think the
blastoderm I have described furnishes a very strong confirma-
tion—not to say verification—of a theory that originated with
Balfour and Rauber ; namely, that this groove represents
a portion of the blastopore. It was this theory that con-
ducted Rauber to a correct interpretation of the marginal
notch. No one had ever discovered any connection between
the notch and the primitive groove, or suggested any such
relation between the two structures. The notch was of rather
rare occurrence, and therefore treated as an irregularity that
required no explanation. Its position directly behind the pri-
mitive groove appeared to Rauber to fit in with the opinion
that this groove extended originally to the very edge of the
blastoderm, precisely as it still does in Elasmobranch Fishes ;
and hence he regarded it as “the ideal hind end of the groove.”
This conjecture is raised to the dignity of an observed fact in
the present case, and its verification is complete, provided the
blastoderm here considered represents an atavistic form.
Comparative embryology must be our guide in a question of
this nature.

All the questions that cluster about the primitive streak are
only so many special sides of a more general question, namely,
How istheembryo formed? On the decision of this ques-
tion hangs that of all the rest. Two different views have

RARE FORM OF THE BLASTODERM OF THE CHICK. 381

been put forward, which may be conveniently distinguished as
the differentiation theory and the concrescence theory.
His and Rauber are the chief exponents of the theory of concres-
cence, and appear to have arrived at similar conclusions inde-
pendently of each other. His developed his view in connection
with investigations on the embryology of Osseous Fishes, and
shortly afterwards extended it to the Elasmobranch and, with
some reserve, to the Chick. With reference to the Salmon,
His announced his conclusion in the following words:

“The structural basis (Uranlage) of the body, then, is a
flat ring, which has its maximum width and thickness at the
future head end, its minimum at the opposite tailend. The
two lateral halves of the ring approach and unite retrogressively
[successiv] as symmetrical body-halves ” (No. 20, pp. 19, 20).
In opposition to Oellacher, His maintains that the head end is
to be regarded as a fixed point, and that the axial concrescence
of the two halves of the ring begins at this end and advances
towards the tail end, which is thus the last formed part of
the embryo. The entire ring is thus brought together along
the axial line, and this process goes hand in hand with the
epibolic expansion of the blastoderm.

The Elasmobranch embryo is formed in a similar manner,
but with this difference, that only a posterior portion of the
ring takes a direct part in forming the embryo, so that the pro-
cess of building up the embryo is completed long before the
final closing in of the yolk by the blastoderm.

In his excellent memoir, entitled ‘ Primitive Streak and
Neurula of the Vertebrates,’ Rauber has given a brief survey
of vertebrate embryology, and summarised the more important
results reached by others and himself, with a view to making
clear their theoretical bearings. Rauber claims that the Avian
embryo arises, in the main, by a longitudinal concrescence of
the two halves of the germinal ring (“ Keimring’’), in the
same manner as the Piscian embryo; and further, that this
process underlies the formation of all vertebrate embryos.

In the case of the Chick, both His and Rauber have called
attention to a lunula-shaped area that appears, after about six

VOL, XXIII.—NEW SER, cc

382 C. O. WHITMAN.

hours’ incubation, in the posterior half of the area pellucida ; and
both agree that this area is continuous at its base with the area
opaca, or germinal ring. According to Rauber, the cells consti-
tuting the lunula, by a centripetal movement, arrange them-
selves along each side of the longitudinal axis of the future
embryo, and thus give rise to a “ bilateral string,” the so-called
primitive streak. The streak represents simply the united
edges of a small portion of the blastoporic rim (“‘ properistome,”’
Haeckel), and this mode of origin of the embryo is called “ sto-
matogenic.” The streak, groove, and marginal notch are re-
garded as so many “ phenomena of conjunction.” The area
opaca is held by Rauber to be homologous with the germinal
ring of the Fishes, and here, as in the Elasmobranch, a posterior
“‘embryoplastic” portion (about one third of the entire ring)
is distinguished from the anterior non-embryoplastic (“ peri-
embryonal”) portion (No. 34).

His distinguishes two concentric rings in the area opaca, or
‘‘ring-area.’ ‘The ring-area of the embryonic disc accord-
ingly now consists of an inner broad and opaque germinal-wall
portion and a thin transparent margin (or secondary marginal
rim) ; both are distinguishable with the naked eye” (No. 22,
p- 165). These two zones of the area opaca are plainly seen
in the blastoderm I have described. The inner thicker zone is
the one which contains the embryoplastic material.

From the above citations it is evident that His and Rauber
are in perfect accord on the main question; and, while both
claim that the vertebrate embryo arises by the concres-
cence ofthe homotypical halves of the germinal ring,
neither has anywhere intimated that intussusceptional growth
could not go on at the same time. On the contrary, Rauber
has expressly stated it as a self-evident fact that such growth
is a concomitant of the process of concrescence (No. 39,
p. 06).

In opposition to the view taken by His and Rauber we
have the differentiation theory put forward by Balfour. In
the second volume of his ‘Treatise on Comparative Em-
bryology’ Balfour has summed up his latest conclusions on this

RARE FORM OF THE BLASTODERM OF THE CHICK. 383

question, and stated at some length his objections to the con-
crescence theory. Balfour takes the Elasmobranch embryo as
a type, a case which His regards as the best illustration of his
own view. ‘The Elasmobranch embryo,” says Balfour,
“arises from a differentiation of the edge of the blastoderm,
which extends inwards from the edge for some little distance.
This differentiation is supposed to contain within
itself the rudiments of the whole embryo, with the ex-
ception of the yolk-sack ; and the hinder extremity of it, at the
edge of the blastoderm, is regarded as corresponding with the
hind end of the body of the adult. The growth in length
takes place by a process of intussusception, and till there are
formed the full number of mesoblastic somites it is effected, as
in Cheetopods, by the continual addition of fresh somites
between the last-formed somite and the hind end of the body”
(p. 254).

The “ His-Rauber view” is introduced as “a somewhat
paradoxical view,” supported by three not very forcible argu-
ments, and a number of minor arguments not worth mention-
ing. These three arguments are stated to be—(1) the continuity
between the thickened edge of the blastoderm and the medul-
lary folds; (2) the embryometrical investigations of His; and
(3) some of the phenomena of double monsters studied by
Rauber. Balfour says that the first argument affords no sup-
port for either theory, and the second appears to prove his own
theory of growth. Rauber’s view of “ pluriradial development”
is passed over without a word of comment, presumably because
it was supposed to be without special importance to the
discussion.

Balfour’s objections to the concrescence theory may be sum-
marised as follows :

1, The medullary groove closes behind earlier than in front,
and the groove does not terminate behind in an acute angle, as
might be expected if the embryo were formed by the coales-
cence of the edges of the blastoderm.

2. The formation of the neurenteric canal would make it
impossible for any further increase in length by concrescence.

384 Cc. O. WHITMAN.

“Tf, therefore, His’s and Rauber’s view is accepted, it will
have to be maintained that only a small part of the body is
formed by concrescence, while the larger posterior part grows
by intussusception.”

3. The blastopore in Amphioxus is not coextensive with
the neural groove, for it is nearly closed before the groove
appears.

4, According to His and Rauber, “ the whole of the dorsal,
as well as of the ventral wall of the embryo, must be formed
by the concrescence of the lips of the blastopore, which is clearly
areductio ad absurdum of the whole theory.”

5. According to Kupffer (No. 31) the epibolic growth of the
blastoderm in Clupea and Gasterosteus is equally rapid on all
sides until the equatorial line of the egg is passed ; and Balfour
considers this to be “absolutely inconsistent with the con-
crescence theory.”

These five arguments probably include all there is of
importance to be said against concrescence. Although some
of them may claim to ke based on comparative embryology,
not one of them, nor all of them together, are broad enough to
cover the ground embraced in Balfour’s theory of the origin of
vertebrates. No attempt is made to discuss the two opposed
theories on the basis of the now generally received view, that
the ancestral form of the vertebrates was an Annelid; and it
seems to me surprising that an advocate of this view should
leave the Annelids almost entirely out of consideration. Bal-
four makes the Elasmobranch embryo the point of departure,
and evidently because the primitive type of development has
probably suffered less modification here than in the amniotic
vertebrate. On similar grounds we may ask, Why not make
the Annelid type of development our starting-point, since the
mode of development may be presumed to have been conserved
in its greatest purity in those animals that have made the
smallest departures from the ancestral form? If there is any
truth in the supposed genealogical relationship between verte-
brates and Annelids we have certainly a right to expect some
fundamental agreement in their modes of development,

RARE FORM OF THE BLASTODERM OF THE CHICK. 3895

share the opinion with some others that such an agreement
does exist, and in it I find one of the most conclusive evidences
of genetic affinity. This agreement consists not only.in the
metameric division of the embryo, which is the only point of
agreement alluded to by Balfour, but also in the formation of
the embryo by concrescence of the two halves of the embryonic
ring.

Accordingly I hold that Balfour has left out of consideration
one of the most important elements of the problem. This is a
criticism from a general point of view; but, as it serves to
make clear the standpoint from which I propose to consider the
question at issue, it comes properly enough before the considera-
tion of the above-named objections, to which we may now
pass.

There is nothing in the first objection which has not been
anticipated and answered in a general way in the writings of
His and Rauber, and I need not here repeat their statements.
It may be worth while, however, to call attention to another
way of meeting the supposed difficulty. It is now quite clear
that the primitive groove and the medullary groove must not
be confounded, and it is equally clear that the concrescence
theory must be able to account for them both in every instance,
in order to maintain itself. But is it possible to keep up the
distinction between them in all cases? What, for example,
can be called the primitive groove of the Hlasmobranch embryo ?
It has been said that here there is no such structure, and by
some the medullary groove has been called primitive groove.
There is some danger of confusion on this point, and I believe
this confusion underlies the objection we are here considering.
If we break loose from all mental pictures suggested by the
word groove, and adhere strictly to the definition of the primi-
tive groove as the plane of junction of the lips of the
blastopore—a definition to which most embryologists will cer-
tainly assent—then it follows, on the concrescence theory, that
something of this nature must be recognised in the Elasmo-
branch embryo. Looking at the question from this standpoint,
it is plain that the medullary groove is not identical with the

386 Cc. O. WHITMAN.

primitive groove. The primitive groove is simply ‘‘a phe-
nomenon of conjunction,” as Rauber terms it; but it differs
from the medullary groove in not being confined to a single
embryonic layer. Both fall in the same median plane, but one
is primary, the other secondary ; one divides the entire embryo
into two homo-typical halves, the other only marks the middle
line of the neural plate. These are, of course, no new facts;
but I have been obliged to state them in order to make clear
my meaning. Now, if we bear in mind that the primitive
groove is only a seam that marks the incomplete coalescence of
two germ bands or halves of the germinal ring, and that the
medullary groove simply marks the axial relations of two folds
on the surface of these bands, we may avoid the confusion that
is sure to arise when this distinction is not kept in view. Re-
garding this distinction as the only one of fundamental import-
ance, I hold that all disputes in regard to the presence or
absence of a primitive groove do not affect the main question
of concrescence. The plane of junction may be marked bya
narrow surface groove or only a faint line, or the coalescence
may be so complete that no plain indication of its position can
be detected. In all Amniota the two grooves are not contem-
poraneous in origin, and one is obliterated as the other takes
its place ; but in Elasmobranchii they are coincident, a circum-
stance that does not alter the fundamental distinction before
insisted upon.

The aim of these remarks, if it need be stated, has been to
show that the manner in which the medullary folds close is
not of primary importance in deciding between the theories of
differentiation and concrescence. It is possible, and perhaps
not improbable, that the formation and closure of the medullary
folds are determined in part by the same causes that concur in
bringing together the two halves of the germinal ring ; but the
two closures are nevertheless quite distinct, both from a mor-
phological and physiological stand-point. The concrescence
theory undertakes to account for the conversion of the germinal
ring into a bilateral embryo; but it is not within its scope to
explain peculiarities in the closure of the medullary folds, since

RARE FORM OF THE BLASTODERM OF THE CHICK. 387

this is only a special closure which follows the main or somatic
closure.

Balfour’s first objection is open to criticism from still another
point, since it is based on a feature which he himself has else-
where declared to be exceptional. In his ‘ Monograph on the
Development of the Elasmobranch Fishes,’ p. 84, he says :—
“The only feature in any respect peculiar to these fishes is the
closing of the medullary canal, first commencing behind, and
then at a second point in the cervical region. In those verte-
brates in which the medullary folds do not unite at approxi-
mately the same time throughout their length, they appear
usually to do so first in the region of the neck.”

The introductory phases of concrescence in the Elasmo-
branch are well shown in figs. 1, 2, and 3, copied after His ;
and they appear to me to meet the first objection with refer-
ence to the angle of coalescence.

The second objection claims that concrescence subsequent to
the establishment of the neurenteric canal would be a simple
impossibility. At first sight this appears to be a fatal argu-
ment against the concrescence theory. Since the general im-
portance of this canal was first insisted on by Kowalevsky, a
similar structure has been discovered in the Birds by Gasser,
and in the Reptiles by Kupffer and Benecke ; and indistinct
traces of itin Mammalia have been reported by Lieberkuhn.
There is still much difference of opinion, not only in regard to
the meaning of this structure and its relation to the allantois
and archenteron, but also in respect to its position. Balfour
places it just in front of the primitive streak; Kupffer and
Benecke contend that it is situated at the posterior end of the
streak; Strahl says it arises near the middle; and finally,
Braun finds several independent canals that appear one after
the other. Strahl’s recent paper (No. 47) is the only one that
offers any assistance in meeting Balfour’s argument. In
Lacerta agilis Strahl finds a small inconspicuous primitive
streak, at the centre of which the neurenteric canal first makes
its appearance. But the point of chief interest here is that
this canal, according to Strahl, travels slowly backs;

388 C. O. WHITMAN.

wards, the fore part closing as the hinder part opens
further back, until ultimately the canal is found in
the extreme hind end of the tail.

That part of the primitive streak that lies originally behind
the canal does not atrophy, but forms the tail and allantois;
while that part that lies before this point is employed in form-
ing the medullary tube, the chorda, and the alimentary canal.
If Strahl’s statements in respect to the change of position of
the neurenteric canal are correct, concrescence beyond its point
of origin is not an impossibility. It is quite possible that this
movement of the canal may explain, what would otherwise be
difficult to understand, the occurrence of several apparently
independent canals reported first by Braun.

The chief difficulty in the case of Amphioxus has already
been considered by Rauber in two papers that appeared in
1877 (Nos. 39, 41), and I do not readily understand why
Balfour should bring up the same point several years later
without alluding to its previous notice. In considering this
objection it must be borne in mind that Amphioxus is no
longer entitled to the rank of ‘ Urwirbelthier,” but is rather
to be classed among the prodigal sons of the vertebrates, to
use a metaphor of Prof. Dohrn. Why, then, should “clear
evidence ”’ be expected from this source rather than from more
respectable types that have not lost their senses by hiding in the
sand? If degenerative simplicity of structure warrants such
expectation it would hardly appear ridiculous to appeal to our
more “ degenerate cousins,” as Lankester calls the Ascidians.
The mere fact that the blastopore nearly closes before the ap-
pearance of the medullary groove cannot be accepted as proof
that the blastopore is not coextensive with the groove, as
Rauber has plainly shown. Rauber’s remarks on this subject
are substantially as follows :—In order to understand the case
of Amphioxus we must not start with the last stage of the
blastopore, when it has been reduced almost to minimum di-
mensions, but with the highest stage of its existence, when it
has its widest expanse. It is about this time that it appears to
lie on the dorsal side of the earlier equatorial line of the egg.

RARE FORM OF THE BLASTODERM OF THE CHICK. 389

Its diameter gradually diminishes as the Gastrula assumes a
more spherical form. Then the double-walled sac elongates in
the direction of the axis of the blastopore, and becomes flat-
tened on the dorsal side, at the rear end of which is seen the
remnant of the blastopore. The lateral edges of the flattened
dorsal surface rise up as medullary folds, which enclose at their
posterior ends the blastoporic remnant. Considering the matter
on the basis that the blastopore has closed up gradually on the
dorsal side, it appears not at all improbable that we have here a
conjunctive form of embryo formation. The investigations of
Kowalevsky thus interpreted in harmony with the concrescence
theory enabled Rauber to anticipate what has since been veri-
fied by observation. Hatschek,! who has paid special attention
to the closure of the blastopore, says, “‘ I came to the conclu-
sion that the original broad Gastrula mouth belongs entirely to
the later dorsal region, and that one point in its rim marks the
hind ‘end of the body.”” And again, “The closing of the
Gastrula mouth begins at its anterior edge, the
hind edge remaining unchanged throughout. The
concrescence of the edges takes place in a line that
forms the later dorsal line. The hindmost remnant
of the Gastrula mouth remains then for some time
as a small dorsal opening at the hind end of the
dorsal surface.”

The reduction to absurdity claimed in the fourth objection
appears to be based on inadmissible premises ; and as soon as
these are stripped of exaggeration the absurdity vanishes.

To say that, according to the view of His and Rauber, the
line of concrescence must be coextensive with the whole of
the dorsal as well as the whole of the ventral wall of the em-
bryo is inaccurate in both particulars. The inaccuracy with
respect to the dorsal wall is comparatively unimportant, since
only the foremost end of the Elasmobranch embryo undergoes
no concrescence, but with respect to the ventral wall the
statement is plainly hyperbolical, and herein lies the whole

1 Hatschek, ‘Studien ii. Entw. d. Amphioxus,” Claus’ ‘ Arbeiten,’ vol. iv,
part. i, pp. 28, 31, 1881.

390 Cd. O. WHITMAN

absurdity. Now, there is no difference of opinion as to the
fact that the final closing in of the yolk takes place “ at some
little distance behind the embryo” in the case of the Elasmo-
branch. A glance at Balfour’s fig. 30 B (‘ Compar. Embryol.,’
p. 52) will show the extent of the line of concrescence. It
will be seen, at the utmost, this line can only be said to
extend from the region of the head along the dorsal side to the
tip of the tail, from this point forward on the ventral side
to the umbilical stalk, and from this stalk backward along
‘* the linear streak ’’ which connects the embryo with the edge
of the blastoderm. The closure of the space (yk) will com-
plete the line of concrescence. It will be seen that there is
an important ventral portion of the body, extending from the
umbilical cord to the head, which is in no sense of the word
formed by concrescence. If we take into consideration the
entire yolk, as Balfour certainly does, it is plain that the line
of concrescence is considerably less than half of the entire cir-
cumference. When we remember that among our invertebrate
vermian relatives the cases are not at all rare in which this
line is much more than half the circumference of the egg, we
must admit that nature delights in just such absurdities as
Balfour has pointed out.

The fifth objection presents the same difficulty that we meet
with in the development of the Bird, where, owing to the pre-
sence of an enormous quantity of food-yolk, the process of con-
crescence has undergone such extreme modification that the
only constant outward manifestation of it is the primitive
groove. But the connection of the primitive streak with the
area opaca, the relations of thickness between different parts
of the blastoderm at successive stages, the occasional appear-
ance of a marginal notch, the extension of the primitive
groove to the notch in rare cases, the occurrence of a neuren-
teric canal, and other considerations of a comparative embryo-
logical nature, appear to me to outweigh the objection, and
compel us to recognise here the same principle of embryonic
formation that characterises the more primitive forms of the
vertebrate stock.

RARE FORM OF THE BLASTODERM OF THE CHICK. 391

Rauber has shown that the phenomena of double monsters
may be explained in perfect harmony with the theory of con-
crescence. A single illustration must here suffice to show the
application of the theory to such cases.

Figs. 10 and 11 are diagrammatic representations of double
formations of the Osseous Fish. In Fig. 10 are seen two inde-
pendent formations (4 and B), separated by a small portion of
the germinal ring (i). Further development leads to the con-
dition seen in fig. 11. The concrescence of the lateral halves
of the germ-ring carries A and B forward, and brings the por-
tions e e together, thus producing a monster that is double in
front and single behind. The intermediate portion (i) now
forms parts of the mesial sides of a and B.

The forms here diagrammatically represented have been
selected with a view to show that the Vertebrates and Annelids
exhibit the same ring-type of development.

Figs. J—3 = early stages of the Elasmobranch (after
His). The manner in which the two halves of the ring are
brought together to form the embryo is well illustrated in
these figures.

Figs. 4—6 = Salmo (after His). Here the line of concres-
cence does not reach to the hind edge of the ring, the two
marginal lobes (m 1) seen in figs. 1—3 being here represented
by a single median lobe. Although the process of concres-
cence is thus partially disguised, it becomes evident enough by
comparing successive stages as represented in fig. 6.

Fig. 7 = normal form of the blastoderm of the Chick.

Fig. 8 = atavistic form of the blastoderm of the Chick, dif-
fering from the normal form only in the extension of the line
of concrescence to the marginal notch (m n).

Fig. 9 = Clepsine, showing that in a case of admitted con-
crescence the relations of the embryo to the germ-ring are
completely analogous to those seen in the blastoderm of the Fish
or the Bird.

Figs. 10, 11 = double formation in the Osseous Fish (after
Rauber).

Fig. 12 = ring-stage of Euaxes (after Kowalevsky). .The

392 C. O. WHITMAN.

two halves of the ring coalesce in the same manner as in
Clepsine.

Let us now consider the two theories of embryo-formation in
amore general way. The theory of concrescence undertakes
to give a rational explanation of a very large body of pheno-
mena; while that of differentiation cannot even make the
slightest pretension to anything of the kind. What is differ-
entiation except an undefined mode of growth and develop-
ment? It is not a mechanical explanation, but simply a vague
name for unknown processes. This theory, even when de-
fended by the sagacity of Balfour, has never offered a sugges-
tion by way of explaining the germinal ring, and the uniform
relations which this ring sustains to the embryo. It cannot
tell us why all the so-called “ phenomena of concrescence” are
situated behind the embryo, instead of before it; nor can it
give any explanation of the general nature of these phenomena,
except when it goes a begging by calling in the aid of the con-
crescence theory. On the other hand, the theory of concre-
scence makes known a law of formation which may, with
reason, be said to hold good throughout the vertebrate group,
and their nearest invertebrate allies. So far as these allies are
concerned, the law is an established fact, about which there is
not, and cannot be, any controversy; and among the lower
vertebrates the appearances are unquestionably in its favour.
Assuming that the immediate ancestor of the Vertebrata was a
segmented worm, it is evident that our theory of the formation
‘of the embryo should include both groups of animals. The
fundamental agreement, which we should naturally expect to
find, appears first of all in the formation of a germinal
ring, composed of two symmetrical halves, which
coalesce along the median neural line from before
backward, thus producing a bilateral embryo. So
far there is essential agreement in the two great classes of
segmented worms, the Cheetopods and Leeches.

I thiik no exception to this statement can be justly taken on the ground

that the germ-bands in some cases contain neuroblastic as well as mesoblastic
elements. But as Balfour has laid considerable stress on this very point, and

RARE FORM OF THE BLASTODERM OF THE CHICK. 393

has even gone so far as to doubt the accuracy of my observations on Clepsine,
on the strength of theoretical convictions, it seems necessary to give it a
moment’s consideration. Balfour remarks: ‘Till more evidence is brought
forward by Whitman or some other observer in support of the view that the
so-called neuroblasts have any share in forming the nervous system, they
must, in my opinion, be regarded as probably forming, in conjunction with the
mesoblasts, two simple mesoblastic bands. Kowalevsky has, moreover, briefly
stated that he has satisfied himself that the nervous system in Clepsine origi-
nates from the epiblast—a statement which certainly could not be brought into
harmony with Whitman’s account” (No. 8, vol. i, p. 289). With reference
to Clepsine, Kowalevsky remarks: “1 preserved only several stages in weak
chromic acid, and from sections of these 1 could only convince myself later of
the origin of the nervous system from the upper layer.” ! This is all he has
said on this point; and I will now show that, if we do not go behind the
verbal statement itself, it does not even require to be brought into harmony
with my account, since it is precisely what I have claimed. The four rows of
neuroblasts in each germ-band lie, at the outset, at the surface, and must
therefore be considered a part of the epiblast, although a specialised part. It
is simply a precocious differentiation of the edge of the epiblast, by which epi-
dermal and neural elements become distinctly marked at an unusually early
stage. In the course of the epibolic growth of the ectoderm the epidermal
portion progresses somewhat more rapidly towards the lower pole than the
géerm-bands, and thus sweeps over the neural portion. But it seems to me
plainly a matter of little importance whether the neural portion looses its sur-
face position during the epiboly, or immediately after the concluson of the
concrescence of the germ-bands; and 1 confess that I do not see wherein this
view requires ‘‘ any special support.” At the time Balfour penned the above
criticism, he evidently was not aware that my observations on the origin of the
nervous system in Clepsine were but little more than a corroboration of those
of an eminent Russian embryologist. I take this opportunity to express my
regret for the same oversight on my own part. It was Professor Metschni-
koff who first determined the precise origin of the nervous system of Clepsine.?
The accord between Metschnikoff and myself extends not only to the facts
observed, but also to the interpretation of these facts. The chief distinction
between Clepsine and other Articulata and Vertebrata with respect to the germ-
lamelle lies, according to Metscbnikoff, in the single fact that “the epidermal
layer separates very early from the basis of the nerve system.” The associa-
tion of neuroblastic with mesoblastic elements in the germ-bands is then a
feature which presents no serious difficulty to the comparison before instituted
between Cheetopods and Leeches.

' Kowalevsky, ‘Embryol. Stud. an Wiirmern u. Arthropoden,’ p. 8.
2 Metschnikoff, ‘ Beitrage zur Entw’gesch, einiger nied. Thiere.’ Mélanges
Biologiques du Bull. de ]’Acad. imp. des. Sci. St. Petersbourg,’ vii, 6, 1871.

394 C. O. WHITMAN.

The second great feature common to the embryos of both
Annelids and Vertebrates is the metameric division of the
body. In the Annelids the division into somites follows closely
in the wake of concrescence, and the two processes are funda-
mental and invariable in sequence. Now, when we see that
the vertebrate embryo arises from a germinal ring, and that
soon after the formation of the embryo begins, the metameric
division sets in and progresses in the same direction as in the
Annelids, there is certainly good ground for inferring that the
phenomena are fundamentally the same in both groups of ani-
mals. Since, in the one case, concrescence and metameric
division are conceded by all to be the two grand formative
processes associated in invariable sequence, it is difficult to be-
lieve that, in the other case, the second process has been pre-
served, while the first in order, and perhaps in importance, has
been entirely suppressed.

The law set up by the elder Milne-Edwards, according to
which fresh somites are intercalated between the last formed
somite and the hind end of the body, does not appear to me to
hold good in the case of embryos that arise by concrescence.
In Clepsine it is perfectly clear that the hindmost segment is
the last formed segment ; and the investigations of Kowalevsky,
Hatschek, and others appear to demonstrate the same for the
Chetopods. The germ-bands lengthen at the expense of pro-
liferating cells; and the additions are made to the hind extre-
mities, so that these are always the youngest portions of the
embryo. The theory of intercalation can therefore be upheld
only by supposing that the proliferating cells represent the
hind end of the body, which they do only prospectively. But
there is still another objection to this theory, which was pointed
out by Kélliker.t In Clepsine the somites are not added one
by one as fast as the material is furnished by the proliferating
cells. The metameric division begins only after the materia.
for the greater number of somites is already present, so that
no one of the successively formed somites, except the ultimate,
can be regarded as the youngest portion of the embryo.

1 Kolliker, ‘Observationes de prima insectorum genesi.’ Ziirich, 1842,

RARE FORM OF THE BLASTODERM OF THE GHIOR. 395

If the theory of concrescence is applicable to the vertebrates,
it is evident that we cannot regard the two ends of the embryo
as equally old portions, which are gradually pushed in opposite
directions by the interpolation of fresh somites successively
budded off from the hind end. The discovery of two “ Pol-
zelien des Mesoderms” at the extreme hind end of the embryo
of Amphioxus by Hatschek, supports the opinion that the
lengthening and the metameric division of the vertebrate em-
bryo take place in fundamentally the same manner as in the
Annelids.

The error of the differentiation theory does not lie in the
assumption of intussusceptional growth, but in excluding the
concomitant process of concrescence. The formation of the
vertebrate embryo does not afford such conspicuous evidences
of concrescence as that of the Annelid embryo; but even in the
more doubtful cases of the Osseous Fish and the Bird, both His
(No. 22) and Kupffer (No. 31) have shown that there is a
general movement of the embryoplastic material which must,
in my opinion, be regarded as a disguised form of concrescence.
No other explanation of the phenomena can claim to bring
them into relation with that form of development still pre-
served in the nearest invertebrate relatives of the vertebrates.

According to the concrescence theory, it will not do to
regard the primitive streak as analogous to the “ linear streak”
behind the Elasmobranch embryo, as has been done by Balfour.
The primitive streak lies within the germinal ring, and is
continuous with it at the hind border; the “linear streak’’ is
located much further back, being entirely behind the embryo-
plastic portion of the ring. Another important difference, not
alluded to by Balfour, consists in the fact that the primitive
streak arises even before the medullary folds; while the
‘linear streak” appears after the embryo is formed, and even
after the process of constriction has narrowed the connection
of the embryo with the yolk to a slender cord, the umbilical
stalk. Nothing analogous to this “linear streak’’ appears in
the normal blastoderm of the Chick. In the exceptional form
of the blastoderm which I have described, the streak connect-

396 Cc. O. WHITMAN.

ing the primitive groove with the marginal notch may be con-
sidered analogous to the “ linear streak” of the Elasmobranch,
while the marginal notch corresponds to the anterior angle of
the “ yolk blastopore,”” marked yf in Balfour’s figure.

LITERATURE ON THE FORMATION AND SIGNIFICANCE OF THE
PRIMITIVE STREAK.

L. Acassiz.—(1) ‘Contributions to the Natural History of the United
States,’ vol. ii, part ili, Boston, 1857. ‘‘ Embryology of the Turtle.”

Batrour.—(2) ‘The Development and Growth of the Layers of the Blasto-
derm,” and

(3) “On the Disappearance of the Primitive Groove in the Embryo
Chick,” ‘ Quart. Journ. of Mie. Sci.,’ vol. xiii, 1878.

(4) ‘Elements of Embryology,’ part. i, London, 1874. By Foster
and Balfour.

(5) “A Comparison of the Karly Stages in the Development of Verte-
brates,” ‘Quart. J. Mic. Sci.,’ vol. xv, 1875.

(6) “A Monograph on the Development of Elasmobranch Fishes,’
London, 1878. Reprinted from the ‘Journal of Anat. and Phys.’ for
1876, 1877, and 1878.

(7) “On the Early Development of the Lacertilia, &.,” ‘Quart. J.
Mic. Sci.,’ vol. xix, 1879. i

(8) “*A Renewed Study of the Germinal Layers of the Chick,”
‘Quart. J. Mic. Sci.,’ vol. xxii, 1882. By Balfour and Deighton.

(9) ‘Comparative Embryology,’ vol. ii, London, 1881.

Braun.—(10) “ Aus der Entwickelungsgeschichte der Papageien,” ‘ Verh. d.
phys.-med. Gesellschaft zu Wiirzburg,’ N. F., Bd. xiv and xv, 1879 and
1880.

(11) ‘‘ Die Entwickelung des Wellenpapageies (Melopsittacus undu-
latus Sh.) mit Beriicksichtigung der Entwickelung anderer Vogel,”
Wiirzburg, 1881.

(12) “ Entwickelungsvorginge am Schwanzende bei einigen Siiuge-
thieren, &c.,” ‘Arch. f. Anat. u. Phys.,’ anat. Abth., H. 4, 5, 6, 1882.

Dursy.—(13) ‘ Der Primitivstreif des Huhnchens.’ 1867.

GassER.—(14) ‘‘ Der Primitivstreifen bei Vogelembryonen,” ‘Schriften d.
Gesell. zu Beford. d. gesammten Naturwiss. zu Marburg.,’ vol. ii, supple-
ment i, 1879.

(15) “Die Entstehung der Cloakenéffnung bei Huhnerembryonen,”
‘Arch. f. Anat. u. Phys.,’ anat. Abth., 1880.

(16) “ Beitrage zur Kenntniss der Vogelkeimscheibe,” ‘ Arch. f. Anat,
u. Phys.,’ anat. Abth., H. 4—6, 1882.

RARE FORM OF THE BLASTODERM OF THE CHICK. 397

GreRLAcH.—(17) “ Ueber d. entodermale Entstehungsweise d. Chorda dorsalis,”
‘ Biol. Centralblatt,’ vol. i, Nos. 1 and 2, 1881.

Gortre.—(18) “ Beitrage zur Entwicklunesgeschichte d. Wirbelthiere. II.
Die Bildung d. Keimblatter u.d. Blutes im Hithnerei,” ‘ Arch. f. mik.
Anat.,’ vol. x, 1874.

His.—(19) “ Untersuchungen ueber die Entwickelung von Knochenfischen
besonders ueber diejenige des Salmens,” ‘ Zeitschr. f. Anat. u. Entwick-
lungsgesch,’ vol. i, H. 1 and 2, 1875.

(20) ‘Untersuchungen iiber die Bildung des Knockenfischembryo
(Salmen),” ‘Arch. f. Auat. u. Phys.,’ anat. Abth., 1878.

(21) “ Ueber die Bildung der Haifischembryonen,” ‘ Zeitschr. f. Anat.
u. Entw.,’ vol. ii, H. 1 and 2, 1876.

(22) “Neue Untersuchungen iiber die Bildung des Hiihnerembryo,”
‘Arch. f. Anat. u. Phys.,’ anat. Abth., H. 2 and 3, 1877.

KOLLIKER.—(23) ‘ Entwicklungsgeschichte d. Menschen u. d. hoheren
Thiere,’ Leipzig, 1879.

Kotter.—(24) ‘ Beitrage zur Kenntniss d. Hiihnerkeims im Beginn der
Bebriitung,”’ ‘ Sitz. d. k. Akad. Wiss.,’ iv Abth., 1879.

(25) ‘Untersuchungen ueber die Blitterbildung im Hihnerkeim,”
* Arch. f. mikr. Anat.,’ vol. xx, 1881.

KowAaLEvsky, OwsJANNIKOFF, AND WaGNER.—(26) “ Die Entw. d. Store,”
Vorlauf. Mittheilung, Mélanges, ‘ Biologiques tirés du Bull. d. Acad.
Imp. St. Pétersbourg,’ vol. vii, 1870.

Kuerrrek AND Brenecke.—(27) ‘ Die ersten Entwicklungsvorginge am Li
der Reptilien,’ Konigsberg, 1878.

(28) “ Photogramme zur Ontogenie d. Vogel,” ‘ Nov. Act. d. k. Leop.-
Carol.-Deutschen Akad. d. Naturforscher,’ vol. xli, 1879.

Kurrrer.—(29) “ Die Entstehung d. Allantois u. d. Gastrula d. Wirbelthiere,”
‘Zoolog. Anzeiger,’ vol. ii, 1879, pp. 520, 598, 612.

(30) ‘‘ Die Gastrulation an den meroblastischen Hiern der Wirbelthiere
und die Bedeutung des Primitivstreifs,” ‘ Arch. f. Anat. u. Phys.,’ anat.
Abth., 1882, pp. 1, 189. Not yet concluded.

(31) “ Die Entw. des Herings im Hi,” ‘ Jahresb. d. Kommission z. wiss.
Untersuchung d. deutschen Meere,” Berlin, 1878.

LieBerktun.—(82) “Ueber die Chorda bei Siugethieren,” * Arch. f. Anat.
u. Phys.,’ anat. Abth., 1882, p. 439.

Oxrtiacner.—(33) “ Beitrage zur Kutwicklungsgeschichte der Knochenfische
nach Beobachtungen am Bachfozellenei,” ‘ Zeitschr. f. wiss. Zool.,’ vol.
xxii, 1872, and xxiii, 1873.

PanpER.—(34) “Beitrage zur Entw. d. Hihnchens im Hie,” Wiirzburg,
1817.

VOL, XXIII.—NEW SER, DD

398 Cc. O. WHITMAN.

RavuBper.—(35) ‘“ Ueber die Embryonalanlage des Huhnchens,” ‘ Centralblatt
f, d. medic, Wissenschaften,’ 1874-75.
(36) “ Ueber die erste Entw. d. Vogel und die Bedeutung der Primi-
tivrinne,” ‘ Sitzungsb. d. naturf. Ges. zu Leipzig,’ 1876, Jan. bis Marz.
(37) “ Ueber d. Stellung d. Hiihnchens im Entwicklungsplan,” Leipzig,
1876.
(38) “ Primitivrinne u. Urmund,” ‘ Morph. Jahrb.,’ vol. ii, 1876.
(39) ‘“Nervencentra der Gliederthiere und Wirbelthiere,” ‘ Leipz.
Sitzungsberichte,’ Jan., 1877.
(40) “Die Theorien der excessiven Monstra,” ‘ Virchow’s Arch, f,
pathol. Anat.,’ vol. 71, Oct., 1877.
(41) ‘ Primitivstreifen und Neurula der Wirbelthiere,’ Leipzig, 1877.
(42) “Die Lage der Keimpforte,” ‘ Zool. Anzeiger,’ No. 38, p. 499,
1879.
(43) ‘* Die Gastrula d. Wirbelthiere und die Allantois,” ‘ Zool. Anzei-
ger, No. 53, p. 180, 1880.
Scnirer.—(44) ‘A Contribution to the History of Development of the
Guinea-pig,” ‘ Jour. of Anat. and Phys.,’ vol. x and xi, 1876-7.

SemPeR.—(45) Verwandtschaft d. geglied. Thiere,” ‘ Arbeiten,’ iii, 1876-7.

Srrant.—(46) “Ueber die Entw. des Canalis myelo-entericus und der
Allantois der Hidechse,” ‘ Arch, f. Anat. u, Phys.,’ anat. Abth., p. 122,
1881,

(47) “ Beitrage zur Entwickelung von Lacerta agilis,” ‘ Arch. f. Anat,
u. Phys.,’ anat. Abth., p. 242, 1882.

Wapryer.—(48) ‘“ Ueber die Keimblatter und den Primitivstreifen bei der

Entwicklung des Hiihnerembryo,” ‘ Zeitschr. f. rationelle Medicin,’ 1869.

DEVELOPMENT OF PELVIC GIRDLE IN THE CHICK. 399

On the Development of the Pelvic Girdle and
Skeleton of the Hind Limb in the Chick.

By

Alice Johnson,
Newnham College, Cambridge.

With Plates XXVI & XXVII.

a

THE investigations described below were undertaken at the
suggestion of the late Professor Balfour, with a view to finding
out, through a study of the development of the pelvic girdle in
the chick, what are the homologies of the pubis in birds with

that of other Vertebrata.
In connection with this question one or two other points,

which appear to me of some importance, have presented
themselves.

The histological development may be briefly considered at
the outset. On the fourth day of incubation the limb is merely
a local exaggeration of the Wolffian ridge, consisting, like it, of
a mass of rounded mesoblastic cells, very closely aggregated to-
gether. The epiblast forms a thickened cap round the free end
of the limb. No differentiation into cartilage or muscle is yet
visible.

The first trace of the skeletal parts appears on the fifth day.
The mesoblastic tissue of the limb is now differentiated into an
axial, or more condensed, and surrounding, or less condensed,
region. Both parts consist of the same rounded cells as before.
They only differ in the degree of concentration of the cells.
These features are shown diagrammatically in fig. 1.

The differentiation of tissue goes on more rapidly in the

400 ALICE JOHNSON.

skeleton of the limb than in the girdle, and more rapidly in the
axial than in the superficial regions of both skeletons. Its main
features are almost the same as those described by Strasser! in
the developing cartilage of the newt. On the sixth day, or
thereabouts, the cells begin to be compressed in the direction of
the long axis of the cartilages. This happens especially in the
tibia and fibula. Dark irregularly-shaped masses—the “ pro-
chondral elements” of Strasser—appear among the cells. They
are apparently derived from the metamorphosed cells, for one
occasionally meets with forms that appear intermediate, in
which the protoplasm has become opaque and stains deeply,
while the nucleus is still visible. I take the prochondral ele-
ments to be cells which have retrograded still further and lost
their nuclei.

Rather later, on the sixth or seventh day, the prochondral
elements have almost disappeared from the central part of the
cartilage. Their place is taken by a homogeneous, slightly-
staining matrix, by means of which the cells gradually become
widely separated from one another. Still later the cells take
on the crescent shape of adult cartilage cells.

Morphology.—Since chicks of the same day vary so much
in their degree of development I have taken the length of the
hind limb as the standard of their age. The following table
shows roughly to what number of days of incubations these
lengths correspond :

Length of hind limb. Number of days of incubation.
0°06 in.—0°1 in. 5—6
0°12 in.—0°2 in. 6—7
0:2 in.—0°25 in. 7—8
0°25 in.—0°3 in. 8—9
0°5 in.—0°8 in. 9—10
1:5 in.—3s° in. 14—20

The Pelvic Girdle on its first appearance (length of hind limb
0:06 in.—see fig. 1) is seen in transverse sections to form one
mass with the skeleton of the limb. It consists of two slight

1 H. Strasser, “ Zur Entwicklung der Extremititenknochen bei Salamandern
und Tritonen,” ‘ Morph. Jahrbuch,’ Band vy, 1879.

DEVELOPMENT OF PELVIC GIRDLE IN THE CHICK. 401

outgrowths of the proximal part of the femur, one being directed
upwards, the other inwards, each, however, hardly extending
beyond the limb itself. The future cartilage is only just dis-
tinguishable from its surroundings of indifferent mesoblastic
cells, since the two tissues pass quite gradually into one
another. .

The next stage (length of hind limb 0:12 in.) is seen in
longitudinal section in fig.2. The series of sections shows the
same perfect continuity of the girdle and femur that existed at
first. Wecan distinguish in the girdle a blunt dorsal prolonga-
tion—the beginning of the ilium—an acetabular region behind
the obturator nerve and a downward process in front of it,
which is obviously to become the pubis. As we go inwards in
the series of sections these two outgrowths, the ilium and the
pubis, disappear, and the central or acetabular region is pro-
longed a little way inwards, being bounded in front by the ob-
turator nerve. The nerve does not appear in the same sections
with the pubis and ilium, but in the figure it is represented as
viewed from the outside, the girdle being supposed to be trans-
parent. At this stage the nerves are remarkable for their large
size in proportion to the skeletal parts. ‘The obturator nerve
coming off from the crural plexus is at this time by far the most
important of its distal branches.

In the next stage (length of hind limb 0°15 in.—see fig. 3)
we can clearly distinguish three elements in the girdle, meeting
in the broad acetabular region, which passes on without a break
into the femur. The region of its junction with the latter is
shown diagrammatically in the figure, but the cartilage of the
femur is continuous with that of the girdle, as are the three
elements of the girdle with one another. The ilium has grown
forwards, arching over the crural nerve, and has given off a
slenderer pointed process backwards. The ischium is directed
almost vertically downwards, but also slightly inwards, being,
as a whole, situated nearer to the middle line of the body than
are the other elements. The main point of interest is the double
nature of the pubis, the anterior branch of which points directly
forwards and slightly outwards, while the posterior is directed

402 ALICE JOHNSON.

downwards and slightly forwards. The obturator nerve passes
between the posterior branch of the pubis and the ischium. A
series of twelve longitudinal sections has been combined to pro-
duce the figure, which is therefore diagrammatic only in so
far as it represents as existing in two dimensions what really
exists in three. The other figures, in which the whole girdle
is represented, were drawn in the same way.

The study of the stages described above shows that the early
development of the Pelvic Girdle of the Chick is similar to
that of the limb-girdles of Elasmobranchs ! in two points: (1)
the skeleton of the limb is developed continuously with the
girdle; (2) the parts of the girdle which are in the immediate
neighbourhood of the skeleton of the limb are first developed,
and the dorsal and ventral outgrowths appear later.

In the next stage (length of hind limb 0°17 in,, see fig. 4)
the posterior branch of the pubis has grown more than the
anterior, and is curved backwards. Its proximal half, how-
ever, retains the direction which the whole posterior branch
had in the earlier stage, and from this we may conclude that
the change of form results from a growth, rather than from a
rotation backwards of the whole cartilage.

A transverse section (see fig. 5) of about the same stage
shows that the girdle is still continuous with the femur. In
the latter, the cartilaginous matrix has begun to be formed
internally, while the peripheral parts (a region of which is cut
through in the middle of the limb) and the girdle still consists
of the condensed tissue described above.

A later stage is shown in fig. 6 (length of hind limb 0°2 in.).
The most striking feature here compared with the preceding
stage is the large development of the posterior part of the
ilium. The ischium has become distally expanded, and the
posterior branch of the pubis is larger still in proportion to the
anterior branch.

About this time the femur begins to be separated from the
girdle by an intervening tract of tissue which has not gone so

1 F, M. Balfour, ‘On the development of the skeleton of the paired fins of
Elasmobranchii,” ‘ Proc. of Zoological Society,’ 1881.

DEVELOPMENT OF PELVIC GIRDLE IN THE CHICK. 403

far on the way to becoming adult cartilage. At first the whole
structure progresses uniformly, except that the girdle always
lags a little behind the femur, but passes off gradually into it.
The “ prochondral elements ” and a small quantity of cartila-
ginous matrix exist across the future line of division, which,
however, develops no further, but retrogrades into the fibrous
tissue of the joint.

Fig. 7 represents a further advance of the girdle towards the
adult form.

In later stages, no important changes take place. The
anterior branch of the pubis, which is always rather behind
the rest of the girdle in histological development, becomes
more and more proportionately insignificant, and forms at last
the pectineal process of the pubis. The posterior branch of
the pubis becomes very slender. Both it and the ischium
grow more and more backwards, passing through the stage
permanent in such forms as Apteryx (where they are much
curved, and their long axes are inclined at an acute angle to
the long axis of the ilium) to the stage found in the adult fowl,
where the pubis and ischium—except the most proximal por-
tions of them—are straight, and point directly backwards, so
that the long axes of all three bones are parallel to one another.

Ossification begins comparatively late, i.e. later than in the
limb. For a long time there is a cartilaginous continuity of
the three elements round the acetabulum. The bones gradu-
ally grow up to and surround the acetabulum. Cartilage
remains also at the free ends of the bones for a long time. A
day or two before hatching (see fig. 18), the acetabulum is
surrounded by bone, except for a small region of its front wall,
continuous with the likewise cartilaginous anterior branch of
the pubis. The position and relations of this latter element,
together with the fact of its remaining cartilaginous so long,
remind one to some extent of the cartilage found in a similar
situation in the Crocodile embryo after the rest of the girdle
has ossified. According to Hoffmann,! this cartilage is homo-

1 C. K. Hoffmann, “ Beitriige zur Kenntniss des Beckens der Amphibien
und Reptilien,” ‘Nied. Archiv f. Zoologie,’ Band iii, 1876.

=

AOA - ALIGE JOHNSON.

logous with the Pubis, while he calls the bone generally
known as the pubis the epi-pubis. But since the acetabular
regions of each bone always remain cartilaginous longer than
the other parts, and since this cartilage is replaced in the
adult by a bony process of the Ischium shutting out the pubis
(Epi-pubis of Hoffman) from the acetabulum, I should be
more inclined to agree with the older view that Hoffmann’s
pubis is merely a part of the ischium. This seems to me quite
consistent with his own account of the ossification. He says:
—(loc. cit. p. 186) ‘‘ Die Verknécherung dieses vorderen Ace-
tabularfortsatzes des Sitzbeines fangt zuerst an der dem Sitz-
bein angrenzenden Partie an und schreitet so allmahlig dem
vorderen Fortsatz des Iliums zu, erreicht diesen aber erst bei
ganz ausgewachsenen alten Thieren.” The fact that the
pubis is moveable in the crocodile is quite sufficient to account
for its being shut out from the acetabulum.

So far as I know, the only literature bearing directly on the
subject of the development of the pelvic girdle in birds is a
paper by Bunge.’ According to him, the pubis and ischium
are at first situated with their long axes in a position vertical
to the vertebral column, and later become rotated backwards,
thus taking on the adult form. This statement has been
generally accepted, but I am unable to agree with Bunge’s
other conclusions. He has omitted to mention the primary
continuity of the femur and girdle and the existence in the
embryo of an anterior branch of the pubis which becomes the
pectineal process. Speaking of the pectineal process in the
adult, he only says that his account of the development proves
that it is a part of the ilium, and he therefore retains the name
** Spina iliaca” given it by the older anatomists. He also
concludes that the avian pubis is homologous with the pubis
of Reptiles. He describes the pubis as originating indepen-
dently of the other elements of the girdle and beginning to
fuse with them about on the eighth day. I find that the
pubis is absolutely continuous with the girdle at the earliest,

1 A, Bunge, “ Untersuchungen zur Entwickelungsgeschichte des Becken-
giirtels der Amphibien, Reptilien, und Vogel,” Dorpat, 1880.

DEVELOPMENT OF PELVIC GIRDLE IN THE CHICK. 405

and all other cartilaginous stages. But, since it lies in a
somewhat different plane from the rest of the girdle, their
junction is only visible in a few sections. In most, the region
of junction is not cut through, as appears in fig. 10, which
represents a section taken from the series out of which fig. 4
was compounded. I think that Bunge must have been misled
by the frequent occurrence of such sections, and so have over-
looked the few in each series in which the junction is really
visible, such as that represented in fig. 11. Fig. 12 again
represents a single section, showing the complete continuity of
pubis and girdle. It is the ossification alone which gives rise
to any want of continuity in any part of the girdle.

Homologies of the pubis in the different Vertebrate
groups.—lIn the pelvic girdle of Ornithorhynchus (see fig. 17)
a large process, whose length is about three quarters of that of
the pubis, projects forwards from the region in front of the ace-
tabulum, in bony continuity with the pubis. The same process
is found in a somewhat reduced form in Echidna, and is still
more reduced in many Marsupials and higher Mammals.
Sometimes it is entirely absent. In embryo birds (see fig. 15),
the process is found in about the same proportionate condition
of development as in Ornithorhynchus. In the adult, it
becomes much reduced, or is absent. Sometimes, as in the
Ostrich, the ilium takes a small share in its formation, but
this appears to be a secondary condition. It is the pectineal
process of the pubis.

In the Dinosaurs, as described first by Marsh,! the embryonic
condition of birds and the adult form of Ornithorhynchus is
preserved in the almost equal development of the two branches
of the pubis, the anterior being shorter and more massive, and
the posterior longer and more slender (fig. 16).

The homologies in these cases seem clear, and have been
generally recognised.

Turning to the reptiles, it is easy to compare the pubis of

10. C. Marsh, “ Principal characters of American Jurassic Dinosaurs,”
‘American Journal of Science and Arts’ (Silliman), vols. xvi and xvii, 1878
and 1879.

406 ALICE JOHNSON.

Lizards with that of Chelonia. In both Lizards and Chelonia
the pubes are directed forwards from the acetabulum, and form
a symphysis. ‘The angle at the symphysis is generally much
greater than thatin Mammals. In some Chelonia it is even
greater than 180°. In both Lizards and Chelonia a process is
given off from the outer side of the pubis. In the latter group
it is often very large (see fig. 13), and is directed forwards, out-
wards, and somewhat downwards. In Lizards it is not so large,
but still considerable, or it may be absent. It is generally
directed outwards and downwards; but in some forms, such as
Cyclodus (see fig. 14), it curves backwards and slightly inwards.
In this case we could hardly compare it with the process found
in Chelonia were it not for the many intermediate forms exist-
ing between these two extremetypes. The process in question
is the processus lateralis pubis. In Crocodiles it is
absent.

In the Urodela the pubes are generally represented by an
unpaired cartilaginous plate, not clearly marked off from the
ischium, which is often ossified. Rarely the pubis itself has a
superficial ossification. The pubic cartilage in Cryptobranchus
is oblong, with a median process in front bearing the epipubis,
and the anterior angles of the oblong are slightly produced.
In Salamandra maculosa these angles form short broad pro-
cesses, which may be compared with the processus lateralis
pubis of Chelonia.

We have, then, in reptiles two branches of the pubis—the
processus lateralis and the main body of the pubis—which
two branches it is possible to derive from the condition found
in Urodela. Also in Dinosaurs, Birds, and Mammals we have
the pectineal process and the main body of the pubis. The
splitting of the pubis into two branches is more complete—
i.e. it approaches nearer to the acetabulum—in the higher
forms.

There is every probability that the two branches correspond
in some way in all these types. Two theories on the subject
are obviously possible. Either (1) the processus lateralis
of reptiles is the pectineal process of the pubis in Dinosaurs,

DEVELOPMENT OF PELVIC GIRDLE IN THE CHICK. 407

birds, and Mammals, and the pubis itself is in both cases homo-
logous, or (2) the pubis of reptiles is the pectineal process, and
the processus lateralis is the pubis of the higher forms.
The first is the view apparently assumed by Huxley.’ Sup-
posing it to be true, the processus lateralis, in becoming the
pectineal process, has retained the forward and outward direc-
tion which it has in the Chelonia. In Dinosaurs the down-
ward direction is also seen. The pubis itself has become
rotated backwards. The mere fact of its pointing forwards in
reptiles and backwards in Dinosaurs, Birds, and Mammals, is
no reason whatever against the theory of its being homologous
in the two cases, for it is generally believed that the whole
girdle has rotated in Mammals through an angle of about 90°
from the position it occupies in reptiles. This would com-
pletely account for the altered position of the pubis. The fact
that the angle formed at the symphysis of the pubes has gene-
rally become more acute in Mammals is a natural consequence
of the transition from the crawling flat-bodied reptiles to the
higher walking forms, in which the body is more laterally
compressed.

In birds the case is somewhat different. The fact of the two
primary sacral vertebre being situated, as Gegenbaur’ has
shown, at a very short distance behind the acetabulum may in-
dicate that the girdle has been rotated backwards to some
extent from the reptilian position. The pubis may thus have
come to point vertically downwards or very slightly backwards,
as in the embryo bird. The adductor muscles passing from the
pubis and ischium to the femur in reptiles are to a great extent
replaced in birds by large muscles, which act as flexors of the
thigh and adductors of the leg. It is evidently advantageous
for these muscles to arise high up, and for their points of origin
to be as rigid as possible. These advantages are attained by
the disposition of the bones in the adult bird’s pelvic girdle, the

1 Huxley, ‘On the Pelvis in Mammalia,” ‘Proceedings of Royal Society,’
vol. xxviii, 1879.

2 Gegenbaur, “ Beitrige zur Kenntniss des Beckens der Vogel,” ‘ Jenaische
Zeitschrift,’ Band vi, 1871.

408 ALICE JOHNSON.

form of which, therefore, may be accounted for in this way.
The pubis, being placed lowest, loses its functional importance
as a point of support for muscles and becomes very slender. Its
middle portion may even abort altogether, as sometimes happens
in ducks and other swimming birds.

So far it appears quite possible to explain the facts by the
first theory.

Turning to the second, we have to imagine a somewhat
different process. The processus lateralis of reptiles, in
becoming the pubis of the higher forms, has retained the posi-
tion which it had already come to occupy in some Lizards (see
fig. 14), and has increased in extent and functional importance.
In Mammals it goes so far as to form a new symphysis, while
in birds the backward direction of the bone is very much exag-
gerated. The part corresponding to the reptilian pubis at first
retains its original situation and almost its original dimensions,
as the anterior branch of the pubis in Dinosaurs, the embryo
bird, and Ornithorhynchus. It gradually dwindles into the
subordinate position of the pectineal process.

This theory again, accounts for all the known facts, and it
agrees, better than does the former view, with the’ relations of
the pubis in Dinosaurs. Marsh! found that the anterior
branch of the pubis in the Stegosauria and Ornithopoda, e. g.
Laosaurus, passed forwards and inwards, ending in a broad
spatulate free extremity. In the Theropoda and Sauropoda,
e.g. Atlantosaurus, no posterior branch of the pubis existed,
but the bone which evidently corresponded to the anterior
branch in Laosaurus formed the symphysis. Judging from
this fact there seems no doubt that the anterior branch is
homologous to the reptilian pubis. I think there can also be
no doubt of the homology between it and the anterior branch,
which, however, no longer forms the symphysis in birds and
mammals.

These conclusions may be tabulated as follows :

_ 1 0.C. Marsh, “ Classification of Dinosaurs,” ‘ American Journal of Science’
(Silliman), 1882.

DEVELOPMENT OF PELVIC GIRDLE IN THE CHICK. 409
Reptiles. | Dinosaurs. |Embryo Bird.| Birds. | Mammals.
1. Pubis Anteriorbranch/Anterior branch Pectinal pro-/Pectinal pro-
of pubis of pubis cess of pubis; cess of pubis.
(‘pubis of ;
Marsh)
2. Processus Pubis (‘ post- |Posterior branch Pubis Pubis.
lateralis pu-| pubis’ of of pubis
bis Marsh)

The development of the skeleton of the limb has
been described by Gegenbaur.' Rosenberg,” has supplemented
Gegenbaur’s accounts by his discovery of the fifth metatarsal,
and quite recently Baur * has published a paper on the Tarsus
of Birds and Dinosaurs. The results of my work on the
development of the bird’s tarsus agree with Baur’s in almost
every detail, so that I will give only a short account of it.

In a five days’ chick (see fig. 1) the tissue of the limb is
condensed axially into a single mass, about three times as long
as it is broad, and extending through the proximal half of the
limb. The skeleton is produced by the subsequent elongation
and segmentation of this mass.

- On the sixth day (length of hind limb 0:14 in., see fig. 8)
we can recognise all the chief elements of the skeleton, though
they are completely continuous. The tarsus forms a broad
transverse band, continuous with the tibia and fibula above,
and with the metatarsals below. Five metatarsals are present,
the first and second being rather closely united. The third is

the longest and the fifth the shortest. No cartilaginous

1 C. Gegenbaur, “ Vergleich.-Aant. Bemerkungen iiber das Fussskelet der
Vogel,” ‘Archiv fiir Anat. und Phys.,’ 1868; and “ Untersuchungen zur
vergleichenden Anatomie der Wirbelthiere,” i Heft, ‘Carpus und Tarsus,’
1864.

? A. Rosenburg, “‘ Ueber die Entwicklung des Extremititen-Skelets bei
einigen Wirbelthieren,” ‘ Zeitschrift f. wiss. Zoologie,’ 1873.

3G. Baur, “Der Tarsus der Vogel und Dinosaurier,” ‘ Morphologisches
Jahrbuch,’ Band viii, Heft iii, 1882.

410 ALICE JOHNSON.

matrix has yet appeared, but the “‘ prochondral elements ” are
visible in the femur, tibia, and fibula.

Soon after—when the limb is 0°17 in. long—separate ele-
ments begin to appear in the tarsus. Of these there are three,
two in the proximal row and one in the distal. The tarsus is
still continuous throughout and continuous also with the tibia,
fibula, and metatarsals. But in these three centres, as well as
in the tibia, fibula, and metatarsals, the differentiation of tissue
has gone further. The outlines of the various parts are indis-
tinct. They all pass gradually into one another by means of
the general groundwork of condensed tissue formed by the
tarsus. The knee-joint is, however, developed at this stage.

A little after this stage, the first metatarsal, which does not
keep step with the others in histological development, begins
to split off from the tarsus and soon hes at some little distance
from it. Baur describes the first metatarsal as originating
quite independently and never coming into any connection
with the tarsus.

The phalanges next begin to appear. When the limb is
about 0:2 in. long, they are marked off by constrictions from
the metatarsals, but are cartilaginously continuous with them.
Later, when the limb is 0°3 in. long, the phalanges are marked
off by intervening tracts of condensed tissue with no matrix in
it. The tip of each toe at this period and for some time to
come consists of a mass of condensed tissue such as always
precedes cartilage (see fig. 9). This appears to be the grow-
ing point of the cartilage. From these facts it seems that the
phalanges are produced by a lengthening and subsequent seg-
mentation of the original distal cartilages of the limb, so that
these cartilages represent the skeleton of the digits as well as
the metatarsals.

On the eighth day (length of hind limb 0:27 in.—see fig. 9)
all the elements of the tarsus are at their most distinct and in-
dependent stage, though they are still united with one another,
with the tibia and fibula, and with the metatarsals by the con-
densed tissue of the groundwork of the tarsus.

Later, the distal and proximal parts of the tarsus become

DEVELOPMENT OF PELVIC GIRDLE IN THE CHICK. 4lLlL

separated, and the two proximal elements fuse together. Next,
the proximal part begins to fuse with the tibia, which has
grown more than the fibula, so that the latter no longer reaches
the tarsus. The posterior lower edge of the tibia first becomes
continuous with the proximal tarsal cartilage, while the ante-
rior face of the latter gives off an upward process, the so-called
‘‘ascending process of the astragalus,” which fits into a
groove in the tibia, and remains for a long time separate from
it. At about the same time the distal part of the tarsus fuses
with the metatarsals, first with the second, next with the
fourth, and lastly with the third. All these processes take
place while the tarsus is still cartilaginous.

Morse! describes, in the tarsus of the embryo bird, an inter-
medium, which at first projects upwards between the distal
ends of the tibia and fibula. Later, the tibiale and fibulare
fuse behind it, while the tibia extends so as to cover the
whole proximal surface of the tarsus, and the intermedium
remains fitting into a groove on the anterior face of the tibia.
It has a separate centre of ossification, but becomes anchylosed
with the tibiale and fibulare, forming what is called the
ascending process of the astragalus.

Both Baur and myself fail to find a separate origin for the
intermedium. Baur describes the ascending process as an out-
growth from the tibiale, in which view I am inclined to concur.
But the deviation of our views from that of Morse may, per-
haps, be explained by the fact that while Baur worked only at
the chick, duck, sparrow, pigeon, and blackbird, and I only at
the chick, Morse investigated some aquatic birds—the tern,
penguin, petrel, gull, &c.

In conclusion, I have to thank Dr. Gadow for his kindness
in giving me help and advice during the course of my work,

' E. 8. Morse, “On the Identity of the Ascending Process of the Astragalus
in Birds with the Intermedium,” ‘Anniversary Memoirs of Boston Society of
Natural History,’ 1880.

412 WALTER HBEAPE.

The Development of the Mole (Talpa Europea.
The Formation of the Germinal Layers, and
Early Development of the Medullary Groove
and Notochord.

By

Walter Heape,
Demonstrator in the Morphological Laboratory, Cambridge.

With Plates XXVIII, XXIX, XXX, XXXI.

In the following paper I propose to commence with a
description of the fully-segmented ovum, leaving the details
of the segmentation for a future communication; thence to
trace the growth of the blastodermic vesicle and the ultimate
formation of the hypoblast, epiblast, and mesobiast of the
embryo; to follow the early stages of the development of the
medullary canal and notochord ; and finally to touch upon the
phenomenon of the inversion of the layers in certain mammals,
and to endeavour to show that in the mole there exists in
development an intermediate condition between the inverted
type, of which the guinea-pig is an example, and the normal
type as exemplified by the rabbit.

Owing to the difficulty of keeping moles alive and the still
greater difficulty of observing their breeding habits when in
captivity, I have found it impossible to determine the exact
age of any embryos, and am obliged to fix their relative age in
accordance with their size, and what appears to me to be the
course of their development.

Under these circumstances it will be convenient to divide

THE DEVELOPMENT OF THE MOLE. 413

embryonic life into periods which I propose shall be regu-
lated by the following conditions:

Stage a. The period of the development of an ovarian into
a fully-segmented ovum.

Stage B. The further development of the fully-segmented
ovum prior to the formation of mesoblast. That is to say, the
stage in which the hypoblast and epiblast are definitely formed.

Stage c. The formation of the mesoblast ; and,

Stage p. The formation of the medullary groove and noto-
chord, and the structure of the neurenteric canal.

Stage A.

I have been hitherto unable to satisfy myself as to the details
of the process of segmentation in the ovum of the mole, but
have been fortunate enough to obtain a fully-segmented ovum
It was found at the upper end of the uterus, and its structure
is as follows ;

It is formed of a number of distinct cells, each surrounded
by a cell-wall, and containing a nucleus. The cells are
arranged in two layers (vide fig. 1), which I propose to name
(1) the outer layer (0. 7.), and (2) the inner mass (7. m.). The
cells of the outer layer are more or less cubical in form, and
are placed side by side in a single row. The main portion of
each cell is composed of hyaline protoplasm, but Toes its inner
border the protoplasm is finely granulated.

The cells of the inner mass are slightly smaller than those
of the outer layer. They are irregulariy polygonal in shape,
and the protoplasm of which they consist is filled with large
and small granules, rendering the cells opaque and dense.

The single row of outer layer cells closely invests and com-
pletely surrounds—except at one point—the inner mass of
segments. At this point (0/. of B.), however, there is a break
in the continuity of the outer layer, and here one of the
granular cells of the inner mass projects on to the surface.

The ovum is surrounded by a thick membrane, the zona
radiata (z.), which is in its turn enclosed in an irregular layer
of hyaline gelatinous material (mm. c.) derived from the uterus.

VOL, XXIII.—NEW SER, EE

414 WALTER HEAPE.

The zona is radially striated, and its outer edge has a
granular appearance, which I have reason to believe, from an
examination of ovarian ova, is due to the irregularity of its
surface caused by the pressure of the follicular epithelium upon
it while still in the ovary.

There is no albumen deposited round the ovum during its
passage down the Fallopian tube, as is the case with the
rabbit’s ovum.

The ovum within the zona measures ‘15 by ‘17mm., while
the inner mass measures ‘1 by ‘12 mm. in diameter ; the outer
layer being about ‘(05 mm. thick. The zona is ‘01 mm. and
the outer coat ‘(014 mm. in thickness.

The size of segmenting ova vary somewhat, but as a rule,
while in the Fallopian tube, they measure between ‘08 and ‘1
mm. in diameter. This fully-segmented ovum shows a con-
siderable increase on that size, and this is probably due to the
absorption by it of nutritive material present in the uterus.
It was examined first of all while fresh, but the details of its
structure were rendered more apparent by treatment with silver
nitrate. The figure was drawn after treatment.

The structure of the fully-segmented mole’s ovum as de-
scribed above is identical with that of the fully-segmented
ovum of the rabbit which van Beneden has described (Nos. 4
and 5). According to this author the result of the first divi-
sion of the ovum of the rabbit is the formation of two cells,
the one of which is smaller and more granular than the other.
The product of these two cells can be distinguished from one
another throughout the process of segmentation, and Beneden
finds the cells derived from the granular segment become invo-
luted within those derived from the larger hyaline segment,
and two layers are thus formed which he terms ‘ entoderm”
and “‘ ectoderm” respectively, according to what he considers
is their respective fate.

Further, the point where the involution took place remains
open in the fully-segmented ovum, and gives rise to the gap
in the outer layer, which is called by Beneden the “ blasto-

THE DEVELOPMENT OF THE MOLE. 415

pore,” and compared by him to the blastopore of other Verte-
brata.

The fully-segmented ovum is therefore considered by this
author to be comparable to the gastrula stage of other verte-
brata.

I have hitherto been unable to confirm the account given by
Beneden of the segmentation, but am by no means therefore
disposed to conclude his careful descriptions are inaccurate.
At the same time it appears to me obvious, from the subse-
quent development of the mole, that his views of the homolo-
gies both of the two layers of segments and of the “ blastopore”
are incorrect.

In the first place, the so-called “ entoderm” segments will
be found to give rise to the greater part of the epiblast of
the embryo; and in the second place, the structure of the
ptimitive streak will be seen strongly to confirm Balfour’s
opinions (vide Nos. 1 and 3) of the homology of that organ
with the blastopore of lower types.

Had Beneden’s interpretations been correct, however, and
had the inner mass really been entodermic, the fully-segmented
mammalian ovum could not even then be compared to the
gastrula condition of Amphioxus; for whereas the enteric
cavity of the latter is within the entoderm cell mass, that of
the former is eventually found to be outside.those cells, and
between them and the ectoderm.

Up to this point in the development no differentiation of
the segmentation spheres into epiblast and hypoblast has yet
taken place, and there is indeed, as I will show later, no
evidence of any differentiation until some considerable time
after the completion of segmentation.

The structure of the mammalian ovum at the close of
Stage A is therefore seen to be, as far as we can now tell,
entirely unlike that of any other animal; and until we have
some knowledge of the steps by which mammals were evolved,
it appears to me useless to attempt to draw any homologies.

It may be interesting to note that although the earliest con-
ditions of mammalian development cannot be compared with

416 WALTER HEAPE.

those of other animals, yet the further development proceeds
(up to a certain point) the more strikingly similar these condi-
tions become, and the usual rule that embryos of various
animals differ from one another less in their earlier than in
their later stages of development is therefore here reversed.

In concluding this section I would draw attention to the facts
treated fully below (1) that the central position of the inner
mass of segmentation spheres in both the rabbit and the mole
is merely temporary, and that subsequently these cells, with
the exception of a very small number, form a portion of the
wall of a vesicle, the “ blastodermie vesicle.”

(2) That the so-called blastopore (Beneden) cannot be simi-
lar to the blastopore present in Amphioxus, and has merely a
secondary origin, its existence being caused by the temporary
involution of a portion of the wall of the blastodermic vesicle.

Stage B.

The Blastodermic Vesicle and the Formation of the
Hypoblast and Epiblast of the Embryo.

The conversion of the fully-segmented ovum into the so-
called blastodermic vesicle takes place shortly after the appear-
ance of the ovum in the uterus. It is due partially to a
flattening-out of the cells of the outer layer, and partially to
the conversion of certain of the cells of the inner mass into outer
layer cells.

The result of these changes is a vesicle the wall of which is
composed of, for the most part, a single row of flattened cells,
the much attenuated zona radiata surrounding the whole.

In the course of its growth the vesicle becomes so large that
the wall of the uterus in the region where it is placed is dis-
tinctly swollen.

It is clearly impossible for the delicate-walled ovum to
expand in the form of a vesicle, and distend the uterine walls
by virtue of the growth of its cells; it must be therefore con-

cluded that it obtains some support. This support is rendered
from within.

THE DEVELOPMENT OF THE MOLE. A417

The vesicle contains a transparent fluid, the nature of which
I am only sufficiently conversant with to say that, after treat-
ment with alcohol a white precipitate is present in the vesicle.

It is equally evident that this fluid can only have been
obtained from the uterus, and that it is present within the
vesicle at a very considerably greater pressure than in the
uterus itself. Such a condition is caused by means of the cells
of the wall of the vesicle; they secrete the fluid within the
vesicle, this function being performed against a pressure which
is greater on their inner than on their outer side, exactly as the
cells of the salivary glands are known to act.

The uterine fluid is secreted by glands, present in great
numbers in the uterine tissue, and is poured through their open
mouths into the cavity of the uterus (vide fig. 51). There is
every probability it has nutritive qualities, since it is thence
taken up into the cavity of the embryonic vesicle, which
eventually functions as a yolk-sac, in the walls of which em-
bryonic blood-vessels ramify.

A specimen showing an early condition of this change of the
segmented ovum into a vesicle has been drawn in optical section
in fig. 2. It differs mainly from fig. 1 in that a crescent-shaped
cavity (d/. cav.) exists between the inner mass and outer layer,
this being the cavity of the blastodermic vesicle.

Van Beneden’s blastopore has entirely disappeared, and I have
no evidence to offer as to the position which it originally occu-
pied ; although there is good reason to believe, from a comparison
of the development of the rabbit and mole with animals which
exhibit the phenomena attending the inversion of the layers,
that Beneden’s statement is correct, viz. that the inner mass
remains attached to that side of the outer layer where the gap
was originally placed.

The appearance of the cells has altered but little; the outer
layer cells are slightly more granular, while the cells of the
inner mass are somewhat smaller and less granular than were
those of the fully-segmented ovum.

The size of the two ova are different, the specimen from which

418 WALTER HEAPE.

fig. 2 was drawn being smaller than the fully-segmented ovum
and not larger, as would have been expected.

I can only attribute this condition either to the variation in
size of different ova of the same age, of which fact I have abun-
dant evidence, or to the effect of the preserving fluid, although
in both instances the objects were treated with silver nitrate
and preserved in glycerine.

However that may be—and this is the point which I wish
to emphasise—the size of the inner mass in fig. 2 is relatively
smaller than that in fig. 1, the diameter of the ovum (fig. 2)
being ‘12 mm., and that of the inner mass ‘06 mm.

The ovum rapidly enlarges, and in fig. 3 the relation in size
of the whole vesicle to the remnant of the inner mass is repre-
sented in an early stage of the development of the blastodermic
vesicle.

The vesicle in this specimen is ‘31 mm. and the inner mass
‘04 mm. in diameter.

This increase in size is due to some extent, without doubt,
to the flattening out and multiplication of outer layer cells
(vide figs. 16—19); but I believe that up to this point in
stage B the cells of the inner mass also contribute to that end.

I have been unable clearly to substantiate this opinion by
means of sections, but the size of the inner mass in this spe-
cimen bears out my views; it is ‘02 mm. less in diameter than
the inner mass in the specimen figured in fig. 2, and ‘07 mm.
less than the inner mass of the fully-segmented ovum. Further,
I have made measurements of a considerable number of speci-
mens of a similar age, and have found this ratio to be almost
uniformly constant.

The structure of the wall of the vesicle and of the inner mass
at this stage is seen in figs. 16—19.

The vesicle wall is formed of much flattened polygonal cells
closely attached to the zona radiata, which bounds them on
their outer side.

The cells contain a large nucleus situated in the centre, and
causing it to bulge towards the cavity of the vesicle.

The nucleus in section appears to be of oval form, while in

THE DEVELOPMENT OF THE MOLE. 419

a surface view (vide fig. 4) it is seen to be rounded. The oval
shape in section is due to it being flattened out, and it is for
this reason also that the nuclei of the outer layer appear in a
surface view larger than those of the inner mass (fig. 4).

The inner mass is solid, more or less rounded in form, and
is attached on one side to the wall of the vesicle. The cells of
which it is made up are always, after treatment with picric acid,
closely adherent to one another, and are sharply marked off
from the cavity of the vesicle (vide figs. 17 and 18).

The specimen drawn in fig. 19, however, was treated with
silver nitrate and preserved in weak glycerine, afterwards being
transferred to spirit, embedded, and cut into sections ; in it the
cells are much more loosely held together, and in another spe-
cimen I have, which was similarly preserved, the same appearance
presents itself.

The irregularly-rounded cells of the inner mass, which are
very considerably smaller than either the cells of the inner
mass in the fully-segmented ovum, or of the specimen drawn in
fig. 2, are composed of granular protoplasm, and many of their
nuclei exhibit the modifications attending cell division.

As the vesicle continues to enlarge the inner mass also now
increases in size, changes its shape, and becomes flattened out
along the side where it adjoins the outer layer; and further,
the cells of which it is now composed become differentiated into
two layers.

The differentiation occurs in the following manner :—Certain
of the cells bordering the blastodermic cavity become separated
off from the main portion of the inner mass, and form a single
layer of cells bounding the mass on its inner side.

This layer is the hypoblast.

The hypoblast is, therefore, derived from cells which result
from the multiplication of the inner cell mass present in the
fully-segmented ovum.

Figs. 20 to 23 adequately represent these changes as they
take place; the cells here and there along the lower border
of the inner mass become more flattened than their fellows,
and stain more deeply with hematoxylin (fig. 20) ; gradually a

420 WALTER HEAPE.

continuous layer exhibits these phenomena (fig. 21), and then
become separated from the remainder of the-mass (fig. 22).

The remaining cells of the inner mass increase in number,
and assume a columnar form, at the same time becoming sepa-
rated by a narrow cavity in the centre from the outer layer.
The cells of the latter layer in the region of the cavity also
increase in number and become thicker than their fellows (fig.
23).

The further development of the hypoblast may be stated in
a few words; it extends laterally by virtue of the multiplica-
tion of its cells, which at the same time become considerably
flattened. Later on, as may be seen in fig. 28, the cells are
again more rounded, and, indeed, at different stages during
the formation of the layers, they assume various proportions.
It is to be noted that this layer, after being once completely
separated off from the inner mass (vide fig. 23), remaitis sepa-
rate until the mesoblast is formed, and increases, therefore,
wholly by the division of its own cells.
~ The hypoblast eventually completely surrounds the whole of
the blastodermic vesicle.

The changes which take place in the remaining portion of
the inner mass and in the outer layer adjoining it are some-
what more complicated.

1. The inner mass increases in size and its columnar cells,
arranged in a double row, form an hemispherical plate, the edge
of which rests upon and is continuous with the cells of the
outer layer. In consequence of this the narrow cavity men-
tioned above assumes considerably greater proportions; it is
bounded below and at the sides by the plate, and above by the
outer layer ; at the same time it becomes partially filled up by
branched stellate cells, which are derived from the cells of the
outer layer.

This cavity may be termed the secondary cavity of the blas-
todermic vesicle in contradistinction to the cavity which is
formed at the close of stage A, and which also arises, although
at the opposite side, between the outer layer and inner mass.

Fig. 5 is a drawing of an ovum at this stage of growth.

THE DEVELOPMENT OF THE MOLE. A21

The opaque inner mass is seen attached to the wall of the
vesicle, and in the centre of the mass a lighter coloured space
indicates the presence of the secondary cavity.

The relations of these parts are, however, more clearly seen
in fig. 24, which represents a section through the centre of the
inner mass of the ovum drawn in fig. 5. The single row of
long columnar cells (fig. 23) has given place to a double row of
more cubical and broader cells which are continuous with the
cells of the outer layer at the circumference of the plate.

The hypoblast lies free below the inner mass and stretches
out laterally beyond the area of the latter.

The cells filling up the secondary cavity are stellate, and are
connected with both the outer layer and inner mass by means
of protoplasmic processes ; the size and general appearance of
the cells and of their nuclei, however, as well as the manner in
which they stain with hematoxylin, leaves little room for doubt
in my mind that they are derived from the former (outer)
layer.

2. The plate of cells now changes its form and becomes
flattened out and applied closely to the zona above, the stellate
cells within the secondary cavity and the outer layer cells
above uniting with it, and the secondary cavity is obliterated.
The structure resulting from these changes is the epiblast
plate of the embryonic area.

Reference to figs. 7, 25, 26 and 27, will, I think, substan-
tiate this view.

Fig. 7 is a surface view of the inner mass represented in
section in fig. 25. The section is cut along the line of the
greatest diameter of the mass, and shows the commencement
of the process of the flattening out of the plate.

The flattening occurs in the first place along one side, the
secondary cavity being there much shallower, while elsewhere
it is as deep as before. This arrangement gives rise to the
appearance seen in fig. 7, in which the light, crescent-shaped
area at one side of the inner mass is the deeper portion of the
cavity (compare figs. 7 and 29).

In all the sections of this inner mass only a few cells were to

422 WALTER JIEAPE.

be seen in the secondary cavity, and in the section here figured
there were none present. I have, however, never found this
to be the case in any other specimen, and imagine they must
have been displaced during the process I then used of cutting
and mounting sections.

The flattening process afterwards extends all round the edge
of the inner mass and the cavity is throughout much shallowed
(fig. 26), cells are present within the secondary cavity, and are
seen in this section becoming incorporated with the plate of
columnar cells. The edge of the plate, as I before mentioned,
is continuous with the cells of the outer layer, and at a slightly
later stage, when the plate is completely flattened out, it occu-
pies the position until then held by that portion of the outer
layer which overlay the inner mass.

At this stage the two layers are indistinguishable from one
another, but wedge-shaped cells can be observed in the upper
portion of the plate (fig. 27 ¢c) which on account of their
shape, the direction of the long axis of their oval nucleus, and
‘the position they occupy appear to me without doubt to
have been derived from the cells of the outer layer.!

Up to this point in the development the blastodermic vesicle
lies free within the cavity of the uterus, and can be obtained
therefrom without difficulty by merely slitting up the uterus
with scissors and transferring the ovum upon the point of a
scalpel to a watch-glass containing the hardening reagent.
This method is, however, no longer possible when the ovum
attains a very slightly older stage. It then becomes still
further enlarged and its walls project into the widely open
mouths of the uterine glands. I find no actual attachment
between the two, and have not been able to distinguish any
outgrowths from the zona such as Bischoff described for the
rabbit and dog (Nos. 6 and 7).

The only method which I have found to enable me to obtain
the fresh vesicle entire, is to sink the uterus, after being cut
open, with the ovum in sitt, slowly in a vessel of salt so-

1 In support of this view see below, an account of a stage in the formation
of the epiblast of the rabbit.

THE DEVELOPMENT OF THE MOLE. 423

lution, it then becomes possible to separate the two without
damage.

Figs. 8 and 9 represent very faithfully the appearance of
ova of this age obtained by the above method.

The wall of the vesicle is bulged out here and there into
papille where it projected into the mouths of the glands. The
elongated condition is due to the fact that when the uterus
was slit open the vesicle fell into that shape in which it
received the greatest amount of support from the surrounding
tissue.

If it is not desired to examine the vesicle in a fresh state it
will be found advantageous to harden the embryo within the
uterus, and to dissect it out afterwards which is an easy
matter.

In order to show the position and condition of the uterine
glands, I have drawn in fig. 51 a transverse section through
that region of the uterus from which the ovum represented in
fig. 8 was obtained.

It will be seen that only on the free non-mesometric side of
the uterus are there any widely open mouths of glands; while
upon reference to figs. 8 and 9, only that side of the ovum
around the embryonic area is seen to be prolonged into papille-
form projections, and as the embryonic area lies against the
non-mesometric side of the uterus in the mole we may con-
clude the projections lie in the mouths of the glands. A
portion of the epithelium of the uterus abstracted fiom the
latter is drawn in fig. 52; it is seen to be prolonged into
hollow finger-like processes which line the uterine glands.

Transverse sections of the embryonic area of this embryo
(fig. 28) show that it is formed throughout of two layers of
cells, epiblast and hypoblast.

One of the prolongations of the vesicle wall has been cut at
one side of the embryonic area (p/) and another is shown in
fig. 29. They are seen to be formed wholly of epiblast, the
hypoblast not being extended into them,

Of the hypoblast layer in the area I have nothing to add to
the account already given; the epiblast, however, has under-

424 WALTER HEAPE.

gone a slight change since we last examined it, inasmuch as it
now consists for the most part of a single row of columnar
cells, which at the sides of the area gradually become less and
less columnar and eventually merge into the flattened epiblast
cells of the wall of the vesicle. ‘This change is, however,
temporary, since in sections of older embryonic areas the epi-
blast is again two layers deep (figs. 33, &c.).

Fig. 30 is a transverse section of the area drawn in fig. 10;
it is very similar to fig. 28; but the edges of the embryonic
area in this case appear to end abruptly, the wall of the vesicle
having been torn away owing to its close attachment to the
uterus.

The condition of the ovum is now considerably changed
from what it was when the blastodermic cavity first appeared ;
it may be divided into two areas, the embryonic and non-
embryonic areas. The embryonic area is throughout com-
posed of an outer thickened layer of columnar epiblast cells
which has been derived partially from a portion of the inner
mass and partially from outer layer cells, and an inner layer
of somewhat rounded hypoblast cells derived entirely from
cells of the inner mass. The non-embryonic portion of the
Ovum may in its turn be divided into two regions.

First, the region immediately surrounding the embryonic
area which is formed of tio layers, an outer of flattened outer
layer cells now known as epiblast cells, continuous with the
epiblast of the area, and an inner of flattened hypoblast cells
continuous with the same layer in the embryonic portion of the
ovum.

Secondly, the region situated at the opposite pole of the
ovum to the embryonic area, where a single row of flat epiblast
alone exists.

Historical.—The details attending the formation of the epi-
blast has given rise to a considerable amount of discussion.
According to Edward van Beneden (No. 5) the fully-segmented
ovum of the rabbit develops into the blastodermic vesicle by a
multiplication and a flattening of the outer layer cells, the inner

THE DEVELOPMENT OF THE MOLE. 425

mass remaining the same in size, and attached to the outer layer
in the region of the now-closed “ blastopore.”” Subsequently the
inner mass flattens out and splits up into two layers, the lower
of which forms the hypoblast and the upper the mesoblast of
the embryonic area, the epiblast being formed solely by the
multiplication of the outer layer cells, which become at the
same time columnar and arranged in a single row.

Beneden, therefore, has not found a stage in which two
layers only exist throughout the area.

Rauber (No. 21), in a previously written paper, finds three
layers present in the embryonic area of a rabbit before the
formation of the primitive streak ; the outer of these (my outer
layer) he calls the ** Deckschicht,” and states that it early dis-
appears, while the middle layer alone forms the epiblast and
the lower the hypoblast of the area; a two-layered area being
thus formed.

Kolliker, in a recent elaborate paper (No. 16), traces the
fate of the three layers, which he also finds in common with
Rauber and Beneden, and declares, in accordance with the
views of the former author, that the outer layer gradually dis-
appears, the middle forming the epiblast and the lower the
hypoblast of the embryo. The details of the gradual disap-
pearance of the Deckschicht occupy much of this paper. Pro-
fessor Kolliker has never seen the cells of this layer assume
a columnar form, as Beneden asserts is the case, and by
means of nitrate of silver staining he satisfies himself they
gradually become broken up, and eventually disappear al-
together.

Lieberkiithn (No. 19) gives an account of the formation of
the epiblast in the dog and mole which is very similar to my
own, in that he considers it is formed of the greater portion of
the inner mass, together with that portion of the outer layer
cells which originally overlaid it. He also draws attention to
the cavity which appears, according to him, within the inner
mass Of cells in the mole, and which he suggests may be com-
parable to the segmentation cavity of other animals.

Hensen (No. 12) for the rabbit, and Schafer (No. 25) for the

426 WALTER HEAPE.

cat, also describe a two-layered stage of the embryonic area
prior to the formation of the primitive streak.

Summary.—With regard to my own work I hold that the
blastodermic vesicle increases in size, not merely on account of
the increase in number and the flattening of the outer layer cells,
as Beneden believes, but by the migration of inner mass cells to
the exterior. This view is supported by the fact that the inner
mass decreases in size during the early development of the
vesicle. I have also satisfied myself of the existence, both in
the mole and rabbit, of a stage in which the embryonic area is
composed of only two layers, the epiblast and hypoblast.

The hypoblast I have shown to be derived from the cells of
the inner mass—a fact which all the observers above mentioned
are agreed upon.

The epiblast I believe to be formed, as does ica of
the remaining portion of the inner mass, after the hypoblast
has been deMehed: together with that portion of the outer layer
which overlies the inner mass.

In the mole this includes also certain cells which we have
seen are derived from the outer layer, and which at one time
lie in a cavity between that layer and the inner mass.' In the
rabbit, however, no such cells exist, and I believe that the epi-
blast is formed of inner and outer layer cells.

With reference to the development of the epiblast in the rabbit
I may saythat since working at the question under the supervi-
sion of the late Professor Balfour (No.3), I have examined more
embryos, and have been fortunate enough to obtain good sec-
tions of the embryonic area of a rabbit embryo of six days four
hours old, which appear to me to be conclusively in favour of
the view we were then inclined to accept. Fig. 49 represents
a section through this area; in it the epiblast plate is seen to
be composed of two entirely different kinds of cells—(1) a
lower more or less columnar or rounded cell, and (2) an
upper flattened or wedge-shaped cell. The latter cells in-
variably occupy a position on the outer side of the plate, across

| In a previous paper (No. 11) I erroneously described these cells as being

derived from the inner mass. y
mo

THE DEVELOPMENT OF THE MOLE. 427

which they form an almost continuous layer, and they are .
distinctly darker stained than are the deeper placed, more
columnar cells. They are generally wide at the top, ending
below in a wedge-shaped base, which grows downwards between
two of the columnar cells lying beneath. Some of the cells
are, however, more flattened, possessing no downward pro-
longation, and some are more columnar, having little or no
expanded upper surface; indeed, there are cells in all stages
of transition, between the flattened outer layer cells of the
previous stage and the columnar cells of the future epiblast
plate (vide fig. 49, ¢. c.).

Kolliker’s valuable paper contains most careful descriptions
and drawings, which, however, appear to me to be capable of
a very different interpretation from that put forward by him ;
in fact, they appear to me to be strongly confirmatory of my
own views. He states that the large nucleated plates which
are visible in surface views of young areas split up in older
embryos into small polygonal areas without nuclei. Now, I
would venture to suggest that the disappearance of the nuclei
of these large outer layer plates can be fully accounted for by
their migration downwards among the cells of the inner mass
(vide fig. 49): and the apparent breaking up of the large cells
may be explained by the actual appearance on the surface of

the epiblast plate, of the polygonal ends of the columnar cells
of which it is now composed.

Stage C.
The Formation of the Mesoblast.

The middle germinal layer has two distinct sources: in the
first place it arises from the epiblast and hypoblast at the hind
end of the embryonic area, in the structure known as the
primitive streak ; and, secondly, from the hypoblast alone in
the anterior region of the area in front of the primitive streak.

428 WALTER HEAPE.

The Primitive Streak Mesoblast.

The primitive streak originally appears at the hind end of
an area similar to the one represented in fig. 10, its presence
being shown in surface view by a slight opacity.

Fig. 31 is a longitudinal section through such an area, along
the middle line. The anterior portion of the area consists of a
layer of columnar epiblast, and a somewhat flattened layer of
hypoblast: at the hind end, however, a passage perforates the
blastoderm and surrounding it the epiblast and hypoblast be-
come continuous with one another, forming the wall of the
perforation. The opening is wider below than above, and
owing, I believe, to the curved condition of this specimen, was
not visible from the surface. The whole length of the area
is not drawn in the figure, and the portion anterior to the spot
at which the reference letters ep. are placed, was bent back, and
underlay the hinder portion of the area.

The cells forming the wall of the passage give rise to the
first mesoblast cells, which are thus derived from epiblast and
hypoblast conjointly ; they extend in front and laterally for a
short distance only as a thin sheet lying free between the
two primary layers, while posteriorly they form a thicker layer
and are united with the epiblast in the middle line.

From this point the primitive streak extends backwards, the
embryonic area itself enlarging in that direction.

Figs. 11 and 12 are surface views of two areas, in which
the primitive streak represented by the dark shading is well
defined. In the former, which is the younger of the two, the
opaque band extends about half way across the oval area,
spreading out behind into two short horns; and down the
centre of the band a lighter streak may be seen, which is
caused by a groove in the epiblast, and is the well-known
primitive groove.

At the front end of the primitive groove there is distinct
evidence in section of the involution of the epiblast, although
no actual perforation of the blastoderm exists. This I consider

THE DEVELOPMENT OF THE MOLE. 429

is the point where in the earlier specimen the blastoderm was
perforated (fig. 31), the increased size of the area being due
hitherto to a growth backwards.

Fig. 12 represents the most advanced condition of the primi-
tive streak. The embryonic area is pyriform, and the primitive
streak is considerably longer than in the former specimen
(fig. 11), and extends relatively further along the area; it is
more opaque, and ends behind in a dark rounded mass or
knob.

I could distinguish no primitive groove by an examination of
the surface, and am obliged, therefore, to rely chiefly upon sec-
tions to determine the relations of the growth. Near the front
end of the streak there is here also distinct evidence of an
involution of the epiblast, although there is no actual perfora-
tion ; and I am inclined to believe this point is identical, both
with the front end of the primitive streak in fig. 11, and with
the point where the perforation exists in the younger embryo
(vide fig. 31). Itisa curious fact, however, that the extent
of the area anterior to the front end of the primitive streak
appears to be less in this area than in the younger one (fig. 11),
while the length of the primitive streak in fig. 12 is greater
than that in the older embryos (figs. 13—14).

The presence of the involuted point at the front end of the
streak appears to me to favour the view that this structure
has not grown forwards, while the addition of the pyriform
hind end is an argument in favour of its backward growth.

The reduction in size of that portion of the area anterior to
the primitive streak may possibly be due to curvature, but this
I am unable definitely to decide.

The eventual reduction in length of the primitive streak is
more easily comprehensible, and is doubtless due to the widen-
ing out of the end knob, this structure having disappeared in
older embryos.

I have frequently observed in surface views a darkly- -shaded
spot at the front end of the primitive streak, which is spoken
of as the node of Hensen, and find that it corresponds with the
spot where the three layers unite. It may also be seen some-

VOL. XXI1I.—NEW SER, FE

430 WALTER HEAPHE.

times when there is no other superficial evidence of the exist-
ence of the primitive streak, but in these cases I have invariably
found by sections that a primitive streak does exist, but that
the mesoblast to which it has given rise is so uniformly dis-
tributed everywhere except at the front end, that it is only
there apparent.

The structure of the primitive streak is different in different
parts, to illustrate which I have figured sections (figs. 33—36)
through various regions of the blastoderm drawn in fig. 12.

The first section (fig. 33) is taken through the anterior portion
of the primitive streak. A plate of columnar epiblast cells
extends across the area; it is thinner at each edge, but of uni-
form thickness elsewhere, except in the middle line, where a
keel-like ridge is formed. ‘The upper half of the keel is wide
and joins the epiblast, with the cells of which it is continuous,
and the lower portion projects into a mass of cells below, but
has no connection with them. These underlying cells I will
deal with later, and will in this place merely draw attention to
the fact that the lower borders of the cells of the keel are
sharply marked off from them, and that these somewhat oval
cells lying below the keel of epiblast are entirely different, both
in shape and character, from the cells above them.

The second section is taken close behind the first; it passes
through the front end of the primitive groove, and is, I believe,
in an analogous position to the point immediately behind the
perforation existing in the embryo, of which fig. 31 is a longi-
tudinal section.

The epiblast is curved in the middle line constituting the
primitive groove, and from the cells of this portion of the epi-
blast, mesoblast is produced.

Immediately below the primitive groove there is no layer of
hypoblast to be distinguished, and here mesoblast is produced
from hypoblast cells. Laterally all three layers are distinct,
but in the middle line they may be said to combine with one
another, and in this region, therefore, the middle layer is
formed from both epiblast and hypoblast. The former does
not here extend beyond the boundary of the embryonic area.

4

THE DEVELOPMENT OF THE MOLE. 43 |

Between the two sections described above, the cord of cells
(fig. 33) joins the front end of the mass of cells formed by the
union of the epiblast and hypoblast in the middle line; and
where this junction occurs there is distinct evidence of an
involution of the epiblast layer.

From the front end of the primitive streak a tongue-shaved
cord of mesoblast cells is projected forwards into the mass of
cells underlying the epiblast in that region, and gives rise to
the lighter shaded prolongation of the primitive streak scen in
fig. 12.

I have been unable to find any complete perforation of
the blastoderm at this stage of growth, although in a some-
what younger embryo and in an older one in which the me-
dullary groove is formed, there is no doubt that it exists.
I have, therefore, either missed the section in which the.
perforation occurred in this embryo, or it has been closed up
by the rapid production of mesoblast which at this stage takes
place.

Although I have only seen a complete perforation of the
blastoderm in one embryo during the primitive streak stage, I
have invariably found at the front end of the primitive streak
evidence of an involution of the epiblast; on this account, as
well as for reasons which will appear in the sequel, I conclude
the front end of the primitive groove is the spot where the
perforation of the blastoderm seen in fig. 31 occurs at an
earlier and later stage. The cord of cells described in the first
section is the front wall of the perforation seen in fig. 31, and
the tongue of mesoblast projecting forwards is homologous
with the anterior growth of mesoblast also seen in the younger
embryo.

This statement is supported by a_ study of sections of
an area but slightly older than that drawn in fig. 11, and
somewhat younger than the one we have been considering.
In this area there was no layer of cells underlying the epiblast
at the anterior end of the primitive streak, and the behaviour
of the forward growth of mesoblast from the front end of that
structure could be more definitely determined. At the front

432 WALTER HWAPE.

end of the primitive streak, at a point relatively similar to that
drawn in section in fig. 34, the epiblast was involuted in the
middle line and a deep pit formed which opened, below into
mesoblast, which is budded off from the lips of the ingrowth.
At this point the epiblast, mesoblast, and hypoblast were united
in the middle line, but in front of it an axial rod of mesoblast
projected forwards for a short distance distinct from both epi-
blast and hypoblast, but soon becoming attached to, and indis-
tinguishable from, the hypoblast. In this condition it may be
spoken of as a thickened axial rod of hypoblast, and as such it
extends forwards for some sections, gradually becoming re-
duced in size and eventually giving place to the single row of
rounded hypoblast which elsewhere existed below the epiblast
in front of the primitive streak.

The third section, (fig. 35) demonstrates the structure of the
area throughout the remainder of the primitive streak in front
of the end knob. It is similar to fig. 34 except that (1) the
hypoblast forms a complete layer across the whole of the area,
and is nowhere combined with the layer of mesoblast ; (2) the
primitive groove is not present, and the number of epiblast cells
concerned in the formation of mesoblast is greater than before,
and (3) the mesoblast extends laterally, lying freely between
the epiblast and hypoblast, outside the limits of the area.
This is a typical section through the middle of the primitive
streak of all the specimens I have examined, and it may be
generally stated that throughout this region the epiblast only
gives rise to mesoblast.

In the knob at the hind end of the primitive streak
(fig. 36) the three layers are again seen to be closely
combined, the hypoblast being indistinguishable from the
mesoblast, and the epiblast throughout nearly the whole
breadth of the area giving rise to mesoblast cells ; there is also
a much greater mass of the latter layer extending some dis-
tance beyond the limits of the area, which in this region is very
narrow.

The junction of hypoblast and mesoblast does not appear to
oeccur in this region in all specimens, although the embryo

THE DEVELOPMENT OF THE MOLE. 435

from which this section was taken is not singular in exhibiting
such a relation between the two.

The Hypoblastic Mesoblast.

A single layer of rounded hypoblast cells similar to those
represented in section in fig. 30 is present throughout the
lighter shaded anterior portion of the area drawn in fig. 11.
At a somewhat later stage, however, these rounded cells in the
region on each side of the thickened axial hypoblast, in front of
the primitive streak, give rise to cells from which they are
themselves indistinguishable ; gradually the hypoblast situated
anteriorly follows suit, and eventually the whole of that
portion of the area in front of the primitive streak consists of a
plate of epiblast below which lies a mass of cells several layers
deep. These cells are rounded and appear throughout as do
the lateral masses of cells below the epiblast in fig. 33.

Fig. 32 represents a section through the anterior region of
the area drawn in fig. 12,a glance at which will, I think,
prove the origin of these cells from the hypoblast.

It appears to me that the continuity of the intermediate
layer with either of the primary layers is a safe guide as to the
origin of the former —by continuity, I mean such relations as
are shown at the node of Henson (fig. 34), where the bound-
aries of the three layers cannot be distinguished ;—and if this
be true I imagine there can be little doubt.as to the origin of
the mass of cells above described.

At a later period of development these cells become split
up laterally into two layers, a lower single layer of flat-
tened hypoblast and an upper layer of mesoblast several
rows deep. This differentiation takes place from behind
forwards, as does the original formation of this layer,
These relations are seen by comparing figs. 32 and 83; in
the former there is no trace-of a separation of the cells
into hypoblast and mesoblast, while further back (fig. 33),
several cells (hy) along the lower border of the mass are more
flattened, their nuclei more elongated, and they stain more
deeply with hematoxylin than do the remainder of the cells;

A 4 WALTER HBAPE.

these become hypoblast cells. In the axial line no such
change occurs, and the mass of cells existing there. is con-
tinuous behind (by means of the axial rod described on p. 430)
with the front end of the primitive streak, and continuous
laterally with both the hypoblast and the mesoblast. Fig. 42,
although it is a section through a considerably older embryo,
represents these relations fairly accurately.

The axial mass of cells eventually gives rise to the noto-
chord. The lateral mesoblast may be called hypoblastic meso-
blast in accordance with its origin, and to distinguish it from
the mesoblast of the primitive streak. The lateral masses of
hypoblastic mesoblast adjoin posteriorly the mesoblast of the
primitive streak, and it does not appear to me to be possible,
with the existing methods of discrimination, to determine the
exact extent of either layer; roughly, however, we may say
that the front end of the primitive streak is the boundary line.

At the stage of development now reached the embryo may be
compared with that of Amphioxus, as far as its structure is
concerned in front of the primitive streak ; two masses of
mesoblast are formed from the hypoblast laterally and the
axial hypoblast thickens and gives rise to the notochord. The
latter is similar to the median diverticulum of the enteric
cavity of Amphioxus, and the lateral masses of mesoblast to the
mesoblast of the united diverticula on each side in that animal;
the lateral diverticula do not, however, appear, but the median
one is, as we shall see, formed later.

With regard to the embryonic vesicle it is much larger than
in the previous stage, and no longer projects into the mouths
of the uterine glands, but is exceedingly closely applied to the
uterine epithelium, so closely that some of the latter is gene-
rally pulled away from the uterus when the ovum is obtained
whole. Fig. 53 is a section of a portion of the vesicle wall
which is formed of flattened epiblast only, and of the uterine
epithelium to which it is closely adherent.

Historical.—Various accounts have been given by different
observers as to the origin of mcsoblast in mammalian embryos,

THE DEVELOPMENT OF THE MOLE. 435

Beneden (No. 5) describes that portion of the inner mass which
remains after the hypoblast is separated from it, as mesoblast,
and states that it retreats to the hinder end of the embryonic
area, becomes secondarily united with the epiblast, and gives
rise to the mesoblast of the embryo.

Rauber (No. 21), Kolliker (No. 16), Hensen (No. 12), and
Lieberkithn (No. 19), argue that the mesoblast arises first in
the primitive streak. Kdlliker considers that the epiblast
alone gives rise to it, and that after being formed in the primi-
tive streak it spreads, eventually, over the whole embryonic
area,and also supplies the mesoblast of the area opaca. From this
author’s statement I gather he considers the primitive streak
arises first in the end knob (Endwuldst), and extends from
thence forwards.

Lieberkiihn differs from Kolliker, and agrees with Hensen,
in that he derives the mesoblast of the primitive streak from
both epiblast and hypoblast; while Hensen differs from the
other observers mentioned, in considering a certain amount of
the mesoblast of the area opaca to be formed in siti from
hypoblast.

Summary.—My own observations lead me to differ entirely
from Beneden as to the formation of mesoblast,and to agree with
the other observers mentioned above in concluding that it is first
‘formed in the primitive streak. I cannot, however, accept
Kolliker’s statement, that the epiblast is alone responsible for its
production, and that it is first formed in the hind knob. I consider
that the hind knob is formed some time after the first portion
of the primitive streak, and that the formation of the latter
takes place from before backwards instead of from behind
forwards, as this observer states; my main reason for this
being the universal presence at the anterior end of the
primitive streak of an indication of the involution of the
epiblast.

Again, I agree most distinctly with Hensen and Lieberkiihn
in regarding the epiblast and hypoblast as the originators of
mesoblast at the front end of the primitive streak, but I must
differ from them, and from KoOlliker (loc. cit.) and Schifer

436 WALTER HWEAPE.

(No. 26), in believing that the primitive streak mesoblast sup-
plies the whole of the embryonic area.

With regard to this point my results are in entire agreement
with those of Balfour and Deighton, expressed in their account
of the development of the-chick (No. 2), who consider that the
anterior portion of the mesoblast is derived as two lateral
plates from the hypoblast, while the axial hypoblast gives rise
to the notochord.

Further, the similarity of the origin of the epiblast, hypo-
blast, and mesoblast of the embryo and of the notochord in the
mammal is so strikingly similar to the relations of the same
organs in Amphioxus that Kolliker’s (loc. cit.) statements as to
the dissimilarity of the germinal layers of mammals with those
of other animals appears to me to require some modification ;
and Repiachoff’s recently expressed opinions (No. 24) that
there is no homology between the germinal layers of higher
Vertebrata and Amphioxus, receives no support from what is
known of mammalian embryology.

Finally, it is very generally believed that mammals are
descended from animals which possessed a large yolk sac, and
it is stated that the blastodermic vesicle is a remnant of this
yolk sac. If this be true (and as far as we know there seems
to me to be no reason to doubt it), the primitive streak of
mammals is homologous with the same structure in birds, and
the existence of such an arrangement, together with the
presence of a complete neurenteric canal (which I shall
describe later) in the mammal, is ancther instance of the mor-
phological facts which led Balfour (No. 1) to conclude that the
primitive streak was homologous with the true vertebrate
blastopore.

The views as to the relations of the layers at the front end
of the primitive streak will be more advantageously noticed in
the following section.

~

THE DEVELOPMENT OF THE MOLKE. A37

Stage D.

The Medullary Groove, Notochord, and Neurenteria
Canal.

The main differences in the superficial appearance in an
embryo of this stage of growth are:

1. The disappearance of the hind knob of the primitive
streak and the widening out of that portion of the area.

2 The great enlargement of the area in front of the primitive
streak, and

3. The appearance in the latter portion of the area ofa broad,
light-coloured band, the limits of which are at first vague, but
which gradually become more emphasised.

This is the medullary groove which first arises near the
anterior end of the primitive streak, and from, there extends
forwards.

Fig. 13 represents an embryonic area, in which a shallow
medullary groove is formed; the extent of the groove is faintly
indicated in front; at the sides it is more definitely marked
off ; while behind it abuts upon the anterior end of the primi-
tive streak, and terminates abruptly.

At the junction of the medullary groove and primitive streak
a deep pit is visible ; this is the dorsal opening of the neuren-
teric passage, which, however, does not appear completely to
perforate the blastoderm at this stage.

The primitive streak extends from the hind end of the me-
dullary groove to the edge of the blastoderm, spreading out
there into two horns.

In fig. 14 the medullary groove is more distinctly indicated ;
a tongue-shaped band extends anteriorly from the front end of
the groove towards the edge of the blastoderm in that direction,
and is the anterior end of the thickened axial mass of cells
underlying the epiblast.

The primitive streak meets the opposite end of the medullary
groove, and sends a forward prolongation between its divergent
walls.

438 WALTER HEAPE.

There was no indication, as far as I could see from the surface,
of adorsal opening to a neurenteric canal in thisembryo. The
growth from the front end of the primitive streak is similar to
what has already been noticed in younger embryos (fig. 31 and
p. 431).

Fig. 15 is a drawing of the hind end of the medullary groove of
a still olderembryo. Herethe walls of the hind end of the groove,
hitherto widely separate, have joined each other, and have en-
closed within the groove the front end of the primitive streak,
and with it the hinder dorsal opening of the neurenteric passage.
From the front end of the primitive streak a prolongation is
sent forward similar to that seen in fig. 14.

The walls of the groove are now distinct.

The structure of the groove of such an embryo as that drawn
in fig. 13 is represented in transverse section in fig. 43. The
plate of epiblast is thin where it is grooved, on each side,
however, becoming about double the thickness, and then
gradually thinning off until it is only a single layer deep
at the edge of the area; here it is curved upwards, and thus
indicates the commencement of the amnion.

The cells underlying the épiblast in this region are divided
in the same manner as we have seen are those of a much
younger embryo (stage c, p. 434), into (1) lateral plates of
mesoblast and hypoblast, and (2) an axial mass of cells, the com-
mencing notochord showing no differentiation into those
layers.

It will be noticed, however, that the axial cells are consider-
ably more isolated from the lateral masses than heretofore,
although still continuous with the latter. The lateral meso-
blast is thick, and at the edge of the area becomes divided into _
two layers, which are the future somatic and_ splanchnic
mesoblasts.

These relations remain the same throughout the medullary
groove, excepting that at the posterior end the axial noto-
chordal cells become thicker and join a forward growth from
the front end of the primitive streak, while at the anterior end
the groove widens, and all the cells underlying the epiblast

THE DEVELOPMENT OF THE MOLE. 439

come into closer relations with one another. Fig. 42 represents
the latter condition in a somewhat similar embryo.

Anterior to the medullary groove the lateral hypoblast and
mesoblast are not yet separated, and a continuous mass of
undifferentiated cells underlies the epiblast plate.

Behind the groove the primitive streak occasions changes
identical with those already described (p. 438).

The mesoblast throughout the embryo projects beyond the
limits of the area, and is there split into somatic and splanchnic
layers. The relations of the neurenteric canal I will describe in
detail in another place (on p. 440) ; for this specimen I will
only say that at the junction of the primitive streak and me-
dullary groove a deep pit is formed by the involution of the
epiblast in the middle line; the pit is widely open above, but
enters a mass of mesoblast below, and is there, as far as I could
see, entirely obliterated.

The groove now deepens, forcing the notochordal cells under-
lying it further downwards, and in this way the latter, while
remaining connected with the hypoblast, becomes separated
from the lateral masses of mesoblast. Such relations are shown
in fig. 44, which is a transverse section through the medullary
groove of an embryo slightly older than fig. 138, taken from the
same relative position as the section in fig. 43.

This is, however, the deepest portion of the medullary
groove, and only in this section and those immediately on each
side of it do the relations hold which are here figured. Both
anteriorly and posteriorly the groove is more shallow, and the
axial hypoblast is continuous with both lateral mesoblast and
hypoblast.

The structure of the remainder of the embryo is identical
with that described above for fig. 13.

The amnion in this embryo is completely formed over the
hind end of the primitive streak, although not so far advanced
at the front end of the area.

In describing the next embryo I will give an account of the
structure of the neurenteric canal.

The arrangement of the layers at the front end of the primi-

440 WALTER HEAPE,

tive streak and the structure of the neurenteric canal at this
stage of growth will readily be understood by a glance at the
drawing of the surface view of an embryo (fig. 15), and com-
paring it with the diagrammatic longitudinal section in fig. 50
and the transverse sections in figs. 37 to 41. The latter are
taken from an embryo of an age between that of fig. 13 and
fig. 14, and the walls of the medullary canal do not yet enclose
the front end of the primitive streak, although the latter is
already placed between them.

The longitudinal section is taken from a younger embryo.

The dorsal hinder opening of the neurenteric canal (figs. 15
and 37) is formed by an involution of the epiblast in the middle
line at the head end of the primitive streak, and the separation
of the lips of the involuted layer.

The passage so formed enters the mass of mesoblast cells,
budded off from epiblast and hypoblast in this region, as it
does in the embryo of which fig. 13 is a drawing; but it does
not end there, it travels forwards almost parallel to the plane
of the layers, and is seen eight sections further forward (fig. 38)
as a canal within the axial mass of mesoblast, which we have
invariably seen to be projected anteriorly from the front end of
the primitive streak. There is no doubt the cells surrounding
the canal at this point are mesoblast cells; they are continuous
with the epiblast in the middle line and with the lateral
mesoblast, and there is a distinct layer of hypoblast below them
(compare fig. 38, and fig. 50 immediately in front of p. sh’);
gradually, however, the canal dips downwards, and as this
prolonged cord of mesoblast joins anteriorly the axial hypo-
blast, the walls of the canal also there, some sixteen sections in
front of fig. 37, become hypoblastic (vide fig. 39). Here the
lateral mesoblast does not join the thickened axial hypoblast,
which is continuous with the lateral hypoblast only.

Three sections further on the axial cells become continuous
with both lateral mesoblast and hypoblast. The lower wall of
the canal now shows signs of becoming thinner (fig. 40), and
five sections beyond this, that is, twenty-four sections from the
hind opening, it becomes divided in the median line, and the

THE DEVELOPMENT OF THE MOLE, 441

neurenteric canal opens below to the cavity of the vesicle
(fig. 41). The arrangement of the layers at the front end of
the primitive streak may be shortly described, therefore, as
fullows:—The epiblast and hypoblast meet and form the hind
wall of the dorsal opening of the neurenteric canal; from the front
portion of this wall a tongue of mesoblast is projected forwards,
separated from the underlying hypoblast, but united with the
lateral mesoblast and with the epiblast in the middle line; it
then joins the thickened axial hypoblast, and becomes freed
from the lateral mesoblast (fig. 39), while anteriorly to this
point the axial cells are continuous with both lateral hypoblast
and mesoblast.

With regard to the structure of the remainder of the embryo,
the medullary groove is shallow and wide, and throughout its
length the axial hypoblast causes a swelling upwards of the
bottom of the groove. The lateral hypoblast and mesoblast
join the notochordal cells throughout the region where the
latter exist.

The notochord is formed of cubical or columnar cells, and is
alternately in the form of an arch and a tube throughout the
whole length of the medullary groove (vide figs. 40 and 41).
Beyond the groove it becomes more flattened out (fig. 42),
and the arrangement there is similar to that described for
fig. 13.

This tubular form of the notochord appears to be very transi-
tory, as I have not met with it in any other embryo except in
that drawn in fig, 14, in which it is not either so definite or so
continuous.

The sections through the area represented in fig. 14 show a
slightly different arrangement. Fig. 47 is taken through
nearly the same region as are the sections from which figs. 43
and 44 were drawn. The medullary groove is much the same
as is represented in fig. 43, but the notochord is less substan-
tial, and the single row of somewhat cubical cells of which it is
composed form an arch whose cavity opens into that of the
vesicle below. It will be obvious that the thickness of the
notochordal cells is much jess than in fig. 41, and that the arch

442 WALTER HEAPE.

is not so completely tubular. The lateral mesoblast is not
continuous with the axial hypoblast (notochordal) cells in the
middle and deeper portion of the groove, but such is the case
both further forwards and backwards.

Immediately beyond the anterior end of the groove the flat-
tened epiblast is thickened to form the medullary plate, and
below it the arched notochordal cells form a complete tube,
the structure of which is similar to that already described
excepting that the lower wall of the tube is thicker.

Fig. 46 is a section through this region of fig. 14, the dark
streak seen in surface view being accounted for by this thick-
ened mass of notochordal cells; it is the only portion of the
notochord in the anterior region which remains thickened at
this stage of growth. The relations of the layers at the hind
end of the area are the same as are described on p.

At a stage slightly older than that represented in fig. 14
the medullary groove is still deeper. A section (fig. 45) taken
through about the same region as are those drawn in figs. 43
and 44 demonstrates this. The epiblast now exhibits a further
change, that portion of it forming the walls of the groove are for
the most part but slightly thicker than heretofore, but imme-
diately on each side it becomes suddenly considerably thicker,
and then gradually becomes thinner again towards the boundary
of the area, where it is turned up to form the commencing
amnion. \

The mesoblast is here completely separated from the axial
cells, being rounded off at the sides bordering the medullary
groove, and at the edges it is split to form splanchnic and
somatic layers. The hypoblast is continuous across the area,
the axial portion exhibiting no increase in thickness to that
situated laterally ; the former being forced by the deep medul-
lary groove into a bow projecting into the vesicle below.

In front of the point from which the section is taken, the
groove first becomes narrower, and then more shallow and
wider, the notochordal cells becoming at the same time thicker
and continuous with the lateral mesoblast. In front of the
groove the relations are similar to what were described for fig.

THE DEVELOPMENY OF THE MOLE. 4A3

14 (p. 442), excepting that the lateral cells below the epiblast
are not in such numbers as before, and the axial cells, instead
of being in the form of a tube, as in fig. 46, are only one row
deep, and are arranged as an arc, the bay of which opens
below (vide fig. 48).

Behind the section (fig. 45) the groove also becomes shallower,
and the notochordal cells thicker, terminating, as in former
specimens, in the front end of the primitive streak; the latter
is, however, now cut off from the remainder of the streak, and
lies within the medullary groove, into which the dorsal pore of
the neurenteric canal now opens.

The section of the groove in fig. 45 exhibits evidence in both
epiblast and hypoblast of an advanced growth on those drawn
in either figs. 43, 44, or 47, although the measurements of the
area are almost exactly similar to the one represented in fig. 14,
from which fig. 47 was taken.

I do not propose to trace the development beyond this point,
but may briefly say, that after the stage just described the
medullary groove becomes much deeper, the epiblast of its wall
being thick, while the epiblast over the lateral portions of the
embryo is composed of only one layer of cubical cells. The
groove deepens first about the hinder portion of the anterior
third of the groove, and from there extends backwards and
forwards.

The further development of the notochord takes place in the
same direction. The flattened notochordal cells seen in fig. 45
become slightly more rounded as the lateral hypoblast and
mesoblast sink to the level of the bottom of the groove, and
then the lateral hypoblast grows inwards, and a small bunch of
cells are isolated and lie between it and the now closed neural
canal.

The amnion is first formed, as I have stated, over the hind
end of the primitive streak, and from there grows forwards a
considerable distance before the head is covered by the ante-
rior fold.

To recapitulate, we may conclude it is probable that

the region of the area in which the main portion of the

4.44 WALTER HEAPE,

embryo is formed is derived by a forward growth. The hind
knob of the primitive streak is lost, and the pyriform hind end
of the area becomes shortened and widened. The medullary
groove appears first as a wide shallow groove in the region
adjoining the head of the primitive streak, from which point it
extends forwards.

The changes undergone by the axial hypoblast are some-
. what complicated. At first the cells of which it is composed are
numerous, they then become fewer, and are arranged first as a
flattened then as an arched plate, which may or may not be
completely closed in to form a tube. Later on the arched
plate becomes flattened out again by the deepening of the
groove, and the notochord is represented by a thin layer of
flattened cells which, as the lateral mesoblast and hypoblast
sink down to a level with the bottom of the groove, become
again more cubical in form. Eventually the lateral hypoblast
grows in from the sides and the axial cells are separated off as
a notochord.

The isolation of the axial cells from the lateral mesoblast
takes place, as does the separation of the notochord from the
hypoblast, from about the middle of the embryo backwards
and forwards.

The same may be said for the medullary groove, which is
first formed from behind forwards ; its conversion into a canal
takes place from about the middle towards the hind and front
ends. ;

The neurenteric canal is complete, opening at first dorsally at
the head end of the primitive streak, and between the latter
and the medullary groove, but eventually becoming enclosed
within the groove and opening at the bottom of its hinder end.

The canal travels forwards in an anterior growth of meso-
blast from the head of the primitive streak, and enters a thick-
ened axial mass of hypoblast, from which it opens downwards
to the cavity of the vesicle.

The amnion is first formed over the hind end of the embryo,
and only at a considerably later period envelops the front end
by a separate formation.

THE DEVELOPMENT OF THE MOLE, 445

Historical.—The arrangement of the layers at the front end
of the primitive streak has been described by Hensen (No.
12), Schafer (No. 26), and recently by Lieberkiithn (No. 20).

According to the former, the axial cells below the medullary
groove at its posterior end are thickened and join the primitive
streak at the node of Hensen, this portion of the primitive
streak being composed of epi-, hypo-, and mesoblast fused
together.

Schafer describes a similar arrangement in a somewhat
different manner. According to him the axis of the embryo in
this region is ‘‘ occupied by a continuous column of cells, which
inseparably connect the epiblast and hypoblast, and, traced
from behind forwards, would appear to be chiefly epiblastic in
origin.”

This author does not appear to believe that the hypoblast
takes part in the formation of the primitive streak, and he
therefore considers, I imagine, that the latter organ begins
where the hypoblast lies free below the mesoblast.

Neither of these observers described any canal perforating
the blastoderm at this point.

Balfour, however, in his ‘ Comparative Embryology’ (No. 3),
has expressed his belief that the axial cord of cells described
by Schifer is the rudiment of the neurenteric canal of Lacertilia
and birds.

Lieberkiihn agrees with Hensen as to the arrangement of
the layers at the front end of the primitive streak, and further
finds a canal present in the mesoblast, which grows forwards
from the front end of the “node.” He states that the canal
arises in the mesoblast, and does not open dorsally through the
epiblast, but that it is prolonged forwards, and opens below
through the hypoblast.

The notochord he believes to be formed from mesoblast,
which secondarily becomes united with the hypoblast. This
author also compares the neurenteric canal, such as he finds in
mammals, with that of birds and lizards, and declares they are
essentially different, inasmuch as in the latter the canal arises

VOL, XXIII,—NEW SER, GG

446 WALTER HEAPE. -

as an inpushing of the epiblast, and connects the neural tube
with the gastric cavity.

Kolliker (No. 17) agrees with Lieberkiihn as to the meso-
blastic origin of the notochord.

Summary.—My own work indicates that a complete neuren-
teric canal is formed similar to that in birds and lizards, first
of all by an inpushing of the epiblast ; secondly, the canal is
conducted to the hypoblast within a tongue of mesoblast, which
grows from the anterior end of the primitive streak ; thirdly, the
canal enters the axial hypoblast, and opens below to the cavity
of the vesicle; and fourthly, the dorsal opening of the neu-
-renteric canal is eventually enclosed within the walls of the
neural tube.

With regard to the notochord, it appears to me evident that it
is an hypoblastic structure, since it arises from an axial mass
of cells, which are themselves derived from the primitive hypo-
blast.

My observations are at variance with Schafer’s, in that
I find no continuous layer of mesoblast in front of the medullary
groove, such as he describes, but a mass of undifferentiated
cells, whose development shows that they are of hypoblastic
origin, and that they split up laterally into sheets of hypoblast
and mesoblast, while axially they remain undifferentiated, and
give rise to the notochord.

Further, that the differentiation of the mass of cells which
gives rise to the notochord takes place, as does the first forma-
tion of the medullary groove, from behind forwards, but that
the separation of the notochord from the hypoblast takes place
first of all somewhat anterior to the middle of the embryo, in
the region where the medullary groove first deepens, and where
the lateral mesoblast first forms protovertebre, and that from
that point the notochord is separated off both backwards and
forwards.

Finally, that the derivation of the notochord from hypoblast
is still further evidence of the incompatibility of Repiachoff’s
views (loc. cit.) with the facts of development.

THE DEVELOPMENT OF THE MOLE. 447

Comparison between the Early Stages of Development
of the Mole and Mouse, &e.

Until a few months ago there had been no satisfactory ex-
planation of the manner in which the extraordinary phenomenon
of the inversion of the layers in the Guinea-pig had been brought
about, although the fact that such an inversion really existed
had been described many years ago by Bischoff (Nos. 8 and 9),
Reichert (No. 23), and Hensen (No. 12).

Kupffer (No. 18), Selenka (No. 27), and Fraser (No. 10),
have, however, recently worked at the development of the
Field Mouse, House Mouse, and Rat, and have found that the
position of the layers in these animals is also inverted.
Further, they each discovered the method by which the inver-
sion was accomplished, and at the same time Hensen (Nos. 13
and 14) arrived at somewhat similar results for the Guinea-
pig.

From these papers and from that of Spee (No. 28) I gather
it is probable that the fully-segmented ovum of these various
animals is similar to that of the Mole.

The changes which take place after segmentation are, how-
ever, somewhat different in each, and show a gradually
increasing difference from the normal type to that one most
specialised, viz. the Guinea-pig ; while the phenomena exhibited
during the development of the mole supply the connecting link
between the .two types.

These facts have not been, as far as I know, hitherto brought
forward, and I venture to think merit some attention.

In the Field Mouse a blastodermic vesicle of flattened outer
layer cells is formed, at one place on the circumference of which
a solid inner mass is attached.

A layer of hypoblast is formed on the lower side of the inner
mass, and the two shortly after flatten out; a thickening of
the outer layer then takes place above the inner mass, and the
flattened plate, with the hypoblast on its inner side, becomes
involuted within the vesicle, and in this way an arched plate
is formed, the circumference of which rests upon the outer layer

448 WALTER HEAPE.

cells. The cavity of the arch (the secondary cavity) is filled up
more or less with cells derived from the outer layer, and thus
a condition is arrived at remarkably like the stage of develop-
ment in the Mole represented in the woodcut, Fig. B, and on

Plate XXIX, fig. 24.

EXPLANATION oF Woopcurt.

Diagrammatic representation of :—a, ovum of Rabbit; B, of Mole; ©, of Field
Mouse (after Kupffer) ; p, of House mouse (after Selenka) ; n, of Guinea-
pig. A,B,C, and D are at a similar stage of development. 8 is at a
much earlier stage, before the formation of a secondary cavity.
41. cav., blastodermic cavity; ep., epiblast ; hy., hypoblast; ¢. m., inner
mass; 0. /., outer layer; sec. c., secondary cavity; ¢., rudimentary
trager.

There is, however, a difference in the future development.

THE DEVELOPMENT OF THE MOLE. 449

The plate of cells in the Mole flattens out again, while in the
Field Mouse it becomes further involuted within the vesicle,
and the lower middle portion becomes the epiblast of the em-
bryo, while the lateral portions form the amnion (Fig. c). In
this manner the secondary cavity is surrounded by inner mass
and outer layer cells, and into this cavity the embryo projects.

In the common House Mouse a layer of hypoblast is first
formed below the rounded inner mass; next above the latter
the outer layer cells become thickened and involuted within the
vesicle, carrying with them the solid inner mass.

A cavity is subsequently formed in the latter, and it elongates
until it nearly reaches the opposite pole of the vesicle, to which
it was originally placed.

Thus the cells of the inner mass alone line the secondary
cavity in this case, and into it the developing embryo projects
(Fig. D).

The Rat develops similarly to the House Mouse, a secondary
cavity forming in the inner mass after it is involuted.

In the Guinea-pig, however (Fig. ©), the solid inner mass
appears to become attached to the opposite pole of the ovum at
a very early stage in the development of the blastodermic vesi-
cle, and the outer layer does not become involuted, if observa-
tions made by Dr. Wilson and myself be trustworthy, until a
considerably later period. A secondary cavity is eventually
formed within the inner mass, and into it the embryo projects.

Thus a complete series of conditions may be traced in these
various animals between the inverted and normal types. In
the Rabbit the solid inner mass flattens out and remains on
the surface of the vesicle; in the Mole it is first formed into a
curved plate, which subsequently becomes flattened out and lies
on the circumference of the vesicle; in the Field Mouse it
flattens out first on the surface and then becomes and remains
involuted ; in the House Mouse it becomes involuted before
becoming flattened; and in the Guinea-pig the inner mass
remains attached at the opposite pole of the ovum before it
becomes a vesicle, and an involution of the outer layer seconda-
rily takes place.

450 WALTER HBEAPE.

A consideration of these facts, together with an examination
of the conditions attending the later stages of development in
some of these animals, leads me to believe that the difference
in the development of normal animals and those in which the
so-called inversion of the layers takes place is one of secondary
importance, and, in fact, that no such fundamental differences
exist as was supposed by the older observers; the temporary
inversion of the layers which occurs in the Mole connecting
the two types very closely.

I do not propose here to enter into a more detailed discus-
sion of the points noticed above, or to attempt to compare the
later stages of development ; the only points which have imme-
diate bearing upon my present work are—

(1) The explanation of the existence of the secondary cavity
and the cells situated within it in the Mole; and (2) the fact
that in the inverted types the epiblast of the embryo is formed
entirely of inner mass cells.

The former may be considered as inversion phenomena of a
temporary nature; while with regard to the second point the
conditions of development appear to me to be sufficient expla-
nation of the difference.

BIBLIOGRAPHY.

(1) F. M. Batrour.—* A Comparison of the Early Stages in the Develop-
ment of Vertebrates,” ‘Quarterly Journal of Microscopical Science,’
1875.
(2) F. M. Batrour and F, Detenton.—* A Renewed Study on the Germinal
Layers of the Chick,’ ‘ Quarterly Journal of Microscopical Science,’
1882.
(3) F. M. Batrour.—‘ Comparative Embryology,’ vol. ii, 1881.
. (4) Epwarp van BenepEN.—“ La Maturation de l’ceuf, la fécondation, et les
premiéres phases du développement Embryonnaire des Mammiféres
Waprés des recherches faites Chez le Lapin,” ‘ Bulletins de l’Académie
Royale de Belgique,’ 1875.
(5) Epwarp van Benepen.—“ Recherches sur l’embryologie des Mammifeéres,
La formation des feuillets Chez le Lapin,” ‘ Archives de Biologie,’ tome
i, fascl. 1, 1880. ‘

(6) Ta. L. W. Biscuorr.— Entwicklungsgeschichte des Kaninchen Lies,’
Braunschweig, 1842.

THE DEVELOPMENT OF THE MOLE. 451

_(7) Ta. L. W. Biscuorr.— Entwicklungsgeschichte des Hunde Kies,’
Braunschweig, 1845.

(8) Tu. L. W. Biscuorr.—‘ Entwicklungsgeschichte des Meerschweinchens,?
Giessen, 1852.

.(9) Tu. L. W. Brscnorr.—‘ Neue Beobachtungen zur Entwicklungsgeschichte
des Meerschweinchens,’ Miinchen, 1866.

(10) Avex. Frasrr.—< On the Inversion of the Blastodermic Layers in the
Rat and Mouse,”’ ‘ Proceedings of Royal Society,’ 1883.

(11) Watrer Heare.— On the Germinal Layers and Early Development of
the Mole,” ‘Proceedings of the Royal Society,’ 1881.

(12) V. Heysrn.—“ Beobachtungen itiber die Befruchtung und Entwicklung
des Kaninchens und Meerschweinchens,” ‘ Zeitschrift fir Anatomie
und Entwicklungsgeschichte,’ 1875-6.

-(13) V. Hensen.— Ueber die Ableitung der Umkehr der Keimblitter der
Meerschweinchens,” ‘ Verhandlungen des physiologischen Vereins in
Kiel,’ 1882.

(14) V. Henspn.— Hin friihes Stadium des im Uterus des Meerschweinchens
festgewachsenen Kies,” ‘Archive fir Anatomie und Physiologie,’
1883.

.(15) V. Henson.—“ Bemerkungen betreffend die Mittheilungen von Selenka
und Kupffer iiber die Entwicklung der Mause,” ‘ Archive fiir Anatomie
und Physiologie’ (Anatomische Abtheilung), 1883.

(16) A. KottrKer.—‘ Die Entwicklung der Keimblatter des Kaninchens,”
‘Festschrift zur Feier des 300 Jahrigen Bestehens der Julius-Maximi-
lians- Universitat zu Wurzburg,’ Leipzig, 1882.

(17) A. Koru1Ker.— Ueber die Chordahdhle und die Bildung der Chorda
beim Kaninchen,” ‘Sitzungsberichten der Wirzburger Phys.-Med.,
Gesellschaft,’ 1883. :

(18) C. Kuprrer.—“ Das Ei von Arvicola arvalis und die vermeintliche
Umkehr der Keimblatter an demselben,” ‘ Sitzungsberichte der K. b.
Akademie der Wissenschaften zu Miinchen,’ 1882.

(19) N. Lirsperxitun.— Ueber die Keimblatter der Saiugethiere,” Zu der
Funfzigjahrigen doctor-jubelfeier des Herrn Hermann Nasse. Marburg,
1879.

(20) N. Lirperxtun.— Ueber die Chorda bei Saugethieren,” ‘ Archiv fiir
Anatomie und Physiologie’ (Anatomie Abtheilung), 1882.

(21) A. Ravger.— Ueber die Erste Entwicklung des Kaninchens,” ‘Sit-
zungsberichte der Naturforschenden Gesellschaft zu Leipzig,’ 1875.

. (22) A. Ravser.— Noch ein Blastoporus,” ‘ Zoologische Auzeiger,’ No.

134-5, 1883.

452 WALTER HEAPE.

23) C. B. Rricurrt.—‘ Beitrage zur Entwickelungsgeschichte des Meer-
schweinchens,’ Berlin, 1862.

(24) W. Repracnorr.—* Bemerkungen iiber die Keimblatter der Wirbel-
thiere,”’ ‘ Zoologischer Anzeiger,’ No. 134, 1883.

(25) E. A. Scnirer.— Description of a Mammalian Ovum in an Early
Condition of Development,” ‘ Proceedings of the Royal Society,’ 1876.

- (26) EH. A. Scuiver.—“ A Contribution to the History of Development of
the Guinea-pig,”’ ‘ Journal of Anatomy and Physiology,’ 1876-7.

(27) Emit Sevenxa.—“ Keimblatter und Gastrulaform der Maus,” ‘ Biolo-
gisches Centralblatt,’ No. 18, 1882.

(28) Grar. Ferpinanp Sprrzu.—“ Beitrag zur Entwickelungsgeschichte der
friiheren Stadien der Meerschweinchens bis zur Vollendung der Keim-
blase,” ‘ Archiv fiir Anatomie und Physiologie’ (Anatomie Abthei-
lung), 1883. ;

TONGUE OF ORNITHORHYNCHUS PARADOXUS. 453

The Tongue of Ornithorhynchus paradoxus:
the Origin of Taste Bulbs and the parts upon
which they occur.

By

Edward B. Poulton, M.A.,
of Jesus and Keble Colleges, Oxford.

With Plate XXXII.

I am indebted to Professor Moseley for kindly giving me a
very perfect tongue of Ornithorhynchus. This animal has never
been brought alive to Europe, and therefore it must be un-
usually difficult to procure the tissues in a condition favorable
for histological investigation. Professor Moseley obtained the
specimen from which the tongue was taken in 1874, and the
animal was one of those mentioned on page 263 of the ‘ Notes
by a Naturalist on the Challenger.’ The organ was hardened
in chromic acid, and subsequently in spirit, and I found it in
excellent condition during my work upon it last Christmas and
again at Easter. There appears to have been some slight
alteration of the most delicate tissues, due to the time that un-
avoidably elapsed before the organ could be hardened, together
with the heat and jolting of a journey by coach. However,
this change was not so great as to prevent me from arriving
at definite conclusions with regard to these tissues, and it is
very unlikely that the terminal organs of the gustatory nerves
(to which I allude) could have been made out perfectly, except
by work upon the fresh specimen.

General Account of the Tongue.—The size and shape of

454 EDWARD B. POULTON.

the organ are shown in figures 1 and 2, which are drawn of the
natural size. There is an obvious division into an anterior and
a raised posterior part.

The Anterior Division is only free from the floor of the
mouth for about one third of its own length, and therefore the
movements of the tongue must be very limited.

The upper surface is covered anteriorly with large papille
directed backwards, becoming gradually smaller posteriorly,
where little more than a rough appearance can be detected by
the naked eye. The sides and lower surface of the free part
are perfectly smooth, and the large papille terminate abruptly
at the lateral and anterior limits of the upper surface. Poste-
riorly, however, the smaller papilla are continuous on to the
sides of the organ, and appear to be present upon the mucous
membrane of the floor of the mouth. There is no trace of a
raphe in the anterior part.

Posterior Division.—The anterior surface of this division
is continuous with the upper surface of the anterior division, and
the former overhangs the latter, making with it an angle of 60°
(fig. 2). This surface is 11 mm. broad where it joins the an-
terior division, and above this its lateral contours slope upwards,
forwards, and inwards, to the most anterior part of the posterior
division. Here the superior and anterior surfaces are continu-
ous, and at this point are two large horny papille or teeth
(deserving this latter name as much as the maxillary teeth of
Ornithorhynchus, which are also epithelial).

The internal sides of the teeth slope inwards, and appear to
be continuous ; the apices are broad and chisel-like. The bases
are surrounded by small hair-like papille, which spring from
the teeth themselves. Posteriorly the upper surface presents a
small and shallow pit in the middle line, and from the anterior
margin of the pit a linear raphe is continued for a short dis-
tance, disappearing as a slightly-marked groove. A fold is
formed at the junction of the mucous membrane of the floor of
the mouth with the side of the posterior part of the tongue.
This fold first appears at about the middle of the side, and runs
backwards and upwards, and on nearly reaching the pit on the

TONGUE OF ORNITHORHYNCHUS PARADOXUS. 455

upper surface it turns suddenly forwards and is lost. There is
a groove in front of the fold and overshadowed by it. This fold
may be considered as the posterior limit of the whole tongue.
The fold and groove, the pit, and the whole of the upper sur-
face of the posterior division, appear quite smooth (although
really papillate for the most part), while the lateral surfaces
and the anterior overhanging surface appear rough from the
presence of small and generally hair-like papilla. Just behind
the teeth are two deep grooves (figs. 1 and 2, ad 0), directed
obliquely to the long axis of the organ. These contain the
structures which bear the taste-bulbs, but the former cannot
be seen from the surface. There is also another gustatory
structure in each groove in front of the fold (figs. 1 and Q,
pbo.), but in this case the convex surface which bears the
bulbs can be seen. I have omitted to give dimensions in this
general description because the size of the parts described is
given in figures 1 and 2.

Histological Account of the Tongue.—For the pur-
poses of this description it is convenient to divide the organ
into an anterior and posterior part ; but the former in this case
should include the anterior surface of the posterior part and
the horny teeth. These two regions are histologically very dis-
tinct. The anterior region (1) contains exclusively tactile and
mechanical papille, while the posterior region (2) bears the
gustatory structures, together with papille of probably me-
chanical function, and different in structure from many of those
on the anterior region.

1. Tue AnrErIoR Recion.—The upper surface of the tongue
at the tip (aud for about 19 mm. behind the tip in the middle
line, and rather more at the sides) is covered by large papille
(fig. 1), and presents many points of difference from the surface
of the more posterior parts. This region is, therefore, conve-
niently subdivided into (A) an anterior subregion of large
papille (easily recognisable in fig. 1) and (B) a posterior sub-
region of small papille (including the horny teeth).

A. The Anterior Subregion.—The papilla, which form
an irregular fringe (one or two deep) on the sides and front of

4.56 EDWARD B. POULTON.

the papillate surface on the tip of the tongue, project horizontally
or slightly downwards. They have swollen rounded ends and
constricted bases, and the superficial layer of cells is not corni-
fied. They are especially large in front, and resemble the
ordinary type of fungiform papilla. These papille contain
especially large medullated nerves, which are accompanied by
blood-vessels. Behind these papille, in front and at the sides,
occur others of a conical shape, with constricted bases and fine
recurved cornified apices. Their anterior surface is convex, and
the cornified layer of the apex extends downwards upon it for a
short distance, while the posterior surface is less convex, and
the cornified cells descend almost to the superficial epithelium
of the tongue. There are many rows of these papille, and they
become gradually lower, broader,and more scale-like posteriorly,
with a sharp crescentic ridge of cornified cells on their apices, di-
rected transversely to the long axis of the tongue. The arrange-
ment is very regularly imbricate. The layer of cornified cells
is now much thicker, and forms the important part of the an-
terior and posterior surfaces, while the shape gradually assumes
that shown in longitudinal section in fig. 6, and is singularly
like that of the teeth of a rasp. This form passes on into the
posterior subregion. The anterior papillz are about 1 mm. in
height, while posteriorly they decrease till at the limits of the
subregion they are not more than ‘3 mm. in height. The epi-
thelium is very simple in structure, a stratum corneum being
entirely absent, and the superficial cells being fusiform with
distinct nuclei. The layer appears to represent the rete Mal-
pighii only. Occasionally little isolated groups of cornified cells
occur at some depth beneath the surface, and surrounded on
all sides by non-cornified cells. Papillary processes are absent
in front, where the papille are thickly placed, occasionally
present behind, as very long and narrow upgrowths between
the less numerous papillz.

The epithelium of the papille is always penetrated by secon-
dary papillary processes, which are sometimes very long and
narrow. ‘There are traces of the existence of a few hair-like
papilla between those of ordinary type at the posterior limits

TONGUE OF ORNITHORHYNCHUS PARADOXUS. 457

of the region, but they are never numerous or distinct. Con-
nective-tissue corpuscles are often found between the cells of
the epithelium, having intruded from the mucosa; but this
phenomenon is better marked in another part of the tongue
(although otherwise similar), where it will be described in
detail. The smooth epithelium beneath the tip of the tongue
is similar to that just described, with few papillary processes.
Its lowest cells contain pigment granules.

This epithelium must be highly sensitive, as the mucosa
beneath it is richly supplied with tactile end-organs, to be de-
scribed in connection with the papille. This anterior sub-
region is the most glandular part of the tongue, and in the
region of the tip the gland-tubes occupy far more space than
the muscle-fibres (fig. 3). The gland-tubes ramify between
the muscle-fibres in the whole thickness of the organ for a dis-
tance of 5 mm. from the tip. Posteriorly to this point they are
not found at the lower surface, but at the posterior limit of this
subregion they form a layer more than 2 mm. thick beneath
the upper surface, becoming slightly thicker in the posterior
subregion, where they finally disappear about 10 mm. in front
of the junction with the overhanging surface. The gland-tubes,
which end czcally without dilatations, are very large, and take
an independent course for long distances among the muscle-
fibre bundles, not branching frequently. They are not united
together into any distinct gland (figs. 3 and 4). The cells have
suffered a little by the time that elapsed before the organ could
be hardened, but they are easily recognisable as belonging to
the ‘‘ mucous” type of Klein. They are transparent tall
columnar cells, staining very slightly in picro-carmine and
borax carmine, deeply in logwood. The walls of the ducts are
for a short distance composed of several layers of cells con-
tinuous with the lower cells of the superficial epithelium. The
lumen of the duct is very narrow during the passage through
the epithelium to the surface, but it rapidly expands below,
and is at once continuous into a gland-tube of usual struc-
ture. The opening on the surface is very slightly funnel-
shaped. The ducts very commonly run in little groups of three

458 EDWARD B. POULTON.

or four, and penetrate the epithelium close together. The
ducts open freely on the lower surface of the tip over the 5 mm.,
where the gland-tubes occur; while above they open compara-
tively rarely between, but very abundantly on the papille, and
especially towards the lower part of the anterior slope (fig. 4).
At this point the great majority of openings are found. . Pos-
teriorly in this subregion, and in the posterior subregion itself,
the gland-ducts only open between the papillae. It seems very
likely that the papille are rendered sticky by the glandular
secretion, and that the aquatic larve, &c., which form this
animal’s food, are thus caused to adhere to them. It is not
likely, however, that the larve are thus captured. A sticky
secretion would be of little use in the mud at the bottom of
ponds, and the tongue has such limited powers of movement,
and is set so far back (the tip is 13 mm. from the anterior
margin of the lower bill), that it is not probable that it can
even be protruded. The prey is caught by the bill, and the
animal is known to rapidly vibrate the lower bill in the water
like a duck, by which means the mud would be washed
away through its lateral grooves. The prey is thus held by
the ridges between the grooves and the flat surface of the
upper bill, and probably crushed to some extent by the pe-
culiar (and, I believe, undescribed) smooth, ridge-like horny
teeth, which are situated at the inner ends of the grooves (two
in each jaw). During all these processes the food is far from
reaching the posterior part of the singularly inflexible mouth.
Hence the importance of these large, adhesive, and (as will be
shown) highly tactile papille on the anterior part of the
tongue. By their means the food can be drawn backwards to
the more effective and corrugated teeth, to be thoroughly
crushed. Hence the importance of this excessive development
of glandular tissue at this particular part of the tongue.

The imbricated arrangement, and sharp points and ridges of
the papille, would also be of great importance in retaining
small insects, &c., which were caught by the adhesive anterior
papillary slopes. Thus an insect attempting to escape would
be met by the hard corneous posterior surfaces of the papille

TONGUE OF ORNITHORHYNCHUS PARADOXUS. 4.59

in front, whose inclination backwards would prove a further
obstacle. Thus the anterior surfaces are tactile and adhesive,
while the posterior are chiefly of mechanical use.

The tactile end-organs mentioned above form a new ter-
minal organ, apparently nearly allied to Pacinian corpuscles.
The shape is oval or fusiform, and the poles are often slightly
flattened. They are very small, as is seen by the highly mag-
nified fig. 5 (405 diameters). The corpuscle is surrounded by
a laminated investment formed of 6—8 extremely thin con-
centric layers. There was doubtless an intervening fluid
between these capsules during life, as they are now found in a
very collapsed and crumpled condition. The number of cap-
sules is very uniform, together with the general appearance
and size of the corpuscles. Between the capsules occur a few
relatively large oval granular nuclei (fig. 5). As in Pacinian
corpuscles the capsules constitute the chief mass of the bodies.
There is an axial, longitudinally, striated fusiform mass, and
an examination of transverse sections with high powers (;4, oil
immersion of Zeiss) showed that this also contains a central
column of different structure. A single medullated nerve-fibre
terminates in each body, losing its medulla on entrance, the
axis cylinder being continuous with the spindle. The cor-
puscles are extremely common in the papille; it is quite usual
to find three in a section of a single papilla, and more than
once I have seen five (fig. 4). Corresponding to this abund-
ance of end organs the papille are richly supplied with medul-
lated nerve-fibres. The corpuscles always occur close to the
lowest layer of the epithelium, and never any distance below
this. They are never situated in the secondary papillary pro-
cesses although they may be close to the bases of these. The
long axis of the body is nearly always parallel to the lower
surface of the nearest epithelium. Groups of two or three
bodies are very common, and sometimes a nerve-fibre appears
to pass through one body into another, although in some
instances careful examination shows that such is not the case
when two corpuscles are arranged in a line, with their apposed
ends almost in contact.

460 EDWARD B. POULTON.

These tactile end organs are very common beneath the epi-
thelium of the lower surface of the tip, and are rarely found
between the papille of the upper surface of this subregion.
They do not occur in any other part of the tongue, and are
not found in the posterior papille of this subregion where the
cornified layer is much developed and the functions are purely
mechanical. They disappear about 4 mm. from the posterior
limits of the anterior subregion.

It is very probable that there are also nerve endings in the
epithelium of the papille, for it is very common to see fibres
continued from the mucous membrane into the epithelium,
especially at the apex of a secondary papillary process. If
these fibres are nervous (which cannot certainly be made out
in this specimen) they are probably bundles of primitive fibrils.
Blood-vessels are very abundant in the tactile papille.

Bs. The Posterior Subregion.—The transition between
this part of the tongue and that just described is shown in fig.
6. It is seen that the rasping papillz are continuous from the
one on to the other without any change of structure except an
increase in thickness of the corneous epithelium. The gland
tubes where present open between, never on, the papille, and
are wider just before they perforate the epithelium than is the
case anteriorly. No tactile end organs appear to be present,
and the functions of this subregion must be exclusively
mechanical. Between the widely separated papille already
described, occur abundant simple pointed papille (figs. 6 and 7),
which are usually much worn down. In all the points men-
tioned there are traces of a transition from the anterior to the
posterior subregion, but a distinct difference between the two
is seen in the structure of the epithelial layer. At the point
of contact the epithelium of this subregion forms a thinner
layer (fig. 6), and this is also true of all other parts of the
tongue where there is a similar change of structure. This
fact ouly holds for the contact, since there may be great varia-
tion in the thickness of both kinds of epithelium distally to
this point. The structural difference is of greater importance,
and can be detected even in an unstained section of the tongue

TONGUE OF ORNITHORHYNCHUS PARADOXUS. 461

without the use of a lens. The epithelium of the posterior
subregion is thus seen to be far denser and to possess an
obvious division into three layers, of which the median one is
much darker than the other two. The use of staining re-
agents and moderate powers shows far greater complexity, and
the various layers of cells behave very differently with different
reagents. Looking at many results it seems that there are
four distinct layers, best shown by aniline black and picro-car-
mine. Beginning from below the stratum Malpighii (1)
stains light slate colour in aniline black, light red in picrocar-
mine; darker at the lowest layer of cells with both reagents.
The upper fusiform cells of the stratum Malpighii form a layer
(2) staining deeply in both fluids. Then follows a layer (3) of
very long thin cells whose outlines are difficult to make out,
staining yellow in picrocarmine and straw-colour in Aniline
black. The uppermost layer (4) stains deeply for the most
part, and the outlines of the cells become distinct and the form
less attenuated except at the surface. Logwood gives the
same results, but in this case it is possible to see the nuclei of
the cells of layer 3, which are usually long and thin, but
sometimes almost spherical, following the shape of the cells
themselves. In patches of variable extent the cells of layer
3 are continued into 4, not often reaching the surface.
The cells of layer 3 are shown by their behaviour with
reagents to be cornified, and the most remarkable thing about
the epithelium is the fact that in upward succession cornified
cells again become noncornified except in rare and isolated
spots. However, these cells are not cornified to the same
extent as those of the horny teeth and the papilla. This
epithelium is shown in figs. 6 and 7, and it is seen to be
thicker in the overhanging surface (fig. 7), where layer 4 con-
tains more cornified cells in larger patches. In all cases the
cornified layer (3) is continued upwards into the papille of
the subregion, but corneous cells are also derived directly from
layer 4. Close to the point of transition into the anterior
subregion the cells above as well as below 3 stain especially
deeply, and at the junction itself 3 ceases altogether, while
VOL, XXIII, —NEW SER, H H

462 EDWARD B. POULTON.

2, and the deeply stained cells above, coalesce and disappear
after persisting for a short distance (fig. 6).

The oral mucous membrane which is continuous with the
sides of this part of the tongue, has a very thick simple epithe-
lium with fine papillee.

The overhanging surface is almost identical in structure
with the rest of the subregion with which it is continuous, and
the papille are of the same structure. The rasp-like papillae
on the horizontal surface are directed backwards, and these
being continuous on to the overhanging surface (which is
directed forwards at an acute angle) are there directed for-
wards. Hence in antero-posterior movement of one surface
upon the other these papillae would work in opposite directions
and would form very efficient agents of attrition, being greatly
aided by the very numerous pointed papille. That this is the
true function of the subregion is rendered likely by the
situation of these two surfaces at the level of the four most
effective teeth. The food after being crushed by the teeth
would be forced inwards and further rubbed down between
these surfaces, the two horny tooth-like papillae no doubt
assisting in the operation by friction against the horizontal
‘surface below. The greater thickness of the epithelium on the
overhanging surface is doubtless due to the greater length of
horizontal surface rubbed against it.

In favour of this view of the action of the subregion is the
great wear shown by the epithelial surface and the papilla
(especially the more numerous simple pointed ones) ; and the
fact that the two surfaces are almost apposed in a tongue pre-
served in the Oxford Museum, while the horny teeth were
directed downwards towards the surface beneath. In another
‘specimen which I was enabled to examine through the kind-
ness of Professor. Moseley, the teeth were almost in contact
with the horizontal surface below, while in this also the over-
hanging and horizontal surfaces were nearly apposed. It is
probable that the chisel-like summits assist in scraping off
particles that are entangled among the papille. The teeth
and anterior part of the horizontal surface may also be rubbed

TONGUE OF ORNITHORHYNCHUS PARADOXUS. 4.63

against the roof of the mouth, which here is very dense and
presents transverse curved ridges. No glands open upon the
overhanging surface. In working at this surface a singularly
difficult structure was met with which I mention, as it seems
likely that others might have a similar experience. At one
point only, a series of specimens showed successive oblique
sections of a single hair which penetrated to the mucous mem-
brane. The epithelium round it was much modified by the
presence of the hair, so that it was long before I could be sure
that the structure was accidental, A hair (probably one of the
animal’s) must have been arrested at this surface and the end
become inserted in some slight cavity between the cells. Thus
the hair gradually worked through the whole thickness of the
epithelium. It was very interesting to note that the super-
ficial epithelial layer (4) was cornified for a considerable distance
round the foreign object and due to its presence. There was a
hollow at the surface filled with fragments of foreign objects,
and probably partially derived from the hair itself. ‘The con-
tour of the hair was rough and frayed where it passed through
the epithelium, and most of its course was greatly twisted.
These were indications of its gradual passage along lines of
least: resistance, and of the great friction caused.

The two horny teeth which form the boundary between the
anterior and posterior regions (here classified with the former)
are covered, except at their apices, with a thick layer of corni-
fied cells (fig. 7). These cells are, however, different from those
of the corneous layer (3) of the overhanging and other sur-
faces, the difference being especially seen in their behaviour
with logwood. The peculiarity belongs to the latter, and the
superficial cells of the horny teeth are normal cornified cells.
Towards the apex a few large rounded deeply staining cells are
sometimes seen in the cornified layer. They may represent
isolated unaltered cells continued from the layers below, as
they are sometimes seen in lines extending from the apex of a
papillary process. I do not feel certain as to their correct
interpretation.

The layer beneath is made up of very granular polygonal

4.64. EDWARD B. POULTON.

cells which seem to be partially cornified. These are again
transitional into the ordinary fusiform cells of the rete Malpi-
ghii. This extremely granular cell is certainly a transition
into a corneous cell similar to that described in the mechanical
papille of Perameles (see this Journal for January, 1883).
Here, however, the cell is very finely and densely granular,
instead of the coarse type observed in Perameles. ‘There is an
exactly similar transition through polygonal finely granular
cells to corneous cells, even better marked in the maxillary
teeth which I hope to describe on a future occasion. Secon-
dary papillary processes enter the rete Malpighii from below,
and in the axis of the horny papille these processes are long
and fine, and from the summit of each a line of granular cells
extends to the very apex. Thus there is no cornified invest-
ment at the centre of the apex, and that which covers the sides
of the papilla terminates in a sharp-edged corneous ring. The
edges are kept sharp by the constant wear of the softer central
epithelium. There is reason to believe that the maxillary
teeth are rendered uneven in a similar manner, i. e. by the
wear caused in parts by the presence of very long papillary
processes, with lines of soft cells extending from their apices.
Similar lines of cells can be detected in the hard investment of
these lingual teeth as well as in the maxillary teeth (fig. 7, /¢),
but in these cases the cells are completely cornified, and their
arrangement in lines is only recognisable by a looser connexion
between them. Small secondary papillae cover the upper
surface of the base of the two horny papille.

2. Tux Posterior ReGion is covered with a thick simple
epithelium resembling the rete Malpighii. The complex epi-
thelium of the overhanging surface ceases at the rounded angle
which separates it from the side of this region, but below, in
the slight groove which separates the posterior part of the
tongue from the floor of the mouth, the epithelium remains
complex. The same structure is continued upwards along
the shallow groove in front of the fold (f, figs. 1 and 2), while
the fold itself is covered with simple epithelium. Similarly at
the sides thin complex epithelium occurs only at the junction

TONGUE OF ORNITHORHYNCHUS PARADOXUS. 465

of the tongue with the oral floor, while the latter is covered
with a thick simple layer with long papillary processes, exactly
like that of the fold (f, fig. 10). All over the tongue of this
animal there is a great tendency for the subepithelial elements
to penetrate between the cells of the epithelium. This is
especially true of the interpapillary processes of the shallow
groove, between the left gustatory area and the fold. Here in
many cases great masses of connective-tissue corpuscles make
up a considerable bulk of the interpapillary process, as is seen
in fig. 9. Outlying corpuscles have processes which extend
toward the mucosa, indicating the direction from which the
intrusion took place. These have been often described before,
but I believe never to such an extent as is here figured. The
corpuscles never ascend above the lowest layer of the complex
epithelium, corresponding to the rete Malpighii. The fine
pointed papille are always much bent as they pass through the
complex layer, while they are quite straight in the simple
epithelium (compare fig. 8, fp and fig. 10, f’ p’). The complex
epithelium ascends along the groove in front of the fold, and is
continued into the pit (fig. 1). Here it is extremely thin
(08 mm.), while the simple epithelium in front of the pit is
many times as thick (over °dmm.). So also behind the pit the
epithelium becomes simple and comparatively thick. Thus the
convex upper surface and sides of the posterior region, covered
with simple epithelium, are completely surrounded by the
dense complex epithelium prolonged backwards from the lower
part of the overhanging surface, skirting the sides of this
region and rising along the groove until it meets in the pit.
Glands are only found in association with the gustatory areas
and beneath the epithelium of the pit; the former are serous,
while the latter appear to be mucous, and are very numerous,
with few openings into the pit. It is probable that there are
m ore numerous openings posteriorly, beyond the limits of the
tongue in my possession, as I inferred from other specimens
that a groove is continued backward from the pit, ending in a
depression in front of the epiglottis.

The whole of the sides and convex upper surface of this

466 EDWARD B. POULTON.

posterior region are covered with fine hairlike papille (fig. 8,
fp, &c.). These are stoutest and longest at the sides, shortest,
smallest, and most crowded in the raphe (fig. 1). In a hori-
zontal section, taken between the anterior gustatory organs,
I calculated that there are over 500 of these papille to the
square millimetre. This is probably a fair average for the
whole surface.

The roof of the mouth exactly fits this convex part of the
tongue, and the former is covered with dense epithelium,
presenting minute ridges against which the fine papille
must rub.

This is the only part of the tongue upon which gustatory
areas occur, there being two pairs, an anterior and posterior.
The anterior pair (a 00, figs. 1 and 2) are situated on the
convex surface behind the horny teeth. All that can be seen
from the surface is an oblique furrow, but it is shown by
sections that the bottom of this furrow is invaginated upwards
into a ridge which bears the taste-bulbs over the whole of its
circumference (fig. 8). The lips of the furrow are surrounded by
comparatively stout and short papille, of which the axial up-
growths of mucosa are continued from the tissue enclosed between
the superficial epithelium and its prolongation downwards to
form the furrow. The inner walls of the furrow are formed of
corneous cells continued from the papille encircling the
opening. This corneous layer ceases below where the furrow
becomes expanded to contain the ridge. One of the most
interesting things about this whole organ is that the furrow
can almost certainly be closed by the contraction of smooth
muscle-fibres arranged as a sphincter. Smooth muscle-fibres
are very difficult to identify with certainty in a section, but I
have no doubt of their presence and arrangement. So effective
is this closure that I have been entirely unable to detect a sign
of the organ in some specimens, by examination of the surface
with a lens. Of course the stout papilla would meet during
approximation of the lips, and act very effectively in prevent-
ing the entrance of particles. Smooth muscle-fibres are pro-
bably present in the posterior organs, but they cannot be nearly

TONGUE OF ORNITHORHYNCHUS PARADOXUS. 4.67

so effective, nor is there the same necessity for such protection.
Beneath the sphincter are cells with a meridional arrangement
which must act as a dilator muscle (fig. 8).

The posterior gustatory areas (po, figs. 1 and 2) are very
similar, but there is less need for protection in a deep furrow,
because of the posterior position and situation ina slight groove
overhung by a fold. Hence the gustatory ridges rise to the
surface (figs. 10 and 11), and bear some resemblance to an
ordinary circumvallate papilla, but the bulbs are placed on the
upper surface and sides, as in the anterior organ, and unlike
any gustatory area yet described in Mammalia (except the iso-
lated bulbs on the fungiform papille). This description only
applies exactly to the right posterior area, for there was a great
lack of symmetry in this specimen: The left area appeared to
be rudimentary, and was only represented by a slight ridge at the
bottom and rather on the anterior side of a furrow, with few
bulbs, and these often placed beneath the epithelium or only
partially embedded in it. Some, however, were situated nor-
mally, and possessed pores. I am unable to state certainly that
this lack of symmetry is abnormal, but it is very probable that
this is the case, considering that the anterior areas entirely re-
semble each other. The structures accompanying the gusta-
tory ridges, in all cases (even the rudimentary left posterior
ridge), are the same as those of other Mammalia. The serous
gland-ducts open as usual into the spaces round the ridges, and
this type of gland is not found elsewhere in the tongue. The
structure of the gland-cells did not seem to be identical with that
of the usual serous type, but this is probably due to post-mortem
changes, especially as the mucous glands have also undergone
alteration. The ducts of the serous glands sometimes contain
nuclei and the débris of cells. Non-medullated nerve-fibres
almost fill up the centre of the ridge and radiate outwards to
end inthe bulbs. I was surprised to find no indications of
ganglion-cells (as described in Perameles), although minute
ganglia occur on the large nerve-branches. Beneath the pos-
terior gustatory areas a tissue resembling adenoid tissue occurs
in rather large amount (fig. 10), and traces of it can be found

4.68 EDWARD B. POULTON.

beneath the anterior areas. ‘The taste-bulbs are, as far as I
have observed, entirely unique in being developed at the ends
of long papillary processes, up which the nerve-fibres can be
seen streaming from the central nervous mass, sometimes
accompanied by capillary blood-vessels.

Sometimes the external surface shows indications of lobation,
the convexities corresponding to the bulbs; but this is un-
common. In rare cases a papillary process may divide and end
in two bulbs. Gustatory pores are present, and are singularly
like those of normal bulbs. The outline of an exceptionally
long and distinct pore is given in fig. 12, from which its length
and diameter are easily calculated. The ordinary length of
the pores in this animal is not more than half that shown in
fig. 12. I never saw any protrusion of cells or processes from the
pore. There appear to be rather under 500 bulbs to the square
millimetre of surface. The bulbs are oval or fusiform, and their
sides rise gradually from those of the papillary processes, of
which they are the expanded ends. The structure of the wall
is of great importance in the organogeny of the gustatory termi-
nation. At the same time it was singularly difficult to be certain
as to interpretations, owing to changes that had taken place in
these delicate structures. However, after comparing immense
numbers of sections, I can confidently assert that many of the
elements are not modified epithelial cells, but are altogether
subepithelial in origin. There are seldom traces of the me-
ridional arrangement of cells that characterises ordinary bulbs.
The elements also differ in being packed loosely, and in being
very heterogeneous.

In many bulbs I have detected the yellow stain that results
from the disintegration of blood in a capillary. Many of the
cells look as if they might have scaled off the sides of the oval
chamber, and thus have been added to elements intruded from
below. Such cells are fusiform in shape, sometimes thick,
sometimes attenuated, but they always stain differently from
the surrounding epithelium. Other cells have many processes
and resemble connective-tissue corpuscles, although they may
be nerve terminations. Others resemble small multipolar gan-

TONGUE OF ORNITHORHYNCHUS PARADOXUS. 4.69

glion cells, while some are spherical. In all the nuclei are
very distinct, and the elements, as a whole, differ from the sur-
rounding epithelium in staining much more deeply, and in their
loose, irregular arrangement. If the looseness is due to shrink-
ing this proves a difference of structure, as the epithelial cells,
exposed to the same conditions, have not shrunk away from one
another. The nerve-fibres enter the bulb, and do not terminate
in a group at the basal pole in the usual way, but are seen
running between the cells of the bulb in various places.

There is more certainty of the nerve-fibres passing to each
bulb in this case, with the easily found papillary upgrowth as
a guide to the bulb, and containing its special fasciculus of
nerve-fibres. I was enabled by teasing to isolate one undoubted
terminal cell with the nerve-fibril still attached to it. It is
shown in fig. 14, and the fibril is seen to branch before termi-
nation. The shape is very simple and fusiform, with no _peri-
pheral process. There are traces of a nucleus and of an axial
line continued from it along the cell. The general appearance
of a bulb is shown in fig. 13, in which the elements are seen to be
separated by considerable intervals, probably due to shrinkage.
There is no doubt that the epithelium has been penetrated by
the bulb, and not merely reflected over it. This is proved by
the non-continuity of the lowest layer of columnar cells over
the bulb (fig. 13). But this intrusion of the bulb does not
mean, as in other mammals, that the elements are formed from
the modified epithelial cells, although these may be present.
A further conclusive proof that the bulb is essentially subepi-
thelial in nature is found in the fact that isolated bulbs in the
abnormal and rudimentary left posterior organ occur beneath
the epithelium, others partially embedded in it, and others,
again, arranged in the usual manner with gustatory pores.

This peculiar form of bulb has an important bearing upon a
theory as to the origin of taste-bulbs suggested by me in the
January number of this Journal (the “ Tongue of Perameles”).

The Origin of Taste-bulbs.—From observations upon
the tongue of Perameles I was led to infer that the usual
mammalian bulb was developed from a group of interpapillary

470 EDWARD B. POULTON.

epithelial cells. At the same time I concluded that this method
of development was of comparatively recent date in Perameles,
while the singularly complete accesssory apparatus suggested
that some other and more primitive form of terminal organ had
not long been supplanted, and had probably coexisted with
these perfect additional structures. Arguing entirely a priori
the suggestion was made that the primitive type of bulb was
papillary in position and subepithelial in structure, and had
gradually given way to a bulb that was interpapillary and
epithelial.

At the time of this suggestion I had little hope that such a
primitive bulb would ever be seen. It seemed probable that
the stage had existed, but that it was incapable of direct proof.

The very next tongue I worked upon—that of the highly
ancestral Ornithorhynchus—supplied a bulb that was at once
papillary and subepithelial. The new bulb, although in some
cases retaining its original position beneath the epithelium, has
usually ascended and acquired epithelial cells, and has finally
penetrated the surface as a gustatory pore. This latter was a
structure which I had not expected to appear until a later
stage. Nevertheless, in these new bulbs and their arrange-
ment we can see a cause why this should not be the permanent
type.

It is obvious that a subepithelial end organ, specialised for
gustatory stimuli and raised until it is in actual relation with
the exterior through an aperture, must be extremely sensitive.
In fact it is probable enough that such an end organ is too
sensitive for the purpose, especially when continual friction
and the mechanical effects of accidental particles are taken
into account. Evidence of this is seen in the entirely unique
protection afforded to the more exposed anterior gustatory
ridges, a protection which must seriously interfere with their
efficacy. It is therefore probable that a less delicate form of
terminal organ arose, which could be brought into closer rela-
tions with the stimuli. This I believe was the cause of the
change of type, and not increased sensitiveness, except in so
far as this is caused by greater exposure. At the same time

TONGUE OF ORNITHORHYNCHUS PARADOXUS. 471

the bulbs of Ornithorhynchus are, in their arrangement on the
summits of the ridges, more exposed than those of any other
animal. But this is made up for in one case by the sinking of
the whole ridge tillit only communicates with the outside by a
deep and narrow chink (fig. 8), and in the other by the posi-
tion and relation to adjacent structures (fig. 10).

The Origin of the Gustatory areas of Mammalia.—
Omitting the fungiform papille, which seem to be primarily
tactile (as they are here), and to have acquired bulbs compara-
tively recently—the gustatory areas are either of the circum-
vallate or foliate type. The former is by far the commoner
type, but foliate areas are not so rare as is generally supposed.

They were first discovered in rodents, but there are indica-
tions of them in many orders, and I find them well developed
in Marsupials (I have found them in Phalangista with many
furrows whose sides were crowded with bulbs). Thus the two
types appear to have arisen together, as we find them both
represented in the lowest order in which they occur. It seems
to me that the ridges of Ornithorhynchus are intermediate
between the two. In both cases the bulbs become confined to
the sides (changing their mode of origin also) as the areas
become more exposed. A circumvallate papilla is then pro-
duced by the shortening of the ridge until it becomes a sub-
circular elevation. At first the base of the papilla would be
constricted as it now is in Perameles. Then the sides would
become straight, and the vallum very deep and narrow (Pha-
langista), and finally the vallum becomes wide and shallow
and of very little value for protection, as in most higher mam-
mals. Conversely the ridge lengthens, rises to the surface,
and two furrows of a foliate organ are produced, over both
sides of which the terminal organs would spread. Just as the
circumvallate papille of marsupials present traces of this origin,
so their foliate organ (as far as I have seen it) consists of a
less number of furrows than other Mammalia, and with a less
regular arrangement.

As a conclusion to these hypotheses, it is well to remember
that Ornithorhynchus cannot show us the exact ancestral form

472 EDWARD B. POULTON.

of any stage, but it is of immense value in affording sugges-
tions as to what the stage has been. This remarkable animal
is doubtless a direct descendant of a type which would give us
sure knowledge as to the origin of many peculiarly mammalian
features. But individual specialisation has accompanied the
long course of descent, so that even in this lowest of living
mammals, structures which are characteristic of the class, and
of which we might fairly expect to see the origin, are assumed
as it were—used as the raw material for further structural
modification. Upon this subject I hope to write on a future
occasion and to give details; I mention it now to show the
uncertainty of interpretating the origin of a structure from
data of Comparative Anatomy only. And yet such data some-
times afford valuable suggestions, capable of verification, and
often of a kind that could not be given by any other study.

F@TAL MEMBRANES OF OPOSSUM AND OTHER MARSUPIALS. 473

Observations upon the Foetal Membranes of the
Opossum and other Marsupials.

By

Hienry F. Osborn, ScD.,
Assistant Professor of Natural Science, Princeton, U.S.A.

With Plate XXXIII.

In 1834 and 1837 Professor Owen published in the ‘ Philo-
sophical Transactions’ and the ‘ Proceedings of the Zoological
Society’ descriptions of the foetal membrane of the kangaroo.
Quite recently his observations have been confirmed by Pro-
- fessor Chapman, of Philadelphia. They are given in full in
the ‘Comparative Anatomy of the Vertebrates,’ and are to be
found, in abstract, in Balfour’s ‘ Comparative Embryology,’
vol, ii. It is surprising that no additions have been made to
our knowledge of these forms during the long period interven-
ing between Owen’s observation and the present time, even
when one is aware of the extreme difficulty of obtaining females
during the period of gestation.

Professor Owen established the following as the distinctive
features of the Marsupial foetal membranes: (1) A large sub-
zonal membrane with folds fitting into the uterine furrows,
but not adhering to the uterus and without villi.
(2) A large and vascular yolk-sac partly flattened out over the
inner surface of the subzonal membrane, and supplied by an
artery and two veins. (3) An allantois of comparatively small
size, not attached to the subzonal membrane, with a blood

1 ¢ Proc. Acad. Nat. Sciences of Phila.,’? 1881.

4.74. HENRY F. OSBORN.

supply of two arteries and one vein. (4) An amnion closely
investing the embryo and reflected over the base of the yolk-
sac and allantoic stalks.

One sees at once that Professor Owen’s observations, valu-
able as they are, still leave us in doubt as to the real relation-
ship existing between the foetal and maternal blood currents,
which, after all, is the main question. By what process does
the embryo, with little or no food yolk to draw upon, support
life during the short but rapid period! of intrauterine growth,
extending not over seventeen days in the opossum, and thirty-
eight days” in the kangaroo ?

My own observations partly confirm and partly contradict
those of Professor Owen; they show that not only does the
yolk-sac in the Marsupials perform the functions of the allan-
‘tois in the placental Mammals, but that the method is the same,
namely, by means of vascular villi developed upon the
subzonal membrane over the attached or chorionic
portion of the yolk-sac.

In the early part of March I had the good fortune to receive
from one of my students*® a female opossum (Didelphys Vir-
giniana), which was found to be in an early stage of preg-
nancy. After opening the animal I found that each horn of
the uterus had a single swelling an inch and a half long and an
inch in diameter. Upon laying one of these open, eight embryos
were seen, lying in a row, partly enveloped in one or two long
furrows. These furrows would extend along the lower internal
wall of each uterus ; if the animal were in its natural position
they would then be horizontal—a fact the importance of which
will appear later. The foetuses varied considerably in develop-
ment, some being nearly twice as large as others. In the
larger embryos there were two visceral clefts, the foetal circula-
tion was completely established, the fore limb was compara-
tively well developed, with the position of the toes faintly

1 Rev. Dr. Bachman, ‘ Proc. Phila. Acad.,’ April, 1848, p. 46. This writer’s
statements have been confirmed by several observers.

2 Owen, ‘Comp. Anat. of the Verts,’ vol. iii, p. 718.

8 Mr. Robert Speir, of South Orange, N. J.

FETAL MEMBRANES OF OPOSSUM AND OTHER MARSUPIALS. 475

outlined, the hind limb was still bud-like. The tail extended
somewhat beyond the hind limb; the cup of the eye was
backward in development, presenting a horseshoe appearance,
like that of a chick in the third day. Altogether by a com-
parison of the older embryos with some newly born opossums
found upon another female, I conjecture that the embryos were
about eight days old, and that the short period of intrauterine
development was about half over. A fetus of median size was
detached by a slight pressure of the needle, and the subzonal
membrane was found to be about 10 mm., or 2 of an inch in
diameter. Through this membrane the embryo could easily be
seen. There was an opaque disc-like area on the subzonal
membrane, and this was found to correspond to the partially
adherent yolk-sac, which was spread over about one third of
the inner surface of the membrane. When a portion of this
area was seen in profile a large number of minute villi
were at once noticed upon the surface of the subzonal
membrane, which was smooth elsewhere. The yolk-sac, as
in the kangaroos described by Professor Owen, had the figure
of a cone, the base attached to the subzonal membrane and the
apex at the umbilicus. At the edge of the area the yolk-sac
was folded back upon itself, as in fig. 1 (woodcut). The umbi-
lical stalk was wide. The attached area was covered with
capillary vessels, and circumscribed by the sinus terminalis ;
this united near one edge of the disc, to form a single vitelline
vein (Pl. XX XIII, fig. 1), and the vitelline arteries were either
double or branched close to the embryo from a single trank.
They were difficult to distinguish.

The allantois was found in the various embryos in all stages
of development, two of which are represented in Plate XX XIII,
figs. land 4). It arises, as in the Placentalia, just behind the
umbilical stalk, and the mesoblast and hypoblast could be readily
distinguished. In later stages it was a small sac with a wide
stalk. In the embryos which were examined no blood-vessels
could be detected, but they undoubtedly develop at a later
period. Compared with the yolk-sac the allantois was ex-
tremely small, nor was it in contact with the subzonal mem-

476 HENRY F. OSBORN.

brane at any point. Still, no opinion could be formed as
to its subsequent relations, for its development is evidently
very rapid, and the embryos were in an early stage of
growth.

The greatest interest naturally was directed to the villous area
of the subzonal membrane. This could be separated with ease
from the subjacent portion of the yolk-sac, revealing the rich
capillary network of the latter. At this point a careful draw-
ing of the foetus was made (Plate XXXIII, fig. 1), magnified
about five diameters, and representing the proportions as nearly
as could be done by the eye. The membrane was composed of
a single layer of polygonal epithelial cells. When seen from
above the villi appeared as rings of thickened epithelium of all
sizes in profile (fig. 3); they were seen to be composed of a
single layer of columnar cells, some of which were produced
into minute processes. The villi varied considerably in height ;
they were hollow, and beneath them was a layer of flattened
epithelium ; whether the latter was derived from the subzonal
membrane or had been torn off from the yolk-sac could not be
ascertained. A portion of the villous area near the sinus ter-
minalis, containing one of the vitelline arteries and a section
of the vena terminalis (Plate XX XIII, fig. 2), shows that at this
period there was no especial enlargement of the capillary
vessels near the villi; in fact, none of the latter showed any
trace of vascularity. Two facts, however, indicated that this
would appear in a subsequent stage:—1l. The villi were appa-
rently beginning a similar line of development to that which
they pursue over the attached allantoic area in the Placentalia.
2. The villous area in each fetus was in close contact
with the uterine furrow, whereas the remainder of the sub-
zonal membrane floated free in the uterine cavity. The word
“attachment”? would be incorrect in this connection, but the
union with the uterine wall was sufficiently close to prevent
separation, even when there was considerable motion in the
water in which the uterus was placed.

By an unfortunate blunder in the laboratory one horn of the
uterus containing the embyos in sitti was thrown away, so

FETAL MEMBRANES OF OPOSSUM AND OTHER MARSUPIALS, 477

that no satisfactory examination of the adjacent uterine wall
could be made.

Although energetic efforts were made, no other pregnant
females were captured,’ so that my observations upon the
Opossum terminated at this period. I was quite convinced,
however, that older Marsupial embryos would show vascular
villi upon close examination.

The opportunity of completing and confirming the above
results was due to the generous assistance of Professor Wilder,
of Cornell University, and Professor Chapman, of the Jefferson
Medical College of Philadelphia. The former most kindly
placed at my disposal a quantity of Marsupial material, which
he had procured from Australia; the latter allowed an exami-
nation of his valuable kangaroo foetus. To both of these gen-
tlemen I wish to express my hearty acknowledgments.

Among Professor Wilder’s material was a fine foetus, which
will be referred to as Specimen 2, because, although it was
labelled “ Removed from an Australian Marsupial,” the memo-
randum giving the generic name was lost. The foetus was
evidently not that of a kangaroo, but probably belonged to one
of the smaller Australian Marsupials. Its external structure? as
well as the character of its membranes left little doubt of this.

Specimen 2 is drawn natural size in fig. 5, and was believed
to be in a somewhat advanced period of intra-uterine life. The
embryo had well-developed fore limbs, in which the fingers
were all distinct; the hind limbs, although much smaller,
showed the division of the toes. The eyes were in a rudi-
mentary condition, but the ear-pits could be plainly seen, while

1 All writers upon this subject refer to the difficulty of procuring females
during the period of gestation. The Rev. Dr. Bachman (loc. cit.) at one time
in the course of three days procured thirty-five opossums, not one of which was a
female. Audubon mentions a still greater proportion of males. At ordinary
periods the sexes are equally numerous.

Characteristic features of advanced Marsupial embryos are the large size
of the tongue, the disproportion between the fore and hind limbs, the large
mouth, and wide nostrils. In the case of Specimen 2, the subsequent examina-

tion of the kangaroo was further confirmation of the fact of its being a
Marsupial.

VOL. XXIII.——-NEW SER. er

478 HENRY F. OSBORN.

the mouth was large, with a much-protruding tongue. The
tail was quite long. As a whole, the embryo in size and
appearance resembled closely that of the oppossum at birth,
except that the snout was shorter, suggesting that the embryo
belonged to one of the short-faced genera—Chironectis,
Petaurus, Phalangista, or Phascolarctos.

Owing to the rupture of the subzonal membrane, as well as
the removal of a portion of the yolk-sac, the precise relations
of the membranes were difficult to determine. As far as they
could be made out the whole was surrounded by a subzonal
membrane, within which the yolk-sac was flattened out over a
larger area than in the case of the opossum, a fact which was
quite consistent with the advanced age of the embryo. The
missing portion of the yolk-sac was largely within the sinus
terminalis, so that the extent and character of the attachment
of the yolk-sae to the subzonal membrane could not be satis-
factorily ascertained. The latter was carefully examined, and
soon a number of low villi were discovered upon it
without the aid of the glass; they were distributed over an
area to which a highly vascular portion of the yolk-sac was
adherent, which was, however, just without the limits of the
sinus terminalis; what their distribution was within the
limits circumscribed by the sinus terminalis could not be
be followed, owing to the removal of the latter. In fig. 5 their
position is indicated by a number of dots (v) ; as the figure repre-
sents the inner view of the yolk-sac, the villi were of course
upon the lower surface, their position being more plainly shown
in the woodcut (fig. 2,v). The villiare shown in fig. 6 as they
appeared in profile under a low objective. They were con-
siderably lower than the subzonal upgrowths of the opossum,
so that the term villus cannot be given them very accurately.
Upon separating the subzonal layer from the yolk-sac, the
former was seen to be composed of somewhat flattened cells,
which, over the summits of the villi (fig. 6, 5), had a truly
squamous character, being quite transparent. The separation
of the subzonal membrane did not leave the surface of the
yolk-sac smooth asin the opossum, but covered with apparently

FETAL MEMBRANES OF OPOSSUM AND OTHER MARSUPIALS. 479

solid papille, derived from the yolk-sac epithelium. Each of
these was found to be provided with a single dilated capillary
vessel branching over its summit (fig. 6,@). ‘These papille,
therefore, with their subzonal caps, recalled at once the simplest
form of allantoic villi, which Professor Turner represents! as
consisting of a vascular core raised upon the surface of the
allantois and covered with a layer of pavement epithelium de-
rived from the subzonal membrane.

The allantois in Specimen 2 consisted of a highly vascular
sac, with quite a long narrow stalk, which was attached to
the embryo just behind the umbilical stalk ; its distal surface
was covered with capillary vessels ramifying in all directions; the
number of main blood-vessels supplying the allantois was not
ascertained. This allantois was proportionately larger than inthe
advanced kangaroo foetus described by Professor Owen; in other
respects it had the same appearance. A more important differ-
ence was seen in the fact that this sac had a disc-like area of
attachment to what was apparently a portionof the subzonal mem-
brane composed of pavement cells. ‘This feature, if confirmed
by later observations, is a highly important addition to our
knowledge of the foetal membrane of the Marsupials. Unfor-
tunately, owing to the torn condition of the subzonal mem-
brane, this point cannot be considered by any means certainly
established.

The uncertainty which existed as to the generic and specific
character of Specimen 2, and the difficulty of a positive deter-
mination of the relation of its membranes, made an examination
of the kangaroo foetus very desirable. According to Professor
Chapman’s record, the mother was killed fourteen days after
impregnation. The embryo was, however, in an earlier stage
of development than that of the opossum ;* the visceral clefts
were still very distinct; the fore limb was elongated, but the
hind limb was amere bud. The yolk-sac, however, was spread
over the inner surface of the subzonal membrane very much as

1 ¢ Journal of Anatomy and Physiology,’ vol. xi, p. 34.

2 H. C. Chapman, ‘ Proceedings of the Philadelphia Academy,’ 1881, part
ili, p. 469.

480 HENRY F. OSBORN.

Diagrams showing the relations of the foetal membranes.—Fig. 1
represents the actual relations of the membranes at the middle period in
the opossum and the kangaroo. The shaded ring (wé.) represents the wall
of the uterus in section, showing how the villous area of the subzonal
membrane (sz.) is in contact. Fig. 2 shows how in Specimen 2 a portion
of the yolk-sac forms an attached villous area beyond the sinus terminalis ;
the dark villi are those actually observed, the remainder are supposed to
have been present when the membranes were complete. The supposed
adherence of the allantois (a/.) to the subzonal membrane is also shown.
st. Sinus terminalis. am. Amnion. Va. Vitelline artery and veins.
z. Villi of younger specimen. v. Vascular villi of older specimen, the
black ones observed, the unshaded ones inferred,

FETAL MEMBRANES OF OPOSSUM AND OTHER MARSUPIALS. 481

in the older opossum embryos, while the allantois was a small
sac supplied by two arteries.

The line of attachment of the yolk-sac to the subzonal mem-
brane was marked, as described by Professor Chapman, by the
sinus terminalis, and this, as in the opussum, seemed to give
rise to a single vein near the edge of the disc; there was a
single artery. The foetus therefore closely resembled the earlier
opossums, and it was very gratifying to discover minute
villi all over the attached area of the yolk-sac, thus
confirming the. previous observations. These villi were so
minute that it is not at all surprising that they were overlooked
by previous observers. I have not yet had an opportunity
of examining them closely ; their external size and appearance
was similar to those found in Specimen 2, although they were
in an early stage of development. Beneath them the disc
formed by the yolk resembled closely that of the opossum, and
it was quite evident that at a later period they would resemble
in internal structure those found in Specimen 2.

I think we may now regard the following facts, in respect
to the foetal membranes of the Marsupials as fairly well
established.

1. That the yolk-sac at an early stage spreads over the inner
surface of the subzonal membrane, forming a disc-like chorion,
which in the kangaroo and opossum is bounded by the sinus
terminalis. This chorion may become extensive in the later
stages. The subzonal epithelium then gives rise to hollow
conical upgrowths of columnar cells. From the epithelium of
the yolk-sac there arise papilla, which become vascular, while
the subzonal cells become very much flattened. The rudi-
mentary villi thus formed, in the early oposssum and kangaroo
embryos, are thickly distributed over the area surrounded by
the sinus terminalis, but in other forms they may extend beyond
this area.

2. The allantois arises in the same way as in the Placentalia
at quite an early stage of development, and soon becomes vas-
cular. In the kangaroo, if it unites with the subzonal mem-
brane at all, it is only in the later period of gestation. In the

482 - HENRY F. OSBORN.

opossum it develops rapidly, so that a brief union with the
subzonal membrane before birth is not improbable. In the
unknown Marsupial (see Specimen 2) this union seems actually
to have taken place.

3. The amnion, as in the Placentalia, in all cases invests the
embryo.

4. One or two long furrows are formed along the lower
internal border of the uterus in the kangaroo and opossum. In
close contact with one of these in the opossum is placed the
villous chorionic disc of each of the numerous foetuses ; the
remaining portions of the subzonal membrane are free. The
embryo is undoubtedly retained in this position throughout
intra-uterine life. During this period the opossum is known
to keep remarkably quiet, so that the uterus is little disturbed,
and is most of the time ina horizontal position... The presence
of foetal villi is strong evidence by analogy of the presence of
minute crypts on the inner wall of the uterus.

It is an undoubted inference from the above facts that in the
early stages of Marsupial development the vessels of the yolk-
sac not only are the channels for conveying the maternal nu-
triment to the foetus, but that this function is performed by
capillaries distributed in low villi, and separated from the
maternal structures, whatever the arrangement of the latter
may be, by an extremely thin layer of subzonal epithelium. It
is evident that these villi are altogether similar in structure to
those which are found over the allantoic chorion of the pig ;?
the difference is merely one of degree. The rudimentary me-
chanism is sufficient to support the rapid growth of the embryo
opossum, which at birth is completely equipped with all the
necessary respiratory and digestive apparatus acquired during
an intra-uterine period barely exceeding two weeks.? This

1 The fact noticed by several observers, that the females are found in plenty
immediately after the birth of the young, would seem to indicate that they
had been in hiding for some time.

2 See Turner, loc. cit.

3 The feebleness of the young at birth has been exaggerated. The opossum
young weigh from four to five grams, and in their bent position are one half

FETAL MEMBRANES OF OPOSSUM AND OTHER MARSUPIALS. 483

could not be effected if the absorbent villous area were shifting
about from one part of the uterus to another. This fixity of
position must have been an important step towards the
establishment of an allantoic placenta.

Although we may now be reasonably certain of the early con-
dition of the foetal membranes in the Marsupials, it must be
borne in mind that all the latter part of their history is still a
blank, and that the allantois in the later stages may enter into
very important relations with the subzonal membrane. Bal-
four, with his usual discernment, suggested a probable condi-
tion among the primitive Placentalia,! in which the yolk-sac
and allantois shared the placental function. I think it is not
improbable from the evidence given by Specimen 2, and from
the rapid growth of the allantois in the opossum, that this con-
dition may yet be found among the Marsupials. The fact that
no foetal membranes are brought forth at birth has, I believe,
been correctly attributed to the very tortuous vaginal passage
through which the young pass in their descent.

The evolution of the placenta is an interesting subject of
speculation, which it is tempting to review, now that we have
more light upon the functions of the yolk-sac.

(1.) In the low reptilian forms which preceded the Mammalsthere
was undoubtedly a substitution of the viviparous for the ovi-
parous mode of reproduction, by the gradual reduction of the food
yolk and the retention of the embryo inthe uterus. The whole
character of Mammalian development points unquestiunably
to the former presence of a mass of food yolk in the ovum, and
there is every reason to suppose that the loss of this source of
supply was gradually and pari passu compensated by the sub-
stitution of the maternal nutrition, so that the embryos were
partly nourished by the yolk, partly by a feeble absorption of

an inch long. All the bodily functions are fully in action, the fore limbs are
strong and provided with claws, the young are taken in the mouth of the
mother from the valva and placed in the pouch, probably close to one of the
nipples, the grasping of which is instinctive. They will retake the nipples after
removal from the pouch and exposure for several hours.

' Comparative Embryology,’ vol, ii, p. 216.

A484. HENRY F. OSBORN.

nutriment from the uterus through the contiguous umbilical
vessels, the allantois retaining solely its reptilian functions.

(2.) With the diminution of food yolk came an increasing
absorption of maternal nutriment through the chorion of the
yolk-sac, upon which villi gradually appeared. The Marsu-
pials may fall into this or the following class.

(3.) A condition in which the allantois and yolk-sac shared
the placental function.

(4.) The primitive Placentalia (Balfour), in which the yolk-
sac formed a large false chorion, and the allantois formed a
small discoid placenta, and in which the maternal parts were
not deciduous.

I hope during the spring of 1884 to be able to follow out
the membranes of the opossum to the later stages. At present,
owing to the hurried preparation of this paper, some valuable
drawings have been omitted, and the study of the kangaroo
was not so complete as I desired, nor have I been able to refer
to all the authorities upon the subject, as I hope to do in a
later paper.

May 14th, 1883.

Observations on the Genus Pythium (Pringsh.).
By

H. Marshall Ward, ™.A.,
Fellow of Christ’s College, Cambridge, Assistant Lecturer in Botany
at the Owens College, Victoria University.

With Plates XXXIV, XXXV, XXXVI.

Amone the numerous species and genera of fungi which
have become known to science of late, there are perhaps none
more important from a biological point of view — unless
Bacteria be excepted—than the minute, and for a long time
ill-understood organisms comprising the genus Pythium,
founded by Pringsheim in 1858 as a group subordinate to the
Saprolegnie. During some recent investigations which I
have lately made for the purpose of obtaining a clearer insight
into certain obscure processes in the vegetable cell, I had the
good fortune to obtain material which afforded an opportunity
for a study of these plant-devouring fungi, under excep-
tionally favorable circumstances. It appears worth while,
therefore, to describe these observations, not only because they
embrace important facts of general biological interest, but also
because the organisms concerned have been apparently almost
ignored in England.!| Whether they are to be regarded as
morphologically of such importance as they are believed to be
may reinain an open question until they have been further
examined.

1 They are not mentioned, for instance, in Cooke’s ‘ Handbook of British
Fungi;’ and the ‘ Micrographical Dictionary,’ 3rd edition, 1875, has a very
insufficient note on the genus.

VOL. XXIII,—NEW SER, KK

486 H. MARSHALL WARD.

One great difficulty experienced by those who have
attempted to define the taxonomic limits of these and similar
organisms, comprising the Saprolegnize, &c., is to determine
whether they are Fungi or Alge; but since true parasitic
members of the group are now known to infest land plants, as
well as green Alge, we may safely put this difficulty aside
when their want of any trace of chlorophyll or starch-forming
substances are also borne in mind. After all, and the opinion
may be abundantly supported, the exact limitation of the lower
Alge and Fungi is of comparatively small importance, since it
may be considered certain that they pass into one another at
one or more points.

The name Pythium was given by Pringsheim to a group
of Saprolegnia-like organisms, because he found that the
process of formation of the ‘‘swarm-spores” differed in this
newly discovered type! from that of Saprolegnia generally ;
De Bary soon afterwards described new species of the genus,”
and other observers gradually added to the list. For many
years, however, great confusion seems to have existed between
Saprolegnia, Achlya, Pythium, and other groups, and it
was not until a comparatively recent date that something like
order was arrived at. This has been accomplished by con-
tinuous and careful observation of the development of
individual forms, and much is due to the indefatigable labours
of De Bary, whose last monograph® is a model for all mor-
phologists.

Several of the known species of Pythium are parasitic on
living plants, though others appear to be habitually sapro-
phytes. In 1874, however, Hesse discovered a species* which
seems to be almost omnivorous, attacking living and dead
plants of widely different kinds, and which can be grown on
animal substances as well. I shall commence by examining
this species in some detail, not only because it is one of the

1 «Jahrb. f. wiss. Bot.,’ B. i.

2 «Jahrb. f. wiss. Bot.,’ B. ii, and literature quoted below.

3 «Beitr. z. Morph. u. Phys. der Pilze,’ R. iv, 1881.
4 «Pythium De Baryanum, ein Endophytischer Schmarotzer,’ Halle, 1874.

OBSERVATIONS ON THE GENUS PYTHIUM. 487

largest and most vigorous forms, but also on account of the
ease with which it may be obtained and cultivated, and further,
on account of its importance as a parasitic enemy of food and
other plants. It is, moreover, in every sense typical.

Pythium De Baryanum.!

In almost any sowing of ordinary cress (Lepidium
sativum), certain of the seedlings after three or four days of
growth especially if kept moist may be observed to become
weak at that part of the stem which joins the root, and to bend
sharply over. In many cases the general rotting and death
of the seedling follow. An examination of such a diseased
seedling shows the part of the stem nearest the ground to
be ina rotten state. Instead of being turgid, semi-translucent,
and of a pale greenish colour, the affected part is seen to be
much contracted; its cells turn brown or yellow, and become
unfit to support the weight of stem, &c., above it. At fig. 1
is a drawing of a seedling in this condition, straightened out
and with the soil wasbed from the roots. The line may be
taken to illustrate the level of the soil at the region where the
damage occurs.

This rotting tissue, and parts of the stem immediately above
it, are seen under the microscope to be full of very delicate
hyphe, branching in all directions in and between the cells.
Closer examination shows that the thin-walled, cylindrical
hyphe are confined to the parenchymatous tissnes, and avoid
the still young fibro-vascular bundles forming the more central
parts of the hypocotyledonary stem of the seedling. It is in
consequence of the flexible axial portion being no longer
supported by the turgid parenchyma, that the “top-heavy ”
seedling bends sharply over at the injured spot.

More minute examination shows that the hyphe run
through the tissues, especially in a longitudinal direction, and
if observation is particularly directed to the portions which are

1 Literature :—Hesse, loc. cit., 1874; De Bary, ‘ Bot. Zeit.,’ 1881; and
‘ Beitrage z. Morph. u. Phys. d. Pilze,’ 1881, R. iv.

488 H. MARSHALIL WARD.

still only partially injured, the hyphe are seen to bore through
the cell-walls, cross the cell-cavities, and send out secondary
branches in all directions, which repeat the same processes in
their turn.

If such a piece of “infected ” tissue be placed in water in
contact with a healthy seedling of cress for 12—24 hours, the
latter will be found firmly attached to the former by means of
hyphe which have grown across the interval and commenced
to bore their way into the healthy tissues. These penetrating
hyphe not only enter any stomata in the epidermis of the
attacked plant, but make perforations through the cell-walls
as before (fig. 2). Similar events follow if a portion of the
infected tissue be placed on the clean, cut surface of a potato
(fig. 3), or on portions of many other plants. These facts will
be referred to later, when we are concerned with the mode of
action of the hyphe. Once inside a suitable tissue, the myce-
lium makes it way through all the thin walls as already
described. After some hours the tissues thus attacked become
reduced to a mere mass of pulp, and the well-nourished myce-
lium begins to form its reproductive organs, at the same
time developing new ramifications in the surrounding water.
The rapidity of these processes, and the extent to which they
go on, are dependent on a number of conditions, apart from
the nature of the host plant. Amongst these, temperature and
the abundance of oxygen are the chief. All being favorable,
certain branches of the mycelium become swollen at the apex
into pear-shaped bodies, large quantities of protoplasm passing
into them. Each of these, having attained a maximum size
and become nearly globular, becomes cut off by a septum from
the rest of the branch, and persists as a nearly spherical thin-
walled cell (fig. 5).

This is a ‘‘conidium”—a simple dilation of the hypha,
rich in protoplasm, and capable of germinating at once (in
fresh water) after separation from the parent branch. These
conidia are formed in immense quantities at the ends of the
numerous branches of the inycelium; not only free in the
surrounding water, but also in the destroyed tissues (fig. 2).

OBSERVATIONS ON THE GENUS PYTHIUM. 489

In any vigorous cultivation they become formed at an early
Stage, and are scattered around in large numbers as develop-
ment proceeds.

Besides these terminal conidia, however, there usually
appear numbers of interstitial conidia (fig. 4), each of which
arises as a simple swelling on the course of a hypha, which
having received much granular protoplasm, becomes at length
cut off by a septum on either side. In fig. 4 are shown the
various changes of the slowly moving vacuoles, &c., noticed
in such a body, observed for some time at intervals during
development.

Each kind of conidium acts as a simple asexual reproduc-
tive cell. If fresh water containing oxygen be added to the
specimen, the conidia soon put forth processes which develop
forthwith into new extensions of the fungus; this happens
whether the conidium be free or still attached (figs. 5 and 6).
If allowed to remain undisturbed, no germination occurs ; the
terminal conidia drop off and remain dormant, and the intersti-
tial ones become free by the decay of the remnants of hyphe on
either side. Under proper conditions their vitality is main-
tained for months,! ready to be called forth in a few hours
when fresh water is added ; the older conidia usually show a
better developed “ exospore” than those which have not been
kept.

Germination consists simply in the extrusion of the ‘“ endo-
spore ’’ into a simple tube (from one or two points) into which
the protoplasmic contents pass, until all is used up in the for-
mation of the germinal tube; the latter grows quickly by
apical growth, enters a suitable nidus by boring through the
cell-walls, or, if none such is present, soon decays. It is note-
worthy that the formation of septa originates when the repro-
ductive organs commence to be developed ; in its young stages,
the mycelium, though copiously branched, consists of a con-
tinuous series of tubes.

1 De Bary says that drought and frost are withstood by these conidia,
* Bot. Zeit.,’ 1881, No. 33, p. 524.

490 H. MARSHALL WARD.

Hesse‘ and De Bary? both describe the formation of zoo-
sporangia in this species ; I have not succeeded in obtaining
zoospores, and can only refer to their descriptions. The
zoosporangium arises, according to these observers, in
exactly the same way as a terminal conidium, and is not to
be distinguished from it by outward characters, until its
further development. This takes place by the lateral outgrowth
of its wall into a kind of beak, into which the contents pass ;
the end then swells up quickly into a gelatinous vesicle which
receives the protoplasm, and this at once becomes divided up
into zoospores. These processes, as figured by Hesse, are
very similar to what occurs in Pythium proliferum, in
which form I have studied them very carefully ; nevertheless the
differences are sufficient to distinguish them. De Bary points
out that the conidia only produce zoospores if sown at once,®
and immediately on separation, in fresh water. My failure
to obtain the zoospores may possibly be attributed to this.
Besides the asexual forms of reproductive bodies, kowever, this
species possesses sexual organs of a simple and typical char-
acter, and from the comparative ease and certainty with which
the process of fertilisation may be studied in these fungi, an
accurate knowledge of the essentials of the sexual process can
be hoped for with some success.

In a well nourished cultivation of Pythium de Baryanum,
_ the appearance of the sexual organs usually occurs in from
forty-eight to sixty hours at latest, and enormous numbers
sometimes arise, after the crop of conidia has considerably
advanced. If a piece of cress-seedling, thoroughly infected
with the fungus, be placed in fresh water in a watch-glass or
open vessel, the oogonia and antheridia are commonly pro-
duced in a few hours: a favourable portion attached to the
lower side of a thin cover-slip in a suspended drop of water,
kept in a damp chamber—allows every step of the development

1 Toc, cit:, p. 19.

2 “Bot. Zeit.,’ 1881, p. 523.

* “Zoosporen sah ich diese Conidien nach mehr als héchstens wenige Tage
langer Aufbewahrung nie bilden,” ‘ Bot. Zeit.,’ 1881, p. 524.

OBSERVATIONS ON THE GENUS PYTHIUM. 491

of sexual organs, and of the sexual act to be followed on one
specimen.! I will describe the different phases as observed on
such a cultivated example.

From a branch which had developed into the surrounding
water, an apical swelling was formed exactly as for a terminal
conidium: much dense granular protoplasm accumulated in
this, and then a septum formed below. Soon afterwards,
a lateral protuberance arose from the hypha immediately
below the septum; this rapidly developed as a somewhat club-
shaped branchlet, also filled with protoplasm, which curved
upwards towards the large spherical terminal body. The first
formed, conidium-like sphere is the oogonium; the smaller,
clavate body, the antheridium. In the case described—a
very common one in this species—the antheridium, having
become separated by a septum from the common parent hypha,
pushed aside the oogonium as its apex came in contact
with it (fig. 9). The antheridium does not, however, al-
ways arise immediately beneath the oogonium; it may even
spring from a different branchlet (fig. 8), and other cases
occur.

On an oogonium favourably situated for examination I
made the observations illustrated at fig. 10. This specimen was
continuously watched from a little before eight o’clock in the
morning till three o’clock in the afternoon, drawings being
made at intervals, when the progress of events was marked by
changes of special interest or importance.

At 8.15 a.m. the oogonium and antheridium had been
completely formed, and in contact for some time, the apex of
the latter being closely attached to the oogonium wall, and
having already comnienced to senda short process into it. The
contents of the antheridium were bright and less dense than
those of the oogonium, with several large brilliant granules
scattered inside; a firm septum marked off this upper part
of the antheridium from the rest. At a period just preceding

1 In many cases the mycelium grows down, across the cavity of the damp
chamber, and, carrying water with it, spreads on the glass slip below: excel-
lent preparations of the sexual organs in all stages may thus be obtained.

492 H. MARSHALL WARD.

the hour named, the contents of the oogonium were coarsely
granular, and marked by many small oily particles, giving the
whole a yellowish-grey appearance. At 8.15 these contents
were beginning to contract towards the centre of the oogonium,
strings and bands of protoplasm being left attached to the
inner wall. This contraction occurred slowly, but was dis-
tinctly attended with amoeboid movements. During the next
three quarters of an hour this withdrawal of the coarsely
granular fatty protoplasm slowly continued, and at 9 a.m.
the condition of affairs was that figured at fig. 10, d. The
beak-like process (which had already commenced to be formed
in a), sent by the antheridium through the oogonium
wall, now became more distinctly evident; the protoplasm
inside the antheridium also seemed to me to have become
paler and more transparent.

The protoplasm of the oogonium was still anchored by
radiating threads to the walls, and seemed to contain fatty
globules ; its slow amceboid movements still continued. These
‘movements became still more decided during the next hour;
and at 10 o’clock (fig. 10, c) the central mass of the oogo-
nium had become almost spherical and free from the walls.
Its fatty and granular contents were also arranged into rather
angular blocks, formed by the gradual flowing together of the
smaller globules and granules. In this condition the naked mass
may be looked upon as an egg, or oosphere, ready for fertili-
sation; for, although the tube from the antheridium extended
right through to the mass in question, I could at this time
detect no passage of substance through it. The brilliant, refrac-
tive granules in the body of the antheridium were observed
to be distinctly undergoing slow changes of position, however,
and the amceboid movements of the oosphere were carrying
it round the inside of the oogonium. That this was a
period of activity, or excitement, so to speak, preceding the
passage of the contents of the antheridium through the
tube into the oosphere, was amply demonstrated by what
followed. Another point, which became clearer afterwards,
was noticed: the oosphere did not comprise the whole of the

OBSERVATIONS ON THE GENUS PYTHIUM. 4.93

protoplasmic contents of the oogonium, part being left
between it and the walls.

Shortly after the stage just described—the period of excite-
ment, so to speak—culminated in the passage of the contents of
the antheridium-tube towards the oosphere, and by 10.55 (cf.
fig. 10, d) this process was fairly commenced. Close and careful
observation of the contents of the antheridium and its “ fer-
tilising tube”’ convinced me that the large, brilliant granules
were gradually being carried through the tube into the
oosphere by means of a slow, more or less continuous, current
of protoplasm.

There is not the slightest doubt of the accuracy of this ob-
servation. A particular granule just about to enter the tube
at 10.55 (fig. 10, d) was observed to pass slowly into the tube,
and disappear at the other end (fig. 10, e) in less than five
minutes. The motion was not rapid, but consisted of a gradual,
steady streaming, as a comparison Of figs. 10, d to f, shows. In
some cases the granules appeared to melt away in the tube.
Meanwhile, the granules in the antheridium were slowly
accumulating at its upper part, each to be carried down the
tube in turn. At 11.10 (fig. 10, f) two distinct vacuoles, sepa-
rated by streaming protoplasm, appeared in the antheridium;
and by 11.50, three or four others had been formed. At this
time, also, nearly all the remaining bright granules were
aggregated at the entrance of the tube (fig. 10, g), and a com-
paratively rapid passage of these through the tube occurred
during the next five minutes (cf. fig. 10, g and A). The slow
revolving motion of the oosphere was still taking place; but the
tube was plunged further into the substance of the oosphere
at 11.30, for instance, than at 11.10, as shown in figs. 10, g
and f respectively.

At 11.45 the last three of the large granules began to pass
over (fig. 10, 2) in the final flow of protoplasm, and the vacuoles
now became very large. The remaining hyaline protoplasm
was still, however, slowly streaming towards the mouth of the
fertilising tube, and much of it passed through during the next
three quarters of an hour, the quantity left in the antheri-

494 H. MARSHALL WARD.

dium at 12.30 being very small (fig. 10, 4). The oosphere,
during the last-named stages, slowly came to the centre of the
oogonium (fig. 10, &), and then ceased to revolve. Mean-
while a very delicate skin had been formed over the now
smooth exterior of the oosphere. This was first detected
about 11.50 (fig. 10, 7), but had become much more evident
at 12.30 (fig. 10, &). At 3 o’clock p.m. the antheridium
contained practically nothing except a trace of granular matter ;
all its remaining protoplasm had passed over into the
oosphere, thus changing it into an oospore. At this hour,
tov, the oospore—as it must now be termed—had become
clothed with a thick envelope. The process of fertilisation was
completed, and the oospore was “ripe” (fig. 10, 7). A com-
parison of figs. & and 7 shows that the protoplasm which
persisted between the oosphere and the oogonium wall at
12.30 had entirely disappeared at 3 o’clock. There can be no
doubt that it became used up to form the thick envelope formed
in the interval around the oospore. Long after the completion
of the latter, the empty antheridium and tube can still be re-
cognised, though the ripe oos pore lies loosely in the oogonium.
In such condition the oospores remain resting for months, as
can be seen in old material (figs. 6 and 7).

It was already known, from the researches of Hesse, that P.
De Baryanum attacks many different kinds of living and
dead plants, and it seems certain that the mycelium and
spores of this fungus are very wide spread in European soils.
Although this fungus is so omnivorous, however, there are
some plants Which it apparently refuses to attack, e.g. fresh
water Alge. Hesse observed it on Trifolium, Spergula,
Panicum, and Zea, besides cultivating it on Camelina and
Lepidium seedlings ; on the other hand, he found that nume-
rous allied plants—among others Brassica, Pisum, Hor-
deum—refused to be infected by P. De Baryanum. Hesse
also failed to cultivate it on Potato seedlings.

The tuber of this latter plant, however, is in reality a very
good medium for the cultivation of the fungus, and my
researches go to show that Hesse’s list of favorable host

OBSERVATIONS ON THE GENUS PYTHIUM. 495

species may be largely extended. I have cultivated this
Pythium successfully on the young buds of Carrot, on cut
slices of Potato and Dahlia tubes, and on the stems of ordinary
greenhouse Pelargoniums.

I do not here propose to go very fully into the details of
these cultivations, since it seems probable that prolonged
research will yield facts of more general significance than may
be safely stated at present. The following observations, how-
ever, are true so far. The mycelium of the Pythium confines
its ravages to the parenchyma cells of the seedlings, stems,
and tubers, and, so far as I could discover, never enters or
crosses a fibro-vascular bundle; in well developed specimens,
however, vigorous branches envelop the vascular bundles, and
possibly obtain nourishment from the young sieve tubes. The
conidia and sexual organs become formed in any invaded
parts of the parenchyma, as well as on the extra-matrical
branches of the mycelium. The mode of action of the myce-
lium seems to be always the same. It consists, speaking
generally, in the absorption of the dissolved materials of the
cell contents, leaving behind a series of empty bags enveloped
—e.g. in the case of a cress-seedling—by the common cuticle ;
these remnants become more gradually the prey of Bacteria
and saprophytes.

Although the above statement is true so far as it goes, there
are some details of importance in the modus operandi of the
parasite, which I can only touch upon here, since I hope
to obtain more information during the course of experiments
now in progress.

In cultivating this mycelium on slices of Potato and Dahlia
tubers, for instance, I noticed particularly that the starch
granules of the former, and the inulin of the latter are not
directly attacked by the mycelium. This is very clear in the
case of the Potato; the starch grains remain intact long after
the hyphe have destroyed all other cell contents. In the
Dahlia, the difficulty consists in deciding whether the inulin
sphere-crystals (precipitated by alcohol in the usual manner)
have diminished in the invaded cells ; I think that such is not

4.96 H. MARSHALL WARD.

the case. Long after the cells had become traversed by the
hyphee, I was able to obtain sphere-crystals of inulin in quan-
tities which seemed to me not less than in cells which were
still uninjured.

Again, in the stems of pelargoniums cut in November,
and into which the Pythium was allowed to grow, the starch
granules could be recognised even after the Pythium had
formed its conidia and oogonia.

It seems highly probable from these observations that P y-
thium De Baryanum—not to extend the generalization too
far at present—is unable to dissolve starch grains unaided; on
the other hand, it is quite certain that it absorbs material from
the cell contents.

My observations so far have led to the following conclusion
as to the mode and sequence of action of this parasite on the
cells of the Potato and Dahlia tubers. Certain portions of
the protoplasm and cell sap are directly attacked, and absorbed
immediately, but of course we cannot say unchanged. The
hypha grows at the expense of these, enters another cell, and
leaves the starch-grains, part of the protoplasm, and the
nucleus untouched.

During these processes the walls of the attacked cells turn
brown, especially where the hyphal tube entered the cell, and the
dead granular remains of the protoplasm and nucleus soon
acquire a similar yellow-brown colour. Starch grains may
often be observed embedded as in a matrix of these yellow-
brown granular remnants, and,if detached from it, one notices the
cavity from which the starch grain has fallen as from a mould.
I have observed a similar phenomenon in potatoes invaded by
Phytophthora infestans, and Prof. De Bary had noticed it
in this case also. It is well known that this Pythium grows
better in very young cress-seedlings than in those more
advanced, and this seems to be in accordance with what is
stated above; unless it is assumed that the more developed
cell-wall of the advanced seedlings simply prove formidable
barriers to the progress of the hyphe, a supposition which does
not seem to cover all the facts.

OBSERVATIONS ON THE GENUS PYTHIUM. 497

How is it then that at a late stage in the development of
such a fungus as this Pythium, the starch grains and all
traces of cell contents—even cell walls—disappear? Without
entering into details, it appears at least highly probable that
the remaining changes in the cell contents are effected by
Bacteria, carried into the invaded tissues by the hyphe of
the Pythium; that these Bacteria reduce the rest of the
protoplasm and nucleus first to a soluble mass, and then cause
the dissolution of the starch grains. But facts point to the
possibility that the action of the Bacteria in such cases is
taken advantage of by the fungus, and that it is not till a very
late period that the mycelium of the latter suffers from the
dominance of the former and eventually becomes in part a prey
to its companion, having meanwhile formed its well protected
oospores and conidia, which lie unhurt among the rotting
débris.

Pythium proliferum.!

The next example of this genus, the life-history of which I
have thoroughly examined, is P. proliferum, discovered by
De Bary on decomposing insects in water about 1860. In
many respects, especially in regard to the sexual organs, &c.,
it resembles P. De Baryanum so closely as to be hardly dis-
tinguishable from it; important differences indeed. can hardly
be said to exist. It occurs, however, only as a saprophyte,
and all attempts to grow it on living plants have failed, though
its cultivation on dead cress-seedlings—the seedlings being
killed by plunging into boiling water—is very easy. Perhaps
the best distinctive character is to be found in the zoospo-
rangia, which are formed in large numbers and retain their
power of forming zoospores for long periods, and are pecu-
liarly beaked. The mycelium resembles that of P. De Bary-
anum in all essential respects, ramifying in the dead tissues
and finally putting forth free branches, which rapidly spread

1 ¢Pringsh. Jahrb. f. wiss. Bot.,’ vol. ii, p. 182; ‘ Beitrage z. Morph. u.
Phys. d. Pilze,’ 1881, R. iv; ‘ Bot. Zeit., Sevt., 1881, p. 558.

498 H. MARSHALL WARD.

into the surrounding water and form racemose systems of
branches; the ends of the main, secondary, and tertiary branches
then swell up as before into pyriform or oval zoosporangia,
each of which becomes cut off by a septum, and then commonly
puts forth a short necklike process or beak before it separates
and falls from the plant (fig. 11). The formation of this beak
may, however, be deferred until after the dormant period.
The branch, at the end of which such a sporangium hasarisen,
then frequently puts forth a lateral branchlet beneath the
sporangium, the latter thus appearing to be placed laterally
on the hypha (fig. 11).

The fallen zoosporangium may either remain dormant for
weeks or months or germinate at once; its behaviour in this
respect apparently depending simply on the fitness of the
environment: germination, or the formation of zoospores, is
not, however, necessarily ‘preceded by the falling of the
sporangium, as will be seen in examples to be studied imme-
diately.

In fig. 111 are depicted zoosporangia which had been kept
dormant many months in a cool cellar; the thick outer wall,
short beak-like prgcess, and large central vacuole are character-
istic. Mingled with these one often observes oospores and
empty zoosporangia (fig. 12), the necks of the latter having
become prolonged and open to admit of the emission of the
contents; soon after adding fresh water to sporangia in the
condition shown at fig. 111, very many of them become thus
emptied of their contents. Others, however, instead of be-
coming thus emptied germinate in the ordinary manner by
throwing out a simple tube as in P. De Baryanum. I have
carefully observed both processes.

The zoosporangium (fig. 13 a) was drawn at 12.30, several
hours after the addition of pure oxygenated water; from the
condition shown in fig. 11! the contents had not much changed
excepting that the vacuole had become broken up into several
smaller ones distributed in the active granular protoplasm; the
beak, which was not formed immediately after the separation
of the sporangium, now commenced to appear as a faint
papilla, laterally situated near the point by which the spo-

OBSERVATIONS ON THE GENUS PYTHIUM. 499

rangium was formerly attached to the parent branch. At
1 o’clock the beak was much longer (fig. 13, 0) and the vacu-
oles had entirely disappeared; and at 2 p.m. (fig. 13, ¢) the
whole contents had become evacuated as zoospores, leaving
the empty beak and case behind. Fig. 14 gives the results of
the examination of another example. The ripe zoosporan-
gium presented at 4.30 p.m. the appearance shown in fig. 14, a.
At 5.5 p.m. the beak was considerably advanced in develop-
ment (0). Soon afterwards the motion of the protoplasm was
much more evident, and a peculiar stage was passed through,
during which the contents partially divided up and again
became granular. At 5.25 p.m. the soft end of the beak gave
way, apparently to the pressure from within, and the contents
flowed out (fig. 14, c) and immediately became divided up into
five actively amoeboid masses, which were soon afterwards
further separated in the jelly-like envelope; during the next
few minutes each acquired a vacuole and two lateral cilia, and
at 0.39 (fig. 14, d) were transformed into five rapidly struggling
zoospores, moving in jerks, and changing the form of their
amoeba-like bodies continuously. At 5.38 they were all free,
escaping rapidly from their mucous prison and swimming about
as reniform zoospores of the well-known type.

Before proceeding to describe the other mode of behaviour of
these bodies, I may record the observations registered in
fig. 16, on the development of the zoospore itself. The
zoosporangium in this case was in the condition shown at
fig. 16, a at 5 o’clock, having lain for several hours in fresh
water. As there shown, the granular contents had become
excavated by vacuoles of various sizes, and a prominent, firm
beak was developed. The vacuoles had not been long formed
when the figure was drawn, and were rapidly changing their
sizes, numbers, and position, as the granular protoplasm became
churned-up, so to speak. The rate at which the changes were
proceeding at this period may be estimated by comparing
fig. 16, a, b and e, all of which were registered within five
minutes.! The condition shown in J was reached in three

1 The plan I pursued in making these drawings was chosen after several

500 H. MARSHALL WARD.

minutes. Two minutes later—i.e. at 5.5 o’clock—the vacuoles
had almost disappeared, a number of minute bright points,
slowly playing in the granular contents, probably representing
them. At this stage, also (c), the tip of the beak became
pale, diffluent, and began to protrude like a gelatinous drop.
Within five minutes later a large clear vacuole appeared in the
protoplasm at the end of the sporangium opposite the beak (d),
and the pale swelling at the apex of the beak, suddenly
began to be inflated like a blown-up bladder. This con-
dition, at 5.10, ushered in the rapid changes depicted in fig. 16,
e toh, allof which occurred within two minutes. The softened
apex of the beak (d) became rapidly distended into a vesicle, into
which the granular protoplasm flowed smoothly and continu-
ously, evidently impelled from behind by the pressure of fluid in
the vacuoles. These vacuoles no doubt contained some soluble
material, excreted by the protoplasm, and an osmotic pressure
was thus established. The details are accurately figured as
they were observed: the rapid flow of the granules through
the axis of the beak (e), the distension of the pale (cellulose?)
apex into a larger and larger vesicle, becoming more and more
tenuous as the contents flowed in (f and g), being very con-
spicuous. At length—at 5.12—the last granules passed slowly
up the axis of the beak, and the former sporangium re-
mained as an exhausted case, in the cavity of which remained
a few minute granules, and a slight residue on the inner
walls, no doubt representing in part excreted material. The
walls of the emptied sporangium collapsed a very little,
and a large number of minute Bacteria could be observed
attached to the outer surface in all cases (k). Even as the
last granules passed slowly up the axis of the beak (A),
the slowly writhing mass of protoplasm began to divide up

different trials: since the granular protoplasm, outer walls of the zoo-
sporangium, and the main part of the beak are practically constant in appear-
ance, I drew a large number of outlines, and left the granules to be filled in
later. My attention could thus be concentrated on the numbers, sizes, and
positions, &c., of the vacuoles, zoospores, and other details ; and these are
accurately represented in all respects.

OBSERVATIONS ON THE GENUS PYTHIUM. 501

into separate blocks. This proceeded very rapidly to the
isolation of the blocks as zoospores. In three minutes
the stage shown at (7) was reached, the individual ameboid
masses becoming quite active at 5.15, tumbling and rolling over
one another meanwhile in a most comical manner. At 5.20
their movements became more active, and the cilia appeared
(cf. fig. 14); and at 5.25 they were vigorously moving in the
extremely tenuous vesicle, the lashing of the two lateral cilia
becoming more and more rapid. One minute later, and the
vesicle burst suddenly, the active zoospores flitting off at
once in all directions. A distinct remnant of the lower third
of the vesicle remained attached to the apex of the widely open
beak (4). The upper parts appeared to become completely
dissolved in the water.

As an illustration of the other mode of behaviour of the
zoosporangia, the various stages shown in fig. 15 may suffice.
Two zoosporangia, which had remained dormant for many
months in a cool cellar, were each observed to put forth the
pale swelling at the apex of the beak, as described above. In-
stead of forming the vesicular swelling and zoospores, how-
ever, the pale apex became prolonged into a tube, the vacuoles
in the sporangium increasing meanwhile as the contents
passed slowly forwards. To give an idea of the rate of growth
of such a germinal tube, the changes at the apex were observed,
as figured (c toz). The germinating sporangium or coni-
dium (fig. 15¢) was drawn at 2.50; at 3.2 the apex had
swollen and perceptibly elongated (d); fig. e represents the
condition four minutes later; f, after another two minutes; g,
after four more minutes—7z.e., at 3.12 o’clock ; 4 was drawn at
3.20; and z, at four o’clock. Although it may be convenient
to distinguish these germinating bodies as conidia, it cannot
be maintained that any perceptible differences between them
and zoosporangia are observable until germination occurs.
Whether the behaviour depends on internal or external influ-
ences cannot be decided at present, though much may perhaps
be said for the latter view.

As already stated, the zoosporangia may germinate either

VOL, XXIII,—NEW SER, LL

502 H. MARSHALL WARD.

forthwith, after separation from the parent, or after a long dor-
mant period; but they also often emit zoospores while still
attached to the parent hypha. I have carefully followed the
phenomena of this process, and select fig. 17 as affording suffi-
cient illustrations of the details. The formation of the zoo-
sporangium requires no minute description (a tod). The
zoosporangium (e) was completely formed, and separated
by a septum as figured, at 12.25, and remained in pretty
much the same condition until after 3 o’clock; at 3.35 several
vacuoles were observed, slowly changing their positions in the
very granular protoplasm (f). Shortly afterwards, the spor-
angium remaining attached, the beak was formed, and by
4.30 (fig. 17, g) was completed. During the next ten minutes
the processes of formation of the zoospores figured in fig. 16
took place as already described, and the zoospores became
developed in the gelatinous vesicle at the apex of the beak (7) ;
the rupture of the vesicle, and escape of the reniform bi-flagellate
zoospores took place as before, and at 5.5 the only remains of
the vesicle were attached to the end of the empty beak (A).
Meanwhile, shortly after the passage of the protoplasm
through the beak into the vesicle, the septum separating the
zoosporangium from the hypha became protruded into the
cavity of the former (¢), and soon attained a considerable
development as a vesicular swelling, in which the granular
contents of the hypha were slowly accumulating (4). From
5.40 to 6.20 (A and 7) this went on gradually and continu-
ously, until a new sporangium had become formed in the
cavity of the old one. In this instance, the new zoospor-
anginm ceased to develop during the night; but in other
examples it followed the usual course. This proliferation of
the hypha is the characteristic which gave P. proliferum its
excellent specific name. It is very common to find the second
zoosporangium thus developed into the old cavity, where it
forms the beak and large central vacuole before passing into
the dormant state, behaving as before on the renewal of favor-
able conditions. The beak of the new sporange is not
always coincident with that of the older one, and may stretch

OBSERVATIONS ON THE GENUS PYTHIUM. 503

the empty membrane of the latter in the manner shown at fig.
18—a not uncommon case.

The oogonia and oospores of Pythium proliferum
were obtained in large quantities, and observed with ease ;
neither in the processes of development nor of fertilization did
I observe any facts of sufficient importance to need descrip-
tion, after what has been said concerning P. De Baryanum.
De Bary has pointed out that the antheridia are shorter and
less curved ; but whether the importance of the distinction can
be insisted upon I will not attempt to decide. The ripe oo-
spores also resemble those of P. De Baryanum very closely,
and need not be further described. Figs. 19, 20, and 21 show
the most important points.

Pythium gracile (De Bary)

may be selected as a further type, and I have had the op-
portunity to observe it closely. A form called P. gracile
had already been discovered by Schenk? in the cells of Alge,
when De Bary found his P. reptans*® with similar habit ;
both these species .are either identical with Pringsheim’s
P. monospernum’, or are so near that with the evidence at
command ‘they cannot be definitely distinguished. For the
present, therefore, we must look upon De Bary’s P. gracile |
as possibly taking the place of these. It has been carefully”
studied by De Bary’, and now stands as one of the best
known types of the genus. The general characters of P. gra-
cile are similar to those of the other forms, except that, as the
name implies, the hyphe are more slender; correlatively, the
oospores and oogonia are more delicate than in the prece- |

ding forms. I obtained oospores from an old cultivation and re

have depicted their structure and germination in fig. 22. The ‘ .

1 ¢ Verhandl. d. Phys. Med. Gesellsch.,’ Wiirzburg, ix, 1857,

2 * Jahrb. f. wiss. Bot.,’ ii, 1858.

3 © Jahrb. f. wiss. Bot.,’ i.

4 «Bot. Zeit.,’ Sept., 1881. De Bary, however, does not consider the
identity of these three forms settled. C.f. also ‘ Beitr. z. Morph. u. Phys. d.
Pilze,’ R. iv, 1881. See also below, p. 510.

504 H. MARSHALL WARD.

ripe oospore (A) differs from that of P. De Baryanum and
P. proliferum, especially in entirely filling up the cavity
of the oogonium, the exospore becoming closely fitted to
the oogonium wall, and being indistinguishable from it,
except under favourable circumstances during development,
&c. After a few hours in fresh water the oospores observed
commenced to germinate in the usual manner (fig. 228), and
the germinal tube either entered the substance of a favourable
matrix—fiies’ legs, meal worms, dead cress seedlings, &c.—or
proceeded to form zoospores (c) in a manner to be described
shortly. |

I made some observations on the particulars of growth and
entry of the germinal tube into a host-plant, which are illus-
trated by fig. 23 (a tof). After some hours, an oospore (a)
was seen to have germinated at some little distance from the
surface of a Cress-seedling (represented by x in the figure),
near which a minute unicellular alga was adherent; this
was at 11.10 a.m. At 11.30 one branch of the germinal tube
had grown rapidly (the others hardly elongating at all) and
extended (6) so as nearly to touch the algal cell near the
cuticle of the seedling. Hlalf-an-hour later, as shown at (c),
the apex of the hypha had touched and glided over-the loose
algal cell, becoming sharply bent at right angles in doing so,
and displacing the cell somewhat from its original position.
So far the observation seems to show clearly that the extension
of the hypha takes place by apical growth only. The free
apex refusing to attach itself to the algal cell, then became
bent towards the surface of the seedling (d), and at 12.30 was
closely appressed to the cuticle. No intercalary growth had
occurred in those parts of the hypha behind the apex, as is
plain from the position of the angle and loose cell in (d);
meanwhile, however, the apex became closely pressed against
the cuticle, apparently lifting the whole hypha slightly in the
process, and by ten minutes past one o’clock (e) it was clearly
making its way into the cell wall. At two p.m. the end had
completely perforated the cuticle and cell wall—not drawn in f
—and had begun to extend vigorously inside. A slight inter-

OBSERVATIONS ON THE GENUS PYTHIUM. 505
calary growth behind the apex appeared to have occurred in
the interval (c, d, and f), but I could not be sure of this since
the change might be due to a straightening out of the tube in
that region. The protoplasm was by this time passing for-
ward towards the apical portion, and only a few granules
remained in the oosphere and proximal part of the hypha.
Once inside the dead seedlings the fungus extends in the
manner already described, and sooner or later begins to form
zoosporangia.

These arise as projecting hyphe, usually vertical or nearly
so from the epidermis of the seedling, and they differ in several
important respects from those already described. In fig. 24
are drawings showing the various stages witnessed in the for-
mation of the complete sporangium. As seen at first, it was
a simple perpendicular branch from the mycelium, filled with
densely granular protoplasm and with a rounded apex; at
11.8 o’clock the slightly swollen rounded apex was extremely
bright, and appeared capped by a hyaline dome, due to some
alteration in the cellulose, During the next ten minutes this
cap of diffluent cellulose commenced to swell up gradually and
at 11.20 presented the appearance seen in the figure. Almost
immediately after this, the finely granular contents streamed
suddenly forward into the centre of a gelatinous vesicle formed
by the bulging out of the diffluent swollen cap, this streaming
resembled very much the rush of endoplasm often noticed in
the protusion of a large pseudopodium by a vigorous ameeba;
the flow of granules was most rapid in the axial portion, and
the last particles followed more slowly. The third figure repre-
sents the moment before this flow; in the last of the series are
depicted the appearances at the instant of its occurrence. In
this example the succeeding stages could not be drawn rapidly
enough; they were the same as shown in figs. 25 and 26, and
a good idea of the rapidity of this process is gained by com-
paring the time-records made ; the last two stages of fig. 2 a@ were
drawn at 11.23; at 11.25, the whole mass of protoplasm had
passed out into the mucous globe, and was already becoming
divided up into zoospores (fig. 26 6). At 11.32 these zoo-

506 H. MARSHALL WARD.

spores, about thirty in number, were rolling over one another
and waving their cilia (fig. 26, d) still enveloped by the
vesicle of mucus, and by 11.37 they were fully formed very
active zoospores. At 11.43 the enveloping vesicle suddenly
gave way and the zoospores passed out free. Fig. 25 shows
drawings of the stage when the protoplasm, having all passed
out of the sporangium into the vesicular enlargement, is
writhing about in an ameeboid manner preparatory to its
simultaneous division into zoospores; @ and b were drawn at
two successive minutes. These stages are between a and 6 of
fig. 26, which may now be described.

The sudden “blowing out” of the hyaline dome into a
vesicle had just been completed at 11.40 (fig. 26, a), having
gone through the stages already described. ‘Two minutes later
the writhing mass of protoplasm, having passed through the
stages figured in fig. 25, contracted towards the centre of the
swollen vesicle, and rapidly divided into about nine blocks (4),
which became separate amceboid masses during the next three
minutes (c). At 11.50—7.e., five minutes after—each mass
was an active reniform zoospore(d); at 11.52, the very
diffluent vesicle, having almost dissolved in the water, gave
way and allowed the zoospores (e) to escape. The complete
zoospore resembles those of the other species of Pythium in
its possession of a reniform amceboid body, two lateral cilia
from the sinus, and a bright, vacuole-like spot near the base of
the cilia.

In figs. 27 and 28 are shown the details observed as to the
germination of the zoospore after coming to rest. At a
(fig. 27) is drawn a zoospore actively moving at 5.10 ina very
minute drop of water. It was watched continuously, and drawn
again at 5.20 (2) and 5.30 (c), when it came to a standstill, and
commenced to withdraw its cilia. At 5.45 (d) the zoospore,
having come to rest, had lost its cilia and vacuole, and had
developed an envelope and several large refractive granules,
From this point, however, apparently owing to a want of oxygen
in the water, its changes were distinctly retrograde. At
6.5 (e) the granules were coarser, and at 6,35 (f) still larger.

OBSERVATIONS ON THE GENUS PYTHIUM. 507

At 8 a.m. next day (g) no signs of germination were apparent ;
the very large granules were now dull, and their protoplasmic
matrix evidently becoming pale and disorganised. Fig. 27 (h)
however, shows the normal course of events in another speci-
men: the zoospore, after about five-and-twenty minutes of
active life, had come to rest (as in fig. 27, d), and at once pro-
truded a short process. Within an hour after this, the stages
2 and & were passed through; the granules becoming used up
in the elongating germinal tube, and a vacuole forming in the
spore, which became larger and larger as its contents were
drawn upon. Soon after the stage /—the tube having reached
its highest state of development at the expense of the proto-
plasmic and granular contents, and having met with no suitable
matrix to enter—the whole perished. Such zoospores,
attached to the cuticle of a cress-seedling killed in hot water,
germinated in the same manner, the germinal tube, however,
entering the cell wall (fig. 28), and extending as a mycelium in
the way described previously.

On cress-seedlings which had been killed by hot water, and
which had been brought into contact with some débris contain-
ing Pythium given to me by Prof. De Bary, I observed the
development of numerous zoosporangia of the typical P.
gracile already described, together with a much smaller
-number of a second, hitherto undescribed form of zoosporan-
*gium. Attempts were at once made to separate the two forms

in the following manner, by a method often successfully em-
ployed in similar cases, and which may be described in detail,
because it is instructive in many ways.

A small portion of the semi-rotten cress-seedling was selected,
on which a young zoosporangium of the required Pythium
was observed to be preparing to emit its zoospores. This
was teased with needles in the hope of removing all the zoo-
sporangia of P. gracile. This done, the cleaned specimen
was placed in contact with a freshly killed cress-seedling in a
drop of pure water on a perfectly clean glass slip. As in all
these experiments, every precaution was taken to avoid acci-
dental infection, by heating the needles, forceps, &c., and of

508 H. MARSHALL WARD.

course the seedling—having been killed by immersion in boiling
water—need not be supposed to contain sources of error. After
lying in contact with the prepared infective material until the
desired zoospores had been observed to escape into the sur-
rounding water, the infective mass was removed immediately.
The new material now lay in a drop of water in which were
the desired zoospores. After a time sufficient for contact and
entry of the germinal tubes, it was quickly removed with per-
fectly clean instruments and allowed to remain for several hours
in pure water, in the hope that it was infected by the desired
zoospores and no others. In all cases—and many trials were
made—the new material developed numerous zoosporangia
of P. gracile before a single specimen of the desired form could
be detected. Not only so: the P. gracile got the upper hand
very quickly, as it had in the original infections. Nay, in many
cases the desired form did not appear at all. The reason was
clear: the superabundant P. gracile not only formed its
zoospores more rapidly and in larger quantities, but they made
better progress in the matrix, and killed off the other form in
the mutual struggle for existence. In every experiment an
odd zoospore of P. gracile got the start, and beat its compe-
titor in the race, and the only result of all the care bestowed
appeared to be the establishment of a purer growth of the
mycelium bearing the P. gracile form of zoosporangia,
prodigious quantities of which, became formed in the course of
a few hours.

During these cultivations I obtained enormous extramatrical
developments of the mycelium of P. gracile, and, both on
cress and on the young buds of carrot, was enabled to watch
the development of the sexual organs with great success.

At fig. 29 are drawn specimens of the extramatrical myce-
lium of well-developed cultures. As shown here, the other-
wise very slender hyphz became swollen up here and there
into knob-like groups of oval or rounded protuberances, into
which the fine grained protoplasm became collected at length.
In fig. 29 a, for example, a firm septum marks off an empty
distal moiety of a hypha, from one full of protoplasm, and with

penne

OBSERVATIONS ON THE GENUS PYTHIUM. . 609

several of the protuberances or lateral outgrowths ; this condi-
tion was observed at 2 o’clock, and remained substantially the
same at 6 o’clock. At 9.50 next morning the change depicted
4’ had occurred; the protoplasm of the hypha had become
further retracted—presumably into the now more developed
protuberances—and a firm septum was formed nearer the
proximal portion of 4, a small commencing outgrowth having
become emptied of its contents also. Two days later, almost
all the protoplasm was thus accumulated into the largest knob
of the outgrowths. At B, a similar accumulation has occurred,
and in this manner the mycelium becomes irregularly septate
independently of the sporangia, conidia, or other reproduc-
tive organs. I think the protuberant outgrowths here described,
must be looked upon as physiologically important for the
accumulation of protoplasm which may serve either for the
immediate need of the mycelium, or for the production of
oogonia; since it is certain that the large protuberances may
do either of two things, they may put forth ordinary hyphe
which merely continue the vegetative growth (fig. 29, p), or
several of them develop into oogonia and antheridia (fig.
31), apparently at the expense of certain of their neighbours,
which they deprive of contents. The above view—that the
difference between a young oogonium and a mere bud or
conidium, is determined by physiological, and not morpho-
logical causes—seems to be almost established when we com-
pare the facts just described, with those figured in figs. 31 and
35, where the oogonium is distinctly beaked, as if it had
begun to grow forth like a conidium, and then been im-
pelled to behave as an oogonium.

The proper zoosporangia of this form could not be dis-
tinguished from those described before (figs. 24 to 27), and
they were produced in equally enormous quantities. In fig.
30, I have drawn two abnormal types or states, one of which -
(6) was rather common in my cultivations. In (a), for
example, the zoosporangium commenced to form, but
proceeded no further than the stage figured; the contents
becoming withdrawn and used up later. In the second case

pio” H. MARSHALL WARD.

(4), the swelling up and passage out of the protoplasm took
place as usual: but the protoplasm was extremely small in
quantity and very watery. Large vacuoles at once became
formed, and the whole faintly granular mass soon fell to
pieces, merely becoming diffused, as it seemed to me, after the
bursting of the absorbent vesicle. These are undoubtedly
pathological phenomena, and cannot be considered as specific-
ally considered of the Pythium.

As to the sexual organs and process of fertilisation, little
need be said here beyond calling attention to figs. 31—36. The
formation and structure of the oogonia and antheridia are
essentially as before ; and the passage of the fertilising material
through the tube (figs. 31a, 32, 33 and 34) needs no further
description—it takes place exactly as before described.

The ripe oospore (fig. 36) presents the peculiar characters
of P. gracile (De Bary), especially in filling up the oogonium,
and I can select no distinguishing features between the two
forms, unless the peculiar knot-like groups of tuberosities and
the extensive extra-matrical growth, &c., be considered more
important than they seem to me.!

It may be, however, that some confusion still: ae between
one or two forms with the simple zoosporangia of this type,
since, as De Bary points out,” the P. gracile so closely studied
by himself is always a saprophyte, and refused to attack
Algz, whereas the earlier forms (P. reptans, De Bary and
P. gracile, Schenk) with which he seeks to identify it, were
found on living Algee. It seems probable that further research,
directed to enquire whether and how far the species named are
distinct, may yield important information as to the limits be-
tween parasitic and saprophytic tendencies; such research,
however, is likely to be valuable only so far as it is made on
carefully isolated cultures, observed during sufficiently long
periods.

1 De Bary, however (‘ Bot. Zeit.,? 1881, p. 570), says the oogonia and
antheridia (of P. gracile) are found only in the substratum, ‘Nur im
Innern des Substrates, inter und intracellular.’

2 * Bot. Zeit.,’? 1881, p. 572.

OBSERVATIONS ON THE GENUS PYTHIUM. 511

Within the cells of a large species of Spirogyra, observed \|
this summer (1882), I found the Pythium with very delicate
hyphe figured in fig. 37. It was not present in any consider-
able quantity, and all attempts to cultivate it in the mass
failed, as did also my endeavours to make it spread to other
Algze or cress-seedlings. I was also unsuccessful in the search
for zoospores and sporangia, and am thus unable to state
exactly what species it was. It is an obvious suggestion that
this was probably the earlier Pythium gracile of Schenk,
which was discovered in similar algal cells, and of which the
sexual organs are not known. If this be the case, it is clear
that De Bary’s P. gracile is a different species, and this would
be in accordance with his failure to cultivate that form on
living Algee—it being, so far as is known, a saprophyte only.!
It may be considered probable, from the evidence at disposal,
that the form here discovered is really Schenk’s P. gracile
(De Bary’s P. reptans); and, at any rate, it were better to
assume this for the present than to assign a new name to my
Pythium until further observations are to hand.

The hyphe of this species are very slender much branched
filaments, which bore through the septa and side walls of the
Spirogyra in all directions, causing the chlorophyll bands to
become contracted into irregular lumps and bands, which
retain their green colour however for a long time before they
slowly decay. The great interest attaching to the specimens
observed was, that oogonia with oospheres and antheridia
were produced in the normal course of the cultivation, and it
is clear that the oospheres and oospores differed consider-
ably from those of P. gracile (De Bary), in that the ripe
oospore is much smaller than the oogonium, whereas in
De Bary’s P. gracile the oospore entirely fills the cavity of
the oogonium; the antheridial cell is also shorter and broader.
The fertilising process was observed and offers nothing specially
worthy of note differing from what has been described. In
fig. 39 the very short antheridium, at the end of a very

! «Bot. Zeit.’ Sept., 1881, p. 572.

512 H. MARSHALL WARD.

long branch from a neighbouring hypha, was in close contact
with the oogonium at 9 a.m., and the process of fertilisation
was already commenced, though no tube could be detected,
owing to the fatty globules of the egg coming close up to the
point of contact. At 12 noon, however, the contents of the
antheridium had passed over, the oospore was already
clothed with a membrane and its contents presented a dif-
ferent appearance, and the “ fertilising-tube” could now be
clearly seen extending between the oospore and the point of
attachment of the oogonium wall. The completely ripe
oospore (fig. 88) has a very thick membrane, and, in some
cases at least, a conspicuous central nucieus-like spot.

It does not seem wise to attempt any further speculations as
to the relations of the three types of Pythium with which we
have been engaged; but it is clear that the one just described
cannot be included in P. gracile (De Bary). It is therefore
either the same as Schenk’s P. gracile, with which it agrees
in habits, &c., or it is entirely new. This can only be decided
after the discovery of the zoosporangia.

It now remains to describe all I was able to observe con-
cerning the fugitive Pythium, the zoosporangia of which
occurred mingled with those of P. gracile, as referred to on
p- 507, but which I could not cultivate separately.

The zoosporangia (figs. 40 and 41) occurred as ovoid or
pytiform swellings of the ends of single, thin hyphe, which
projected verticaily from the cress-seedlings, considerably over-
topping the sporangia of P. gracile (De Bary), with which
they were mingled. Each appeared in some respects similar to
those of P. proliferum already described, a resemblance
which might be insisted upon, if the mycelium of this form were
not so much more delicate, and if the zoosporangia were
developed upon racemose branchings, as in P. proliferum.
Such, however, was not the case, and I must regard this
slender form as distinct for the present. The development of
the sporange asa terminal swelling of the free hypha, which
then becomes separated off by a septum and develops a beak as
a continuation of the long axis, are sufficiently shown in fig. 40,

OBSERVATIONS ON THE GENUS PYTHIUM. 513

and, after what was said before, need not be further dwelt
upon.

The development of the zoospores may be more fully de-
scribed, since it affords further distinctions for separating
this species. At 10 o’clock a.m. the ovoid zoosporangium
was fully formed, and had developed its short vertical beak
from the distal extremity or apex (fig. 41,4). It remained
almost unchanged during an hour, the only recognisable changes
being the movements of the numerous minute granules, and
the formation of ten or twelve small vacuoles (fig. 42, 2) in
the protoplasm. ‘These soon disappeared, and the end of the
beak became more transparent and its walls marked by fine
longitudinal strie. At11.40 a.m. the granular contents passed
out slowly, inflating the substance of the beak into a delicate
gelatinous vesicle in the mode already described in P. proli-
ferum. Here, also, the mass commenced to divide into
zoospores, passing through similar stages, and finally be-
coming free (figs. 41, B-p). All these processes occupied a
perceptibly longer time; in this case, however, nearly twenty
minutes having elapsed between the emission of the protoplasm
and the completion of the zoospores within the vesicle. At
12 o’clock the young zoospores were moving independently
(p), tumbling one over the other with active amceboid move-
ments, and soon afterwards the cilia appeared, at first short and
slowly waving, then soon elongated, apparently at the expense
of the knobs at their extremities, and by eight minutes past
12 the ten zoospores were rapidly flitting about. One minute
later (at 12.9) the vesicle gave way, and the free zoospores
escaped in the usual manner.

The gemination of the zoospore took place in the usual
manner (fig. 43, a) after the swarming, and in one case I
observed the entry through the cuticle of a dead cress-seedling
(fig. 43, B). But the process of zoospore-formation may
occupy even a longer period than above described. In the
specimen drawn at fig. 42 (4-6), the separated zoospores
moved in an active amoeboid manner for more than an hour
before the cilia were developed. Of course, that this Pythium

514. H. MARSHALL WARD.

was struggling under unfavourable circumstances must not be
forgotten—since we have seen that it apparently became
driven out of the field eventually by its successful rival, P.
gracile, De Bary—and such circumstances may have affected
the normal course of its development, even in details. Never-
theless, there was no direct evidence to show that such was the
case in the specimens figured.

Another peculiarity which renders the separation of this
Pythium necessary was the behaviour of the sub-sporangial
portion of the hypha. As shown in fig. 44, the end of the
hypha grows through the empty sporangium, and becomes
continued as a long hypha which can certainly bear a zoo-
sporangium again at its apex, though I only once saw one
such, and that not fully developed.

All things considered, it seems necessary to regard the above
form as distinct from any yet described in this paper. Whe-
ther it is the form called P. ferax by De Bary! cannot be
decided until the oospores, &c. are observed, though it
appears so far to answer to the descriptions given of that
species.

The last representative of this remarkable genus which I
have actually studied, is P.intermedium, De Bary,’ a form
which occurs mingled with P. De Baryanum both on dead
and living plants. My specimens were obtained through the
kindness of Prof. De Bary, and cultivated on killed cress-
seedlings as before. The sexual organs have never been
observed, and its only claim to be regarded as a distinct species
seems to be its habit of producing the conidia in vertical
series (figs. 45, 46), and the many similarities between these
conidia and those of the Peronosporex; facts of which
advantage has been taken in comparing the two groups. The
chief points are as follows, according to De Bary: the conidia

1 «Bot. Zeit.,? 1881, p. 562. I have only once seen the zoosporangia
of P. ferax, through the kindness of Prof. De Bary, and cannot decide
that the two forms are identical.

2 * Bot. Zeit.,’ 1881, p. 553.

* « Bot. Zeitung,’ 188], loc. cit.

OBSERVATIONS OF THE GENUS PYTHIUM. 515

may be formed in series of four or five, the apical one always
being the older, and will even become developed free in the
damp air of a moist chamber. The conidium may pro-
duce either a germinal tube or zoos pores on germination, much
as in the Peronosporee: after drying, however, the conidia
appear to lose their power of germinating altogether.

As to the claim of this form to be considered a species
nothing further can be said, so long as the sexual organs are
undiscovered.

516 ALFRED ©. HADDON.

On Budding in Polyzoa.
By

Alfred C. Haddon, M.A.,
Professor of Zoology in the Royal College of Science, Dublin.

With Plates XXXVII and XXXVIII.

INTRODUCTORY.

Mr. Hincks, in his valuable monograph on the British Marine
Polyzoa (1), after discussing the nature of the “ brown body ”
found in the old zocecia of Polyzoa, and its relation to the de-
veloping bud closes with these words (p. ]xiii): “ There seems,
therefore, to be grounds (pro tanto) for desiring some further
investigation of the subject.” This, then, must be my excuse
for offering these somewhat imperfect observations.

The investigations on Flustra carbasea were made in the
month of May, 1879, whilst I was occupying the table belong-
ing to the University of Cambridge, in Dr. Dohrn’s Zoological
Station at Naples. The other observations were made on ~
species obtained from Dublin Bay during 1881-2.

Of the nature of the brown body itself 1 do not propose to
treat, as the evidence of other observers as well as of my own
studies is perfectly satisfactory in favour of its being, as Hincks
says, “derived from the polypide, and is the result of its de-
cline;’’ but I will limit myself solely to the origin and develop-
ment of the bud.

Own OBSERVATIONS.

Flustra carbasea, E. and 8. (The following observations,
when not otherwise stated, apply to the living state only) In

ON BUDDING IN POLYZOA. 517

most of the empty zooecia which had been previously inhabited,
a brown body was observed situated towards its lower end,
this was surrounded by funicular tissue (‘‘ endosarc,” Joliet),
which sent out irregular strands to the walls of the zooecium,
some of them being connected with the band of flexible endocyst
which stretched across the mouth of the zooecium. In the centre
of this band, and therefore connected with the endocyst on the
one hand and with the funicular tissue on the other, was
situated, in the earliest observed stages, a small rounded mass of
cells yellowish in colour, surrounded by a sheath of transparent
cells, which together constitute the nascent bud.

The bud soon acquires a well marked central cavity (Pl.
XXXVIL, fig. 1), then becomes oval in form, and depends from
the anterior band of endocyst. A further elongation next takes
place, this process being more rapid above than below, result-
ing in a pyriform body, of which the upper and narrower part
consists of a thin double walled sac, the outer wall being the
sheath and the inner one the attenuated internal layer. The
lower and wider portion consists of the thin outer sheath
enclosing the active internal cells (P]. XX XVII, fig. 2). To
anticipate—the inner cells will form the external layer of the
tentacular sheath, the external epithelium of the tentacles, and
the internal epithelium of the alimentary canal of the new
polypide, while the outer layer or “sheath” will form the
inner layer of the tentacular sheath, the inner epithelium of
the tentacles and the tissue which surrounds the digestive tract.

A series of somewhat complicated changes now takes place
in the lower moiety of the inner layer. (It should here be
premised that the outer layer is perfectly passive throughout,
merely adapting itself in such a manner as to wrap itself round
the active inner layer.) One side of this portion of the bud
protrudes, the protrusion becomes constricted off in such a
manner as to produce a blind sac, depending by the side of the
remainder of the bud, the constriction is quite complete except
at the uppermost point, this being the spot where the rectum
will be connected with the lophophore; fig. 3, Pl. XX XVII,
which is drawn from a preparation, illustrates the commence-

VOL, XXI1I,—NEW SER. MM

518 ALFRED ©. HADDON.

ment of this process. It will be noticed that the inferior por-
tion of the area which is being constricted off is connected with
funicular tissue; as a matter of fact, there is often a slight
difference in the character of those cells which occupy a corre-
sponding position in yet earlier buds. While this has been
happening, the other portion, which has a plate-like form,
becomes crenulated along its margin, the crenulations, which
point upwards and rather inwards, increase in size and we
have some twenty-two incipient oval tentacles formed. These
tentacles are thus early ranged into a circular lophophore,
continuous except in that region from which the above-
mentioned pouch is hanging. This gives an appearance of
bilateral symmetry to the lophophore, as was noticed by Allman
in Paludicella (19), and Nitschein Flustra membranacea
(13) (Pl. XX XVII, figs. 4 and 5.). |

The developing polypide now consists of a disc-like body
(lophophore), surrounded by twenty-two oval tentacles, emar-
ginated at one spot from which depends a cecal pear-shaped
bag—the future stomach and intestine.

A circular depression occurs towards one end of the disc of
the lophophore which rapidly deepens into a rounded sack
(Pl. XX XVII, figs.6 and 7). The fundus of this sac impinges

Fic. 1.—Diagram showing the relations of the devoloping bud in a Marine
Polyzoan. a, anus; J.4., brown body; ¢. p., communication plate ;
ect., ectocyst ; ep., epiblast ; 7, funiculus ; ézé., intestine; m., mesoblast ;
@., esophagus; 7. m., retractor muscle; s¢., stomach; ¢., tentacles ;
t. s., tentacular sheath.

ON BUDDING IN POLYZOA. 519

upon the cecal stomach, the two organs coalesce and their
lumens become continuous by wall-absorption. This diver-
ticulum is the wsophagus, and the polypide has now the
characteristic form of the adult.

The woodcut (fig. 1) indicates the relation of the developing
bud to the zooecium far better than a verbal description can do.

Very shortly after this, owing partly to absolute increase in
size and also to the elongation of the tentacular sheath, the
polypide has come into contact with the brown body, which, as
was before mentioned, lies near to the bottom of the zooecium
(Pl. XXXVITI, fig. 8). The walls of the stomach, or more
strictly, that portion of the stomach which forms the gastric
cecum, grow round and envelop the brown body, so that the
brown body passes as a whole into the alimentary tract of
the young Flustra, which now has the form depicted in
Pl. XXXVII, fig. 9.

The brown body immediately commences to undergo disin-
tegration, and, previously to passing into the intestine, the
remains are whirled round and round within the globular pyloric
portion of the stomach by the action of the minute cilia with
which the latter is clothed (Pl. XX XVII, figs. 10 and 11).
Ultimately all trace of the brown body, as such, is lost save a
small quantity of feecal matter in the intestine, and by this time
the gastric glands become very apparent (Pl. XX XVIII, fig. 12).

The reason why these buds, at this stage, appear of a dif-
ferent colour from the ordinary marginal buds of the colony,
as was first noticed by Hincks and animadverted upon by
Joliet (17) (still later, see Hincks (1) pp. lvii, Ixii), is pro-
bably owing to the digestion of the brown body with the
concurrent development of digestive glands, the other buds
gaining the whole of their nutriment directly from the parent
tissues, and thus not requiring a distinct digestive apparatus.

Annulations of the stomach at this stage indicate the exist-
ence of circular muscles, the walls of the pylorus become mus-
cular and much thicker, and, as before mentioned, its lumen is
ciliated, the cecum is a wide pouch lined with secretory cells.
The intestine is swollen, while the rectum is a short very

520 ALFRED C. HADDON.

narrow tube still retaining its primitive connection with the
tentacular crown, to which it is attached about one third from
the base.

While the changes described above have been taking place,
the tentacles have been gradually lengthening, at first, they
are short finger-like processes from the periphery of the lopho-
phore, closed above, open below, containing within their
cavities an extension of the original outer layer of the bud
which here forms an epithelial lining (Pl. XX XVII, figs. 8 and
9, a, and also woodcut, fig. 1); not till comparatively late do
cilia arise on the outer epithelium, only certain aspects of the
surface of the tentacles are clothed with cilia (see Pl. XX XVII,
fig. 9, @).

vathe tentacular sheath ultimately becomes continuous with
that portion of the endocyst of the zooecium which surrounds
the mouth of the cell as was insisted upon by Nitsche (13,
p- 463).

The retractor muscles of the body and lophophore arise, as
noticed by Repiachoff (15) from the peritoneal lining of the
polypide. }

The funiculus early becomes prominent and is probably
derived from the irregular strands of funicular tissue which
occur in the parent zooecium; it appears as a thickish cord
stretching from the fundus of the developing polypide to the
base of the zooecium, and, almost invariably, it is in direct
connection with the brown body, so that it serves to direct
the developing alimentary tract to that nutritive mass, thereby
ensuring the better nutrition of the growing bud.

Abnormalities extremely rarely occur in which there may
be two buds developed, or more than one brown body, or the
polypide may not come into contact with the brown body.
The second abnormality probably being the result of the third.

It is thus clear that the bud in Flustra carbasea is deve-
loped at a distance from the brown body, that it approaches
the latter, envelopes it, and extracts nutriment therefrom. As
was pointed out by Repiachoff (16) the same occurs in several
genera of Polyzoa (Tendra, Lepralia, Membranipora,

ON BUDDING IN POLYZOA. O21

&e.). Joliet also witnessed the ingestion of the brown body in
Eucratea chelata, which passed through the alimentary
canal-of the developing polypide, but owing to its resistant
membrane the brown body suffered no alteration; but in
Lepralia granifera the very thin envelope of the brown body
is destroyed, not being able to resist the action of the juices of
the stomach, the movements caused by the cilia, and the con-
tractions of the intestinal walls: thus the brown granules
which it contains are set at liberty, whirled about and shortly
evacuated by the rectum. Hincks, himself (I.c. p. lxii, footnote)
noticed the formation of a polypide-bud quite separate from the
brown body in Bugula calathus.

After the able discussion of the subject by Joliet (17), it
seems quite superfluous to reopen the controversy as to the
probable origin of the bud from the brown-body (“ germ-
capsule”’); but Hincks (1. ¢. p. 1xiii) has still left it a slightly
open question. According to Smitt (and Hincks), there would
be at least two modes of bud-formation amongst the Polyzoa:
1. In the old zooecia (a) formed quite close to the brown-body,
and arising directly from it, (2) formed at a distance from the
brown-body and not arising: from it. 2. In the new zooecia,
also arising de novo. From the accounts of other observers,
one method of bud-formation serves in all cases, the origin
in an old or a new zooecium being always from the same
tissue, though they are by no means agreed as to what that
tissue is. It is merely a question as to how close to or how far
from the brown-body the bud shall arise.

Taking all the evidence we possess, it seems to be quite
evident that the generally received account is the correct one,
but that the approximation of the undeveloped bud to the
brown-body may mask its real distinctness in a few instances.

A further observation on a living specimen (Naples, Sept.,
1881) is represented on Pl. XXVIII, fig. 13. The bud had
reached the stage of Pl. XX XVII, fig. 1; it was suspended in
the anterior band of endocyst, and was connected with the
parent polypide by the tentacular sheath of the latter, and
probably also by some funicular tissue. In this example the

522 ALFRED C. HADDON.

older polypide was rapidly histolysing into the brown-body ;
thus in this case the bud was formed before the complete de-
gradation of the parent, and at a slight distance from it.

A prepared specimen (Pl. XX XVIII, fig. 14) indicates the
origin of part at least of the bud from the endocyst of the
opercular opening ; the original occupant of this zooecium had
scarcely commenced to decay.

Flustra securifrons, Pall.—Another prepared specimen
from Naples (Pl. XX XVIII, fig. 15) shows a possible double
origin for the lophophore and stomach in a young marginal zooe-
cium. It will be seen that anterior band of endocyst has just
been formed, and slung upon this is an undoubtedly epiblastic
invagination or proliferation, coated by mesoblast. On one side
is a mass of cells, which is continuous with what appears to be
the incipient funiculus. This mass of cells, I take it, will form
the future stomach and intestine; it soon ceases to exist as a
distinct group of cells. 1 have several times noticed this stage.

Flustra papyracea, E. & S,—In new zooecia the buds
may be seen to arise in close contact with the endocyst of the
floor or of the wall of the cell, according to whether they may
be terminal or lateral additions. Very shortly they assume a
more central position, and are more or less thickly enveloped
in a funicular plexus, from which latter there is every appear-
ance of additions being made to their substance. The develop-
ment of the polypide is exactly as described above.

In old zooecia the buds are developed in the anterior portion
of the cell.

Pl. XXXVIII, fig. 16, shows a bud which is partly formed
of columnar cells and partly of rounded. The latter appear to
be produced at the expense of the funicular tissue; the former
probably arose from the epiblastic layer of the endocyst.

Bugula flabellata, J. V. Thompson.—Pl. XXXVIII,
fig. 17, shows a new zooecium, within which is the young bud,
which has a well-marked bilobed appearance. Closely applied
to the fundus of the stomach-sac is an ovary, which has been sup-
plied ready-made to the bud. It is invested by the funicular
tissue, which organically connects all the members of a Polyzoan

ON BUDDING IN POLYZOA. 523

colony. Figs. 18 and 19 are consecutive sections of a similar
bud at a later stage, and illustrating the same point. Fig. 20
is a longitudinal section of a slightly later stage, showing the
cesophageal invagination inpinging upon the stomach.

Eucratea chelata, L.—lIn old zooecia the bud is derived
from a small mass of cells, which is situated just below the
hinge of the operculum, and from the first is apparently in
equally intimate connection with both the endocyst and
strands of funicular tissue (Pl. XX XVIII, fig.21); subsequently
it occupies a central position just above the brown-body, and
then it commences to go through the characteristic develop-
ment. It is this stage which has, I imagine, deceived Joliet.
into believing that the bud arises from the funiculus itself.

In new zooecia the bud has a similar origin, only in this
case from the base of the zooecium. P].X XXVIII, fig. 22, shows
the lophophore to be quite distinct from the digestive tract,
while the latter is closely connected with the funiculus.

Aleyonidium gelatinosum, L.—A portion of the bud, at
all events, arises by invagination of theendocyst. P].X XXVIII,
fig. 23, clearly shows that both the epiblastic and the meso-
blastic layers of that tissue are equally implicated.

Fig. 24 isa longitudinal section, corresponding to figs. 7, 20,
22, &e.

I have, in fact, observed the distinctness of the lophophore
from the alimentary tract in the following forms:—Bugula
avicularia, B. flabellata, Flustra carbasea, F. papy-
racea, F. securifrons, Eucratea chelata, Diachoris
magellanica, Alceyonidium gelatinosum, Vesicularia
spinosa,

LarvaL GEMMATION.

The phenomenon of budding is generally supposed to take
place during the embryological history of a Polyzoan. The
following very brief summary of what is known on the subject
is abstracted from the late Prof. Balfour’s ‘ Elements of Com-
parative Embryology,’ vol. i.: the sentences within inverted
commas being transferred from that invaluable work.

or
ho
Ne

ALFRED C. HADDON.

Entoprocta.

The larval gemmation of Pedicellina is, for convenience
sake, noticed a few pages further on.

Ectoprocta—Gymnolemata.

At the stage of thirty-two segmentation spheres the archen-
teron is formed by the invagination and subsequent sub-division
of four (Barrois) or eight (Repiachoff ) middle cells of the oral
surface, but it does not appear that this archenteron is ever
functional, and there is every probability in favour of the view
that this functionless organ gives rise to a bud, the so-called
“ dorsal organ’ (=‘ pharynx’ of Barrois), as the archenteron
in Pedicellina has been shown to do by Hatschek (see below,
p- 531). It is worth noticing that “ according to Hatschek it
develops as a solid outgrowth of the hypoblastic walls of the
mesenteron shortly before the mesenteron joins the cesophagus
(fig. 129, B, v),” p. 244. “A nearly similar organ to this is
found in the embryo of Loxosoma [Vogt, 6, and Barrois, 1*].
Here, however, it is double, and forms a kind of disc connected
with two eye spots,” p. 245.

The greater part of the internal organs of the larva now de-
generates and forms a nutritive or yolk-mass. “ The skin of
the larva after these changes gives rise to the ectocyst or cell
of the future polype. The future polype itself appears to
originate, in part at any rate, from the so-called dorsal
organ.”

“The first distinct rudiment of the polype appears as a white
body, which gradually develops into the alimentary canal and
lophophore. While this is developing the ectocyst grows rapidly
larger, and the yolk in its interior separates from the walls and
occupies a position in the body cavity of the future polype,
usually behind the developing alimentary canal. According to
Nitsche it is attached to a protoplasmic cord (funiculus) which
connects the fundus of the stomach with the wall of the cell.
It is probably (Nitsche, &c.) simply employed as nutritive
material; but, according to Barrois, becomes converted into
muscles, especially the retractor muscles.”

ON BUDDING IN POILYZOA. 525

* Adopting the hypothesis already suggested in the case of
the Entoprocta, the metamorphosis just described would seem
to be a case of budding accompanied by the destrucion of the
original larva.”

“This view of the nature of the post-embryonic metamor-
phosis is apparently that of Claparéde and Salensky, and is
supported by Claparéde’s statement (see below, p. 538) that the
formation of the first polype ‘resembles to a hair’ that of the
subsequent buds,” p. 249.

Dr. W. Repiachoff (14) in his study of the development of
Tendra zostericola, says that he cannot with certainty
say how the inner epithelium of the middle and hind gut arises,
but his figures clearly show that this tissue is intimately con-
nected with the ‘“ brown-mass.” Several figures in his
plate viii, indicate the occurrence of an epiblastic involution
at the pole of the embryo, opposite to that where the blastopore
has closed up. This invagination will form the external layer
of the tentacular sheath, the outer epithelium of the tentacles
and the cesophagus of the primary zooid; it is in fact the
stomodeum. The pedicle of invagination of the archenteron
is absorbed, the latter being the rounded body, which he calls
the ‘* brown-mass.”” From one end of this a U-shaped promi-
nence is produced, which is apparently hollow from the very
commencement of its formation, the remainder of the mass

being solid: this is the future intestine. The outer cells of the

“brown mass ”’ differentiate into the inner epithelium of the
stomach, which soon acquires a free communciation with the
exterior through the cesophagus. The central residual portion
of the “ brown-mass” is digested within the stomach like any
other food-yolk. The ‘‘ brown-mass ” is surrounded by a deli-
cate membrane, the splanchnopleure. To render the above
account more clear, I reproduce his fig. 7, woodcut No. 2,
which compare with woodcut No. 1.

It is clear, if the above be a correct interpretation, that the
initial individual of a colony, in this species at all events, passes
through a development which is normal in all its essentials, nor
does there appear to be any histolysis of the primary larya.

526 ALFRED C. HADDON.

To recapitulate—omitting the purely secondary phenomena of
the external form and the behaviour of the body-wall—the
blastopore closes up and the pedicle of invagination forms

Fre. 2.—Primary zoccium of Tendra zostericola. The tentacles, though
present, are not shown. (After Repiachoff.) Jph., lophophore; ¢zz., in-
testine; 2. m., retractor muscle; s¢., stomach.

neither the oesophagus nor the intestine. The archenteron is
at first solid ; a portion of its substance is prolonged to form
the intestine, which subsequently opens to the exterior outside
the tentacles. The tentacular sheath and the tentacles are
derived from an epiblastic depression, from the floor of which
the esophagus is evaginated, which then fuses with the stomach.
The inner face of all these organs is coated with mesoblast.
The details of the later development are perfectly normal.

It is possible that, in some cases, the indifferent character of
the cells of the archenteron and the stomodeal invagination,
have misled observers into the belief that the embryo has un-
dergone histolysis, and that the first zooid of the colony is
produced by larval gemmation, for the view of the total
formation of a bud (‘ polypide’ of authors) from the endocyst
has been so firm that a well-marked involution, such as the
stomodeeum, would be interpreted as a bud rather than as a
portion of the embryo. For myself, I am inclined to believe,
with Barrois, that the occurrence of the destruction of the
primitive larva is not necessarily universal amongst Polyzoa.

ON BUDDING IN POLYZOA. 5a

Ectoprocta—Phylactolemata.

I have treated of these later on.

OruEeR AuTHoRS’ OBSERVATIONS ON ADULT GEMMATION.
Entoprocta.

According to Prof. Carl Vogt, in Loxosoma phascoloso-
matum (6), the bud is formed by a rising of the outer cellular
layer of the parent, carrying its cuticle along withit. The cavity
thus produced is filled not with “cells”? but with an undivided
sarcodic mass, which very soon breaks up into homogeneous
non-nucleated masses of protoplasm. This anomalous material
at first groups itself into three masses, superiorly the hood
(capuchon) or lophophore, which from the first possesses a
central cavity, the vestibule ; below this is a small solid mass of
cells, the stomach, and inferiorly lies the pedal gland. Other
differentiations of these protoplasmic masses produce the tran-
sitory pedal body between the stomach and pedal gland, the
reproductive organs between the lophophore and stomach, and
the general parenchyma of the body. The stomach acquires
a central lumen and the intestine and rectum now make their
appearance, also, by him, derived from the protoplasmic mass.
They, too, are at first solid. The cesophagus is a diverticulum
from the hood. The tentacles are the last organs to make
their appearance, then the vestibule first opens to the outer
world, and the rectum into it, and the bud becomes detached.
In this form the pedal gland atrophies. The author informs
us that he has tried the effect of various reagents and also
section-cutting, but has “abandoned these methods, which de-
mand so much time and care, and in the present case could
give me no positive information” upon points which direct
observation of the living organism had failed to solve.” It
appears to me that the formative elements of the bud are true
cells, as all other observers maintain, and that the earliest
stages were incorrectly determined. The difficulties of the
homologies of the parts vanish, if taking a somewhat later
stage, we look upon the anterior mass with its central hollow

528 ALFRED C. HADDON.

to be an invagination from the epiblastic cells (“ hypodermal
layer ”) at the apex of the developing bud, whilst the underlying
originally solid mass of cells which have primitively proliferated
from the parental stomach and the pedal gland with the other
internal structures, are modifications of migrated mesoblast.
The development of the cesophagus as a depression from the
hood also favours the interpretation, as does Vogt’s aceount,
if we except his earliest stages.

Subsequently, Prof. M. Salensky examined the gemmation
of Loxosoma crassicauda (9). He describes the first stage
as consisting of a small group of cells surrounding one central
one. The latter by division forms a central mass which attaches
itself to the anterior end of the lengthening and pedunculated
outer wall of the bud. A slight longitudinal fissure appears in
the ectoderm (epiblast of the bud), which is the rudiment of
the orifice of the hood. The central mass becomes hollow and
forms the hood and the whole of the digestive tract. ‘The
rudiment of the digestive tube presents itself under the form of
a cul-de-sac, in which two parts can be distinguished. ... The
superior part is the rudiment of the intra-tentacular depression,
the inferior part is the rudiment of the digestive tube and of
the rectum. ... The superior part appears as a sac open in front.
The edges of the aperture by which the sac opens now consists
of ectoderm and endoderm which are completely united,” p. 21.
By “endoderm ” Salensky means the inner layer of the double-
layered bud, which tissue, according to him, forms the inner
epithelium of the alimentary tract, the intra-tentacular space
and the inner surfaces of the tentacles, their outer surface being
formed at the expense of the ectoderm, the tentacles themselves
arising just where these two layers fuse. On p. 19, he says,
‘the ectoderm and endoderm have arisen from the ectoderm or
the integument of the mother. ‘This fact is so clear to anyone
who observes the profile of young buds of Loxosoma, that there
cannot exist any doubt as to its reality. From the analogy
which exists between all the species of Loxosoma, I may affirm
that the described phenomena should be common to all the
species.”

ON BUDDING IN POLYZOA. 529

The history of the other organs need not detain us; it is not
stated from which of the two primitive tissues they are second-
arily derived.

Salensky’s account presents us with fewer difficulties than
that of Vogt, but while agreeing with him as to the epiblastic
nature of the outer layer af veils, I would suggest that the central
cell, which he thinks is of the same value but does not prove it,
is really derived from the alimentary canal of the parent, and
is therefore hypoblastic. It is also possible that the involution
which he describes, but on which he does not lay much stress,
really forms the intra-tenacular space, as his account of the
formation and position of that cavity appears to me to warrant
that supposition, and that his (epiblastic) endoderm pie
only into the stomach and intestine.

Hincks in his abstract of this paper (10), says: “I am quite
unable to harmonise the account given by the author of this
portion of the developmental history with that which we have
from Vogt.”

It will probably be found that the harmonizing of these
and other accounts is possible according to the views stated
above.

Prof. Oscar Schmidt (5) has propounded the original view
that the bud in Loxosma cochliaris formed parthenogeneti-
cally from an egg, and that it is therefore not a true bud but an
embryo! His paper is accompanied by a plate which is too
sketchy to be of any value whatsoever. Nitsche and Salensky
overthrow this theory, and the latter points out that buds in
which no ovaries are developed may give rise to secondary buds,
thus precluding any possibility ofa parallelism between the bud
of a Loxosoma and the ovicell of one of the Ecto procta.

Nitsche (4) has studied the gemmation of Loxosoma
Kefersteinii. In this form he asserts that the bud originates
from a grouping together of one or two ectoderm cells, these
divide and form a single layered ring round one central cell.
This latter, which he calls the ‘“‘ Endodermzelle,” divides into
two, then into four, and altimately forms a mass of cells which
acquires a central lumen, and by subsequent constriction differ-

530 ALFRED C. HADDON.-

entiates into the cavity of the hood and into the alimentary
tract; the generative organs arise as a pair of lateral protuber-
ances between the hood and the stomach ; the external orifice
of the hood is formed comparatively late. The muscle cells
and gelatinous connective tissue of the bud are derived from
two or three “ Mesoderm” cells which make their first appear-
ance when there are some half dozen “‘ Endoderm ” cells, and
which are probably segmented off from ectoderm cells of the
bud. He is unable to say from which layer the foot gland is
derived. In this form the bud is not attached to the parent
by the aboral extremity of the stem but at a spot where the
body and the stem unite.

It is thus quite clear that this able investigator regards the
whole bud as being derived from the epiblast of the parent.

I have cut a large number of sections to elucidate the
question of the budding in Loxosoma. The form I worked at
was L. tethye,so abundant on sponges of the genus Tethya,
at Naples. Most of my specimens were killed with osmic acid
and stained in picro carmine. Unfortunately my results are not
so exhaustive as they might be. Pl. XX XVIII, fig. 25, shows
an epiblastic down-growth from the apex of the bud, which will
form the cavity of the hood; below this is a small group of
cells the nature of which I am unable to state definitely ; they
may be mesoblastic, or they may partly or wholly be hypo-
blastic, for there is no reason why the closely lying hypoblast
cells of the stomach should not proliferate to supply its com-
plement towards the bud, but it must be distinctly borne in
mind that I have no direct evidence at present in favour of this
view. Pl]. XX XVIII, fig. 26, is a slightly later stage. Ata much ~
later stage (P]. XXXVIII, fig. 27), below the hood cavity
lies the small circular stomach which contains a central cavity
and which is continued into a short blind intestine which
already possesses its normal curvature. I could discover no
connection between the stomach and the cavity of the hood.
Woodcut 3 would represent a diagram of such a stage. There
is no need to point out its parallelism with a similar stage in
so many other Polyzoa. The subsequent formation of an cso-

ON BUDDING IN POLYZOA. 531

phagus, and the later development of the bud may be passed
over.

Fic. 3.—Diagram to illustrate the probable relations of the lophophore and
stomach in a Loxosoma bud. The upper invagination is the lophophore
cavity, the lower is the foot-gland, the compressed body within the bud
is the stomach.

Uljanin (3) describes the development of the bud in Pedi-
cellina. A protuberance of the cuticle contains some round
clear cells; the outer soon arrange themselves as an epithe-
lium, and a constriction divides off the bud from the stem;
meanwhile two cavities appear in the central parenchyma; the
lower and larger one he rightly regards as the stomach, the
upper one he calls the “ brood pouch,” whereas it really is the
lophophore cavity, the lumen of which is at first quite distinct
from that of the stomach. There is nothing of further interest
to us in his paper.

Prof. Salensky (9) has also studied the development of the
bud in Pedicellina echinata, on p. 32 he says, “ At the
summit of the bud, several of the ectodermal cells elongate and
sink within; probably these cells give rise to the endoderm.”
The further development of the bud follows almost precisely
the same course as that which he gives for Loxosoma. As we
shall immediately see, Hatschek gives a different rendering of
the phenomenon, and I would point out that Salensky’s figure
of his earliest stage (Salensky pl. xiv, fig. 26) would very
well bear the former’s interpretation.

The fullest account of the budding Pedicellina is in the
very careful researches of Hatschek (8), in which he shows
that at the growing point of the stolon there is a single-layered
tubular mass of cells lying close beneath the external epithelium,

532 ALFRED C. HADDON.

which is continually dividing into two by transverse constric-
tion. Of these the anterior portion separates from the posterior,
and becomes connected with a solid mass of cells, which have
proliferated off from the external epithelium. This latter soon
acquires a lumen, and we have a pair of single-layered closed
sacs occupying a distinct prominence of the stolon, which
are beginning to be shut off from the general cavity of the
stolon by the neighbouriug fusiform mesoderm cells arranging
themselves into a diaphragm. Some of the scattered fusiform
cells of the stolon become cut off, and so pass into the bud ;
but at the junction of the primitive closed sac, with the pro-
liferating epithelium, there is one single mesoderm cell, which
by division soon forms a small rounded mass, and is apparently
concerned with the formation of the generative organs. The
larger anterior sac forms the intra-tentacular space, and by the
involution of its walls produces the tentacles, and of its floor
the cesophagus and the hind gut. A central solid invagination,
which shortly becomes a hollow sac, is the rudiment of the
nervous system. ‘The posterior smaller sac is prolonged and
bent upon itself, and becomes converted into the stomach and
intestine, communication taking place between the invagina-
tions which form the fore and the hind gut.

In the embryo Hatschek has discovered that a couple of
cells separate themselves from the oral side of the endoderm.
These form a single-layered sac, which becomes quite detached
from the alimentary tract of the embryo, and is connected with
a small ciliated invagination of the lateral epiblast; it also
possesses a mesoderm coating. This remarkable structure is
regarded by Hatschek, with great probability, as the first bud ;
and it will be noticed that it contains the three germinal layers
of the embryo. Unfortunately, there is a gap between this
stage and the earliest of his true stolon buds ; but it seems
pretty evident that the primitive single-layered tubular mass
of cells mentioned above is the persistent structure derived
from the stomach of the embryo. Assuming this to be the
case, we have then in every Pedicellina-bud the three
embryonic layers, each one of which gives rise to its traditional

ON BUDDING IN POLYZOA. 533

organs, viz. the epiblast, to the external skin, the lophophore,
the intra-tentacular space, the esophagus, rectum, and nervous
system; the hypoblast, to the stomach with its digestive cells,
and the intestines; the mesoblast, to the muscles and general
parenchyma of the body. The generative organs apparently
arise from a special mass of mesoderm cells, which very early
appear as a single cell, which may arise from the primitive
hypoblast, or may be one of the primitive embryonic mesoderm
cells. After describing the embryonic bud, Hatschek says
(p. 515):—** The whole formation, which we have just studied,
gives, as will be shown further on, the material for the con-
struction of all the secondary individuals of the stock, whilst
the whole of the remainder of the larva goes directly over to
the primary oldest individual.”

It appears that Prof. J. Reid (2) was one of the first (1845)
to point out the fact that new buds form on the stems of
Pedicellina echinata when the polypides die; it has also
been noted by several observers since. It would be a most
interesting fact if this process were found to take place when
no remnant of the polypide was left. The histology and mor-
phology of this phenomenon require to be elucidated.

Ecroprocra—GYMNOLAMATA
(Marine Polyzoa).

Nitsche (13) makes a distinction between the outer epithelium
of the endocyst and the inner muscular layer, and he derives the
outer epithelium of the lophophore and tentacles and the inner
epithelium of the alimentary canal of the bud in Flustra
membranacea from the former (‘ Epithelialschicht ”?)—in
other words, for him, the lophophore and alimentary canal of
the young bud have a purely epiblastic origin. The tentacular
sheath, the muscles and peritoneal lining, &c., of the polypide
being derived from the inner muscular tunic of the endocyst,
i.e. from mesoblastic tissue.

In Fl. membranacea all the changes in the decadence of
a polyp into a brown-body can be seen; this is yet more
clearly manifest in Aleyonidium hispidum. Here single

VOL. XXIII.——-NEW SER. NN

534 ALFRED C. HADDON.

zooecia likewise very frequently lose their polyps by decay,
but long before the polyps have lost their characteristic form,
and have become brown-bodies, the endocyst of the upper end
of this zoecium begins to form a new polyp by budding in-
wardly. In the same zowcium we very frequently find a
decaying polyp, which very distinctly shows its original
nature, together with a new young bud, which does not differ
in the least from the polyp-buds in the zocecium-buds at the
edge of the colony. Here, also, the new polyp is formed, just
as the old one, by the budding of the endocyst of the zocecium
inwardly (p. 466). By “polyp ” Nitsche, of course, refers to
the alimentary tract, as he accepts the dual nature of the
zooecium and its contained digestive apparatus.

Salensky (9) states (p. 55) that the internal tissue of the
two-layered bud is derived from the external epithelium of the
endocyst of the parent, and the outer from the internal layer.
Here again the lophophore and digestive tract are epiblastic
structures, while the mesoblast of the bud is derived from the
mesoblast of the parent.

Joliet (17) refers all the buds to his ‘‘endosare.” Under
the term ‘ endosarc ’ Joliet includes “all the formations which
one calls under the names of colonial nervous system, funiculus,
fusiform layer of the endocyst.” ‘It is this which constitutes
the muscular tunic of the fresh-water Bryozoa, the paren-
chyma of the stems of the stolons of the Pedicelline, and of
the feet of Loxosoma.”

He says that it is a “ provisional name,” “ which I shall be
quite disposed to change for another more general term as soon
as I shall have seen, or someone has shown me, its homologue
with the ectoderm or the endoderm of allied animals or of the
embryo.” Surely neither alternative is necessary! It will he
seen from what follows that I do not regard the funiculus as a
simple structure, nor the bud as entirely derived from the funi-
culus, therefore I cannot class all the contents of a zocecium,
save the outer layer of the endocyst, as being formed from an
homologous tissue. The tissue, as he describes it, answers in
position, structure, and generally in function, to the mesoblast

ON BUDDING IN POLYZOA. 535

of all other animals, and therefore it seems to be to be super-
fluous to coin a new term to express mesoblastic tissue.

Joliet asserts that in some cases the bud is entirely derived
from the funiculuu—HEucratea chelata, Vesicularia spi-
nosa (young zoecia), Beania mirabilis, Lepralia
Martyi, and L. granifera; in others, the bud is apparently
in intimate relation with the ‘ endocyst,’ but always connected
with a funiculus— Membranipora membranacea, M.
pilosa, and possibly the old zowcia of Vesicularia spinosa.
It must be remembered that Joliet limits the term ‘ endocyst’
simply to the external epithelium (epiblast) of the body wall.
He says (pp. 221-2): “ When a bud forms anew upon the
endocyst of an old cell, one generally sees that it is very early
provided with a funiculus, which, even then, almost attains its
(proper) diameter; and ever since my attention has been drawn
to this point I have never seen a bud formed under these con-
ditions which lacked this attachment. I am thus driven to
believe that the buds develop by preference upon the points
of the endocyst where the strands, of which I have spoken, are
fixed, and thus from their earliest state they are naturally
provided with a funicuJus.”

Thus Joliet is driven to admit, apparently against his in-
clination, that in some cases the ‘endocyst,’ outer epithelium
(epiblast), may participate in the formation of the bud, he
goes on to say (p. 247): “I should almost be tempted to
generalise and to say, to terminate, that in all the Bryozoa the
development of the polypide is made at the expense of the
pretended colonial nervous system, if the Pedicellinze did not
constitute, according to Salensky, a very serious and very
striking exception. This author, in a recent work (9), seeks
to demonstrate that the budding of the digestive tract, which
he compares to the Polypide, is made at the expense of the
endocyst. I here produce a figure certainly strongly resem-
bling his, and in which the bud is still reduced to five cells ;
but these cells do not appear to me to be directly united to
those of the endocyst, and have always appeared to me to have
more resemblance with the fusiform cells of the parenchyma.

536 ALFRED ©. HADDON.

Even supposing that the opinion of Salensky is justified, as we
shall see immediately, the tissue called nervous is directly
derived from the endocyst and that in the young buds of Vesi-
cularia spinosa the granules, at the expense of which the
bud is formed, belong to the colonial nervous system, and
are only the cells of the endocyst recently detached from
the walls, —one may say that the two cases are closely
allied.”

But, as in the case of Pedicellina, he begs the question
by some such argument as the following :—That, according to
Salensky, the new bud arises from cells proliferated from the
endocyst; that the endosarc similarly arises from the endocyst;
therefore we may say that in this case the bud arises from the
endosare !

It does not appear to me that Joliet’s figures illustrating the
proliferation of the endosarc from the “endocyst” are perfectly
conclusive. The figure of the bud of Pedicellina, which he
refers to above (his pl. xii, fig. 9) really proves nothing. His
fig. 1, pl. xii, is possibly more to the point, but then Hatschek
(accepting his statements to be true) has disposed of this most
thoroughly. The only other figure he gives us is that of the
vegetative extremity of a stolon of Bowerbankia imbricata
(pl. xii, fig. 2) ; it remains to be proved whether this case falls
with Pedicellina, or, if it exists, what is the exact inter-
pretation of this proliferation.

Dr. E. Ehlers (16) describes the phenomena of budding in
the form he has more particularly examined, Hypophorella
expansa (Ehlers). In the lateral branch from the stolon which
is about to form a new animal, and which we may term the
bud, he finds externally a cuticle within a nucleated blastema
(kernhaltiges Blastem); he does not find the two layers
which Nitsche describes in Flustra, but has thought, though
he cannot prove it, that the outer layer which forms the cuticle
may be a Syncytium, though ‘‘I have never succeeded in
showing nuclei in it,” corresponding to the cylindrical layer
found in Flustra. The bud increases greatly in size and early
assumes a form much like the adult zoocium. ‘The cuticle

ON BUDDING IN POLYZOA. 5d

passes into the ectocyst of the adult internally. Owing to the
rapid growth there is a large central cavity, the walls are un-
doubtedly lined with the syncytium and with a portion of the
original central blastema, while from the apex depends, icicle-
like, the remainder of the blastema. This latter is to form the
alimentary tract, its peritonenal lining and some of the muscles,
the rest are formed as processes from the body-wall, i.e.
from its inner layer. The tentacle-sheath is formed by an
invagination of the tissue at the apex of the bud, so that the
syncytium forms its inner lining ; that is really the future outer
layer of the tentacular sheath. ‘“‘ With regard to this homo-
geneous outer layer of the Hypophorella bud, in which one
would like to see the homologue of a cell-layer, I must remark
that I have never seen it continued into the first rudiment of
the gut” (p. 109). “The endoderm appears as a separate
development of the tissue of the indifferent body-wall at the
spot distinguished by the above-mentioned invagination.” The
tentacle disc (Tentakelscheibe) or incipient lophophore is
formed from the endoderm. There is very early a cavity in
the formative material of the alimentary tract; this is the
stomach. The tentacles grow out from the edge of the tentacle
disc, at first only eight. ‘The remaining two or three appear
a little later. [It is usually stated that the permanent number
of tentacles arise from the first, but in several forms, e.g. F].
papyracea, Diachoris magellanica, &c., I have observed
that four lateral ones usually appear first, the more central
being the larger, but I have not yet satisfied myself as to the
exact rhythm ; ‘lateral’ has, of course, relation to the median
line as marked by the mouth and anus.—A. C. H.]

As far as I can discover, Ehlers speaks of the blastema which
forms the alimentary canal as “endoderm,” because it does
produce that structure, and naturally speaks of the remainder
as mesoderm, while he really has no doubt that the outer homo-
geneous layer is the ectoderm. In this all critics will probably
agree with him, but the exact origin of this blastema has yet
to be demonstrated. I would, however, join issue with our
author on one point, and that is the origin of the tentacle-disc.

538 ALFRED C. HADDON.

On Taf. IV, fig. 34, he figures the invagination of the
tentacle-sheath, and in the bottom of this depression he shows
two large cells. These he imagines form the inner portion of
the sheath; apparently they have the same optical character,
as the incipient tentacle-disc, and from what I can make out
that organ has not yet appeared. It is strange that the tentacle-
sheath should so early differ in optical characters from the
remainder of the alimentary tract, when it is derived, by him,
from the same tissue, so I am strongly inclined to suspect that
this careful observer has fallen into an error, and that the
lophophore like the outer portion of the tentacular-sheath, is
really an epiblastic derivative, which later on acquires con-
tinuity, as far as its cavity is concerned, with the remainder of
the digestive apparatus. This correction will give morpho-
logical completeness to the whole process.

Claparéde (12) derives the buds from the endocyst both in
the larve and in the adults. In his description of the larva,
he says: ‘‘From a certain spot of the endocyst an oval
mass projects towards the interior, in which a cavity soon
appears. This hollow structure entirely corresponds with the
invaginated sac of a young Bugula-bud, the development con-
tinues henceforth in a perfect parallel with that of the bud.
It is very probable that this sac arises from the primitive
mouth-furrow of the larva, but I have not directly observed
that it does so. I need not describe the formation of the
polypide within the sac, as it resembles to a hair that of the
polypides of the bud,” p. 169.

From an examination of Smitt’s Plates (11), it would seem
that the lophophore and cesophagus are at first distinct from
the digestive tract in Tubulipora serpens (pl. iv, fig. 9),
and in Alcyonidium parasiticum (pl. v, fig. 13-14), and
that they subsequently become united. Hincks, who is the
English exponent of Smitt, clearly states Smitt’s opinion that
the buds are derived from the brown-body (‘‘ germ-capsule ”)
at all events in many cases; but this view has been so fully
discussed and combatted by all subsequent writers, that I need
not dwell on it further,

ON BUDDING IN POLYZOA. 539

Hincks, in his admirable Monograph (1), adds his testimony
to that of Smitt, but is willing to admit that in many cases
the buds may be derived from the endocyst or funicular tissue.
He does not really go into the question of gemmation, nor
does he give any perfectly satisfactory observations of his own,
neither does he discuss the morphology of the phenomenon.

EctoPprocrAa—PHYLACTOLEMATA.
Freshwater Polyzoa.

Allman in his beautiful monograph (19) says: “ With the
exception of some peculiar forms of gemme (statoblasts) to be
presently described, these bodies (gemmz) always originate in
the endocyst.” Hethen goes on to describe the process of
gemmation in Paludicella and in Lophopus. The figures
which he gives bear out his view, but all his observations were
made from living examples, and thus he has not seen the cells
implicated in the process, nor verified his results by means of
sections. It is thus left uncertain what exact part is played by

Fic. 4.—Diagram of embryo of Alcyonella, modified from Allman.
e, Ciliated epiblast. m. Mesoblast. 4. Hypoblast. 4. c. Body-cavity.

the external cells (epiblast) and the inner network of muscular
fibres (mesoblast) of the endocyst, but judging from pl. xi,
figs. 5, 9, and 10—14, it would seem that the epiblast of the
parent gives rise to all the alimentary organs of the bud, while
the mesoblast of the mother, passes into the mesoblast of the

540 ALFRED ©. HADDON.

bud. I would point out that his pl. xi, figs. 7—9, and 12—
15, suggest a double origin of the alimentary organs, and that
the connection between the cavity of the lophophore and the
lumen of the stomach occurs comparatively late.

The account of the development of Alcyonella by Dr.
Allman is unfortunately far from satisfactory, and I would
venture to suggest another interpretation (fig. 4) of the stage
represented in pl. xi, fig. 30, and by No. 4, fig. 5, p. 34,
which is, that the polypide is developed from the remains of
the archenteron of the embryo, probably by a direct conversion
of the walls and of the lumen of the archenteron into those of
the alimentary tract of the young polypide. The lophophore
and cesophagus would be derived from the overlying epiblast.
The remainder of the body wall of the embryo, consisting of
epiblast and peripheral mesoblast, “ becomes enveloped in an
ectocyst, to constitute the cell of the adult polyzoon. The sub-
sequent changes are produced by the gemmation of new poly-
pides, with their proper ectocysts and endocysts” (p. 35). In
other words the embryo passes over entirely into the first adult
of the colony. Ata very early stage, between figs. 30 and 31,
a second polypide makes its appearance ; there is little doubt
that this second bud is constricted off from the older polypide,
although Allman leaves one to suppose that it, like the former,
“appears to take place in a manner quite similar to that by
which new polypides are produced by gemmation from the
walls of the endocystal cavity in the adult” (p. 34). It is
very unfortunate that Allman should have derived the alimen-
tary canal from the epiblast, when hypoblast already was
present in the embryo,

“Plumatella fruticosa presents similar developmental
phenomena ; the ciliated larva, however, in this species, differs
from that just described, in having its polypide single.”

I do not propose to discuss the morphology of the statoblasts
at present. Allman (I. c. p. 38) describes how they take their
origin entirely from the funiculus. Ultimately ‘a young
polyzoan gradually emerges and floats away. . . . At the
period of its escape it possesses all the essential organization

ON. BUDDING IN POLYZOA. 541

of the adult. . . It loses no time, however, in developing
gemme, which soon change it to the compound form of the
adult” (p. 39).

Metschnikoff (21) in his studies on Alcyonella describes
the formation of the bud in the embryo. Allman in his Pre-
sidential Address to the Linnean Society (24) thus narrates it.
The segmentation of the egg produces “a central cavity sur-
rounded by a double layer of cells. This constitutes the cyst
of the well known Alcyonella-larva, within which two poly-
pides subsequently make their appearance by budding. In
this budding both lamina of the cyst-walls participate. The
outer lamina serves for the formation of the outer epithelium
of the tentacles, and the inner epithelium of the alimentary
canal ; while the central nervous system, which in the larva is
very large, is also most probably derived from it. The inner
lamina, on the other hand, forms all the muscles of the body,
as well as the genitalia and the inner epithelium of the body
cavity.”

Nitsche (22) has also studied the budding in Alcyonella
fungosa andin Cristatella mucedo. I again quote from
Allman’s address, p. 499, “ He (Nitsche) had already shown
that the wall of the cystid or zoecium of Aleyonella con-
sists of three different layers besides the externally excreted
ectocyst or cuticula. ‘These are an outer epithelium, an inner
epithelium, and a tunica muscularis lying between the two
and consisting of a structureless supporting membrane on
which lie transverse and longitudinal muscular fibres. The
first indication of the polypide-bud shows: itself as a sac-like
bulging inwards of the cystid wall. In this bulging the
tunica muscularis, however, takes no part, but seems to be
absorbed at the spot where the bud occurs. The polypide-bud
consists therefore at this stage of a two-layered cellular sac,
whose inner layer, bounding its central cavity, passes continu-
ously into the outer epithelium of the cystid wall, while the
outer layer is continuous with the inner epithelium of the
cystid. Nitsche follows Metschnikoff in regarding the outer
epithelium of the cystid as the outer germinal layer or ecto-

542 ALFRED C. HADDON.

derm, the inner epithelium as the inner germinal layer or
endoderm ; and if we further regard the tunica muscularis as a
middle germinal layer or mesoderm, we may view the young
polypide-bud as composed of two concentric cellular layers,
the internal derived from the ectoderm, the external from the
endoderm of the cystid, while the mesoderm of the cystid takes
no part in the formation of the bud. . . Folds and secon-
dary introversions of this two-layered cellular sac give to the
young polypide its definite form. . . The inner epithelium
of the alimentary canal is derived from the ectoderm of the
cystid, while the outer is derived from the endoderm. The
two layers of the tentacular sheath have a precisely similar
derivation.” A detailed account of the further development is
given, and on p. 001 we read, ‘‘as Nitsche suggests, we must
not in the present instance lose sight of the fact that the inner
layer of the bud, though arising from the ectoderm of the
cystid, has fundamentally different relations from those of an
ordinary ectoderm, for there proceeds from it, at the same time
with the nervous substance of the ganglion, the entire epithe-
lial lining of the intestinal tract.” While the ‘‘ endoderm ” of
the cystid behaves in all respects like an ordinary Mesoblastic
tissue. What could have been the conditions which in
process of time have so upset the traditional functions of the
germinal layers !

Hatschek (8) describes, in Cristatella, the relations of the
polypides to the colony and the increase by gemmation. In
the lateral growing points he finds that there is a single-
layered sac, depending within the cavity of the stolon and
slung by a mesodermic layer. ‘This sac constricts into two
unequal portions: the portion constricted off becoming the
inner epithelium of the alimentary tract of a new polypide, the
tentacular portions being supplied by an epiblastic involution
(Hatschek, fig. 3, p. 539). This process being repeated,
Hatschek says that Nitsche’s figures (22 and 23) do not prove
his statement that in Aleyonella the inner layer of the poly-
pide sac is derived from the ectoderm of the cystid, and goes on
to say that Nitsche’s figures will bear his (Hatschek’s) inter-

ON BUDDING LN POLYZOA. 543

pretation of the similar process in Cristatella. This con-
tinually constricting sac described by Hatschek lies between
the epiblast and the mesoblast of the stolon, and it is quite
open to us to discuss its morphological value. If we look upon
this sac as hypoblastic tissue derived from the archenteron of
the embryo, the budding of these fresh-water Polyzoa would
present no difficulty. Allman describes the initiatory steps of
the formation of a colony. If we accept a different (i.e. a
hypoblastic) origin of the internal epithelium of the alimentary
tract of the earliest polypides than that which Allman indicates,
then the two accounts mutually assist one another.

Reinhardt (25) states that in Aleyonella fungosa and
Cristatella mucedo: “ After segmentation the mass is con-
verted into a gastrula by invagination; the gastrula-mouth
closes and the segmentation cavity disappears.’” :

“The cystid (in Cristatella) consists, as in Alcyonella,
of an ectoderm, a median layer (the tunica muscularis), and an
entoderm. ‘Thus, Hatsckek must be wrong when he names
the inner layer of the bud mescderm, and his description of
the budding is inexplicable by comparison with the above-
mentioned details, though these may perhaps correspond with
his second unknown process of budding. The bud develops by
a thickening of the ectoderm into which the entodermic cells
are pushed; there is no indentation of the former. The
tunica-muscularis is very early formed; the cavity of the
tentacle-sheath is separated later from the alimentary canal,
and the lophophore is formed by an invagination into this
tentacle-sheath. The later development of the buds corre-
sponds with that described by Nitsche in Alcyonella.”
(From the English abstract.)

Reinhardt clearly gives us the three germinal layers, it is
difficult to understand him perfectly as he gives no figures, but,
accepting his statements, we apparently have an embryo in
which the alimentary canal has a retarded development, an
embryo which is, in fact, all body-cavity, such an embryo can
easily result from an ordinary enterocoelous form, such as
Argiope, Sagitta, &c., by an exaggeration of the ccelomic

544 ALFRED C. HADDON.

diverticula and a simultaneous arrest in the formation of the
permanent alimentary canal. The following diagram (fig. 5)
will sufficiently explain my meaning. This would make the
body-cavity of these forms an enterocoel. The musculature of
the body-wall appears to develop prior to the formation of
these diverticula.

This author brings into harmony the observations of Allman,
Metschnikoff and Nitsche, for we have only to concede that
the epithelial lining of the body-cavity of the embryo (cystid)
in my interpretation of Allman, (which, by the bye, was made
before I had seen the accounts by Nitsche and Reinhardt), is
derived from archenteric diverticula ; an earlier or later deve-
lopment of the tunica muscularis between the two layers, is
really of little consequence.

Fie. 5.—Diagram representing a possible degradation in the formation of the
alimentary tract from an originally enterocoelous larva.

Salensky (9) states that his own observations on Paludi-
cella have convinced him that the superior layer of the
zooecium gives rise to the lophophore and the internal epithe-
lium of the polypide, while the ‘inferior Jayer is transformed
into the interior layer of the zocwcium, the tentacular sheath,
and, at the same time, into the muscles. He says (p. 56) :—
“Tt is impossible not to remark the interesting analogy exist-
ing between these two layers and the embryonic layers of other
animals.” . . . . ‘It is acknowledged by several recent
embryological researches that the endoderm of many animals
forms from the ectoderm, sometimes as a thickening of this
latter, sometimes as an invagination.”

This “ analogy,”
not be relied on as giving any true insight into the nature of
the phenomenon of budding, for we cannot look upon the
epiblastic layer of the endocyst as being morphologically

which other authors have remarked, must

ON BUDDING IN POLYZOA. 645

equivalent to the embryonic cells (segmentation spheres) of
larvee in the blastula stage. Moreover, the invaginated or
grown-over hypoblast cells of the gastrula stage are not
derived from the epiblast (ectoderm). Before this difference
in position these two layers are usually optically, and they
certainly are morphologically and physiologically, quite dis-
tinct—apparent optical similarity can never constitute morpho-
logical identity. The fundamental difference between the
epiblast and the hypoblast is shown by the usually very early
distinction between these two layers. For instance, the two
layers are often practically distinguishable in the stage of
eight segmentation spheres, and even in some cases the first
segmentation furrow marks their distinctness. We must,
therefore, disallow the use of the term “ endoderm” for that
mass of cells derived from the epiblast of the parent zooecium,
which is supposed to give rise to the alimentary canal of the
new polypide.

We can agree with Salensky when he continues :—* From
its position and from the formations which it produces, the
inferior layer of the zocecium presents a great resemblance to
the mesoderm.” He might have added that they are one and
the same.

This method of bud-formation he believes to be common to
the whole group of the Polyzoa.

A. Hyatt, in his elaborate memoir on the ‘ Sub-order Phylac-
tolemata’ (20), scarcely alludes to the phenomenon of gemma-
tion. On p. 221 he says :—* The free part of the endocyst of
the cell on the abdominal side, bringing forth true buds.’
And on p. 218 he gives a sketchy account of how “ the stato-
blastic polypide begins to multiply by the process of budding.
An internal swelling of the endocyst, on the lower side, in. the
vicinity of the bases of the anterior retractor muscles, first
shows the position of the coming polypide. This elongates
into a little hollow sac with a thickened rim, upon the upper
edge of which, in the Hippocrepian Polyzoa, a slight notch is
formed by the duplication and pushing out of its sides into two
loops joined along the centre’ (the lophophore). . . . “A

546 ALFRED C. HADDON.

transverse constriction of the body of the little sac draws the
line between the cesophagus and the stomach; and the subse-
quent deepening of this constriction divides off the internal
cavity, establishing the cardiac and pyloric valves.” ‘The
figures which Mr. Hyatt gives are most unsatisfactory, nor
does he appear to have checked his observations by means of
sections. The minute outlines which he gives of Cristatella
ophidioidea (Hyatt) will equally well bear Hatschek’s in-
terpretation. The incipient bud of Fredericella regina
(Leidy) (pl. vii, fig. 5, v) is made to depend from the endo-
cyst, but we are not informed as to the significance of the two
layers which he there depicts. The figure of Plumatella
arethusa (Hyatt) may have any interpretation. The only
point which is quite clear is that Hyatt believes that the
polypide buds are entirely derived from the endocyst of the
parent.

Dumortier (18) made some interesting observations on
Lophopus crystallinus. He states that he has seen balls
of mucus floating in the general fluid of the body become
attached to the body-wall. ‘‘I have said that these globules
appear to be of the nature of mucus, for, besides that they are
formed by the stomach, an eminently mucous organ, their sub-
stance does not permit the supposition that they are endowed
with any organised tissue” (p. 49). ‘‘The adventitious bud
_ once disposed of, it soon establishes a focus of irritation in that
place, which will excite the development in the interior of its
mass, and a protuberance at the exterior so as to make a
bump.” Dumortier states that the alimentary tract is de-
veloped from this ball of mucus, while the tentacles are
developed from the body-wall. The tendency of his statements
is towards the view of the hypoblastic origin of the digestive
portion and an epibiastic origin of the lophophore, though, of
course, this view of the case could not present itself to him

(1836).

GENERAL CONCLUSIONS.

In all cases of budding in the animal kingdom, so far as I am

ON BUDDING IN POLYZOA. 547

aware, it has been shown that representatives of the three
primary germinal layers enter into the bud, and there form
corresponding tissues, but, strangely enough, the Polyzoa form
an apparent exception to this rule, as the buds are said to arise
solely from the endocyst (Nitsche, &c.), or from the endosare
(Joliet). Assuming the generally received opinion of the nature
of these tissues, in neither case would the bud have any hypo-
blast in its composition. It is inconceivable to me how a bud
could originate unless it possessed an offshoot from all the essen-
tial organs of its parent: that is to say, the bud should possess
a portion of the parental epiblast, mesoblast, and hypoblast ;
for how could either the epiblast or the mesoblast suddenly
depart from its ancestral traditions and take upon itself the
function of digestion? It is conceivable that, in process of
time, the method of gemmation should be considerably modi-
fied, but hardly that one of the most important of the three
primary embryonic tissues should not be represented at all.
Embryologists are fully conversant with variations in the deve-
lopment of organs, and with the masking of the origin of certain
organs, as in the case of “‘ precocious segregation,” but they
nevertheless have firm faith in the essential ‘‘conservatism” of
the layers themselves.

The question now before us is: Are the three germinal
layers represented in the buds of Polyzoa? The following are
my reasons for answering this in the affirmative.

Nitsche and others, as we have noticed above, would derive
the whole of the bud from the endocyst—that is, from the
epiblast and peripheral mesoblast. Joliet, in combating this
view, points out that in Kucratea chelata and in all the
Cheilostomata which he has studied he has reason to believe
that the bud is really formed on some portion of the endosare,
and not on the endocyst. In Hypophorella, Ehlers, the bud
‘is produced on the funiculus in the centre of the cell, as in
Eucratea. In many cases it is developed at the very base of the
zocecium, immediately over the communication plate or septum
and the orifices through which the connective threads pass, and
therefore probably in connexion with the endosare. 1[Hincks]

548 ALFRED C. HADDON.

have observed it in this position in the young cell of Beania
mirabilis; and in this species Joliet has convinced himself
that the polypide is actually derived from the endosarcal cord.
In the rudimentary zooecium of Victorella pavida the
forming polypide seems to me [Hincks] to be enveloped in the
endosarcal plexus, and to be (in all probability) produced by
it. . . . It may be, as Joliet suggests, that the authors
who have referred it to the endocyst have not been sufficiently
alive to the distinction between these two tissues. It may be
that the function is to some extent shared by the endocyst ”
(Hincks, pp. l, hi).

Joliet thus clearly believes in the endosarcal origin of the
bud. This ‘endosarc’ is, by him, derived from the endocyst.
In his use of the term ‘endocyst’ one must not understand
both layers, but only the outer. . This seems to be clear from
Joliet’s account, and from his deriving migratory cells also
from the growing end of the outer layer. The inner layer of
the endocyst (peripheral mesoblast or somatopleure) is com-
posed, according to all authors, of fusiform cells—i. e. cells
similar to the characteristic cells of Joliet’s endosare. It is
thus certain that Joliet would consider the buds of the Polyzoa
as composed solely of mesoblastic tissue, or possibly of some
modified epiblast as well.

We have seen that in Flustra carbasea the tentacles and
the mouth area arise from one mass of tissue, and from the
latter an invagination takes place forming the mouth and
cesophagus (Stomodeeum); whereas the stomach and intestine
arise from another mass of tissue. These two closed sacs
(Stomodeeum and stomach) later on unite to form a continuous
tube. It was this well marked double origin of the digestive
tract which first led me, when in Naples, in 1879, to study
the question of Polyzoan gemmation. I have already enume-
rated some of the forms in which I have since seen the same
phenomenon.

The resemblance of the above to the formation of similar
structures in the embryos of so many animals is most striking,
and seems to suggest that we have here to deal with an epi-

ON BUDDING IN POLYZOA. 549

blastic derivative which forms the outer layer of the tentacular
sheath, the outer epithelium of the tentacles, the mouth area,
and the lining of the cesophagus; and with a hypoblastic deri-
vative which occupies itself with the inner lining of the stomach
and intestine. We may safely assert that the outer layer of
the incipient polypide is mesoblastic as it develops into the
inner layer of the tentacular sheath, the inner epithelium of
the tentacles (somatopleure), and into the investing sheath of
the alimentary canal (splanchnopleure), as well as into the
muscles of the future polypide. Nitsche and Hatschek show
for the Phylactolemata, and the latter for Pedicellina, that
the nervous system is derived from an epiblastic invagination.
There are no observations for the Gymnolemata, as to the
ganglion, but it is in such close contact with the lophophore
that we may safely assume its origin from that body. This
would, of course, give it an epiblastic derivation.

Prof. G. J. Allman was, I believe, the first to promulgate the
view that the zooecium and the polypide are distinct individuals,
at all events this statement is very generally accepted; but it
seems rather incredible that generations of individuals
solely composed of a digestive canal and its appurte-
nances, such as muscles and nerve ganglion, seg-
mental organ, and possibly generative organs, should
live within the body cavity of one persistent indi-
vidual which lacks these organs and only possesses
a body wall, funiculus (?) and body cavity.

1 The analogies which have been drawn between this supposed phenomenon,
and the undoubted cases of physiological and structural differentiation amongst
the Hydrozoa, will not really hold good; for in these the buds, though still
connected, are all external, and their specialisation can readily be accounted
for ; whereas, in the other case, each successive internal so-called bud develops
within the body-cavity of iis parent in such a manner as to have precisely the
same relations as if it really was its alimentary tract, and not a bud. It is
not easy to conceive how this could come about, nor is it rendered any easier
if we yet farther follow the distinguished author of this view, and regard the
zooecium as the host not only of a nutritive polypide, but also of male and
female individuals; for Prof. Allman suggests that even the testis and the
ovary are, save sexually, aborted polypides !

VOL, XX1II.—NEW SER, 0oO

000 ALFRED ©. HADDON.

It is impossible to regard the body-wall and the alimentary
canal of the Entoprocta as distinct individuals, and their
gemmation resembles, in its essentials, that of those animals
which can multiply by budding (e.g. Ascidians). The budding
of some of the Phylactolemata, too, does not necessitate this
strange commensalism. Why, then, should it only occur
amongst the Gymnolemata ?

Let us admit that the previous inhabitant of a zocecium dies
away altogether, but before doing so gives rise toa bud in a
normal manner, which bud is primitively located on the oral
wall of the zocecium of its parent. The future history of the
bud would present no startling peculiarities if its growth were
to take place in two directions; if some of the epiblastic and
mesoblastic portions of the bud tended to form the body-wall
of the new Polyzoan: as it is already provided with an ectocyst
there would be no need to form a new one, so the new body-
wall would simply be applied to the dead cyst. Meanwhile,
an epiblastic involution depends into the body cavity of the
newly-formed individual, carrying down with it the hypoblastic
derivation from the parent, both being coated with a meso-
blastic sheath. This is the structure which has been regarded
by authors as a whole bud, and which has been variously termed
“bud,” © zooid,” “ polypide,” and “ polyp,”’ but which I make
bold to say is merely a portion of the new bud. (It will be
noticed that in the preceding pages I refer to this structure
under the generally received terminology, I purposely do so to
prevent any confusion.) I have already detailed the future
history of this part of the bud, so it would be superfluous
to repeat it again here, and that of the body-wall has no especial
interest.

It might be objected that the funicular tissue extends through-
out the entire colony, and that it does not die with the tempo-
rary inhabitants of the zocecium; assuming this to be the case,
there is nothing to prevent this tissue being enclosed by the
body-wall of the growing bud, without its being a primitive
portion of that bud, after being thus enclosed it would serve
to connect the new member with the rest of the colony, and

ON BUDDING IN POLYZOA. 55

by this means the bud would be engrafted into the life of the
whole, for, undoubtedly, without being histologically nervous,
this tissue can transmit stimuli, and it certainly possesses
other important functions. It is difficult to conceive of portions
dying and being renewed de novo, besides, having such
undifferentiated functions it would a priori have greater
vitality and be less likely to die with each individual, especially
as it is all the time protected from external damage by the
walls of the zocecium.

I have shown that in Eucratea chelata the bud partly
arises from the endocyst, and therefore we must be cautious in
accepting Joliet’s statement as to the universality of the origin
of the buds from the endosare.

We, have, however, just seen that Joliet and Hincks lead
one to imagine the possiblity of different tissues, the endocyst
and the endosare (funicular tissue) being implicated in the
gemmation of certain forms, and my own observations very
strongly incline me to this view. Hatschek’s beautiful inves-
tigations are very clear as to the complex origin of the bud, and
practically prove that all the three germinal layers are con-
cerned in the budding of Pedicellina and Cristatella.

To recapitulate :—In the Entoprocta, Hatschek’s observa-
tions prove the process of gemmation to be normal in Pedicel-
lina. My own on Loxosoma indicate that no real anomaly
exists in that form.

The discrepancies of most observers, combined with the
errors of some in their interpretation of the phenomena in
Pedicellina, will allow us some latitude in dealing with the
generally received views on the budding in Loxosoma.

In the Phylactolematous Ectoprocta, Hatschek’s account of
Cristatella gives a clue as to what will possibly prove to be
the characteristic method of gemmation in the group, and it is
one which has every morphological probability.

The absence (?) of statoblasts in Paludicella may perhaps
be accounted for by supposing that, compared with the true
Phylactolemata, this form is a late immigrant into fresh water,
and that it still retains most of the structural characteristics of

552 ALFRED ©. HADDON.

the Marine Ectoprocta. If this be the case, it is probable
that the mode of gemmation in this Polyzoan will be found to
resemble that in the latter rather than that in the former.

The Gymnolematous Ectoprocta present us with the great-
est difficulty, and it must be remembered that we have here to
deal with a highly specialised and at the same time degraded
group—the degradation being mainly caused by the sessile
habits and by the secretion of a strong protective covering,
resulting not only in the loss or diminution of certain organs,
such as a muscular body-wall, nervous system, sense organs,
excretory ogans, &c., but also in the simplication of certain
tissues. This is especially noticeable in the body-wall and in
the mesoblastic tissues generally, the tendency apparently
being for these tissues to lose their distinctive cellular character
and to form syncytia or even plasmodia ; for the vagrant protean
funiculus is more comparable with a plasmodium, in which
the fusiform cells described by Joliet are immersed, than with
an ordinary cellular tissue.

In many forms of this group both the endocyst and the
funiculus appear to take part in the gemmation. I would
again draw attention to the marginal buds of Flustra (PI.
XXXVIII, fig. 16) and Bugula flabellata (Pl. XXXVIII,
fig. 17), in the latter of which the ovary lies in such close
contact with the fundus of the developing stomach that it
suggests something more than a secondary attachment. The
ovary (shown by Huxley to be developed from the funiculus in
Bugula avicularia), as is well known, passes ready formed
into some buds imbedded in certain funicular tissue. Might
we not assume that the stomach tissue also has a similar
origin? Indeed, some still earlier buds exhibit a very close
connection between the stomach and the funiculus. In most
of the forms enumerated on p. 523 I have seen the stomach
intimately united with the funiculus in early buds, and,
though I have not yet been able to prove that the stomach
mass does absolutely and entirely arise from the funicular
tissue, yet the evidence in favour of that view is, to my mind,
very strong.

ON BUDDING IN POLYZOA. 503

There is, however, a certain amount of direct evidence that
a portion of the bud is derived either from an invagination or
from a proliferation from the outer layer of the endocyst—in
other words, from the epiblast of the parent organism (see Pl.
XXXVIII, fig. 23, &c.).

Every one will agree that the bud contains mesoblastic
elements directly derived from the parent.

Assuming, then, that the digestive tissue of the bud is
derived from the funiculus of the parent, a new construction
must be put upon this important organ of the Polyzoa, neces-
sitating a hypoblastic origin for a part at least of this much
discussed tissue. I would venture to suggest that, at all events
in the Gymnolemata, a portion of the cord is indirectly
derived from the archenteron of the embryo which initiated
the colony. This derivative may be plasmodic rather than
cellular, and probably is more or less clothed with degenerate
mesoblast. If subsequent investigations can demonstrate this,
then the anomalous character of Polyzoan gemmation will be
taken away, and the phenomenon reduced to a more normal
method.

Whatever value the suggestions just put forth may possess,
this paper will at least indicate the lines upon which this
question must be approached in the future.

BIBLIOGRAPHY.
General.

(1) T. Hixcxs.—* A History of the British Marine Polyzoa.” London, Van
Voorst, 1880.
(1*) J, Barxois.—* Recherches sur l’emb. des Bryozoaires,”’ Lille, 1877.
Entoprocta.

(2) J. Rerp.— On the Anatomy and Physiology of some Zoophytes,” ‘* Ann.
and Mag. Nat. Hist.,’ vol. xvi, 1845.

(3) B. Uusanin.—“ Zur Anatomie u. Entwickl. der Pedicellina,” No. 2,
* Bull. Soc. Imp. des Natural,’ Moscow, 1870,

(4) H. Nitscur.—“ Beitrage zur Kenntniss der Bryozoen,” “ Ueber den Bau
u. d. Kuospung v. Loxosoma Kefersteinii,” ‘ Zeit. f. Wiss. Zool.,’? Bd.
xxv, Suppl. Bd., p. 361, 1875.

554. ALFRED C. HADDON.

(5) O. Scumipt.— Die Gattung Loxosoma,” ‘Archiv. f. Mik. Anat.,’ xiii,
1876.

(6) C. Voer.— Sur le Loxosome des Phascolosomes,” ‘Archiv. de Zool.
expér. et génér.,’ Tom. v, 1876.

(7) T. Hincxs.— Loxosoma, by Carl Vogt,” Translation and Condensation
of above, ‘ Quart. Journ. of Mic. Sci.,’ vol. xvii, New Ser., 1877.

(8) B. Harscurx.—‘“ Embryonalentwick. u. Knospung d. Pedicellina echi-
nata,’’ ‘ Zeit. f. Wiss. Zool.,’ Bd. xxix, 1877.

(9) M. SaLensky.—“ Etudes sur les Bryozoaires Entoproctes,” ‘ Ann. des
Sci. Nat.,’ 6 sér. Zool., Tom. v, 1877.

(10) T. Hivcxs.—“ Salensky on the Polyzoa Entoprocta,” Abstract of above,
“* Quart. Journ. of Mic. Sci.,’ vol. xviii, New Ser., 1878.

Ectoprocta Gymnolemata.

(11) F. A. Smirr.—“* Om Hafs-Bryozoernas Utveckling och Fettkroppar,”
‘ Aifvers af Kongl. Vet.-Akad. Forh.,’ 1865, No. 1. Stockholm.

(12) E. Crararnpse.— Beitrage zur Anat. u. Entwick. d. Seebryzoen,”
‘Zeit. f. Wiss. Zool.,’ Bd. xxi, 1871.

(18) H. Nitrscue.— Beitrage zur Kenntniss der Bryozoen,” «Zeit. f. Wiss.
Zool.,’ Bd. xxi, 1871.

(14) W. Repracnorr.—‘ Zur Entwick. der Tendra zostericola,” ‘ Zeit. f.
Wiss. Zool.,’ Bd. xxv, 1875.

(15) W. Repracnorr.—* Zur Naturgeschichte der chilostomen Seebryozoen,”
‘ Zeit. f. Wiss. Zool.,’ Bd. xxvi, 1876.

(16) E. Exturs.—‘ Hypophorella expansa. Ein Beitrag. zur Kenntniss der
minirenden Bryozoen.” Gottingen, 1876. (‘Abhandl. d. Kénigl.
Gessells. d. Wiss. zu Gottingen,’ Bd. xxi, 1876.)

(17) L. Jotrnr.—“ Contributions a Phist. nat. d. Bryozoaires des cétes de
France,” ‘Arch. d. Zool. Exp.,’ vol. vi, 1877.

Ectoprocta Phylactolemata.
(18) B. C. Dumortrer.—‘ Mém. sur l’anat. et la phys. des Polypiers com-
posés d’eau douce nommés Lophopodes.’ Tournay, 1836.
(19) G. J. Artman.—' A Monograph of the Tresh-water Polyzoa, Ray
Society, London, 1856.

(20) A. Hyarr.— Observations on Polyzoa, Suborder Phylactoleemata,”
‘ Proc. of the Essex Inst.,’ vol. iv, No. 8, 1865. Salem, Mass.

(21) Murscunixorr.— Bull. de l’Acad. de St. Pétersbourg,’ Tom. xv, 1871.

ON BUDDING IN POLYZOA. . 559

(22) H. Nitscuz.—“ Untersuch. u. d. Knospung d. Siisswasserbryozoen, insbe-
sondere der Alcyonella,” ‘ Sitzungsberichte der naturforschenden Ge-
selischaft zu Leipzig,’ 1874.

(23) H. Nrtscuz.—* Ueb. d. Knospung der Polypide d. Phylactoleemen Siiss-
wasserbryozoen,” ‘Zeit. f. Wiss. Zool.,’ Bd. xxv, Suppl. Bd., p. 343,
1875.

(24) Atuman..—“ Recent Progress in our Knowledge of the Structure and
Development of the Phylactolematous Polyzoa,” Anniversary Address,
‘Jour. Linn, Soc.—Zool.,’ vol. xiv, 1879.

(25) W. Reinnarpt.—* Zur Kenntnis der Siiswasser-Bryozoen,” ‘ Zool.

Anzieger,’ iii, No, 54, 1880, Abstract in ‘Journ. of Roy. Mic. Soc.,’
ili, No. 4, 1880.

556 SYDNEY J. HICKSON.

The Structure and Relations of Tubipora.
By
Sydney J. Hickson, B.A., B.Sc.,

Scholar of Downing College, Cambridge, and Assistant to the Linacre Pro-
fessor, Oxford.

With Plates XXXIX and XL.

Amoncst a most valuable series of specimens collected at
Zanzibar, by Mr. G. Gulliver, M.A., M.D., late Naturalist to the
Transit of Venus expedition, was a fine specimen of Tubi-
pora purpurea, which had been carefully and admirably
preserved in strong methylated spirit. Having the good fortune
to have this material at my disposal and likewise some fine dried
specimens in the Oxford University Museum, I undertook, at
the suggestion of Professor Moseley, a renewed examination of
both the hard and soft parts in this important and interesting
genus. My main objects were (Ist) to examine and record
the varieties of the structures known as the “ infundibuliform
tabule ” in the dried specimens; (2nd) to discover the true
meaning of these tabule by sections made through spirit
specimens of the corallites; and (8rd) to clear up, as far as
possible, certain of the other doubtful points in the anatomy
and histology of the genus. For the first of these I found
that a lump of Tubipora musica given to me by Professor
Moseley, afforded the most fruitful results. My second object
was fully achieved upon the specimen of Tubipora purpurea
brought from Zanzibar by Mr. Gulliver. The same specimen
also gave me some excellent results in studying the anatomy

THE STRUCTURE AND RELATIONS OF TUBIPORA,. 557

of the polyps, but at the same time many important histological
details must remain doubtful until some further material can
be procured in which the polyps have been killed in the fully
expanded condition.

Brief history of our knowledge of Tubipora.—The
first mention I can find of Tubipora is by Aldrovande, (1) who,
in 1648, describing it under the name “ Pseudo-corallium
rubrum calamites,” considered it could not be a true coral,
because, as he asserts, it does not adhere to rocks. ‘* Hujus-
modi coralium caret cortice in fundo Maris crescit non autem
saxis adheret more aliorum coraliorum.” In 1651 Bauhin (2)
described it under the name “Coralliis affine alcyonium
rubrum.” Subsequently, it was referred to by Imperato (9)
Rumphius (23) and Tournefort (25), the last named being the
first to give it the common name “ organ-pipe coral.” “ Tubu-
laria est plante genus, feré lapidem ex pluribus tubulis con-
stans, organi musici aemulum.” Pallas, (21) writing in 1776,
was the first to give it the name Tubipora. He seems to have
made a much more accurate examination of it, and to have
been thoroughly convinced of its animal nature. In _ his
definition, he refers to the tabule in these words: “ Tubuli
articulati, siphunculo continuo ad orificium stellato communi-
cantes.” The first good figure of the coral was published by
Ellis and Solander (6) in 1786, and this was improved upon
by Lamouroux (13) in 1821, both of whom figure and describe
the tabule. The first attempt to give a figure of the polype
was made by Quoy and Gaimard (22) in 1833.

Coming down to more recent times, our literature is still
very meagre. Percival Wright, (26) in 1869, first discovered and
described the formation of the skeleton of the tubes by the fusion
of spicules, and also described many details of the anatomy
of the soft parts which were previously unknown. In 1874
G. von Koch (10) published his dissertation on the anatomy
of the organ-pipe coral, but he seems to have been unacquainted
with the paper previously published by Wright on the same
coral. He gave figures for the first time of the mesenteries,
muscles, transverse sections through polyps in various regions,

558 SYDNEY J. HICKSON.

&c., but he overlooked entirely the tabule and the “‘ siphono-
glyphe.”

A. Anatomy of Tubipora.

Throughout this memoir I shall employ the following terms :
The encrusting lamina attached to stones, &c., from which the
young colony springs, I shall call the ‘‘stolon.” The individual
tubes I shall refer to as the “ corallites,” the laminz connecting
the tubes together I shall call the “ platforms” (Bricke of v.
Koch). The inner tubes, funnel-shaped tabule or flat tabule,
in whatsoever form they occur I shall call the ‘‘ tabule.”” The
points at which the platforms meet the corallites I shall call
the ‘‘nodes,” and for the ciliated groove on the ventral side
of the stomodeum I shall use the term I have elsewhere
proposed for it (8), namely, the “‘ siphonoglyphe.”

I. The Skeleton of Tubipora.—The hard parts of the
“organ-pipe coral”? have already been described by several
authors from the time of Pallas, but as many points still
remain obscure and others entirely undescribed, I propose to
give here a further account of them from the examination of
a large number of different specimens belonging to different
species.

In the Oxford Museum there is a specimen of a young
colony of Tubipora purpurea growing upon a piece of a
madreporarian coral. ‘The corallites are seen to spring from
a flat lamina, the stolon (fig. 1, s¢) which, creeping over the
surface of the support, gives origin as it goes to new corallites
(fig. 1,a@a). The presence of this stolon in the young colonies
of Tubipora seems to have been overlooked by previous
authors. This may be accounted for by the fact that as the
colony increases in size the stolon ceases to grow; the colony,
however, continuing to increase in size by the origin of new
corallites from the platforms (fig. 1, 65), soon completely hides
the stolon and the area to which it is attached. Moreover,
when the colony dies and is broken off the stolon remains
attached to the rock to which it was attached, so that none of
the large pieces in our museums (as far as I have been able

THE STRUCTURE AND RELATIONS OF TUBIPORA. 559

to observe), nor the pieces offered for sale by the dealers
possess the stolon at all. The stolon does not follow all the
fluctuations of its support, but in many places may be seen
to skip over large crevices. This point is, I think, of some
importance when we compare the stolon of Tubipora with
the creeping network of tubes from which the corallites of
Syringopora spring, the two being, I consider, homologous.
That the stolon should cease to grow at an early stage in the
growth of the colony is not to be wondered at, as its function
(namely, that of giving origin to new corallites) is completely
taken up by the platforms situated in the more peripheral
regions of the colony.

The individual corallites then, originating either from the
stolon or from a platform, pass up through a varying num-
ber of platforms towards the periphery. They are usually
straight, but occasionally I have observed them deviating
considerably in their course. Their power of growth is not
unlimited ; in one case I have traced a single corallite passing
through as many as seventeen platforms, but the average number
does not exceed twelve or thirteen in Tubipora musica,
Towards the termination of the corallites the walls become
thin, lose their deep red colour, and end in pale jagged edges.
In the course of the growth of the colony these free ends of the
dead corallites become covered over by neighbouring platforms,
and lost to view.

The platforms are formed as outgrowths from the lips of the
growing corallites, in a manner which will be more fully de-
scribed when I come to describe the soft parts of the animal. At
first they are exceedingly thin, and their skeleton composed only
of a few scattered spicules in the mesoderm. Consequently
in dried specimens the young platforms are entirely lost.

As the platform becomes older it increases in thickness, and
the spicules unite together to form a firm lamina. Each of
these older platforms may be seen in section to be really com-
posed of two delicate lamine, between which numerous canals
ramify in all directions (fig. 2).

Some of the most important and interesting structures in the

560 SYDNEY J. HIQKSON.

skeleton of Tubipora are the tabule. These may consist either
of simple flat partitions in the cavity of the corallite (fig. 6), or
they may be concave or convex, or cup-shaped (fig. 6) ; they
may be in the form of long drawn out funnels, or in the form
of axial tubules within the corallites (fig. 3), or assume much
more complicated shapes and forms. They were first referred
to by Pallas (21), and subsequently figured by Ellis (6) and
Lamouroux (13). These authors called them the “siphunculi,”
and seem to have considered that their normal if not their only
condition was that of hollow tubes open at both ends. Curiously
Professor Nicholson has only recently fallen into the same error,
for he says (20, p. 221): “‘ The axial tube itself, so far as I
have seen, is always open along its entire length. . . .”
As, however, every intermediate condition can be found
between axial tubes open at both ends and simple flat parti-
tions exactly similar to the tabule of the Favositidz, I shall
call them throughout the “ tabule.”’

The simple flat tabula is a condition which is very frequently
met with, but very often the tabula is not complete, but
stretches only part of the way across the cavity of the corallite.

Sometimes only a small strand could be found, reminding
one of the “ tongue-shaped ” tabulz of the Favositide (Nichol-
son, 19, p. 41), and in a few instances the tabulz were still
further reduced to mere spiniform projections of the walls of the
corallite, reminding one of the condition found in Pachypora.

Complete tabulz were also found slightly convex or concave,
as in Michelinia and other Favositidz ; others were cup-shaped
(fig. 6), and others were funnel-shaped with the narrow end
drawn out to a fine point.

A very common condition, however, is that in which the
tabula takes the form of an axial or inner tube, bulging at the
nodes and giving off a varying number of short tubules to the
platform (fig. 3) or, as Ellis (6) puts it, “ siphunculis continuis
geniculatis, ad genicula radiatis.”

Frequently, however, the condition is much more compli-
cated. One tabula, starting from a node, is drawn out into the
shape of a long funnel, and passing downwards is entirely

THE STRUCTURE AND RELATIONS OF TUBIPORA. 561

enclosed by a similar tabula proceeding from the next lowest
node, and the two pass some distance down the corallite as a
tube within a tube (fig. 4). This condition, which was first
discovered by Professor Moseley, (18) is of great importance in
the consideration of the relations between Tubipora and Syrin-
gopora. In one instance, I have observed the two tabula, after
passing some distance down the corallite apart from one another,
fuse together to form one tube, as they do in Syringolites.
Another example! I have met with is of some interest, from
its giving a superficial resemblance to certain Zoantharian
corals. A tabula, in the form of an axial tube, completely
closed above by a convex tabula, gives off eight delicate tubuli
to the platform at the node. When viewed from above, this
has exactly the appearance of a central solid columella con-
nected to the wall of the corallite by eight septa. (See wood-

Fic. 1.—Diagram of septiform tubule of Tubipora. (q@), axial tabula from
which eight processes radiate.

cut, fig. 1.) When we remember that in many instances the
presence of septa-like structures is the only reason for placing
certain fossil forms amongst the Zoantharia, examples such as
the one just mentioned showing structures which might have
been mistaken for septa and columella, had Tubipora been
known only in a fossil condition, become of importance.

It would be, however, an endless task were I to attempt to
describe the varieties of tabulee met with in Tubipora, and I
think it is only necessary for me further to mention that the
conditions met are often of the greatest possible complication.
In fig. 5 is represented a condition illustrating this statement ;

1 This example, and many others illustrating the varieties of the tabule of
Tubipora, may be seen in the Oxford University Museum.

562 SYDNEY J. HICKSON.

but many more could have been figured showing features quite
as complicated as this.

Before leaving the tabule, 1 may mention that at first I
had some difficulty in determining the exact shapes of the
tabula. They are frequently exceedingly delicate, crumbling
at the slightest touch or blown away by the slightest breath.
The only satisfactory way of exposing them is to carefully file
away the wall of the corallite until a small hole is formed, and
then, with a fine pair of forceps, break away the parts of the
wall which have been thinned by the filing process.

When the corallites or platforms of Tubipora are examined
with a hand lens, the coral is seen to be covered by numerous
round holes (fig. 2), and thin transverse sections reveal the fact
that these holes completely penetrate through the walls of the
corallites and the platforms. These perforations have already
been described and figured by Professor Nicholson (20, fig. 2),
and, as he points out, they are not always in the form of simple
tubuli, but are very often branched. The examination of thin
sections reveals the fact that the corallites are built up of a
number of spicules, which are so firmly bound together that it
is impossible to separate them without injury. Both Professor
Wright, who first discovered this, the true nature of theskeleton,
and Professor Nicholson, speak of the spicules as being “fused”
together. I think the employment of this word is likely to
lead to misunderstanding, The spicules are not really fused
together, but firmly bound together by means of minute serra-
tions fitting into minute serrations, just like the membrane
bones of the skull. The sutures between the spicules can
always be seen (fig. 9, Sp.), and I have no evidence at com-
mand to prove that they are ever obliterated. In the walls of
the corallites the sutures have a tendency to run across—that
is to say, in a direction parallel with radii drawn from the
centre of the corallite, and similarly, the longest axes of the
spicules are usually disposed in the same direction. ‘This fact
is of some importance, as occasionally individual spicules will
project out radially into the cavity of the corallite in a manner
exactly similar to the so-called “ septa” of Syringopora.

THE STRUCTURE AND RELATIONS OF TUBIPORA. 563

Towards the free end of the corallite the walls become much
thinner, and above this, in spirit specimens, perfectly free
spicules can be found scattered in the mesoderm. I have
nothing further to add to the excellent description of these free
spicules given by Professor Percival Wright (26).

The tabule consist of a simple network of spicules, the extent
to which the spicules are joined together being much less than
it is in the wall of the corallite (fig. 7).

II. Anatomy of the soft parts of Tubipora purpurea.
—When a transverse section is made through the wall of a
polype of 'Tubipora from which the skeleton has been removed
by decalcification, it is seen to be composed of three principal
layers (figs. 9, 10, and 12). Externally there is an ectoderm
(Ep,) composed of a single row of oval cells situated with their
longest diameter parallel with the layer of mesoderm upon
which they rest ; internally there is an endoderm (Ep, com-
posed of two or three rows of spherical cells, and between the
endoderm and ectoderm a mesoderm composed of a homo-
geneous jelly-like matrix containing a few scattered mesoderm
cells (c) and fibres (f); and in this mesoderm are seen a
number of large spaces (s s) occupied by the spicules before
decalcification. The ectoderm is composed all over the surface
of the polypes of a single row of oval cells in which, even after
prolonged staining in borax carmine, I have been unable to
discover any nucleus. In places, especially in the older parts
of the polyps, the ectoderm is entirely destroyed by parasites
of which endless different kinds both animal and vegetable
may be found (see figs. 9 and 10 d a.}).

Where the ectoderm is invaginated into the corallite (fig. 8.,)
its character changes. In the first place the cells are consider-
ably smaller in size. Whilst outside their longest and shortest
diameters are respectively ‘03 and ‘02 mm., in the invaginated
portion they are never larger than ‘01 by ‘003 mm. In the
second place, there are two or three rows of cells instead of only
one, and cells of the most superficial row by being elongated

1 For a description of the Foraminifera infesting Tubipora see Carter (3).

564 SYDNEY J. HICKSON.

vertically and closely approximated, give the appearance of a
columnar epithelium. It is impossible to say from the study
of non-living material only whether these cells are really
ciliated, but I am inclined to think that they are not, as in
some places I could distinguish a delicate membrane, like a
thin cuticle, covering the free edges of these cells. In the
above description of the histology of this portion of the ecto-
derm, I have described what I believe to be the true nature of
it after a careful examination of numerous sections; but as the
cells are here so very small, a renewed examination of speci-
mens specially preserved for histology is desirable.

The ectoderm covering the tentacles is composed of two or
three rows of cells the most superficial of which is distinctly
ciliated. These cilia are probably in the living animal long
and powerful and produce currents by their action which bring
food to the polype.

The mesoderm consists of a homogeneous matrix in which
may be found cells and fibres. The cells are usually pyra-
midal in shape but sometimes spherical or bipolar (fig. 12, ¢).
The angles of the cells are usually drawn out into long pro-
cesses lying in the matrix. Fibres are seen spreading through
the matrix in various directions just as described and figured
by Kélliker (11) in other Alcyonarians. In the actively
growing mesoderm, such as is found in the young platforms,
groups of small cells may be seen, which, budding off from the
ectoderm, sink into matrix of the mesoderm. These groups of
cells give rise to the spicules in the following manner. At
first a small calcareous particle is seen lying in the midst of
these cells, and as this increases in size the cells become more
and more flattened around it until only a delicate membrane
with two or three nuclei can be seen covering the spicule,
After a time even this membrane disappears and the spicule
lies freely in the matrix.

The endoderm consists of a layer of loose spherical cells
varying considerably in size and appearance (fig. 12, Jp).
The cells which lie next to the mesoderm are the smallest and
youngest, and they stain well in hematoxylin and borax

THE STRUOTURE AND RELATIONS OF TUBIPORA. 565

carmine ; the more peripheral cells are much larger, and after
even prolonged soaking in various staining fluids retain their
peculiar brown colour. I have noticed the same peculiarity in
the endoderm cells of the Gorgonide. These cells are more-
over filled with highly refracting bodies, which make it a
matter of great difficulty to determine whether they possess a
nucleus or not. In Alcyonium the endoderm cells are con-
stantly being shed, especially when the colony is in a sickly
condition, and then they exhibit a slow and irregular ameeboid
movement. From the similarity that exists between the endo-
derm cells of spirit specimens of Aleyonium and Tubipora I am
inclined to think that in the latter genus also they exhibit
amoeboid movement in the living condition.

The tentacles, eight in number, stand, in the retracted con-
dition of the animal, side by side in front of the stomodeeum.
They are not withdrawn into tentacular pouches at the side of
the stomodzum as they are in Paragorgia, Sarcophyton, and
certain other Alcyonarians, nor introverted as they are in
Corallium (Lacaze-Duthiers) and Heliopora (Moseley), but
they simply remain, as they are withdrawn, in front of the
stomodzeum and parallel with one another (fig. 8, 7’).

Each tentacle is provided with 14—16 pinnule on each side,
arranged ina single series, but both the number and arrange-
ment of these pinnule varies in the different species. Each
tentacle is covered, as previously described, by a ciliated ecto-
derm, and internally an irregular cavity communicating with
the general body-cavity is lined by endoderm (fig. 8). Be-
tween the ectoderm and endoderm there is a thick layer of
mesoderm which contains a number of scattered spicules as
first described by Prof. Wright (26).

The stomodeum is, in the retracted condition, thrown into a
number of folds, as it is in so many other Alcyonarians,
Heliopora (Moseley), Pennatula, &c. (Marshall), (fig. 8,
Stom.). Its epithelium is columnar and ciliated, the cilia over
the general surface of the stomodzum being very small and
difficult to see in spirit specimens. At first I had some diffi-
culty in finding any trace of the siphonoglyphe owing to the

VOL. XXIII.—NEW SER, EP

566 SYDNEY J. HICKSON,

numerous and close folds into which the stomodzeum is thrown,
but after a careful examination of a number of sections, the
characteristic long cilia and the thickened epithelium on the
ventral side of the stomodeum were found (figs. 8 and 9, Sz),
and I have no doubt that when the polype is fully expanded a
siphonoglyphe is present of the characteristic shape and
nature.

The stomodeum is held in position by eight mesenteries
which bear the powerful retractor muscles (figs. 8 and 9, 7.
m.). The muscular bundles exhibit the same arrangement
as in other Alcyonarians (von Koch 10, fig. 6), being
placed in all cases on the side of the mesentery which faces
the ventral side of the polype. The protractor muscles (fig.
8, pm) are exceedingly delicate structures, consisting of but
a few parallel fibres situated on the parts of the mesen-
teries in front of the stomodeum. The remarkable difference
in size between the retractor and protractor muscles may be
accounted for by the fact that strong muscles are required to
suddenly retract the polypes, when irritated, and to drive out
the water contained in the body-cavity at the same time,
whereas the expansion of the polypes is always in Alcyonarians
a very slow process and is probably aided, to a considerable
extent, by the ciliary action filling all the cavities of the polype
with water and thus helping to drive the polype out of the
tube.

The ova are attached to the sides of the dorsal and dorso-
lateral mesenteries immediately below the termination of the
stomodeeum (fig. 8, ov.). Hach ovum is enclosed in a capsule
and attached by a short stalk to the side of the mesentery. I
have never seen the stalk of an ovum attached to the mesen-
terial filament as von Koch (10) describes it to be, nor can I
find more than exceptionally that an ovum is attached either
to the ventral or ventro-lateral mesenteries (conf. von Koch 10,
fig) 7).

There are only two mesenterial filaments, the dorsal ones, as
in the siphonozooids of Pennatulidae (Kolliker) and Sarco-
phyton (Moseley), and these extend for a considerable distance

THE STRUCTURE AND RELATIONS OF TUBIPORA. 567

down the tube (fig. 10, M@. F.). Each mesenterial filament
consists (fig. 10, M@. F.) of an enormously thickened, columnar,
ciliated epithelium supported by a portion of the mesoderm of
the mesentery, but my specimens were not sufficiently well
preserved to allow me to enter into any further details of thei
histology.

Formation of the Platforms.—Professor Wright, in his
description of Tubipora, says: “ I think it is pretty evident that
the external tabule (i. e. platforms) are formed in the first
instance as flattened offshoots from the upper edges of the
tubes.” I am able entirely to confirm this view, as many of
the polypes in my spirit specimen exhibit the earliest indications
of a platform in the form of a thin rim spreading out from the
lip of a polype. This thin rim, as it increases in size, either
meets and fuses with other similar rims proceeding from neigh-
bouring polypes, or else it simply surrounds the lips of the
adjacent polypes and fuses with them (fig. 11). At first the
young platforms are quite pale, but soon delicate pink spots
may be seen scattered over their surface, and as the platform
increases in size and thickness these spots unite together into
a delicate network, and eventually the whole surface assumes,
to the naked eye, a deep red homogeneous colour. An exami-
nation of sections through these young platforms shows that at
first the rim consists of a fold of ectoderm containing a thin
lamina of mesoderm ; subsequently, however, as the lamina of
mesoderm becomes thicker, canals lined by endoderm are pushed
into it, and soon ramify in its substance, forming the canal
systems of the platforms (vide von Koch 10, fig. 10).

The ectoderm of the young platform is of the same nature as
the ectoderm of the invaginated portion of the retracted polype
(vide supra, p. 563, fig. 8.,.), consisting of a number of small
cells arranged in more than one row, and giving the appearance
of being in a condition of rapid multiplication and growth ; the
mesoderm, too, does not contain the characteristic pyramidal
cells and fibres of the other parts of the polypes, but contains
numerous groups of small round cells, which have sunk down
into the matrix from the ectoderm.

568 SYDNEY J. HICKSON.

Even in exceedingly young platforms small white thickened
spots may be seen on the upper surface, and these are young
buds. The first sign of a young bud is a proliferation of endo-
dermal cells on the upper side of the cavity of one of the canals
of the platform ; this is followed by an invagination of the ecto-
derm above it, which soon takes the form of a wide bag with a
narrow mouth. Around this bag eight lobate folds of the canal
with its thickened endoderm grow up, the thin laminz of me-
soderm remaining as the eight mesenteries. Subsequently a
communication is established between the ectodermic invagi-
nation and the canal, but I have been unable to trace the
growth of the bud further.

Formation of the Tabule.—As Stewart (24) suggested
some years ago, the tabule are formed by a shrinking of the
endoderm and its accompanying lamina of mesoderm away from
the calcareous wall of the corallite, and the reformation of
spicules upon it (fig. 10). As the corallite increases in length
and the polype recedes farther and farther away from the lower
parts of the tube the calcareous wall becomes thicker and
thicker. This increase in thickness of the wall of the corallite
is accompanied probably by a certain loss of vitality of the
mesoderm, and this causes the thin strands (fig. 12) connecting
the ectodermic and endodermic mesodermic lamine to break,
and consequently the endoderm and endodermic lamina of meso-
derm shrink towards the axis of the tube. Having shrunk,
the mesoderm forms a fresh layer of spicules, which, uniting
together, form the tabulz of the dried coral. Having under-
gone one process of shrinking, it is quite possible for it to un-
dergo asecond and to form a second deposit of spicules ; in this
manner the condition in which two axial tubes are found may
be accounted for. At the nodes the endoderm runs out in the
form of canals into the platforms, and these canals, when the
shrinking occurs, would have a certain restraining action, and
hence the bulging of the axial tubules at the nodes as described
above (fig 3), and the formation of the delicate radial tubules
figured in woodcut, fig. 1, and in figs. 3, 4, and 5.

THE STRUCTURE AND RELATIONS OF TUBIPORA. 569

B. The Relation between Tubipora and Fossil
forms.

If it be borne in mind that the only known living forms at
all allied to Tubipora (De Blainville (5), von Koch (10),
Hickson (8) ) are Cornularia and Clavularia, the former possess-
ing no skeleton at all, and the latter but a few scattered spicules,
it is evident that a long series of intermediate forms, some of
which must have possessed skeletons suitable in every way for
geological preservation, must have become extinct. Formerly
it was considered that the extinct Syringopora was a near ally
of Tubipora, and the older naturalists, such as Ellis (6) and
Cuvier (‘Regne Animal’), placed them in the same family ;
and this view is held now by such authorities as Zittel (‘ Hand-
buch der Palaeontologie’), G. von Koch(10), Moseley (17), and
others. Certain eminent paleontologists, however, have re-
cently maintained, on grounds which I can hardly consider to
be entirely satisfactory, that Syringopora is not really allied to
Tubipora. Dr. Lindstrom (14) places Syringopora amongst the
Rugosa and Verrill, Nicholson and others place it amongst the
Zoantharia perforata. The renewed examination I have
made of the skeleton of Tubipora, carried on side by side with
the examination of the soft parts, leads me to believe that the
view of the older naturalists is the correct one, and that
Syringopora is really an Alcyonarian closely allied to 'Tubipora.

When Professor Nicholson published his book on ‘'Tabulate
Corals’ (19) he seems to have considered that the position of
Syringopora amongst the Zoantharia perforata was defi-
nitely settled, for he says (p. 213): “‘ As to the recent genus
Tubipora it seems unnecessary to enter into any detailed dis-
cussion, as the known facts as to the internal structure of
Syringopora render any direct affinity between the two out of
the question.” More recently, however, he has published a
paper which specially discusses the relationship between these
genera (20), and he urges the following three differences
between them as being of special importance :

570 SYDNEY J. HICKSON.

‘“‘(a) In the first place there is the very important and remark-
able difference in the minute structure of the calcareous
skeleton in the two types in question. In Tubipora the
corallum is made up of fused calcareous spicules, which are
so disposed as to give rise to a universally distributed system of
minute canaliculi or tubuli, which open on both the outer and
inner surfaces of the skeleton by well-marked apertures. The
size of these tubuli is comparatively so great that it is quite
impossible that their presence could be overlooked in thin sec-
tions of Syringopora, if they really existed in this genus. On
the other hand, the skeleton of Syringopora, as regards its
minute structure, is quite compact, and shows no signs what-
ever, either of being penetrated by a system of tubuli, or of
being formed by the fusion of ectodermal spicules.” It is
difficult to see why this difference should be considered of any
great morphological importance. The size of the pores or
* tubuli,” as Professor Nicholson calls them, varies consider-
ably in the different regions of the corallite, being at the younger
ends much larger than they are at the older ends, so that it is
evident that as the corallite grows older the tubuli have a ten-
dency to be filled up, and a still further continuation of this
process would make the wall of the corallite quite aporous. I
have no evidence to prove that the complete filling up of these
perforations in the walls ever does occur in Tubipora, but
should an example be found in which this has occurred I should
certainly not consider it sufficient reason for the formation of a
new genus or even a new species. That the skeleton of Syrin-
gopora “ shows no signs of being formed by the fusion of ecto-
dermal spicules” is not to be wondered at, as we possess no
means of studying either the development or the growth of the
skeleton of this form, since the delicate growing ends would be
broken down and destroyed; and even in recent genera (such
as Corallium, Lacaze-Duthiers), in which the skeleton is known
by an examination of its growth to be composed of fused
spicules, no evidence of them can be seen in thin transverse
section through the hard parts.

The second difference urged by Professor Nicholson as being

THE STRUCTURE AND RELATIONS OF TUBIPORA.,. 571

of importance is that ‘‘(d) True tabule are always present in
Syringopora, and in all the typical forms of the genus have
the character of a series of invaginated cones, which gives rise
centrally to an axial tube. In no specimens that I have ever
seen can there be recognised any similar series of funnel-shaped
tabule in Tubipora. I cannot, in fact, recognise that any
true tabule are present in Tubipora, so far as my own ob-
servations enable me to come to a conclusion on this point, and,
as already stated, I do not regard the axial tube of Syringo-
pora as being formed in the same way as the somewhat similar
looking structure in Tubipora, or as being really homologous
with it.” The description I have given in this paper of the
tabulz of Tubipora proves, | think, that the difference between
Syringopora and Tubipora is only one of degree and not one of
kind, and I cannot see that there is any evidence to prove that
the tabule in the twoare not homologous. There are perfectly
flat tabule in Tubipora, as in some examples of Syringopora,
and there are cone-shaped tabule fitting one into another as
in Syringopora (fig. 13, ¢). In fact, the only striking dif-
ference between the two in respect to the tabule is that in Tubi-
pora they are more sparsely scattered in the corallites, and are
more frequently of the form of axial tubes, open at both ends.
The third point of difference urged by Professor Nicholson is
that (c) “ the corallites of Syringopora are provided with a
well-developed septal system, of which absolutely no traces can
be recognised in 'Tubipora. Moreover, the septa of Syringo-
pora are not mere marginal plice, such as form the ‘ pseudo-
septa’ of Heliopora, but they are in the shape of vertically-
arranged rows of spines, which may be well compared with the
septal spines of such undoubted Zoantharians as Porites.” In
answer to this the third and last objection of Professor
Nicholson | must point out that in some cases, as I have men-
tioned above, spicules do project into the corallites of Tubipora,
giving an appearance in transverse section exactly similar to the
individual septal spines of Syringopora, and that in many cases
the septal spines of Syringopora are exceedingly sparse and
reduced in size to a minimum, so that when a specimen of

572 SYDNEY J. HICKSON.

Syringopora is examined with spines in this rudimentary con-
dition and compared with a specimen of Tubipora, no difference
of importance can be distinguished between the two genera in
this respect.

These, then, are the three principal objections to the rela-
tionship between Tubipora and Syringopora, and, as I have
endeavoured to show, none of them is by any means insuperable.
When this is borne in mind, and the numerous points con-
sidered in which the two genera resemble one another, I think
that the Zoantharian affinities of Syringopora must at least be
considered very doubtful.

The corallum in both Syringopora and Tubipora consists of
a number of tubular parallel corallites separated from one
another by spaces, which are bridged over by hollow tubular
processes in Syringopora or platforms containing a network of
canals in Tubipora. In both genera new buds are formed on
these connections between the corallites, a striking similarity
which has been quite recently dwelt upon by G. von Koch
(10 a), in a paper which came into my hands since my plates
were sent to the lithographer.

I have drawn in fig. 13a a corallite springing from one of
the tubular connections between the corallites in Syringopora,
and if this be compared with the corallites springing from the
platforms in fig. 1a, or in von Koch’s fig. 20, the striking
similarity between the two genera in this respect will be
seen.

When the surface of a corallite of Syringopora is examined
with a lens it is seen to be covered with a number of small pits
which bear a striking resemblance to the mouths of the per-
forations of Tubipora (fig. 2, 2), and this pitting can be seen
quite as plainly and distinctly on the inner side as on the
outer side of the corallites. Although these pits do not pene-
trate the walls in Syringopora as this coral is presented to us
after centuries of fossilisation, yet I think they afford some
confirmation of the opinion that its corallites are really of the
same nature, i.e. spicular, as in Tubipora. ‘This opinion is,
moreover, still further confirmed by an examination of the

THE STRUCTURE AND RELATIONS OF TUBIPORA. 573

tabule of Syringopora, for they are found to be exceedingly
friable and perforated by numerous small holes, just as they
are in Tubipora.

The corallum of Syringopora springs from a prostrate net-
work of tubes ; in Tubipora it springs from a prostrate stolon,
which, as I have shown, contains a network of canals. The
prostrate network of tubes in Syringopora was in contact with
rocks and stones at certain places ; the stolon of Tubipora is in
contact with its support only in certain places, and at others
rests upon no support, so that I think we have sufficient
evidence to assume that these two structures are really
homologous.

The presence of tabule in Syringopora, which was formerly
considered to be a strong point in favour of the Zoantharian
affinities of the genus, has become, since our knowledge of the
structure of Tubipora has increased, an argument in favour of
its Aleyonarian affinities. In addition to the striking similarity
in shape between the tabule of Syringopora and Tubipora,
the general shrunken appearance they have in the former genus,
as shown in the numerous figures given by Professor Nicholson
(19, 20), goes to prove that they are due in this genus too toa
shrinking of the mesoderm, and that consequently they are in
every respect homologous.

A very great difficulty, however, that is found by Professor
Nicholson and others in regard to the Alcyonarian affinities of
Syringopora, is the undoubted relationship which exists between
this genus and the family Favositide, which family is considered
by them to be undoubtedly Zoantharian, and closely allied to
Porites. But the evidence in favour of the Favositide being
really Zoantharians is by no means conclusive, for it seems to rest
entirely upon the presence, in some cases only, of spiniform
septa iy the corallites.' The presence of tabule is no longer
evidence against their being Alcyonaria, as Professor Moseley
has shown that Heliopora, which possesses tabule, is un-
doubtedly an Alcyonarian and Tubipora, also is now known
to possess true tabule. In fact, when we remember that tabule

1 In Stenopora septal spines are absent.

574 SYDNEY J. HICKSON.

are occasionally found in Tubipora, exactly similar to the flat
tabule of Favosites, the convex tabule of Michelinia, the
lappet- or tongue-shaped incomplete tabule of Pachypora,
aud the funnel-shaped tabulz of Syringolites, and that tabule
are quite unknown amongst the Poritide—the family which is
supposed to be most nearly related to the Favositida—we must
look upon these structures as evidence, if any, in favour of
the Favositidz being Alcyonarians. But the evidence in favour
of the Alcyonarian affinities of the Favositide does not come
entirely from the side of Tubipora, for, as Professor Moseley
(17) has pointed out, there is strong evidence of close relation-
ship between Heliopora and Favosites.

Professor Moseley also points out that the signs of Favosites
forbesi being dimorphic is also in favour of this affinity; and
this point becomes of greater importance as our knowledge of
the Alcyonarians increases. Kolliker has pointed out that in
such closely allied genera as Heteroxenia and Xenia, the former
is dimorphic, the latter is not. Again, in such closely allied
genera as Paragorgia and Briareus the former is dimorphic
and the latter is not; and I have just been informed by Mr. 8.
Ridley, who has been investigating the Alcyonarians brought
back by H.M.S. “Alert,” that the genus Melitodes is also
dimorphic, whilst, as far as I am aware, the genera most
nearly allied to it are not dimorphic. ‘Thus, a tendency for
various genera to become dimorphic seems to be a characteristic
feature of the Alcyonaria, a feature, moreover, which, as far as
I am aware, is entirely unknown amongst the Zoantharia, and
consequently the evidence bearing upon the question afforded
by Favosites forbesi, which was apparently dimorphic, is not
without considerable weight.

Thus, I think that taking all things into consideration, the
evidence at our command tends to prove that the Favositide
are really Alcyonariaus, and that Syringopora is also an
Alcyonarian allied to Tubipora.

THE STRUCTURE AND RELATIONS OF TUBIPORA. 9575

c. Remarks on the Zoological position of
Tubipora.

In De Blainville’s ‘Manuel d’Actinologie,’ 1834, Tubipora
is placed in the same family, namely, “les Tubipores,” witn
Cornularia and Clavularia. In modern text-books of zoology
there is a tendency to place the genus Tubipora in a separate
family, and to classify Cornularia and Clavularia with the
Alcyonide. Von Koch (10 and 10 a), however, agrees with
the older view, and he says (10 p. 6): “ Von den lebenden
Formen stehen ihr wohl die Cornulariden am nachsten, und
scheint es, dass diese Familie einen sehr urspriinglichen
Zustand der Octokorallen reprasentirt.” I am inclined to
agree with von Koch and the older naturalists, and have else-
where (8) proposed that Tubipora should be included with
Cornularia, Clavularia, Sarcodictyon, and allied forms, in one
family, which may be called the Stolonifera.

In order to arrive at any conclusion as to the grouping of a
family of animals it is necessary to take into consideration the
lines upon which the phylogeny of that group probably pro-
ceeded, and consequently, I propose to give certain speculations
concerning the phylogeny of Tubipora, to which I have been
led in the course of my investigation.

As I pointed out above, there is every reason to suppose that
a long series of intermediate forms between the recent 'Tubipora
and a Cornularia-like ancestor must have become extinct. As
long as the walls remained unsupported by skeletal structures
as they are in Cornularia, it was a matter of impossibility
for the corallites to attain any great length. When, however,
owing to the formation of a skeleton the corallites increased
in length, communications between the individual corallites
would be of immense value to the colony for keeping up a
continuous and sufficient circulation. If we suppose that the
corallites stood as near to one another on the stolon as the
polyps do in the recent Cornularia, fusion of the walls of adja-
cent corallites and the formation of pores after the manner of
the mural pores of Favosites would be not impossible but even

576 SYDNEY J. HICKSON.

probable. Having formed communications of this nature
between the individual corallites, every intermediate condition
between this and the condition in which the pores are drawn
out into the form of hollow tubular communications or perfo-
rated platforms would be of advantage to the colony as afford-
ing more room for the development and growth of buds; and,
indeed, as Professor Nicholson (19) has shown, every inter-
mediate condition can be found between the mural pores of
Favosites and the tubular communications of Syringopora or
the platforms of Chonostegites.

Previously, however, to the Cornularia-like ancestor there
must have been a solitary ancestor similar to Haimea or Hartea.
The genus Sarcodictyon (Gosse, 7) consists of simple polyps
united together only by very delicate tubular threads. Froma
condition of the stolon so simple as this there may have been
in the course of evolution two principal variations. Hither the
thread-like communications increased in number and size and
underwent various anastomoses, forming a retiform stolon like
Syringopora, or else the threads may have become broader by
the growth of coenenchym, and eventually formed a lamellar
stolon containing a network of canals, as in Cornularia and
Tubipora. Thus arranged in a tabular form the phylogeny
of the Stolonifera might be represented as follows :

Tubipora

ae

Syringopora Favosites-like ancestor

Cornularia, Clavularia
— Sarcodictyon

Solitary ancestor like Haimea.

Before concluding this paper, I must acknowledge my ia-
debtedness to Professor Moseley for much valuable aid and
advice, to my sister, Miss A. W. Hickson, for the most excellent
and valuable drawings represented by figs. 1 and 2, and to

THE STRUCTURE AND RELATIONS OF TUBIPORA. 577

my sister, Miss C. M. Hickson, for the careful drawing of the
spicules of the tabule of Tubipora, represented by Fig. 7.

BIBLIOGRAPHY.

(1) U. Atprovanpr.—‘ Museum metallicum,’ 1648, p. 290.

(2) J. Baunin.—‘ Historia plantarum,’ 1651, t. iii, p. 808.

(3) H. J. Carrer.—“On Foraminifera found in and about Tubipora
musica,” ‘Annals and Mag. of Nat. History,’ 4th series, vol. xix,
p- 209.

(4) J. D. Dana.—‘ Exploring expedition of the United States Zoophytes,’
1849, pl. xix.

(5) H. M. De Buainvittz.—‘ Manuel D’Actinologie,’ Paris, 1834, p. 500.

(6) J. Extis anp D. Soranper,—‘ Natural History of the Zoophytes,’ 1786,
p. 143.

(7) P. H. Gossz.—‘‘ On Sarcodictyon catenata,” ‘Annals and Mag. of
Nat. History,’ 3rd ser., vol. ii, p. 276.

(8) 8. J. Hicxson.—“ On the Ciliated Groove (Siphonoglyphe) in the Stomo-
deeum of Alcyonarians,” Abstract in ‘Proceedings Royal Soc.,’ No.
226, 1883.

(9) F. Imperato.— Historia naturale,’ 1672, p. 631.

(10) G. von Kocu.—‘ Anatomie der Orgelkoralle (Tubipora hemprichii),’
Jena, 1874.

(10a) G. von Kocu.—‘ Die ungeschlechtliche Vermehrung einiger Palaeozoi-
schen Korallen,’ Kassel, 1883.

(11) A. Kontrker.—‘ Icones histiologicee oder Atlas der vergleichenden
Gewebelehre,’ 2 Abth., 1 Heft, Leipsig, 1865, p. 167.

(12) J. B. Lamarcx.—‘ Histoire des animaux sans vertébres,’ 1836, t. ii.

(13) Lamovroux.—‘ Exposition methodique des Polypiers,’ 1821, p. 66.

(14) G. Linpstrom.—‘‘On the Affinities of the Anthozoa tabulata,”
‘Annals and Mag. of Nat. History,’ 4th series, vol. xviii, pp. 1-17.

(15) C. Linyzus.— Systema nature,’ 10th ed., t. i, p. 789.

(16) H. Mitne-Epwarps.— Histoire naturelles des Corallaires,’ Paris, p.
130, pl: B, 1857.

(17) H. N. Moserey.—‘‘ Report on the Scientific Results of the Voyage of
H.M.S. ‘Challenger’ during the years 1873-76,” vol. ii, part. iii;
‘Report on certain Hydroid, Aleyonarian, and Madreporarian Corals,’

. 125.

(18) IN. MosrLtey.—* Notes on the Structure of Seriatopora, Pocillopora,
Corallium, and Tubipora,’ ‘Quarterly Journal Microscop. Sci.,’
October, 1882, p. 398.

(19) H. A. Nicnotson.— On the Structure and Affinities of the Tabulate
Corals of the Paleozoic Period,’ Edinburgh, 1876.

578 SYDNEY J. HICKSON.

(20) H. A. NicHotson.—“ On the Structure of the Skeleton of Tubipora
musica, and on the Relation of the Genus Tubipora to Syringopora,”
‘ Proceedings of the Roy. Soc., Edinburgh, 1880-81, p. 219.

(21) P. S. Patras.—‘ Elenchus Zoophytorum,’ 1776, p. 337.

(22) Quoy et Gaimary.—‘ Voyage de L’Astrolabe,’ Zool., 1833, pl. 21, t. iv,

p. 257.

(23) G. E. Rumpnrus.—‘ D’Amboinsche Rariteitkamer,’ 1705.

(24) C. Srewart.— On a New Species of Stylaster, with a Note on Tubi-
pora,” ‘Journal of the Micro. Soc.,’ 1878.

(25) J. P. TourneFrort.—‘ Institutiones rei herbariz,’ t. i, 1719.

(26) P. Wrieut.—* Notes on the Animal of the Organ-pipe Coral,” ‘ Annals
and Mag. of Nat. History,’ 4th ser., t. ii, 1869, p. 377.

ON THE MALLEUS OF THE LACERTILIA, ETC. 579

On the Malleus of the Lacertilia, and the Malar
and Quadrate Bones of Mammalia.

By

M. L. Dollo,

Assistant-Naturalist in the Royal Museum of Natural History, Brussels.

With Plate XLI.

THE apparent absence of a quadrate bone in Mammals has
given rise to a large number of theories, differing in details, but
all, or nearly all, tending to discover this bone in the chain of
ossicula auditus of these animals. A recent work, published
by my friend Professor P. Albrecht,! departs, however, from
the view ordinarily taken, and, as I believe, advances our
knowledge considerably on the subject. I have been led by
the study of this memoir to a discovery which I believe to be
of considerable interest, of which it is the object of the present
communication te give an account.

Before passing to the consideration of my own observations
I will briefly review the state of the question. This cannot be
better accomplished than by reproducing the excellent recapi-
tulative tables given by M. Albrecht.

1 Pp. Albrecht, ‘“ Sur la valeur morphologique de l’articulation mandibulaire,
du cartilage de Meckel et des Osselets de louie, avec essai de prouver que
Vecaille du temporal des Mammifers est composée primitivement d’un squa-
mosal et d’un quadratum *’ (Mayolez, Bruxelles, 1883).

580 M. L. DOLLO.

TABLE I.)

Theories published up to the present time on the Morphological
Differences between the Mandibular Articulation of the

Lower Gnathostomous” Vertebrata and the Mammalia.

| Mandibular Articulation of the Lower |

Gnathostomous Vertebrata | Mandibular Articulation of Mammalia.

Quadrato-articular articula- culation.
tion. . . . . . .)|Gegenbaurt . Dentary articulation.

| ; |
Huxley? . . Se ae me
| _v. K6lliker® . Ditto, ditto. |

1 P. Albrecht, op. cit., p. 249.

* HK. Haekel, ‘Anthropogenie,’ Leipzig, 1874, p. 425.

3 Huxley, ‘A Manual of the Anatomy of Vertebrate Animals,’ London,
1871, p. 84. Wiedersheim, ‘ Lehrbuch der vergleichenden Anatomie der
Wirbelthiere,’ Jena, 1882, T. i, p. 189 and 155.

* Gegenbaur, ‘ Grundziige der vergleichenden Anatomic,’ 2* Auflage, Leip-
zig, 1870, p. 662.

6 V. Kolliker, ‘Entwicklungsgeschichte des Menschen und der héheren
Thiere,’ 2° Auflage, 1879, p. 486.

THE MALLEUS OF THE LACERTILIA, ETC.

TABLE

| Fe

58]

Theories on the Development and the Morphology of the

Reichert?
Ginther®.

Gegenbaur? .

.| Malleus, Incus . .
.| Malleus, Incus, Stapes

Ossicula Auditus

Mandibular Arch.

Malleus (Articular).
Incus (Quadrate) .

Huxley® ., Malleus (Quadrate) {
W. K. Parker® .| Ditto

W. K. Parker Ditt

and Bettany? re ;
Salensky® . Malleus, Incus, Stapes

V. Kolliker’.

Wiedersheim”® .

Malleus (Articular) .

Processus gracilis (An-
gular). .

Incus (Quadrate)

Ditto sues:

1 P, Albrecht, loc. cit., p. 250.

2 Reichert,
Archiv,’ 1837.

“Ueber die Visceralbégen der

.|Stapes.

of Mammalia.

|
Hyoidean Arch.

Os lenticulare (Sym-
lectic). |
‘Stapes (Hyomandibular)
Incus (Hyomandibular) |

Os lenticulare, Stapes.

. Ditto.
. Incus (Hyomandibular) Stapes.

Wirbelthiere,”

Auditory
Capsule.

‘Stapes.

|Ditto.

‘ Miller’s

3 Giinther, ‘ Beobachtungen iiber die Entwicklung des Gehororgans,’ Leip-

zig, 1842.

4 Gegenbaur, loc. cit., pp. 662 and 663.

5 Huxley, loc. cit., and ‘ Proc. Zool. Soc. London,’ 1869.
6 P. Albrecht, loc. cit., p. 250.
7 Parker and Bettany, ‘The Morphology of the Skull,’ London, 1877,

8 Salensky, ‘‘ Zur Entwicklungsgeschichte,”

Jahrgang il, p. 250.
9 V. Kolliker, loc. cit., pp. 471—473, 475—478, 480—487,

10 Wiedersheim, loc. cit.
vi, p. 415; and Fraser,

VOL. XXIII_—NEW SER,

‘Zoologischer Anzeiger, Leipzig,

See also Salensky, ‘ Morpholog. Jahrbuch,’ vol,
‘ Philos. Trausact.,’ 1883, p. 901.

582 M. L. DOLLO.

TaB.e III.}

Theories ou the Morphological value of the Incudo-mallear and
the Mandibular Articulations of Gnathostomous Vertebrates.

Articulations. Gegenbaur, vy. Kolliker. | Huxley, Parker and Bettany. |
Incudo-mallear of Mam- Quadrato-articular. | Hyomandibular-
malia. : | quadratic.
Mandibular of Lower Quadrato-articular | Quadrato-articular
Gnathostomous Verte-| (viz. Incudo-mallear). | (viz. Malleo and
brata. Articular).
Mandibular of Mammalia. Squamoso-dentary. _ Squamoso-articular.

From the contents of these tables it follows :

lst. That according to the anatomists who have preceded
M. Albrecht, the lower gnathostomous Vertebrata have their
jaws hinged by means of a quadrato-articular articulation,
whilst Mammalia are provided with an entirely different kind
of articulation, concerning the exact nature of which these
anatomists are at variance.

2nd. That according to the same authors the Promammalia?
must have possessed originally a quadrato-articular articulation
of the lower jaw, but that they have lost it (Huxley, Parker,
and Bettany), or at least have given up its use in the act
of mastication (Gegenbaur and von Kolliker), and that at the
same time their quadrate (Huxley, Parker, and Bettany), and
perhaps also the articular (Gegenbaur and von Kolliker) and
angular elements (von Kolliker) of their mandible have become
included in the chain of the ossicula auditus.

After this statement M. Albrecht continues :

Ist. That he is persuaded that all of the ossicula auditus
of Mammalia are represented by homologues in the Amphibia,
and by all the ossicula auditus of the Sauropsida, and that
they correspond to the suspensorium of fish. The facts may
be stated as follows :

1 P, Albrecht, loc. cit., p. 12.

> KE, Haeckel, ‘ Natiirliche Schopfungsgeschichte,’ Berlin, 1872, p. 538.
3 P. Albrecht, loc. cit., p. 251.

THE MALLEUS OF THE LACERTILIA, ETC. 583

]
| Fenestra oe Interfenestral Chain, Albrecht, extend- |

: : s 3 é - Fenestra
panica. ing, without interruption, from one Feuestra :
| Albrecht. to another : DS
{_——_—_____ LE a ee Sa oe a Bie ee ee
| Mammalia, Membrana | Malleus + Incus + Os lenticulare -Membrana
| tympanica.| + Stapes. | ovalis.
2 . : | Albrecht.
Sauropsida. | Ditto. Columella auris. Ditto.
= |
iS Urodela. | Ditto. Ditto. Ditto,
|=: ( Anoura. | Ditto. Ist ossicle + 2nd ossicle + 3rd ossicle | Ditto.
|S + 4th ossicle.
Further :

Mammalia. | wandible. | Extramandibular portion of Mec- |Otic Region
Sauropsida, | Ditto. | kel’s Cartilage and Interfenestral | of Cranium,
Amphibia. Ditto. | Chain. Ditto,
Pisces | Ditto. | Suspensorium. |Ditto,

Thus, since the quadrate does not form any part at all of the
interfenestral chain of Sauropsida, it cannot enter into the
composition of that of Mammalia.

2nd. That it is not possible to understand (especially on the
theory of Gegenbaur, v. Kolliker, and others) how the Mam-
malia could have acquired an articulation of the lower jaw
different from that of other gnathostomous Vertebrata, and that,
therefore, he is of opinion that the glenoid cavity of Mammalia
ought still to be found in the quadrate bone.

M. Albrecht has in his hands at the present time the skull
of a new-born child, in which the squamous portion of the
temporal bone is divided into two parts, viz. :

a. The squamous portion of the temporal bone, properly so
called, the homologue, in his opinion, of the squamosal of
Sauropsida.

B. The zygomatic portion bearing the glenoid cavity, and
thus homologous, with the quadratum of the Sauropsida.

Other cases of the complete separation of the squamosal and
the quadrate are mentioned in literature, and are cited by M.
Albrecht. In addition this distinguished anatomist states that
he has observed traces of a squamoso-quadratic suture in several

584. M. L. DOLLO.

skulls of apes in the Royal Belgian Museum of Natural History,
and that he will shortly publish an account of them.

To sum up, M. Albrecht’s arguments lead, on the whole, to
the following two conclusions :

ist. The quadrate cannot form part of the interfenestral
chain of bones of Mammalia.

2nd. One of the two bones (the zygomatic portion) formed
by the division of the so-called squamosal is doubtless the
homologue of the quadrate of Sauropsida.

I shall examine these two statements successively, in order
to show whether the results of my own observations tend to
confirm or invalidate M. Albrecht’s conclusions.

I.

Can the quadrate form a part of the interfenestral chain ?

It is evident that if there were found simultaneously existing
in the same animal a mandible composed of six normal elements,
a true quadrate, and a malleus, it would immediately follow
that it was impossible that that quadrate could form any part
of the chain of ossicula auditus, for—

Ist. It could not be confounded with the malleus, because
there would already be one there.

2nd. It would be still more impossible to identify it with
the remainder of the interfenestral chain, because it would be
situated outside the malleus, and would not touch any of the
remaining ossicula.

Everything depends, therefore, on the discovery of a malleus
in the condition described above. Now, I have found in several
Lacertilia (Leiolepis guttatus, Ctenosaura pectinata,
Uromastix spinipes, Lophyrus dilophus, Basiliscus
vittatus) a small bone which appears to answer the question.
I shall endeavour to prove that it has really the morphological
value of a malleus.

a. Firstly, it has the form of a malleus; it being possible to
distinguish in it :

THE MALLEUS OF THE LACERTILIA, ETC. 585

a, A eapitulum.

b. A cervix.

ce, A manubrium.

d. A processus gracilis.

(3. It has the same connections; that is to say:

a. It is applied along the tympanic membrane in such
a manner that the manubrium is parallel to the mem-
brane.

6. It is united besides to the remainder of the inter-
fenestral chain by means of a cartilage attached to it in
the region of the cervix.

c. It lies in contact with the quadrate in exactly the
same manner as that in which the malleus of Mammalia
is in contact with the quadrate of M. Albrecht.

y. It is connected with the articular element of the man-
dible by means of a malleo-articular ligament, which M. Albrecht
identifies with the extra-mandibular portion of Meckel’s carti-
lage.

6. There is little doubt, there appears to me, that this mal-
leus is identical with that described by Peters as existing in
the crocodile! and with the “ suprastapedial extrastapedial
(manubrium) ” of Mr. W. K. Parker.?

This being admitted, the first of the naturalists cited has
demonstrated the continuity of the malleus with Meckel’s
cartilage, as in Mammalia,’ an observation confirmed by the
English naturalist. This result, therefore, is in support of
the opinion expressed in 6. .

e. The malleus of Mammals serves for the insertion of a
small muscle (tensor tympani), the origin of which is in the otic
region of the cranium. The same occurs in the case of the
malleus of Lacertilians.?

1 W. Peters, ‘ Monatsberichte d. K. p. Akademie d. Wissenschaften zu
Berlin,’ 1868, p. 592.

2 W. K. Parker, ‘Phil, Trans. Roy. Soe.,’ London, part. ii, 1879, pl. 438,
fig. iti and vi.

3 W. K. Parker, ‘ Nature,’ July 13th, 1881, p. 253.

4 W. K. Parker, ‘Phil. Trans.’ (v. supra), fig. vi, st. m. (the so-called
“ stapedius ” of the author).

586 M. L. DOLLO.

To sum up, then, I believe that I have discovered in Lacer-
tilia a real malleus, the homologue of that of Mammalia, and
with that fact as a starting-point it appears that the conclu-
sion may be formed that the quadrate of Mammalia is not to
to be sought for amongst their ossicula auditus. In this
respect, therefore, M. Albrecht’s theory receives confirmation,
although it will be necessary that it should be slightly modi-
fied. Instead of the statement with regard to reptiles :!

Columella = malleus + incus + os lenticulare + stapes, the
matter must stand thus:

Columella = incus + os lenticulare + stapes.

Il.

Is the quadrate of M. Albrecht really the homologue of the
quadrate of Sauropsida ?

Such is the second question to be considered. Before going
further it will be well to consider what are reasons given by
M. Albrecht in favour of his interpretation. They are three
in number.

Ist. The quadrate of Mammalia, since it could not find
place in their interfenestral chain, must be to be found else-
where.

2nd. The glenoid cavity ought still be found in the quadrate
bone.

3rd. There are a certain number of instances known in
which the squamous portion of the temporal bone of man, or
of the primates, is divided into two parts.

a. The squamous bone properly so called.
3. The zygomatic portion.

It appears to me that two further methods of inquiry should
be employed to make certain whether the quadrate of M.
Albrecht is a real quadrate or not. These are:

lst. The examination of its connections.

2nd. The study of its development.

' P. Albrecht, loe. cit., p. 15.

THE MALLEUS OF THE LACERTILIA, ETC. 587

Leaving aside the latter question, to which M. Albrecht pro-
poses to return shortly, the attempt will be made to demon-
strate the correctness of his views by the former method.

Nevertheless, before this interesting subject is commenced,
it is indispensable to determine the morphological value of the
malar bone of Mammalia, since this determination will be of
the greatest utility in the sequel.

In a recent work! M. Albrecht has proved that the malar
bone of Mammalia ought to be considered as formed of three
parts, which he calls

Ist. A premalar.
2nd. A postmalar. | See the Figs. 1, 2, 3, 4, below.
3rd. A hypomalar.

>

Fic. 1.—Diagram of the malar bone divided into two parts by a horizontal
suture. (The Japanese bone.) (After P. Albrecht).
«+ y. Post-malar (Posterior postfrontal, Albrecht.) (Post-
frontal, Dollo). + Premalar (Anterior postfrontal, Albrecht.)
(Jugal, Dollo). z. Hypomalar. (Quadrato-jugal, Albrecht).

1 P. Albrecht, “Sur le Crane d’une Idiote de 21 ans,” ‘ Bull. Soc. Anthro-
pologie d. Bruxelles, T. I., p. 163.

588 M. L. DOLLO.

Fic. 2.—Diagram of the malar bone divided into two parts by a vertical
suture. (The right malar bone of the cranium of Albrecht’s idiot.)
(After P. Albrecht.)

y. Premalar (Anterior postfrontal, Albrecht.) (Jugal, Dollo).
z+z. Postmalar (Posterior postfrontal, Albrecht). (Post-
frontal, Dollo.) + Hypomalar (Quadrato-jugal, Albrecht).

Fic. 8.—Diagram of the malar bone divided into three parts, constructed from
the bipartite malar with the horizontal suture (the Japanese bone), and
the bipartite malar with the vertical suture (right malar of Albrecht’s
idiot). After P. Albrecht.

x. Postmalar (Posterior postfrontal, Albrecht) (Postfrontal,
Dollo). y. Premalar (Anterior postfrontal, Albrecht.) (Jugal,
Dollo.) 2. Hypomalar (Quadrato-jugal, Albrecht).

THE MALLEUS OF THE LACERTILIA, ETC. 589

-
Br+y+Z ‘

Fie. 4.—Diagram of the normal malar. (After P. Albrecht).
a+y-+z. Postmalar (Posterior postfrontal, Albrecht.) (Post-
frontal, Dollo). + Premalar (Anterior postfrontal, Albrecht.)
(Jugal, Dollo). + Hypomalar (Quadrato-jugal, Albrecht).

By comparison with the Sauropsida M. Albrecht arrives at

the following interpretation of these parts:
Premalar—Anterior postfrontal.
Postmalar— Posterior postfrontal.
Hypomalar—Quadrato jugal.

Lastly, in his opinion a small isolated bone discovered in a
young Cynocephalus appears to represent the jugal, which
is usually co-ossified with the supramaxillary bone.

I regret that I cannot agree with his conclusions, for in fact :

Ist. The bone of the Cynocephalus appears to me to be
simply a wormian bone, which has appeared at this spot as if
for the very purpose of leading to the construction of a theory.

2nd. In M. Albrecht’s comparison, the jugal is excluded
from taking part in the contour of the orbit, an arrangement
of which I know no other example, not even in the Lacer-
tilians with the double postfrontal.

3rd. Why should it be desired to assign to the human sub-
ject two postfrontals, when in the Sauropsida this structure is
only to be met with amongst the Lacertilia, and even amongst
these an exception rather than the rule?

(Continued on page 592.)

*xAUOLLY ‘T

ee ee eOpuysoy, ‘T
= ty = ——
; es 3 o—fo—c
psX{morydory ‘ft | a BS
Fe bs TY4Sq b
‘TA ‘ON ut gdaoxe ®o—%o—e x 8 Sd
VI[VMUBY “T > pe -
ca et ee | = ry
gp SOAV “T 0 — te ft SP
SS aed bg catia gosos ,sk{uq ‘T
one | el ale {| pe
| bg Ws wd ae S oe Oran ae
: ee aan pe a St aoe = eal j
bg 148g <—"s a Bog gE. a | }
SSS wp p ee] 22 bg | //3¥d
eer /
bo ae Ae ae ill
3 $2 3 ; ;
Oz ® EF IN AN Itq
So 7
as Oren sone Z 2
SE SE RE FE PE typos
ob Sa Sa Bs a verreyd
O° Pe 33 Se 2. -oo0qoukyy “Tt ‘RIMOL2YO “T
2 ‘slay ° BE <a ee 5 ¢ 6 3 ) = leeaor o> — % —
= Et [0 °9 ae F aeee ae L f t ic
a: {means ‘G 38 IIA TA ; g 5 “ ! }
5 Seba THEA LH (0 Tea 2 d S 2. ‘tesnl a e4 ee ' | / \ |
5 “eyoqulwosy *¢E ‘X ursvondeoxa ayy SL : 3 5 es
5 puuaqesrydmy a4 puv viseqyey ydaoxa 0 S Be ees S 2, bg sq bg 148q
etiprydg ‘Ty LBITTZAIOVT “T ~s peqrun equo.yys0g Go
Ts rae % nh Sed “II 10d
a ee ~ s —$ $< $<
| | ‘yesnl ‘jeSnl-oxvapeng
-oJeIpEny 949 03 pues ayvapenty 043 ‘uado y40gq “pasopd 430g
bg Lqsq bg ISq peziun you [vsomenbg 0} poy jesouenbg ‘Tl T
SS SSS ——c_—_—————— —_—_—_—_—_—_— --

“essoy [wi1odu1oq-vadns
® yo suryeiedes ‘youe
[Byuorz3s0d-osouruubs oN —fuguoayysod - osomunbs y
x *xXI
a ———

“nssoz [vrodure9-vadns
ou atojoroyy pus ‘youre

“WESOJ [BIOd MI9}-0104R] OU OIOJOIOY} PUB [BSN L-0vIpENny ON

Ee

“VSsOJ [BLOC UI94-0.109¥[-BAdNs v OF ASTI “wssoy [erodurg-0.104
Surais “Guonguoo assoz [v1oduie4-0109¥] ~v] put -vadns oyearedes a10zo10q3
pur -vidns ‘oe yujU01y3s0d-osomenbs on pues ‘yore [eju0173s0d-osomeunbs y
tn

[Bsnp-oyvipuny y

“VITVWAV], GNV VdIsSdOUNVG ONILSIXY AO #SSOY TVUOdWaAT-OUALVT GNV IVUOdNAT-Vudng

THE MALLEUS OF THE LAOERTILIA, BTC. 591

(Notes to the tabular statement on page 590.)

' Lobserve the following arrangement in Rhynchosuchus Schlegelii :

TT
Pstfr ———-———_————-. Sq

jJ——— Qj ——~a

Rhynchosuchus Schlegelii.

This fact will be made use of in a moment.

? Cuvier, ‘ Ossemens fossiles,’ 1836, pl. 239, figs. 9—12.

% Cuvier, loc. cit., figs. 17—20.

4 A. Milne-Edwards, “Mémoire sur le type d’une nouvelle famille de
Pordre des Ronguers,” ‘Nouvelles Archives du Museum de Paris,’ 1867,
p. 66 and pl. vii, figs. 1, 2, 3, 4.

5 The formula, as well as the preceding one, assumes that the quadrate
of M. Albrecht is really a quadrate. Otherwise we have :

irae i
J Qj
Mammalia. |

that is to say, the structure of a Chelonian, in which the supra-temporal fossa
has remained open, and in which the quadrate has been withdrawn and transe
formed into an auditory ossicle.
6 fT. H. Huxley, ‘A Manual, &.,’ p. 282.
7 T. H. Huxley, ‘A Manual, &c.,’ p. 220.
® T. H. Huxley, ‘A Manual, &.,’ p. 233—238.
9 T. H. Huxley, ‘A Manual, &c.,’ p. 230.
1 T, H. Huxley, ‘A Manual, &c.,’ p. 225.
u T, H. Huxley, ‘A Manual, &.,’ p. 228.
12 T, H. Huxley, ‘A Manual, &.,’ p. 228.
'3 Cuvier, loc. cit., figs. 21—34.

592 M. L. DOLLO.

4th. M. Albrecht’s * former interpretation, viz. that the jugal
when it is not ossified by a particular centre is ossified either
by the supra-maxillary or by the quadrato-jugal, appears to me
even less capable of defence, for among the living Sauropsida
possessing a quadrato-jugal Hatteriais the only one in which
this bone is fused to the jugal. In Hatteria we have to deal
with an exception. Therefore, why should it be desired to
render the condition in the human subject a matter of excep-
tion rather than to bring it within the rule? In other words,
it appears to me more rational to assign to Man a separate
quadrato-jugal, such as occurs in Birds, Chelonians, and
Crocodilia.

The explanation which follows seems to me a more correct
one, because by its application all the objections we have just
passed in review disappear.

Premalar= Jugal.
Postmalar= Postfrontal.
Hy pomalar= Quadrato-jugal.

This, if M. Albrecht’s quadrate be admitted, gives us the
formula :

_—_—_—_————e

Pstfr Sq

J ——— aj ——— Q

Homo.

Let us attempt to discover in the Sauropsida an analogous
arrangement, in order to better justify our theory.

If a skull of Hatteria be examined, it is seen that three
temporal fesse can be distinguished,” a latero-temporal, a
supra-temporal, and a post-temporal, I have discussed the
variations of the latter at length in another memoir,’ and it

1 P, Albrecht, ‘Sur la valeur morphologique de l’articulation mandibulaire,
&e.’ Hypomalar = Jugal + Quadrato-jugal.

2 T. H. Huxley, ‘A Manual, &c.,’ p. 220.

3 1. Dollo, ‘‘ Quatriéme note sur les Dinosauriens de Bernissart,” ‘ Bull.
Mus. Roy. Hist. Nat. Belg.,’ t. ii, p. 242.

THE MALLEUS OF THE LACERTILIA, ETC. 593

is not therefore necessary to return to the subject here. Let
us study the variations of the two former; they are compre-
hended in the table on page 590.

Let us now compare the formula given for man with that of
Rhynchosuchus Schlegelii.

——_

Pstfr Sq Pstfr - Sq
: SS 5

J Qj Q J Qj Q

——————————_—— —_——_—_—_—— $< —

Homo sapiens. Rhynchosuchus Schlegelii.

Only two things are necessary to pass from one to the
other, viz. :

Ist. The disjunction of the squamoso-postfrontal arch.

2nd. The closing together of all the other bones under con-
sideration.

Now, the former condition is very common amongst the
Sauropsida. It is to be observed, as has been already stated,
in the Ophidians, the Amphisbeenide, the Ascalobotide, the
Chalcidea and Ophidiiform Scincoidea, the Chelonia, and in
Birds; whilst the latter is by no means rare amongst the
Chelonia.

In short, my explanation of the malar bone of Mammalia
avoids entirely the objections raised by M. Albrecht’s theory,
and exhibits a simple combination of the known phenomena
in Lizards:

Malar = postfrontal + jugal + quadrato-jugal.

Before resuming the discussion concerning the quadrate, I
may again express my opinion, based on the facts which follow,
that—

Ist. The Lacertilia and Ophidia possessed originally an
ossified quadrato-jugal. .

Qnd. That therefore there existed originally in all the Sau-
ropsida a latero-temporal fossa bounded thus :

594. M. L. DOLLO.

Sauropsida.

My reasons are as follows:

lst. The ancient Reptiles (Ichthyosauria, Plesiosauria, Dino-
sauria, &c.) possessed an ossified quadrato-jugal

2nd Amongst the existing Sauropsida, Birds, Crocodiles,
and Chelonians, as well as Hatteria, possess one still,

3rd. This latter animal represents amongst the Lacertilia a
primitive type, as it is closely allied to the Triassic Rhyncho-
saurus.

4th. The quadrato-jugal exists still in a ligamentous con.
dition in the Ophidia and Lacertilia, and this is certainly an
indication of rudimentation since in all the Chalcidez and ophi-
diiform Scincoidea the squamoso-postfrontal arch likewise
becomes ligamentous

Let us now consider by what peculiarities in its connections
the quadrate of Sauropsida is distinguished, and then examine
whether these peculiarities are also to be found in M. Albrecht’s
quadrate.

The quadrate of Sauropsida may be in relation— j

Ist. At its proximal extremity with the squamosal, the pa-
rietal, and the parotic process (Iguana).

2nd. At its proximal extremity again with the malleus
(Uromastix).

3rd. At its distal extremity with the mandible and the
pterygoids (Iguana).

4th. At its distal extremity again with the quadrato-jugal
(Crocodile).

5th. Occasionally with the postfrontal (Rhynchosuchus).

6th. Lastly, with the tympanic membrane, which is inserted
into it in a posteriorly placed concavity, which might be termed
the tympanic concavity (Iguana).

THE MALLEUS OF THE LACERTILIA, ETC. 595

Now, all these relations, excepting the one with the ptery-
goid, occur in the case of M. Albrecht’s quadrate. If, then,
it is borne in mind that it is solely because of the enormous
development (compared to that in the Sauropsida) of the alis-
phenoid in man that the separation of the quadrate and pterygoid
is due, it will be admitted that M. Albrecht had good grounds
for holding the quadrate of the lower gnathostomous Verte-
brata as homologous with the zygomatic portion of the squa-
mous part of the temporal bone of Mammalia.

SUMMARY.
To resume:

I. I believe that I have discovered in Lacertilia a true mal-
leus homologous with that of Mammalia, and that this circum-
stance allows of a modification of the table of homologies given
by M. Albrecht! in the following manner.

ee

| | Interfenestral Chain.
1 | A \

| ea Sibi Tae mA

feUNSNES! I CeNELAle dives cywiacenes ecto. Suspensorium.
Teleosteans ......... he LAE Symplectic... + Hyomandibular. |
| Sauropsida............ M. tymp. ... Malleus + Columella............ M. ovalis.
| Mammalia ............ _M. tymp. ... Malleus + Incus + Os |
| nee | 3 Aiea + Stapes......... vs Sate:
ny ae rodela.| IM. tymp, 1.0: Columellay ici. .pcipinn some cgeticn M. ovalis.
Sue bess | M. aie ... Lstossicle + 2nd ossicle

+ 8rd ossicle + 4th

| OssiGletseen Ves eneee + M. ovalis.

|
|

II. I believe that I have determined the morphological
value of the malar bone of Mammalia. It is as follows:
Malar = postfrontal + jugal + quadrato-jugal.

1 P, Albrecht, ‘ Sur la valeur morphologique, &.,’ p. 253, m.

Albrecht has also communicated to me his idea that the Mallens is nothing
else than the symplectic of Teleosteans. The grounds on which he supports
this opinion are, that according to Stannius the extra-mandibular portion of
Meckel’s cartilage fixes itself upon the symplectic. I am the more disposed to
admit the correctness of this homology because, in the Lacertilia, it is to be
clearly observed that the Malleus (in regard to which M. Albrecht agrees with
me) is the intermediate bone between the quadrate aud the columella (hyoman-
dibular), just as in the osseous fishes the symplectic is interposed between
the quadrate and the hyomandibular.

596 M, 4h. “DOL:

This may be expressed in the following formula :

relations which are to found amongst t he Sauropsida also.

III. Lastly, I have by the study of the connections of the
bones materially strengthened M. Albrecht’s theory, which
holds that the zygomatic portion of the squamous part of the
the temporal bone of Mammalia is the homologue of the quad-
rate of the lower gnathostomous Vertebrata.

In conclusion, it is my duty, which is very agreable to me,
to thank my learned friend, Professor P. Albrecht, for the
liberality with which he has placed at my disposal the wood
blocks which he has used in the illustration of his two recent
memoirs, and all the more so because they are to be used to
attack one of the theories of the distinguished anatomist
himself.

NOTES ON ECHINODERM MORPHOLOGY. 597

Notes on Echinoderm Morphology, No. VI. On
the Anatomical Relations of the Vascular
System.

By

P. Herbert Carpenter, Mi.A.,
Assistant Master at Eton College.

I pReEw attention in my last note! to the striking variations
in the results of observations upon the vascular system of the
Echinoderms, which have been made by different Continental
naturalists. The leading member of the French School, Pro-
fessor E. Perrier, asserts positively that the so-called ‘ heart ”
of the Echinozoa is an excretory gland, which communicates
with the exterior through the madreporite, and is entirely free
at its inner end, no vessel proceeding from it to join an oral
ring; in fact, it is not a part of the blood-vascular system at
all. Professor Perrier had arrived at this conclusion as the
result of his own observations upon Urchins and Starfishes,
together with those of Apostolidés upon the Ophiurids. He
totally denies the existence of the radial blood-vessels described
by Ludwig in the Asterids ; and as regards the Urchins he, like
Hoffmann, is able to find but one vascular ring around the
mouth. This is described as connected with the water-tube
(stone-canal) and the radial vessels supplying the tentacles, and
also as the ring in which the ventral or internal marginal vessel
of the intestine originates. ‘Il y a donc bien réellement com-
munication entre l’appareil vasculaire intestinal et le prétendu
appareil aquifére.”’ *

1 This Journal, vol. xxii, pp. 371—386, October, 1882.

2 «Sur Vlappareil circulatoire des Oursins,’ ‘Comptes rendus,’ 1874, t. 79,
pp. 1128—1132.

VOL. XXIII.—NEW SER. RR

598 P. HERBERT OARPENTER.

This question of the communication or independence of the
water-vascular and blood-vascular systems is one of funda-
mental importance in the morphology of the Urchins, and
indeed of all Echinoderms. Hoffmann, Agassiz, and Perrier
have expressed their belief in the former; while the more
recent work of Teuscher and Koehler seems to indicate that
the water-vascular and blood-vascular systems of the regular
Urchins, at any rate, have no communication with one
another, except through the spongy tissue of the Polian
vesicles.

I have pointed out in previous notes! that Ludwig’s obser-
vations upon the Stellerids led him to regard the so-called
heart as a plexiform portion of the vascular system, connected
both with an oral and with an aboral blood-vascular ring, the
former giving rise to radial trunks which lie between the water-
vessels and the ambulacral nerves.. I have, as I have said
before, considerable faith in the accuracy of Ludwig’s observa-
tions; and this led me to suggest the possibility that the
connection of the “heart” or “ central plexus” with a second
oral ring, other than that of the water-vascular system, might ~
have been overlooked by the French naturalists. I little
thought, however, that, as regards the Urchins, a few months
would bring a striking confirmation of this suggestion in the
complete work of Mons. Koehler bimself, from whose earlier
writings, unaccompanied by figures, it was difficult to gain a
clear conception of his results.

He has recently published an elaborate memoir,’ enriched
with seven beautifully executed plates, which illustrate the
minute anatomy of the Urchins in a manner that ‘has never
before been attempted. Working, like Hoffmann and Perrier,
by the injection-method, he has been able largely to extend
the results of his predecessors; and he has also added very
considerably to our histological knowledge. His observations
have not been limited to the regular Urchins, but have been

1 This Journal, vol. xxi, 1881, pp. 170—180; vol. xxii, 1882, pp. 372—375.

2 “* Recherches sur les Hehinides des Cotes de Provence,” ‘Ann. du Mus.
d’ Hist. Nat. de Marseille,’ Zoologie. Mémoire No. 3, pp. 1—167, pl. i—vii,

NOTES ON THE ECHINODERM MORPHOLOGY. 599

extended to the Spatangidz, which appear to present certain
anomalous characters, as yet but imperfectly understood.

It will be remembered that Perrier, like Hoffmann, could
only find one vascular ring round the mouth of an Echinus;
and he described it as connected, not only with the water-tube
and radial water-vessels, but also with the ventral vessel of
the gut. The results of his injections of the so-called heart
led him to assert that it is “ trés nettement terminé vers le bas
et qu’il n’en sort aucun canal. Il ne peut donc étre question
d’un vaisseau inférieur paralléle au canal du sable et aboutis-
sant a l’un des deux cercles vasculaires que l’on suppose exister
autour de l’cesophage.”

Koehler finds, however, that by inserting his cannula into the
lower end of this organ, which, like Perrier, he calls the
“ ovoid gland,” he is able to inject a vessel lying alongside the
water-tube, but totally distinct from it. He calls it the
“ olandular canal,’”’ and finds it to be connected with an oral
ring in which radial vessels originate. These are not the
water-vascular ring and radial water-vessels, which can be
injected from the water-tube altogether independently of those
described by Koehler. He says that the glandular canal can
be readily followed with the aid of a lens, right up from the
lantern to the apex of the ovoid gland. He speaks of it as
containing blood, and has frequently found coagulum both in
it and in its ramifications over the peripheral part of the ovoid
gland. When this glandular canal is carefully injected, the
fluid enters the circumoral ring before mentioned, together with
the internal marginal vessel which is connected with it, and
not with the water-vascular ring as formerly described. The
injection also passes into arborescent ramifications within the
Polian vesicles; but if pressure is used, it fills these organs
and enters the water-tube and radial vessels.”

This second oral ring seems to have been previously seen by

1 © Recherches sur lapparcil circulatoire des Oursins,” ‘Arch. de. Zool.
exp. et. gén.,’ . iv, 1875, p. 613.
2 Op. cit., pp. 65-70.

600 P. HERBERT CARPENTER.

Teuscher,! who failed, however, to distinguish the radial blood-
vessels in the ambulacra, confounding them with the so-called
perineural spaces. The two vessels are readily visible on the
inner aspect of the ambulacra, each sending branches to the
ampulle, and may be injected with different materials ; while
their distinctness, though not apparent in Teuscher’s section
of the ambulacra, is quite evident in that figured by Koehler.

According to Teuscher, the water-vessels of the ambulacra
ascend the outside of the lantern to join the ring at its summit,
just as Perrier described them; while the blood-vessels pass
over the actinal membrane towards the mouth, ascend along
the pharynx, between it and the interpyramidal muscles, and
join the superior of the two rings, in which the intestinal vessel
originates. Koehler, however, also working by the section
method, has been unable to find these radial vessels described
by Teuscher on the pharynx, and his injections have not led
him to suspect their presence.” In fact, he says distinctly that
at the edge of the peristome “les vaisseaux ambulacraires, de
doubles qu’ils étaient, deviennent simples et forment alors les
cing branches qui montent sur la face-externe de la lanterne,
et vont aboutir au cercle péricesophagien inférieur” (?.e. water-
vascular ring).

It is noteworthy, however, that in the Spatangids either of
the two oral rings, internal or external, may be injected from
the corresponding radial vessel, and Koehler says distinctly
that each of these rings sends branches into the ambulacra.®
This leads to the suspicion that each radial vessel of an
Echinus communicates directly with a corresponding vascular
ring, just as was described by Teuscher, and that the water-
vascular and blood-vascular systems are distinct, at any rate in
the peristome and ambulacra.

Koehler’s memoir would have been still more valuable than
it is, had he entered more fully into a comparison of his results

1 “ Beitraige zur Anatomie der Echinodermen,” iv, ‘ Echinidae. Jen. Zeitschr.,’
Bd. x, pp. 520—521, Taf. xx, fig. 6.

2. Op. cit., pp. 70, 73.

3 Op. cit., p.p. 93, 94, 100,

NOTES ON ECHINODERM MORPHOLOGY. 601

with those of other workers, especially Teuscher and Ludwig,
both in this group and in other Echinoderms. He entirely
ignores the fact that blood-vascular and water-vascular systems,
consisting respectively of oral rings and radial vessels, have
been described as distinct and totally independent in Asterids,
Ophiurids, and Crinoids, apart from Teuscher’s work on the
Urchins. He continually speaks of the ambulacral vessels as
being ‘‘ double,” meaning that both of them belong to what
Perrier called the respiratory portion of the vascular system,
that, namely, which communicates with the exterior by the
water-tube. ‘The so-called absorbent portion of the vascular
system consists of the intestinal vessels, the ventral one of which
is connected with the superior oral ring.

Koehler’s position, therefore, involves the curious anomaly
that Echinus has two oral rings, both containing blood, but
the one is “absorbent” and the other “ respiratory ;’’ while
there are also two sets of radial vessels, both, however, belong-
ing to the respiratory system, and connected exclusively with
one ring (water-vascular, auct); whereas in Spatangus, as
he himself describes, each oral ring is connected with a cor-
responding set of radial vessels. This is also the case in all
the other Echinoderms; and [ cannot but think that the great
difficulties involved by the presence of the lantern in Echinus
may have caused the connection of the smaller radial vessels
with the superior oral ring to have escaped Mons. Koehler’s
notice. He gives no figure showing their union with the
water-vessels proper before joining the inferior or water-
vascular ring, which suggests the possibility that the supposed
union is an inference and not an observed fact. But if an
inference was necessary, it would surely have been more natural
to suppose that a character which has been described by four
different observers in Asterids, Ophiurids, Crinoids, and
Spatangids, is also’ to be found in Echinus.

The next point to be considered is the connection of the so-
called heart or ovoid gland with the vascular system, which
has been so positively denied by Perrier. ‘The glandular canal
discovered by Koehler which rises from the superior oral ring

602 P. HERBERT CARPENTER.

(blood-vascular, auct) is frequently described by him as con-
tinued upwards through the ovoid gland by means of its
excretory canal to the level of the madreporite. He says, for
example, that it is “ en communication avec la glande ovoide
et lui permet de recevoir le sang en assez grande quantité.”
Further on he speaks of “ Vinterposition sur un certain point
du trajet des vaisseaux d’un organe glandulaire, destiné sans
doute a débarrasser le sang de produits inutiles et a les laisser
s’échapper au dehors 4 travers la plaque madréporique.”?
There can, therefore, be no question about the intimate rela-
tion of this organ with the blood-vascular system of the
Urchins, just as in other Echinoderms, as described by Ludwig,
Teuscher, myself, and others.

This relation, however, has been denied by Perrier and
Apostolidés, who suppose the ovoid gland not only to open by
its excretory duct into the sinus beneath the madreporite, as
described by Koehler, but also to be entirely independent of
any vessels whatever. Their observations on Asterids and
Ophiurids are in direct conflict with those of Ludwig; but
Koehler’s discoveries in the Urchins furnish a strong argument
in favour of Ludwig’s views.

The minute structure of the ovoid gland in Urchins and
Spatangids is briefly described by Koehler in the following
terms :— A reticulum of connective tissue, supporting cellular
elements that undergo a peculiar degeneration, the final result
of which is the formation of numerous pigment masses.” This
network of connective-tissue fibres is most regular near the
periphery of the gland, the fibres being more numerous and also
better defined towards the centre. It encloses alveolar spaces,
within which are groups of from one to four naked masses of
protoplasm, of an irregular stellate form. Each contains a
nucleus, which is of variable size. It occurs in all stages, from
a finely granular condition to one in which it is little else than
a mass of brown pigment-granules, surrounded by a thin pro-
toplasmic envelope, and in the peripheral region the alveoli
sometimes contain nothing but these pigment groups.

Op. cit., pp. 72, 102

NOTES ON ECHINODERM MORPHOLOGY. 603

These cells were supposed by Perrier, and also by Aposto-
lidés, in the case of the Ophiurids, to be grouped into columnar
acini, as in the case of compact glands like the liver. Koehler
was at first inclined to take the same view of their arrangement,
but subsequent investigations led him to give the description
of the structure of this organ, which has been summarised
above. Several so-called lacune are visible in sections through
the gland. Some of them correspond to the numerous anas-
tomosing canaliculi which unite into its excretory duct. But
others, especially those in the peripheral and lowest parts of
the gland, are filled with coagulum. These represent the
lumina of the vessels which ramify on its surface, and are con-
nected with the “ glandular canal” rising from the oral blood-
vascular ring. The structure of the upper part of this canal
is essentially similar to that of the gland itself; but towards
the peristome the trabecular structure becomes less marked,
and lacunar spaces lined by an epithelium relatively more
prominent. Various transitional stages may be found between
the cellular elements of the ovoid gland and the epithelial cells
lining the lower portion of the glandular canal, and Koehler
regards the former as a modification of the latter.

The glandular canal and parts of the ovoid gland of the
Urchins thus consist of numerous small vessels with an epithe-
lial lining, while the cells of other parts are more irregularly
arranged among fibres of connective tissue. This is not so
very different from Ludwig’s account of the same organ in
Asterids and Ophiurids, the correctness of which has been
questioned by Perrier and Apostolidés; and it would seem,
therefore, that the designation “ central plexus” is not so very
inapplicable after all. ‘‘ Plexiform gland” would, perhaps, be
better; but any appellations based upon its form, such as
“ovoid gland” (Perrier) or “ piriform gland” (Apostolidés),
are inconvenient when applied to other types of Echinoderms,
such as the Crinoids, in which the so-called ovoid gland breaks
up into a number of loosely-connected lobules. The organ in
question is most certainly not a heart, and may very probably
have something to do with the production of the brown

604. P. HERBERT CARPENTER.

pigment-bodies, which are so familiar to all workers on
Kchinoderms.

Koehler makes little reference to the observations of Geddes,!
who found ameeboid corpuscles containing brown pigment-
granules in the perivisceral fluid, and considered them as
respiratory in character. These cannot be very different from
the pigmented “ cellules 4 protoplasma irrégulier émettant de
fins prolongements,” which Koebler found in his osmic-acid
preparations of the ovoid gland; while the larger pigment
masses, devoid of a nucleus but surrounded by a thin proto-
plasmic envelope, which Koehler found in the ovoid gland,
seem to have been also met with by Geddes in the intestinal
vessels. He also describes them as occurring at certain times
in the ambulacral vesicles ; and he speaks of these yellow-brown
granules as being developed at the expense of the nuclei of their
epithelial lining, which recalls the relation of similar granules
to the nuclei of cells derived from the epithelium of the so-
called glandular canal described by Koehler. The Polian vesicles
have essentially the same structure as the “ ovoid gland,” and
are evidently the source of many of these pigmented cells. ‘The
colouring matter of the latter contains iron, and undergoes
changes of tint when exposed to the atmosphere, so that it is
doubtless of a respiratory nature, as supposed by Geddes,
though Foettinger’s discovery? of hemoglobin-tinted cor-
puscles within the water-vessels must not be forgotten.

Considering that the production of these pigment granules
involves the destruction of the cells of the ovoid gland and
Polian vesicles, Koehler regards these organs as excretory in
function, as had been previously done by those who denied the
connection of the ovoid gland with the vascular system, though
positively asserting its communication with the exterior by a
duct. It is likely enough that this organ is one of the facto-
ries of respiratory pigment, the necessary material reaching it

1 “Qbservations sur le fluide périviseéral des Oursins,” ‘Arch. de zool.
exp. et. gén., t. viii, pp. 483—496, pls. xxxvii, xxxviii.

9

2 “Sur lexistence de ’hémoglobine chez les Echinodermes,” ‘Arch. de
Biol.,’ vol. i, pp. 405—412, pl. xvii

NOTES ON ECHINODERM MORPHOLOGY. 605

from the oral blood-vascular ring through the “ giandular
canal,” and the fluid being forced on by the contractions
of the ventral intestinal vessel which joins the oral ring. One
would wish, however, for something more than the mere asser-
tion of Koehler that a liquid is secreted by the gland. He
describes a coagulum found in its excretory duct as the
“coagulated secretion” of the organ.! But he finds the
same coagulum in the vessels ramifying on the surface of its
lower portion and in the internal vessel of the intestine.? If
the one is a glandular secretion, why not the other too? ‘They
are both described in exactly the same terms, and are evi-
dently of the same nature as the coagulum so often found in
the blood-vessels of other Echinoderms. I have met with it
over and over again in the radial, and also in the intervisceral
vessels of Crinoids. Is this a coagulated secretion too? Not
even Professor Perrier himself,who has recently published a brief
note® ‘On the Organisation of Crinoids,” has ventured to assert
that the so-called ovoid gland of these creatures communicates
with the exterior, though he will not admit the existence of
any intervisceral vessels connected with it. But as regards
the Echinozoa, I would still ask for proofs, based not merely
on the results of injection, but also on the section method, that
the “excretory duct” of the ovoid gland communicates with
the exterior, and not with an aboral ring in which the genital
vessels originate, as described by Ludwig in the Stellerids.
The direct communication with the exterior of the blood-
vascular as well as of the water-vascular system would, if
established beyond dispute, be a somewhat important mor-
phological fact ; and before accepting the apparently well-
grounded conclusions of Koehler and Perrier respecting the
regular Urchins, one would like to know what Ludwig has to
say on the subject. ‘This will appear, I hope, ere many
months are past.

' Op. cif., pp. 98, 99.
2 Op. cit. pp. 77, 90.
3 «Comptes rendus,’ t. xevil, No. 3, pp. 1S7—189.

606 P. HERBERT CARPENTER.

In the Spatangids, as in the Urchins, Hoffmann! was only
able to find one vascular ring around the mouth, that of the
ambulacral system. He described the ventral vessel as not
originating in this ring, but as communicating with it by a
special connecting branch; while he was unable to differen-
tiate the water-tube from the so-called “ heart,” of which,
however, he recognised the glandular nature.

Teuscher ? satisfied himself of the existence of another vessel
than the water-tube at the side of the gullet, but failed to make
out its connections in the peristome. It corresponds to the
‘‘ glandular canal” of Echinus. Working downwards from
the madreporite, he found the “ heart ” or gland with the water-
tube close toit. The latter is readily recognisable by its lining
of columnar epithelium, and lies closer to the gland than in
Echinus. At the lowest part of the gland, where the water-
tube dilates slightly, Teuscher finds ‘‘ seine innere dem Herzen
anliegende Wand immer stark verdunnt.. .. Der Stein-
canal nachdem er am Herzen voribergegangen ist, wird von
einem oder zwei feinen Gefassen begleitet.’” It then passes
downwards underneath the diverticulum to reach the gullet.
There is some obscurity in this description, as there also is in
that given by Koehler, but the two partially interpret one
another. InSpatangus, asin Echinus, Koehler found two
vascular rings round the mouth, each with radial extensions
into the ambulacra. But he describes the communicating
branch from the internal marginal vessel as connected with
both rings. ‘“‘ Simple sur presque toute sa longueur, elle se
bifurque a son extrémité et chacune des deux branches se jette
dans un des cercles peribuccaux.” This is improbable, to say
the least of it, for in no other Echinoderm are the visceral
vessels known to communicate with the water-vascular ring.
Doubtless, however, it does not surprise Koehler, who does
not distinguish between blood-vascular and water-vascular
systems. But as this point is important, one regrets that he

1 «Zur Anatomie der Echinen und Spatangen,” ‘ Niederl. Arch. de Zool.,’
Bd. i, pp. 1O—112, Taf. iii—x.

2 Loc cit., pp. 681—534.

NOTES ON ECHINODERM MORPHOLOGY. 607

did not illustrate it by another of his admirable figures. In
any case, however, he is to be credited with the discovery of
the blood-vascular ring in Spatangus, though not regarding
it in the same way as most other authors would

In accordance with his peculiar views he described the
water-tube as double along the whole length of the gullet,
meaning that there are two canals, as stated by Teuscher.
One of these, that farther from the gullet, is somewat irregular
and sinuous in form, being lined by large cells with volumi-
nous granular nuclei and masses of pigment, while the more
uniform vessel, closer to the gullet, is lined by small epithelial
cells. ‘“‘Au point ot ’cesophage se termine, le canal sinueux
s’amincit peu a peu et cesse d’étre distinct ; il se confond avec
le deuxiéme canal qui reste dés lors unique et continue son
chemin jusqu’a ’extrémité de la courbure inférieure et de la
jusqu’a Vorgane d’excrétion.”? After the disappearance of this
sinuous pigmented canal the more uniform one remains of the
same character until it reaches the diverticulum. Its lumen then
becomes divided up by partitions, and its cellular lining con-
sists of larger elements with granular nuclei, so that it passes
gradually into the so-called excretory organ. Koehler speaks
ofit as the water-tube!, not only communicating with the gland,
but in perfect structural continuity with it; and it is doubtless
the one described by Teuscher as the water-tube between the
gland and the gullet. But I much doubt its really belonging
to the water-vascular system; and it appears to me to
correspond to the ‘ glandular canal” of Echinus, which
likewise connects the gland with the oral, blood-vascular
ring.

Towards the apical extremity of the gland, where its charac-
teristic parenchyma becomes less developed, two special canals
differentiate themselves. 1. The ‘“ madreporic canal,’ more
centrally placed, often containing coagulum, and the first to
appear at the tapering upper end of the gland, which it resem-
bles in structure. 2%. A more peripheral one lined by a
regular columnar epithelium. The glandular tissue finally

' Op. cit., pp. 92—99. °

608 P. HERBERT CARPENTER.

disappears, and these two canals pass up side by side towards
the apical pole.

The results of Koehler’s injections lead him to identify the
second canal with one or two canals which pass from the sur-
face of the gland on to its supporting mesentery, where they
lose themselves in a large and irregular network in the inter-
stices of the connective tissue. According to Koehler, there-
fore, the two vessels which together make up the stone-canal
at the level of the gullet, each communicating with one of the
peribuccal rings, soon fuse into a single canal that terminates
in a gland placed at the extremity of the diverticulum, while
the intestinal vessel communicates through its connecting
branch with both the peribuccal rings. “The fusion of the
two systems is thus as complete as possible, and it would be
difficult to admit the existence of a distinction between a
water-vascular and a blood-vascular system.

To this I would remark: (1) The fact that each peribuccal
ring sends a separate branch into the ambulacra, does not
look like a more complete fusion of the two systems than exists
in Echinus; where Koehler, though I believe wrongly, des-
cribes both the radial vessels as communicating only with the
water-vascular ring. He cannot surely mean that each ring
in Spatangus communicates with both the radial vessels.
As regards the connecting branch joining both the peribuccal
rings, I should wish, as I have pointed out above, for more
distinct proof than a mere assertion.

(2) Koehler regards the organ which is commonly called the
stone-canal of Spatangus, as homologous with the glandular
canal of Echinus, on account of its relations to the excretory
gland. It appears to me, however, that Teuscher’s identifi-
cation of the canal with columnar epithelium, found by him at
the apical pole, as the water-tube, is more correct than
Koehler’s view of what is evidently the same structure. A
canal of this kind would not be likely to end in a vascular
network in the mesentery; and I cannot but think that
Koehler, who does not mention Teuscher’s opinion, is here
in error. Both authors agree in its peripheral position as

NOTES ON ECHINODERM MORPHOLOGY. 609

regards the gland, and also that it is not continuous, indepen-
dently of the “ glandular canal ” with either of the two canals
arising from the peribuccal rings. I will assume, though Koehler
nowhere says so, that the more sinuous pigmented canal which
“ disappears ” higher up, starts from the water-vascular ring ;
while the other, which eventually becomes glandular, is con-
nected with the blood-vascular ring as in Echinus. If this
be the case, it would seem that the Spatangids, like the
Crinoids, and possibly also the Ophiurids, have an interrupted
water-tube arising from the water-vascular ring; though in-
direct communication between the two is effected by the body-
cavity into which both tubes open by their inner ends. This
seems tome much more probable and in better accordance with
the morphology of Echinoderms generally than the position
taken up by Koehler. According to him the “ madreporic
canal” of Spatangus is to be considered as the water-tube,
though it lacks the characteristic columnar epithelium. It is
further comparable in every respect, as Koehler himself admits,
to the so-called excretory duct of the ovoid glandin Echinus,
and therefore, also in other regular Echinoderms. In both Aste-
rids and Ophiurids this “ excretory duct ” is said by Ludwig to
join an aboral ring in which the genital vessels arise; and
certain points in Koehler’s memoir lead me to think that the
same may be the case in the Urchins. Various facts mentioned
by Hoffmann, Perrier, Teuscher, and Koehler, more especially
by the two latter, who are the only ones to figure sections of
the ambulacra, also indicate that the system of perihemal
spaces originally derived from the ccelom, which have been
described under various names in the Asterids and Ophiurids,
really occur in the Urchins too. But on this point, as on many
others, we shall doubtless be enlightened by Ludwig himself.
I have already alluded to the recently published note by
Professor Perrier, ‘‘On the organisation of Crinoids.” He,
of course, uses the somewhat inappropriate term “ ovoid
gland” in speaking of the central plexus, but says nothing
whatever about its having any communication with the exterior.
This is not very surprising, for the water-pores of a Crinoid

610 P. HERBERT CARPENTER.

are scattered about over the ventral surface of the disc, and
not collected into a single plate as in most Echinozoa. He
does not seem to believe in the connection of the central plexus
with the intervisceral blood-vessels, which has been described
by Ludwig and myself. For he says, “ Les vaisseaux qui
paraissent en partir ne sont autre chose que les ramifications
de la glande, se terminant d’ordinaire par des renflements
ayant l’aspect de culs-de-sac. Ces ramifications courent au
milieu des innombrables trabécules du tissu conjonctif de la
cavité générale, qui peuvent eux-mémes parfois prendre |’appa-
rence de vaisseaux.” In the early larva, the central plexus is
“un corps fusiforme plein, allant du cercle oral au pédoncule
dont il continue le cordon axial. Ce corps n’emet aucune
ramification: il ne saurait par conséquent étre question a ce
moment d’appareil vasculaire. Le corps ovoide s’implante
chez la Comatule adulte sur l'un des planchers horizontaux
de l’organe cloisonné.”

Professor Perrier does not definitely name the species which
has afforded him material for his observations. But, as he
says that they have principally been made on young indivi-
duals, and on Pentacrinoid larvee, it is ‘tolerably evident that
the common Ant. rosacea of the English Channel and the
Mediterranean has been the subject of his researches. I do
not know whether he has ever made preparations of Ant.
eschrichti, or of any species of Actinometra or Penta-
crinus. But unless he has done so, he appears to me to be
somewhat rash in denying the conclusions reached by other
workers who have had these opportunities, on the strength of
observations made on a single species.

Some years ago Professor Perrier, who had worked by one
method only, was led not only to deny the existence of a
particular canal in the arms of Ant. rosacea, which had been
described by Dr. Carpenter, but also to predict that no one
else would find it. It has, however, been described by Greeff,
Teuscher, Ludwig, myself, and finally by Perrier himself;
and I cannot help suspecting, therefore, that he has been again
misled by the limited nature of his observations, The mere

NOTES ON ECHINODERM MORPHOLOGY. 611

fact that any one fails to demonstrate the existence of a certain
anatomical structure is no proof of the non-existence of that
structure. Two instances already mentioned, viz. the connec-
tion of the ovoid gland in the Urchins with an oral ring, and
the existence of both cceliac and subtentacular canals in the
arms of Crinoids are cases in point. Another, which will be
noticed immediately in more detail, is that Professor Perrier,
working with young and fresh material, has seen the connec-
tion between branches from the axial cords of the arms and the
muscle-fibres, which I have long sought for in vain in spirit
specimens.

Both Ludwig and myself have found that certain anatomical
points are not easily demonstrated in Ant. rosacea, whereas
they are much more evident in the larger Ant. eschrichti.
I have five series of sections through the disc of this species ;
and have also cut four other species of Antedon, four of
Actinometra, and three of Pentacrinus, together with
Promachocrinus, Rhizocrinus, and Bathycrinus. I
venture to think, therefore, that [ am on the whole in a better
position than Prof. Perrier for forming a judgment respecting
the anatomical relations of the central plexus. In one respect,
however, my opportunities have been inferior to his. All my
work has been done on material which has been some time in
spirit; and though this affects anatomical structure but little,
it makes a vast amount of difference in histological work.
Prof. Perrier, on the other, hand has had access to an abundant
supply of fresh Ant. rosacea of all ages; and his statement
that the histological structure of the “ ovoid gland” of this
type is identical with that of the same organ in other Echino-
derms! must therefore be received as authoritative. But
when he says that the intervisceral blood-vessels described by
Ludwig and myself as originating in this organ are merely
ramifications of the gland ending in apparently blind dilata-
tions, I must totally disagree with him. That the walls of the
central plexus are of a glandular nature must be apparent to
everyone who has examined a section of it. But I have also

' As described by himself, Apostolidés, or Koehler ?

612 P. HERBERT CARPENTER.

no doubt whatever as to the connection of its cavities with
those of the chambered organ, and through it, with the vas-
cular axis of the stem in stalked Crinoids. Ludwig! has
given excellent figures in illustration of these points; and my
own observations have repeatedly demonstrated their accuracy,
not only in Ant. rosacea, but in other genera and species.
The chambered organ is an enlargement at the top of the
vascular axis of the larval stem, which Perrier describes as
continuous with the ovoid gland. But he nowhere mentions
the vessels contained in this axis which expand into the
cavities of the chambered organ above, just as in the stalked
Crinoids; and he does not appear to consider these chambers as
part of any blood-vascular system. If, as he seems to imply,
they are disconnected from the ovoid gland in the adult, how
does he explain the connection of the latter with the axis of
the larval stem ?

It will be noted that Prof. Perrier tacitly admits the con-
nection of apparently vascular structures with the ovoid gland,
though he speaks of them as its ramifications and as seemingly
blind. I am well aware of their apparent blindness, but it is
simply due to the impossibility of any single section showing
more than a very small pcrtion of their winding course. This
is a difficulty familiar to all workers. But a careful study of a
good dissection, or of a moderately thick transparent section,
especially with a binocular, or an accurate plotting out on
paper of a series of thin sections by means of a camera, will
reveal much that is totally unrecognisable in other ways. The
diagrammatic figures which I have given of transverse and
longitudinal sections through the disc of Actinometra? were
made by thus plotting out with a camera.

The intervisceral blood-vessels of this and other types have
no glandular structure whatever, as they should on Professor
Perrier’s theory. They are simple tubes as described by
Ludwig, and lined. by an epithelium which is more delicate
than that within the extensions of the body-cavity into the

1 * Zeitschr. f. wiss. Zool., Bd. xxviii, Taf. xiv—xviii.
2 This Journal, vol. xxi, Pl. xii, figs. 14, 15.

NOTES ON ECHINODERM MORPHOLOGY. 613

arms. They often contain coagulum, and with @ little prac-
tice may be readily distinguished from connective tissue.
Professor Perrier gives us hardly any information respecting
the relations of the upper end of the ovoid gland. He admits
its connection with an oral ring in the Pentacrinoid, but he
does not say whether the functions of this ring are water-
vascular, blood-vascular, or the two combined; and when he
says that the gland gives off no ramifications, he must have
forgotten Dr. Carpenter’s description of its subdivision “into
diverging branches, of which one passes to each ray.”” This is
perhaps a stage which has not come under Professor Perrier’s
observation. The radial branches of the ovoid gland develope
into the genital vessels which form a plexus beneath the ambu-
lacra of the disc, and eventually extend into the arms and
pinnules. Professor Perrier does not mention this plexus,
though it cannot well have escaped his notice ; nor does he
enter at all into the question of the ventral termination of the
ovoid gland in the adult. I regret his silence the more, as this
is especially a point on which more extended observations are
wanted. Both Ludwig and myself have experienced consider-
able difficulty with Ant. rosacea; but I have found A.
Eschrichti, A. carinata, Pentacrinus decorus, and
Promachocrinus kerguelenensis much more favorable
subjects of study. In the first-named species the ventral
branches of the central plexus end in a spongy organ with well-
defined limits, which has somewhat the appearance of a lym-
phatic gland. It is especially developed between the mouth
and anus, and is connected both with the oral blood-vascular
ring and with the genital vessels of the rays. An organ of
essentially the same character, though less prominent, occurs
in Ant. rosacea; and J had hoped for some account of it from
Professor Perrier, who, unfortunately, does not mention it. I
trust, however, that in the complete memoir which he is prepar-
ing he will remedy this omission. Meanwhile I am searching for
this spongy organ in as many different species as possible, and

1 *Pree. Roy. Soc.,” vol. xxiv, p. 221.

VOL, XXIII.— NEW SER. ss

614 P. HERBERT CARPENTER.

propose eventually to describe the comparative anatomy of this
portion of the vascular system in the various types of Crinoids.
At present, I would emphasise two points strongly, viz. the
connection of the central plexus with the oral ring and genital
vessels above, and with the vascular axis of the stem at its
other end, which does not communicate with the exterior, as
the corresponding (?) part of the ovoid gland is said to do in
the Echinozoa.

It will be evident from what has been written above that, so
far as the vascular system is concerned, I am inclined to adopt
Ludwig’s views rather than those held by Professor Perrier
and his colleagues. The French author, however, is returning
good for evil, and sides with me ia one, if not both, of the two
cardinal points wherein I disagree with Ludwig, viz. the
nervous system of the Crinoids, and the homologues of their
basal plates in Starfishes. It is with the first of these questions
only that I am now concerned. For the past six years I have
been continually advocating my father’s view respecting the
nervous nature of the fibrillar envelope of the chambered organ
of the Crinoids, and its extensions into the rays and arms.
Ludwig, however, expressed his total dissent from this doctrine;
and it has consequently been ignored or dismissed with the
briefest possible mention in the various German text-books on
comparative anatomy. I have reason to believe that a few
teachers have assented to it; but, so far as I know, Professor
Perrier is the first continental worker on Echinoderms who has
publicly adopted it. This is the more important, as he for-
merly expressed his inability to do so; and he has been able
to strengthen it in two important points. For he has not only
seen the ultimate branches of the axial cords, which altogether
escaped the notice of the German observers, but he has also
traced a connection between some of them and the muscle-
fibres through the intervention of stellate cells; while he has
followed others into the tentacles, and describes them as enter-
ing the papille borne by these organs. I have lately found
that these ramifications of the axial cord occur in the stem of
Pentacrinus and Bathycrinus, and that they are greatly

NOTES ON ECHINODERM MORPHOLOGY. 615

developed in the arms of the latter, bipolar cells being inter-
calated in their course.

I have also seen these bipolar cells in Ant. Eschrichti,
which type has further yielded me another important piece of
evidence. The existence of a fibrillar plexus, derived from the
axial cord, within the connective tissue forming the perisome
at the sides of the brachial ambulacra has long been known to
me. But until about a year ago I searched in vain for any
similar structure on the disc. At last, however, I succeeded
in following this plexus from the arm-bases down on to the
disc. It is extensively developed among the sacculi at the sides
of the ambulacra, and forms an annular network in the con-
nective tissue occupying the lip, but of course much further
from the mouth than the subepithelial ring discovered by
Ludwig. I detected this plexus first in a specimen which had
been stained with borax-carmine, and subsequently found it
also, though less readily visible, in hematoxylin preparations.
Last of all, on looking over the remains of my earliest sections,
made in Professor Semper’s laboratory at Wiurzburg in the
winter of 1875-76, and stained with Beale’s carmine, I was
able to make out traces of the same parambulacral network
which had originally escaped my notice.! I have likewise
found it in the disc of A. antarctica, and even of A. rosa-
cea; and I have no doubt that the action of gold-chloride or
osmic acid on fresh material would bring it out in a more strik-
ing manner.

I am strongly inclined to believe that extensions of this
plexus are in direct connection with the fibrils of the subepi-
thelial band, which is regarded by Ludwig as the sole nervous
apparatus in the Crinoid organisation. In fact, some histolo-
gists who have seen my preparations have expressed themselves
as having no doubt that this is the case. I hope, however, to
obtain some still better sections than those upon which this
Opinion was based before finally adopting it as my own. In

1 Preparations illustrating this structure were exhibited at the meeting of

the Zoological Society on December 19th, 1882, and will be figured in the
“* Challenger ” report.

616 P. HERBERT CARPENTER.

the minute details of the disc structure of the Crinoids there is
still very much to be worked out; and having plenty of material
I trust in the course of time to be able to add considerably to
our present knowledge, both of the nervous and of the vascular
systems.

ORIGIN OF THE SEXUAL CELLS IN HYDROIDS. 617

Recent Researches upon the Origin of the Sexual
Cells in Hydroids.' A Review.

By
A. G. Bourne, B.8Sc., Loma.

THE question as to the place of origin of the sexual products
of Hydroids is one upon which very various opinions have of
late years been current. The ectoderm and endoderm have,
in turn, been put forward as giving rise to either, or both, eggs
and spermatozoa.

Kleinenberg, in speaking of Hydra, and F. E. Schultze, of
Cordylophora, state that both products are derived from the
ectoderm, a result with which Weissmann agrees.

Grobben has observed the same origin in Podocoryne
carnea, and F. E. Schultze in Sarsia tubulosa. The
Hertwigs have shown the same ectodermal origin of both
elements in numerous meduse; and lastly, Ciamician has
shown the same origin in Tubularians, Weissmann having
arrived at the same conclusions even before Ciamician’s pub-
lications.

On the other hand Weissmann has clearly demonstrated
both products to have an endodermic origin in Plumularia,
Sertularella, and Eudendrium.

1 Weissmann, A. ‘ Observations sur les cellules Sexuelles des Hydroides.”
‘Bibliothéque de l’école des Hautes Etudes, Section des Sciences Naturelles,’
xxiv, No. 3.

VaRENNE, A. De. “Sur l’origine des spermatozoides chez les Hydraires,”
‘ Comptes Rendus,’ xciii, pp. 1032—1034. “ Développement de lceuf de la
Podocoryne carnea,” ‘ Comptes Rendus,’ xciv, pp. 892—894. ‘ Recherches
sur les Polypes hydraires, (Reproduction et Développement),’ Paris, 1882,
8vo, (104 pp., 10 pls.).

618 A. G. BOURNE.

The spermatozoa have been stated in some cases to arise
from the ectoderm while the ova arise from the endoderm. KE.
von Beneden has shown this to be the case in Hydractinia,
Fraipont in Campanularia, and Weissmann in Gonothy-
rea, These various modes of origin have been described as
existing in the same family.

Weissmann’s recent observations advance our knowledge a
considerable step in a slightly different direction; he has
shown that there are a large number of species, of genera, and
even of families, in which the generative products do not
originate in reproductive individuals—gonophores—but in the
parenchyma of the trophosome, the ceenenchyme of Milne
Edwards and Haime, the coenosarc of Allmann, and that they
afterwards migrate to a “ maturing bud” (gonangium). Such
an origin Weissmann terms ccenosarcal, in contradistinction
to a blastoidal origin; and he would recognise two types of
Hydroids—ccenogenous (ccnosarcogenous), and blas-
togenous.

The whole process especially with regard to the formation of
the gonangium and the migration of the sexual elements
into it is a remarkable one, and we may cite Weissmann’s
observations upon Plumularia echinulata (Lam.).

Both varieties of sexual cells form in the endoderm of
the coenosarc, usually in the trunk of the colony, often
at the base of the lateral branches. At the time that the
sexual cells appear, there is no trace of gonangia. These
form in the ccenosare in the neighbourhood of sexual cells.
The sexual cells arise in a similar manner in both sexes, ova
and spermatospores (sperm mother-cells) arising by metamor-
phosis of ordinary endoderm cells : this Weissmann has observed
with great certainty.

The gonangia develope with perfect regularity from below,
upwards, so that we can determine beforehand the precise spot
where a gonangium will later on develope.

Where a gonangium is about to develope the first change
takes place in the ectoderm, which in the trunk consists of
several layers of cells.

ORIGIN OF TH SEXUAL CELLS IN HYDROIDS. 619

The cells of the outer layers which have primitively an
irregular polygonal form become elongated and placed perpen-
dicularly to the plane of the basal lamella, at the same time
changing in character, losing their granules and becoming
clear. Such a modification takes place just above a group of
sexual cells.

The modified ectodermal cells now form a rounded tubercle
which becomes nipped at the base by a circular groove and
will now penetrate the perisarc. The perisare in fact is
gradually eaten away by a chemical action. That there is no
mechanical pushing and gradual thinning out of its substance
can be seen by tracing the parallel strize which stop short at
the edges of the window which is formed. This eating away
of the perisarc is a most remarkable process. The perisarc is
a chitinous substance and only dissolves with the greatest
difficulty in concentrated acids or alkalies.

Experimenting upon Plumularia echinulata, Weiss-
mann found that the perisare entirely resists the action of
sulphuric and hydrochloric acids for five days, as well as that
of caustic potash: it did, however, dissolve in the latter at the
end of a month.

Numerous organic bodies resist the action of strong acids or
alkalies while they are attacked by weaker solutions, Weiss-
mann has therefore tried all stages of dilution. In a ‘1 per
cent. solution of potash the perisarc had not, however, com-
pletely dissolved at the end of a fortnight.

There is a further curious point ; even while these cells are
dissolving the outer layers of perisare they do not attack the
youngest (innermost) layer which they push before them. This
young layer, which Weissmann terms the cambium layer,
present differences both chemical and physical from the older
layers, it stains more strongly and more readily with carmine,]
and, moreover, is soft and elastic, as may be seen during the

' T have observed a similar difference in the chitinous layers of the lens of
the central eye of a young Limulus, one portion although here occupying
not an internal but a central position, stains deeply when treated with borax-
carmine while the remainder remains unstained.

620 A. G. BOURNE.

further growth of the gonangium. After the formation of this
circular window in the perisarc, the ectodermic tubercle passes
outwards, and the endoderm commences to grow into its lumen
and to line its walls.

The gonangium is now ready for the penetration of the
sexual cells. The entrance is partly passive, the result of dis-
placement and growth, but partly active, resulting from the
actual migration of the ova or spermatospores.

The time of the descent of the sexual cells is not the same in
both sexes. In the males this takes place when the endoderm
sends its prolongation into the ectodermic tubercle, the mass
of spermatospores then glides slowly upwards towards the
Opening in the perisare. In the females the corresponding
movement often only takes place somewhat later.

In certain gonangia the peripheral portion retires from the
perisarc, and the cellular mass of a gonangium separates as
a blastostyle, properly so called, upon which develope the true
gonophores. While the blastostyle is growing in length there
forms at the spot where the ovules are, a cul-de-sac, which
becomes a gonophore; the latter separates more or less from
the blastostyle, and, finally, remains attached by a short
pedicle only.

In the interior of the gonophore the endoderm growing
more rapidly than the ectoderm becomes plicated, and the
ovules come to lie in the niches so formed.

As soon as the eggs arrive at maturity fertilisation takes
place, and at the same time the endodermic tube slowly retires
from the gonophore at its extremity, and during the maturity
of the first gonophore a second is formed. Whence come the
ova in the second gonophore? Are they derived from the
ceenosarc or from the blastostyle? Weissmann supports the
latter hypothesis. A third gonophore may be formed, but such
rarely happens.

Weissmann describes similar processes in Plumularia
setacea and Sertularella polyzonias and S. gayi.

In Gonothyrea Loveni Weissmann states that the sexual
cells do not arise in the coenosarc, but in the gonophores.

ORIGIN OF THE SEXUAL CELLS IN HYDROIDS. 621

Weissmann also describes the process in EKudendrium
ramosum where the sexual cells are ccenosarcal in origin.

It is important to note that in no case have sexual cells been
seen to arise in the hydranth, but always in the trunk of the
colony.

De Varenne has come independently to somewhat similar
conclusions, but he goes further and describes a ccenosarcal
origin where Weissmann had not observed it.

He shows that in all the forms which he has studied Cam-
panularia flexuosa, Plumularia echinulata, Sertu-
laria pumila, Gonothyrea Loveni, Podocory ne carnea
and Obelia geniculata, the ova and spermatozoa alike
develope in the ccenosare of the trophosome, and, moreover,
originate from endodermal cells, whether they are matured in
fixed sporosacs, in medusoids which remain attached (Gono-
thyrza Loveni), or in medusoids which have a free existence ;
whether they remain in obvious connection with the endoderm
or migrate so as apparently to lie in the ectoderm.

De Varenne believes that previous observers who have put
forward the ectoderm as the place of origin of the generative
products of either sex have been misled by the fact that in
many cases, although actually originating as stated above, such
cells migrate even after they have reached the gonangium. The
endoderm reforms itself beneath the migrated ova or spermato-
zoa, a new homogeneous membrane (Stiitzlamella) is secreted
by the newly-formed endodermal cells, which might easily be
mistaken for the original structureless lamella, the latter as well
as the ectodermic layer having become reduced on account of
the pressure exercised by the developing egg or spermatozoa
to an extremely thin layer, which, however, remains outside
the sexual products.

It has been usual to consider the gonophores, whether these
remain fixed (sporosacs) or become free-swimming medusoids,
as the sexual persons, the trophosome polyps being asexual
persons, but if the observations above recorded are true and
should be proved to have the universal application that
De Varenne seems to consider they have, the view now usually

622 A. G. BOURNE.

held that there is a true alternation of generations among the
gymnoblastic and calyptoblastic hydroids must be abandoned.
Asexual buds and sexual generative products both arise in the
trophosome, the latter, however, become collected into specialised
buds which may, to the obvious advantage of the species,
become detached and actively locomotor, and even acquire a
higher organisation than the fixed trophosome.

OSTROLOGY ETC., OF SYNGNATHUS PECKIANUS. 623

On the Osteology and Development of Syng-
nathus Peckianus (Storer).

By

J. Piayfair McMurrich, ™.A.,
Professor in the Ontario Agricultural College, Guelph, Canada,

With Plates XLII and XLIII.

Ir is now some time since I began the study of Syngna-
thus, recognising the fact that little, if anything, had been
done towards elucidating many points in the anatomy and de-
velopment of the grotesque group of the Lophobranchs. Other
matters have prevented as rapid work, and as close application
to the subject in hand, as would have been wished for, and
even now the work is exceedingly incomplete. Nevertheless
it has seemed to me to be wise to publish the observations as
far as completed, inasmuch as I have been able to throw some
light on certain points, and to correct some mistakes made by
authors who have preceded me.

The material for study was obtained at Beaufort, N.C.,
where it was quite abundant among the seaweed near the
shore. The presence of the brood-pouch on the under surface
of the postabdominal region of the male renders the collection
of young stages comparatively easy, but the intermediate stages
between the newly hatched young and the adult were less
easily obtainable, and I have been unable to bridge over this
gap in my observations. The material under observation may
be divided into five stages, the relative characters of which may
be described as follows:

624 J. PLAYFAIR MCMURRICH.

A. Length 3—4 mm.: cartilages not quite fully formed.

B. Length 6—7 mm.; cartilages fully formed ; considerable
amount of yolk still present.

c. Length 8—9 mm.; yolk-sac very much diminished in
size.

p. Length 10—11 mm.; yolk completely absorbed; ready to
leave the brood-pouch.

E. Adult.

I. DEVELOPMENT AND STRUCTURE OF THE CRANIUM.

In Stage a (Pl. XLII, fig. 1) the head is round and small,
with the lower jaw bent up, and closely applied to the under
and anterior surface of the cranium, there being as yet no dif-
ferentiation of the upper jaw and snout. Below the medulla
oblongata (M. 0.) the notochord (N.C.) bends abruptly
downwards, and becomes ensheathed on either side by the
parachordal cartilages, which are continued forward to unite
with the trabecule cranii (Tr. Cr.), the extremity of which
is somewhat bent upwards, and as yet, as it were, within the
cranium, reaching no farther forwards than a point between
the eyes. The auditory capsules (Au.) are formed in carti-
lage, and are apparently united to the parachordals, but possess
no rudiment of the semicircular canals, though two otoliths are
plainly visible.

In Stage B (fig. 2) the snout makes its appearance, but it is
as yet bent upwards, and lies closely applied to the front of the
cranium, owing to the upward flexion of the trabecule men-
tioned above, which has increased considerably. At their co-
alesced extremities, and articulating with their sides, is on
either side a cartilage (EK. Pa.), whose signification will be
discussed when treating of the visceral arches. When the
skull in this stage is viewed from above (fig. 4), the nares (Ol.)
may be seen lying on the upper surface of the snout, slightly
anterior to the eyes. They have a thickened margin, and be-
hind them, extending down from the centre towards the sides
of the snout or rostrum, are two cartilaginous bars (Na. C.),

OSTEOLOGY ETC., OF SYNGNATHUS PECKIANUS. 625

while above these is a cartilaginous plate (Tg. Cr.) extending
upwards to between the eyes, and produced to a point between
the nares, forming the tegmen cranii. This and the two
nasal rods have apparently no connection at present with the
coalesced trabecule, but are apparently independent cartilages :
the tissue lying between the nares and the extremity of the
trabecule is somewhat fibrous. The auditory region is now
more fully developed, and the semicircular canals are repre-
sented by fibrous bands. The parachordals posteriorly have
extended upwards and around the nervous cord to form the
occipital region of the skull.

Stages c and Dp are very slightly different. ‘They present an
increase in size from Stage B, and the fibrous semicircular
canals have become cartilaginous. The snout has increased
somewhat in length in c, and still more in Stage p, the growth
taking place at first by an increase in length of the horizontal
portion of the cartilaginous “rostral plate,” as the elongated coal-
esced trabecule may be called, and, latterly, by a straightening
out of its bent up end, which even in Stage c (fig. 5) formsa
right angle with the horizontal portion. The posterior region
of the skull is almost completely encased in cartilage, its very
summit only being of dense membrane ; more anteriorly the
membrane becomes thinner, and extends further down the
sides of the skull.

At the base of the skull, and extending from its posterior
portion forward for a considerable distance in the median line,
is a dense membrane (fig. 15, Pa. S.), identical in appearance
with that roofing in the cranial cavity, and also with the mem-
brane surrounding the notochord at this stage: this is the
commencement of the formation of a bone, which from its
position must be the parasphenoid. At the sides of the
vertex of the anterior portion of the cranium proper, and ex-
tending back nearly to its posterior portion, are two similar
condensations of tissue (Fr.8.). They consist, as is seen on
section, of plates lying in the connective tissue enclosing the
cranial cavity, one on either side, and at the centre of each
plate and perpendicular to it a ridge passes along its entire

626 J. PLAYFAIR MCOMURRICH.

length, projecting out into the integument. Still more poste-
riorly only one such plate of membrane is seen, which occupies
a median position at the vertex. These appear to be the
membranous rudiments of the frontals and dermo-supra-
occipital.

The Adult.—In the adult cranium the cartilage persists to a
great extent, although surrounded almost completely by bone.
No enchondroses, as in the higher vertebrates, appear ; ectostoses
aud parostoses form the cranium. One point noticeable at the
first glance is the elongation of the occipital regions, and the
compactness of the region immediately behind the orbits;
separating these two portions there is a membranous space,
closed in by the parasphenoid.

The occipital region ossifies below as the basi-occiptal (fig.
9, B.O.). Posteriorly this is round, forming an articular facet
for the first vertebra, but anteriorly it becomes flattened out,
and expanded into a thin plate—being, in fact, fan-shaped.
Posteriorly, on section, the rapidly diminishing notochord is
plainly visible, and on either side of this are the parachordal
cartilages. More anteriorly the chorda does not appear, and
the parachordals unite to form a single plate. Below it is
deeply grooved for the reception of the parasphenoid, which
underlies it, and is almost enclosed by it. Laterally it articu-
lates with the exoccipitals, and more anteriorly with the
pterotics.

The exoccipitals (Ex. O.) are well-developed bones, forming
the postero-lateral floor of the skull, and extending somewhat
upwards upon its sides to articulate with the epiotics, and in
front with the pterotics. Above them on either side is a
parosteal bone (fig. 8, S.Tp.), upon whose homologies I am
undecided. I am inclined, however, to consider it homologous,
to a certain extent at any rate, with the supra-temporal of
Amia.

Above, at the vertex of the skull, is the supra-occipital
(S. O.), a large bone, extending forward nearly to the sphenotics.
It appears to consist of two portions; (1) a parostosis, which
may be termed the dermo-supra-occipital, and which de-

OSTEOLOGY, ETC., OF SYNGNATHUS PECKIANUS. 627

velops apparently in a manner similar to what occurs in the
placoid scales of the body, from the membranous plate already
spoken of ; and (2) an ectosteal portion, which may be termed
the auto-supra-occipital. The supra-occipital is somewhat
narrowed in front, projecting between the posterior portions of
the frontals.

Laterally from this bone we have on either side the parie-
tals (Pa.), small bones which do not meet in a sagittal suture,
but are separated throughout their entire length by the supra-
occipitals, as in the pike and the salmon. Anteriorly they
articulate with the frontals. The name dermo-epiotics,
which Bridge has suggested for these bones in Amia,! is here
very applicable, for they directly overlie the epiotics, appearing
like a parostosis formed upon these bones.

The pterotics (Pt.O.) extend far forward to a point just
behind the sphenotics, and form, along with the pro-otics and
sphenotics, the articulating surface for the hyomandibular.
They descend pretty well to the base of the skull, their poste-
rior parts overlapping the anterior lateral portion of the
basi-occipital, and articulating with the exoccipitals. They
form the lateral boundaries of the membranous space in front
of that bone. Anteriorly they articulate with the pro-otics.

These bones (fig. 9, Pr. O.) bound the membranous space in
front, and extend forward nearly to the anterior limit of the
postorbital region of the cranium. They extend only slightly
upwards on the sides of the skull to articulate with the pterotics
and sphenotics, and unite across the middle line of the skull
anteriorly, but are separated posteriorly, leaving a space closed
in only by the parasphenoid. At the sides in this region are
the sphenotics (Sp. O.). There seem to be no distinct ali-
sphenoids, the part of the skull having the usual relations of
these bones being ossified by the pro-otics. There is apparently
also no basisphenoid.

The frontals (fig. 8, Fr.) are membrane bones of compara-
tively large size, extending from the parietals posteriorly to

1 T. W. Bridge, “ The Cranial Osteology of Amia calva,” ‘Journ. of Anat!
and Phys.,’ vol. xi, 1877.

628 J. PLAYFAIR MCMURRICH.

slightly in front of the ectethmoids anteriorly. They are some-
what club-shaped, broadening out posteriorly. On viewing
them from the surface they appear to be unsymmetrical, one
forming a projection which fits into a corresponding indentation
on the other. This want of symmetry is apparent rather than
real, for a section (fig. 11, Fr.) shows it to be caused by an
overlapping, the portion of bone overlapped being equal to that
which overlaps it. Opposite the sphenotics, from the inner
surface of each frontal, a process passes down, which articulates
with the front edge of the ascending or pro-otic process of ther
parasphenoid.

The parasphenoid (fig. 9, Pas.) is a long parostosis, ex-
tending from the basi-occipital nearly to the anterior extremity
of the cranium on its under surface. Behind it is round, and
almost enclosed in the basi-occipital, lying in a deep groove in
that bone. More anteriorly it widens out to form a partial
floor for the membranous space in front of the basi-occipital, the
cartilaginous trabecule lying immediately above it. At the
anterior limit of the postorbital region of the skull (fig. 10) it
sends up on either side a process, somewhat triangular in shape,
when viewed laterally. These articulate with the pro-otics, and
along their anterior edges with the descending processes of the
frontals. These processes may be termed the ascending or
pro-otic processes of the parasphenoid. The space bounded
laterally by these two bars, above by the frontals, and below
by the parasphenoid, serves for the passage of the orbital
muscles and nerves. Anterior to this, in the orbital region, the
parasphenoid becomes rectangular, and finally triangular (fig.
11, Pa. S.); the apex being directed upwards, having attached
to it the lower edge of the interorbital membrane. In front of
the orbit it articulates on either side with an ectethmoid, and
still more anteriorly (fig. 12) becomes convex, being deeply
grooved on the under surface, in which groove lies the vomer
(Vo.), almost enclosed, and presenting a similar appearance to
the parasphenoid when lying in the groove in the basi-occipital.

There is no orbitosphenoid, the passage for the orbital
muscles and nerves being very large, appearing, in fact, almost

OSTEOLOGY, ETC., OF SYNGNATHUS PECKIANUS. 629

as if the whole anterior wall of the cranial cavity were wanting.
The interorbital septum (fig. 11, J. S.) is merely membranous,
as in the Siluroids and Cyprinoids.

Immediately in front of the orbits, on either side, is an
ectethmoid (fig. 8, Ect. E.), the ossification of the preorbital
process of the ethmoidal cartilage. They extend down the sides
of that cartilage, and articulate below with the parasphenoid,
and above with the frontals.

At this point or slightly anterior the frontals terminate, the
vemainder of the rostrum being formed by the ethmoid, with
the vomer lying along its under surface. There appear to be
no membrane bones occupying the position of the nasals of most
fishes.

The ethmoid (fig. 12, Eth.) consists posteriorly largely of
cartilage surrounded by a certain amount of ectosteal bone.
In the cartilage, on either side, is a canal, in which the olfac-
tory nerves and vessels run, passing to the olfactory capsules,
which form deep indentations in the sides of the cartilage.
Immediately in front of the orbits the cartilage appears to con-
sist of two portions, an upper, originally the tegmen cranil, and
a lower, the coalesced trabeculae, and between these the olfac-
tory nerves and vessels at first run. ‘This distinctness of the
two parts obtains, however, only for a short distance, the two
halves soon uniting and becoming indistinguishable, the olfac-
tory structures becoming enclosed in a canal. In front of the
olfactory organs the cartilaginous part of the ethmoid rapidly
diminishes, there being a nearly corresponding increase in bone.
In this a canal appears, which passes towards the surface as
one traces it forwards. I take this to be the continuation of
the main slime canals, which also traverses the frontals. Ante-
riorly the ethmoid becomes thinner, but remains nearly of the
same breadth, and its cartilaginous portion entirely disappears ;
still more anteriorly (fig. 13) it becomes almost scale-like, the
vomer (Vo.), which hitherto has been round and small, now
becoming larger and triangular, and forming the greater part
of the thickness of the rostrum in this region, and this relation
persists to its extremity.

VOL, XXITI—NEW SER, TT

630 J. PLAYFAIR MCMURRICH.

Summary.—The most noticeable feature in the configura,
tion of the embryonic cranium is the bending up of the facial
portion against the front of the skull, this being due to the
bending up of the coalesced trabecule cranii at their extremity.
As development proceeds this bending up does not diminish-
but the mouth becomes carried forward by the growth of the
horizontal portion of the trabecule, and it is not until Stage p
is arrived at that the elongation of the snout is dependent upon
the straightening out of the rostral cartilage.

In the adult the first thing to be noticed is the forward
extension of the occipital region and the compactness of the pro-
otic region. The absence of any cartilaginous sphenoid bone,
the wide opening for the passage of the orbital muscles and
nerves consequent upon this, the absence of an osseous inter-
orbital septum as in the Siluroids and Cyprinoids, the want of
nasals, and the structure of the ethmoids, are also points of
considerable importance.

II. THE ViscERAL SKELETON.

In the youngest stage observed (A) most of the visceral
arches were apparently fully formed, and consisted of seven
cartilaginous bars, some, however, being more differentiated
than others, the first three already showing a specialisation
into their future parts.

The first postoral or mandibular arch consists, on either
side, of a well-developed mandibular portion (fig. 1, Mck.),
bounding the gape below, extending forwards and upwards,
and curving slightly inwards towards its fellow of the opposite
‘side. Articulating with the proximal extremity of each of
these Meckelian cartilages is a single rod-shaped portion
(Pt. Qu.), extending backwards and slightly upwards, but
lying quite free in the tissue of the preevertebral portion of the
skull, except for the articulation with Meckel’s cartilage.
Subsequent development shows this to be the rudiment of the
pterygo-quadrate portion of the jaw, and hence it may be
denominated the pterygo-quadrate cartilage.

The second postoral or hyomandibular arch is represented by

OSTEOLOGY, ETC., OF SYNGNATHUS PECKIANUS. 631

a curved rod (H. M.), extending from the skull downwards and
forwards towards the proximal extremity of Meckel’s cartilage»
with which it appears to articulate. Anteriorly it lies below,
and almost parallel to, the pterygo-quadrate cartilage, and
becomes somewhat broader posteriorly. It presents as yet no
differentiation into the hyomandibular and symplectic portions,
to which this segment of the arch corresponds, nor in fact does
it so separate afterwards, the line of division between the two
being indicated merely by the articulation of the second seg-
ment of the arch, i.e. hyoid portion. ‘This consists at present
of a rod lying below, and inclined towards, the upper cartilage,
which it meets a little in front of the dilated portion.

The remaining four arches constitute the branchial arches
(Brs.), present at this early stage to the same number as in
the adult. No azygos, median hyoid cartilage could be de-
tected at this stage, nor is there a median branchial cartilage.

In Stage B (figs. 2 and 3) the mandible has become bent up
upon the skull, and the pterygo-quadrate cartilage retains the
same characters as in the preceding stage. The hyomandibular
has, however, undergone some modifications. It has become
distinctly angulated, the symplectic portion (Sym.) extending
forward parallel with the axis of the skull, while the hyoman-
dibular (H. M.) moiety is bent up almost at right angles to it,
slightly posterior to the eye, and articulates with the lateral
under-surface of the auditory capsule. Near the anterior
border of this portion is an oval foramen, and at the angle of
junction with the symplectic a separate cartilaginous centre
apnears, uniting the hyoid to the hyomandibular portion of the
arch, and representing the interhyal. The hyoid arch (Hy.)
extends from this downwards towards the median line, and
represents the combined ceratohyalsandhypohyals. About
half way from the extremity of each portion of it is a process
extending upwards towards the symplectic, which is notched
to receive it. A median hyoid rod (G. H.) is now easily seen,
extending from in front of the hyoid cartilages back to the
branchial region. It is not quite straight, but curves slightly
to the side, and at the junction of the hyoid and hyomandibular

632 J. PLAYFAIR MCMURRICH.

turns abruptly upwards towards the base of the skull, parallel
to the first branchial arch. This is probably the genio- or
basihyal.

The branchial arches (Brs.) remain the same practically as
in the last stage, and throughout show very little differentiation.
Another cartilage, however, makes its appearance at this stage,
which, though not at present connected with the visceral
cartilages, eventually unites with them. It consists (fig. 2,
E. Pa.), on either side, of a cartilaginous rod, articulating with
the sides of the rostral cartilage, near its anterior extremity,
and extends backwards and slightly downwards. It is the
commencement of the ethmopalatine cartilage.

The principal change to be noted in Stage c (fig. 5) is the
growth of the pterygo-quadrate cartilage. It has now grown
upwards and expanded at the extremity with an anterior and
posterior process. The anterior is clearly connected with the
ethmo-palatine by a band of connective tissue, and represents
the pterygoid portion, while the posterior one has no connec~
tions, but probably is the future metapterygoid. Another
point is clearly noticeable at this stage, which helps in no
small degree to indicate the homologies of the cartilages. The
symplectic does not meet the basal portion of Meckel’s carti-
lage; the only cartilage articulating with this being the
quadrate portion of the pterygo-quadrate. In previous stages
this meeting and articulation seems to exist, or, at all events,
the two cartilages are almost in contact, but now their want of
union can be clearly seen.

The growth of the pterygo-yuadrate constitutes again the
most noticeable feature when we examine Stage p. In this it
is seen that the pterygoid process has grown so far forward,
and the ethmopalatine so far backward, that they are now
separated only by a very small portion of connective tissue.
The growth has been mainly, however, on the part of the
pterygoid process. The metapterygoid process has grown back-
wards only a short distance, and at this stage does not form a
buttress to the hyomandibular, as in the salmon. In the pre-
ceding stage, but more clearly to be seen in this, one can

OSTEOLOGY ETC., OF SYNGNATHUS PECKIANUS. 633

notice on section a dense, somewhat irregularly-shaped mem-
brane (fig. 15, Inf. O.8.) lying on the outer surface of the sym-
plectic. It appears to correspond to a membrane bone in the
adult, which I have named the infra-orbital.

It will be well to compare the cartilages of the young
Syngnathus with those of a typical Teleost, and for this
purpose no better choice can be made than the salmon, which
has been so admirably worked out by Prof. Parker.!

On comparing Parker’s fifth stage with my Stage p, the re-
semblance will at once be seen. The hyomandibular and sym-
plectic elements are not separated. The former is broad and
stout, tapering towards the point of articulation with the
interhyal, and a little below this there is a slight bend. In
Syngnathus the bend is greater and a little further back,
being exactly at the point when the interhyal articulates, i. e.,
at the point where the two elements meet ; and with this greater
angulation there is a consequent elongation of the symplectic
element. The quadrate element in the Salmon presents the
same relations as in Syngnathus, but its metapterygoid
portion is much larger, and lies upon the symplectic, forming
a buttress to it, while the pterygoid process is shorter.

The relations, however, of the pterygoid and _ palatine
portions will be more readily recognized on examining an
earlier stage of the Salmon. It will then be seen that the
palatine portion is originally distinct from the pterygo-quadrate,
consisting of a rod bounding the gape above, and extending
from the trabeculz almost to the angle of the mouth. Anteriorly
it is large and stout, tapering gradually posteriorly, being in
fact club-shaped. This is denominated by Parker the second
visceral arch, and evidently is the same as the cartilage I have
described as the ethmopalatine in Syngnathus.

In Clarias capensis the palatine or ethmopalatine rod
is longer than in the Salmon, and overlaps the pterygoid,?

1 W. K. Parker, “On the Structure and Development of the Skull in the
Salmon,” ‘ Phil. Trans.,’ 1873.

* W. K. Parker, “On the Structure of the Skull in Sharks and Skates,”
‘Trans. Zool. Soc.,’ 1878.

654. J. PLAYFAIR MCMURRICH.

while in a young Eel (Anguillula acutirostris) it is entirely
wanting.| We can thus from these four types construct a
scale, passing upward from Anguillula with no ethmo-
palatine to Syngnathus with a comparatively small one, then
to Salmo, in which it is more developed, and finally to
Clarias, in which it extends back to overlap the pterygoid.
The splint bone, which corresponds to and ensheaths this car-
tilage, is the maxillary ; and I regret that I have been unable
to collect data from which conclusions as to the relative extent
of the gape in these forms might be drawn.

Interesting comparisons can also be made between Stage Din
Syngnathus and the cartilaginous skull and arches of
Acipenser. In this the relative angulation of the hyoman-
dibular is the same, and at the same point asin Syngnathus;
the symplectic runs horizontally forward, but is not quite so
long, and a yery strong resemblance obtains between the
pterygo-quadrate in the two forms. In Acipenser, as in
Syngnathus, this is represented by a hammer-shaped carti-
lage, the basal portion or handle corresponding to the quadrate,
the anterior process of the head to the pterygoid, and the pos-
terior to the metapterygoid, whig¢h does not pass back to form
a buttress to the hyomandibular, The ethmopalatine seems
to be wanting.”

The phylogenetic significance of the teleostean ethmo-
palatine is apparently doubtful. Parker and Bridge con-
sidered it a structure with no representative in the Selachian
jaw, while Balfour’ points out the possibility of its being “an
element, primitively belonging to the upper arcade of the
mandibular arch, which has become secondarily independent
in its development.” This suggestion I do not think tenable,
and would prefer to side with the later view of Parker and of
Marshall, that it represents a preoral visceral arch, to which
the lachrymal cleft and the third nerve correspond.

1 Tbid.

2 This description has been taken from fig. 241, in Gegenbaur’s ‘ Elements
of Comparative Anatomy,’ London, 1878.

3 ¥. M. Balfour, ‘ Comparative Embryology,’ vol. ii, p. 478, London, 1881,

OSTEOLOGY, ETC., OF SYNGNATHUS PECKIANUS. 635

The only literature referring in any important way to the
development of the Lophobranchii which I have been able
to find is confined to three papers; of these, Calberla’s! treats
of a much earlier stage than I have been able to study, and
de Quatrefages” is of little or no importance, leaving only J. A.
Ryder’s* from which I could obtain any information Dr.
Ryder’s observations were unfortunately limited to a single
specimen of Hippocampus antiquorum, which had just
left the brood-pouch, and corresponds almost to my Stage pv,
being very slightly older. From the rather peculiar arrange-
ment of the mandibular skeleton at this stage no little difficulty
would no doubt be experienced in determining the homologies
of the cartilages from a single specimen, since it is only by
tracing their development that one can be certain of the signi-
fication of abnormalities. Accordingly, in Ryder’s paper, there
are certain statements which I am convinced are errors, partly
of observation and partly of mistaken homology.

These remarks apply especially to his description of the
mandibular and hyoid arches. On comparing Dr. Ryder’s
figure with my drawing of the cartilages in Stage c (fig. 5) it
will be seen that what I have represented merely as a foramen
in the hyomandibular portion of the second postoral arch is in
his figure pourtrayed as a dividing line, completely separating
the hyomandibular cartilage into two parallel portions, the an-
terior of which he terms the metapterygoid. He was dealing
with Hippocampus antiquorum, while my observations
were made on Syngnathus; but these two forms are very
closely allied, and the arrangement of the cartilages are appa-
rently identical, so that a statement made regarding one is no
doubt tenable for the other. Accordingly, since the develop-

1 B. Calberla, “Zur Entw. des Medullarrohres u, d. Chorda dorsalis der
Teleostier u. d. Petromyzonten,” ‘ Morph. Jahrb.,’ ili, 1879.
2 A. de Quatrefages, “‘ Memoire sur les Syngnathus,” ‘Ann. des Sci. Nat..’

1842.
3 J. A. Ryder, “ A Contribution to the Development and Morphology of
the Lophobranchiates (Hippocampus antiquorum, the Sea Horse),”

‘Bull. U.S. Fish Comm.,’ 1581.

636 J. PLAYFAIR MCMURRICH.

ment shows the metapterygoid at this stage to be an entirely
different cartilage, and since, by repeated cbservations, I haye
satisfied myself of the absence at this point of a chondrification
distinct from the hyomandibular, and of the existence of an
elliptical foramen, which, being covered by a band of muscular
tissue, might easily be mistaken for a line of separation, I take
exception to Ryder’s identification.

Having thus started on the wrong path, the homologies con-
tinue to be erroneous. Thus the distal horizontal portion of
the hyomandibular arch is termed the quadrate, whereas, from
its relations and from the evident absence of articulation between
it and Meckel’s cartilage, it must be the symplectic.

But it is in the homologies of the two upper cartilages of the
arch that Dr. Ryder errs chiefly. He says: ‘‘ Above the articu-
lation of the quadrate (i. e. the symplectic of my figures) with
Meckel’s cartilage a curious bent element (x) appears to repre-
sent the superior maxillary. Just in front of the expanded
upper extremity of the maxillary lies the posterior extremity
of the upper labial or intermaxillary element (Ja), which is
continuous with a similar piece on the opposite side; this in-
termaxillary bar curves over the anterior upward bend of the
rostral cartilage (7. ¢.). It contributes the skeletal boundary
of the upper part of the oral opening (m’), and is not seg-
mented in the median line, so as to articulate with its fellow
of the opposite side, like Meckel’s cartilage of the lower jaw.”
Now, in the first place, I differ from him in regard to his
assertion that the intermaxillary bar, as he calls it, curves
over the rostral cartilage. In Syngnathus these bars, as seen
from a surface view (fig. 4), certainly articulate with the sides
of the rostral cartilage, though when viewed from the side the
turned-up extremity of the latter gives at first sight an im-
pression of their continuity across the front of the skull. In
the second place I differ as to the identification of the car-
tilages. The terms ‘ maxillary” and “ intermaxillary” are
misnomers, the bones so denominated in the adult being,
without exception, membrane bones ; and further, the develop-
ment of the lower cartilage, and the fact that it is the only

OSTEOLOGY, ETC., OF SYNGNATHUS PECKIANUS. 657

cartilage which articulates with the mandibular, shows con-
clusively that it can be no other than the pterygo-quadrate.

Other discrepancies between Ryder’s paper and my own,
such as the transverse segmentation of the hyomandibular
arch, the existence of a distinct symplectic (which probably is
my geniohyal), and the division of the hyoid element into
cerato- and hypohyal portions, will be noticed at once on com-
parison, and certain of them may be due to the difference in
the age of the forms compared.

Adult.—I must here repeat the statement made previously
that intermediate stages between p and the adult were not
obtained. As a consequence some of the points in the follow-
ing description are merely conjectures. There is a certain
amount of complication, owing to the excessive elongation of
the symplectic, and the presence of membrane bones, which
study of the intermediate stages can alone satisfactorily
unravel.

The hyomandibular articulates with the skull immedi-
ately behind the orbit, the articular surface being afforded it
apparently by the sphenotic, pro-otic, and pterotic. Its direc-
tion is perpendicular to the longitudinal axis of the skull, and
it is composed of a central cartilage surrounded by ectosteal
bone (fig. 10, H. M.).

Below it articulates with the interhyal (I. H.), a some-
what triangular ossicle, in which the original cartilage still
persists. This forms the connecting link between the hyo-
mandibular and hyoid element proper. The latter (Hy.)
consists apparently of two portions, a cerato- and hypohyal,
the epihyal being apparently absent. The elements of either
side approach each other in the median line below, and are
moveable upon the interhyal. Two muscles, inserted by a
single tendon, pass from their distal extremities backwards,
and anteriorly they are connected with the mandible. When
the muscle is in a state of repose the hyoids are bent up on
the under surface of the skull, and lie between the symplec-
tics ; by its contraction their extremities are drawn downwards,
and consequent upon this there is a similar downward move-

638 J. PLAYFAIR MCMURRICH.

ment of the mandible, and at the same time an enlargement of,
the buccal cavity, whereby water for respiration is drawn in.
These bones form consequently an important portion of the
respiratory pump, acting as it were the parts ‘of handles,
whereby the force of the muscles is transmitted to the man-
dible and to the buccal cavity. The geniohyoid element
does not appear to ossify and in fact has disappeared.

Extending forwards horizontally from the hyomandibular
almost to the extremity of the elongated snout, is the sym-
plectic, the cartilages of the two portions of the arch being
continuous. Posteriorly (fig. 11, Sym.) the bone is oval in
shape, and consists mainly of the original cartilaginous bar,
the osseous portion being small. In the anterior portion of
the orbital region, however, the latter becomes greater in pro-
portion to the cartilage, and the bone assumes a hammer
shaped appearance on section, the two heads of the hammer
articulating with a membrane bone afterwards to be described,
and enclosing with it a space. The handle is directed up-
wards towards the orbit. In the nasal region (fig. 12, Sym.)
the handle has extended upwards to articulate with a mem-
brane bone bounding the nasal cavity below, and the hammer
shape has entirely disappeared. Anteriorly (fig. 15, Sym.) the
symplectic becomes completely surrounded by the quadrate
(Qu.), and its cartilage becomes entirely absorbed, its position
being indicated by a round foramen in the section.

The quadrate (Qu.) extends backwards to the region of the
nasal capsules. It then consists (fig. 12) of a small oval portion
lying below the symplectic. Anteriorly (fig. 13), however, it
enlarges and grows upwards on the outer side of that bone, and
finally completely encloses it as mentioned above. The cartilage
is entirely absent from the bone posteriorly, and it is only in
its most anterior part that it is present.

Lying outside the symplectic, and slightly bending round so
as to bound it below, there is to be seen in the orbital region a
membrane bone (Inf. Or.). It bounds the orbit below, and -
corresponds to a certain extent with a bone, or a series of bones,
in other Teleosts, which usually receive the name of infra-

OSTEOLOGY, ETC., OF SYNGNATHUS PECKIANUS. 639

orbitals. To this bone I propose to give the same name,
although it is much larger, and extends farther forward, leaving
the orbital region altogether, and thus loses to a certain extent
the claim to the name. It is largest in the orbital region (fig.
11), and then constitutes the outer bone of the snout; more
anteriorly (fig. 12), however, it becomes, as it were, thrown
down towards the under surface of the snout, and gradually
becomes smaller, terminating slightly in front of the posterior
extremity of the quadrate.

Appearing in the series of transverse sections! in the nasal
region, about the same time as the quadrate, is a bone (fig. 12,
Mpt.), which [ am inclined to refer to the pterygoid series,
probably the metapterygoid. Posteriorly, as in the bones
above described, no trace of cartilage is present, but near its
anterior extremity it appears, and is there seen to be continuous
with that of the quadrate, so that the ectosteal bone would
correspond to the posterior process of the pterygo-quadrate
cartilage of the young stages. It passes forward as far as the
mouth, becoming gradually smaller anteriorly, just as the
quadrate becomes broader. Posteriorly it articulates with the
upper portion of the symplectic in its broad region. It has the
appearance of a scale, separated from the symplectic by a
quantity of muscular tissue. At first sight one would not be
inclined to identify this bone as the metapterygoid, on account of
its great separation from the hyomandibular, since it is generally
described in connection with that bone to which it usually
forms a buttress. From its origin, however, one would be rather
inclined to imagine it having a much closer relation with the
quadrate than the hyomandibular, which belongs to the suc-
ceeding arch, and this we find to be the case here. The great
elongation of the symplectic has carried the quadrate far forward,
and the metapterygoid has accompanied that bone, and iost all
connection with the hyomandibular.

Overlapping the lower portion of the under surface of the

! This series, consisting of 165 sections, was prepared by Giesbrecht’s

shellac method, the object being first decalcified by a 3 per-cent. solution of
HCl in 96 per cent. alcohol, and then stained in toto with Bismarck brown.

640 J. PLAYFAIR MCMURRICH.

infra-orbital posteriorly, and the quadrate for some distance
anteriorly, is, on either side, a scale-like bone (figs. 12 and 13,
M.), which has no special representative in other Teleosts, and
is merely a membrane bone formed in the dense integument
closing in the buccal cavity below.

The anterior process of the pterygo-quadrate cartilage ossifies
as the pterygoid, its cartilage apparently passing forward to
become continuous with the ethmo-palatine. Anterior to the
pterygoid bone, bounding the mouth above, is the palatine,
an ectosteal bone developed upon the ethmo-palatine cartilage,
and on the outer side of this is a splint-bone, which bounds
the gape above, and which, from its relation to the ethmo-
palatine cartilage, must be the maxillary. The premaxille
appear to be wanting, the maxille bounding the gape.

My sections of the mandibular region not being very good, I
am unable to make any statements concerning the ectosteal
bones of the mandible. The opercular bones are not present
in as great numbers as in typical Teleosts. There is a very
large operculum, somewhat scale-like and convex outwardly,
which articulates with the hyomandibular. The przoper-
culum, a very constant bone in the Teleostei, here appears to
be absent, or at any rate very rudimentary. A membranous
suboperculum bounds the operculum below, and is continued
up behind it as far as the spiracular branchial cleft; there is
no interoperculum. The outer surface of all bones upon
the surface is beautifully sculptured, some of the thinner ones
presenting an elegant fenestrated appearance.

Summary.—In the first place, the great elongation of the
symplectic is very noticeable, and as a consequence there is a
wide separation of the hyomandibular and metapterygoid ele-
ments. The ethmo-palatine articulates with the sides of the
anterior extremity of the ethmoid or rostral cartilage, and
grows backwards to unite with the pterygoid process of the
pterygo-quadrate. In the adult the elongation of the posterior
mandibular region, and the concentration of the anterior por-
tion, are well marked. The great elongation backwards of
the metapterygoid and quadrate bones, the absence of any

OSTEOLOGY, ETC., OF SYNGNATHUS PECKIANUS. 641

apposition of the former to the hyomandibular, the abortion of
the geniohyoid, the absence of the intermaxillaries and the
preopercula, are the most important points.

During the past few years, chiefly through the investigations
of Parker, Balfour, Gegenbaur, and Marshall, much light has
been thrown upon the segmentation of the skull and the rela-
tions of the visceral arches. There has been an increasing
tendency to refer the various subcranial arcades to the category
of visceral arches, and to increase the number of segments
which have become coalesced to form the cranium.

It is now well recognised that all the visceral arches behind
the mouth, including the mandibular and hyoid arches as such,
are merely the modification of a series of cartilaginous bars situ-
ated in the walls of the pharynx, and originally supporting
branchiz. The posterior visceral arches, to the number of five,
increasing to six or seven in the Notidani, and diminishing to
fourin Syngnathus, still retain this original function, but it
is not so with the two anterior postoral arches ; these have be-
come modified to subserve other purposes, and have lost toa
large extent their original structure and appearance. The distri-
bution of cranial nerves to these, however, resembles that of the
branchial arches, the trigeminal being referable to the mandibu-
lar, and the facial and the*auditory (which from embryological
facts may be considered as one) to the hyoid arcade and the adja-
cent structures. The visceral clefts point to the same conclu-
sion. Each visceral arch bounds posteriorly a visceral cleft ;
so for the hyoid arch we have the spiracular cleft, and for the
mandibular the mouth, since Dohrn’s late researches! show
that this is formed by the coalescence of two hypoblastic out-
growths, the median opening only forming later. From the rela-
tions of the head cavities one would deduce the same conclusions.
Balfour, who has worked them out very thoroughly in the
Elasmobranchs, thus speaks of them :*—“ As the rudiments of

1 A. Dohrn, “ Studien zur Urgeschichte des Wirbelthierkorpers,” ‘ Mith, a

d. Zool. Station zu Neaple,’ Bd. iii, 1882.
2 F. M. Balfour, loc. cit., p. 558. See also ‘Monograph of Elasmobranch.

Fishes,

642 J. PLAYFAIR MCMURRICH.

the successive visceral clefts are formed, the posterior part of
the head-cavity becomes divided into successive sections, there
being one section for each arch. Thus, the whole head-cavity
becomes on each side divided into—(1) a premandibular sec-
tion; (2) a mandibular section; (3) a hyoid section; (4)
sections in each of the branchial arches.”

The question arises—How many are the preoral visceral
arches, and what are their relations to the cranial nerves, vis-
ceral clefts, and head-cavities ?

If we consider Amphioxus as an ancestral type for the
vertebrates, we find that originally there were no preoral vis-
ceral arches, and the Ascidians point to the same fact. Ac-
cordingly one must suppose that the anterior cleft, which,
functioned as the mouth in the ancestral forms, gradually lost
that use, its function being as gradually assumed by succeeding
arches, since we have strong grounds for supposing that there
are in the Craniota arches in front of the mouth. As to the
number of these arches we have several theories—some main-
taining that there is only one, others deciding for two, and
others for several. As far as our present knowledge goes, I
consider that we are entitled to recognise two.

The authorities upon this subject are by no means few;
from them the views of Parker and Marshall may be cited as
illustrating what has been done in this line. Parker, in his
paper on the skull of the Salmon,! seems inclined to accept the
trabecule cranii as a preoral arch, but later rejects this theory.
He says :—“‘When we consider. . . . that the walls of the fore-
part of the cranium are formed by growth from the trabecule,
just as posteriorly the walls are formed by growth from the
parachordals ; that nerves are similarly emitted through the
trabecular and occipital walls ; when it is seen, in short, that
the trabecule are neural in their relations, as completely as,
and in similar fashion to, the parachordals, it seems impossible
to resist the conclusion that the trabecule and the para-

| W. K. Parker, “On the Structure and Development of the Skull in the
Salmon,” ‘ Phil. Trans.,’ 1873. .

OSTEOLOGY; ETC., OF SYNGNATHUS PECKIANUS. 643

chordals must be placed in one and the same category.”? Still
later, he speaks for the presence of three przoral arches, which
he names epipterygoid, ethmo-palatine, and prorhinal.”

Marshall’s view is that only two arches are represented in
the preoral region, namely, the ethmo-palatine, or lachrymal,
and the olfactory.’

Upon what grounds is it possible to base conclusions as to
this point? It seems to me that before we can come to any
definite conclusion as to whether a structure is really a visceral
arch, we must be able to show the presence or rudiments of an
arch, a cleft, a nerve supplying that arch, and a head-cavity.
Let us now apply these tests.

In front of the trigeminal nerve which supplies the man-
dibular arch, we have the olfactory, optic, oculomotor, pathe-
ticus ; and the abducens may also be enumerated here, as it has
not yet been accounted for. It has been shown by Marshall
and others that of these the optic, patheticus, and abducens
are probably not entitled to be ranked in the same category as
the other cranial nerves, leaving only the oculomotories (i11)
and olfactory (i) to be accounted for. The former is referable
to the region of the ethmo-palatine cartilage, and we have also
a cleft, the lachrymal, and a head-cavity, the premandibular,
the walls of which, according to Marshall,‘ become transformed
with the superior, inferior, and internal recti, and inferior
oblique muscles of the eye, to which the third cranial nerve acts
as a supply.

All the necessary parts are present, then, for forming a firm
basis on which the conclusion as to the validity of the ethmo-
palatine or lachrymal arch may be vested. The olfactory arch
is, however, much more complicated.

1 Parker and Bettany, ‘On the Morphology of the Skull,’ London, 1877.

2 W. K. Parker, ‘‘ On the Evolution of the Vertebrates,” Hunterian Lec-
tures, ‘ Nature,’ vol. xx, 1879.

3 A. Milnes Marshall, “The Morphology of the Vertebrate Olfactory Organ,”
© Quart. Journ. Micr. Sci.,’ vol. xix, 1879.

4 A, Milnes Marshall, “On the Head-cavities and associated Nerves of
Elasmobranchs,” ‘ Quart. Journ. Mic. Sci.,’ vol. xxi, 1881.

644 J. PLAYFAIR MCMURRICH.

Marshall’s observations are apparently conclusive as to the
segmental nature of the olfactory nerve, and the homology of
the Schneiderian membrane of the olfactory capsules with the
gills of the posterior arches. Dohrn’s! researches on the
pituitary body of fishes tends to refer this structure also to a
visceral cleft, and there is a union between it and the nasal
cavity in Petromyzon, which, however, Dohrn states is
merely secondary. We have here, then, a nerve, and probably
a cleft, but the remaining structures are apparently absent.
As to the head-cavities in Elasmobranchs, there is only one
premandibular; and in Teleosts Ganin states ? that this is the
only one present, evidently showing that there is a tendency
for these structures to disappear ; and we may conjecture that
there was originally a second premandibular cavity, which has
disappeared even in the Elasmobranchs.

The visceral arch corresponding to this segment, I am in-
clined to think, is represented by the trabecule cranii. The
lowest type of Vertebrate presents no prolongation of the
cerebral nervous system beyond the extremity of the notochord.
The portion in the higher vertebrates anterior to that point is
merely an overgrowth, and we must consider the pituitary
region as corresponding to the extremity of the under-surface
of the brain in Amphioxus. Accordingly we have a pre-
veriebral portion of the cranium which is supported by the
coalesced trabecule cranii. These structures at first are two
cartilaginous bars, articulating with the extremities of the
parachordals, and extending down parallel to the posterior
visceral arches. Eventually they become bent upwards so as
to run parallel with the long axis of the skull, unite anteriorly,
and finally send up lateral outgrowths to form the side wall of
the anterior region of the skull. At first they present no
differences in appearance or relations from the visceral arches.
They are curved so as to approximate below, and articulate
with the vertebral region of the skull. This represents the

1 A. Dohrn, loc. cit. See also ‘ Zool. Anz.,’ vol. v, 1882.
? Ganin, “Ueb. die Entw. des Kopfskelets bei Kuockenfische (Rhodens
Gasterostens),” ‘Zool. Anz.,’ No. 51.

OSTEOLOGY, ETC., OF SYNGNATHUS PECKIANUS. 645

first stage in the phylogenetic history of the group. The
second stage of union is also well marked, and it is not until
comparatively late in life that the commencement of the third
stage, represented by the beginning of the lateral upward
growth, makes it appearance; and it may be explained by the
function the trabecule have in the adult of supporting the
facial region, and protecting the olfactory sense organs. This
specialization is only a degree greater than that found, for
instance, in the second postoral arch, where a longitudinal
division occurs, the posterior rod forming the hyoid, for the
support of the tongue and aiding in the respiratory process,
and the anterior one the symplectic, uniting the mandible to
the skull, and serving for its support.

The relation, too, of these trabeculze to the pituitary fossa is
also important. They articulate with the vertebral region of
the skull immediately behind that structure, and, in fact, have
the same relation to it as the other arches have to their clefts.
By being bent up, and by the atrophy of the pituitary cleft,
they eventually lose their original relations, and enclose them
instead of being their posterior boundaries.

The following table, which is a modification of Marshall’s
table,! itself a modification of one by Balfour,” will show at a
glance the relations which I consider probable as existing be-
tween the various segments. Asin Marshall’s table, clefts, &c.,
have been allowed for the Notidani.

1 A, Milnes Marshall, “The Morphology of the Vertebrate Olfactory Or-
gan,” ‘ Quart. Journ. Micr. Sci.,’ vol. xix, 1879.

2 F. M. Balfour, “On the Development of Elasmobranch Fishes,” ‘ Journ.
of Anat. and Phys.,’ vol. xi. See also ‘Monogr. of Hlasmobranchii.’

VOL. XXIIIL——NEW SER. UU

646 J. PLAYFAIR MCMURRICH.

Branchial Clefts. | Visceral Arches. Head-cavities. Cranial Nerves.
Pituitary or olfac-| Trabecul cranii. | Aborted (?). ‘Olfactory.

tory.
Lachrymal. Kthmo-palatine. | Preemandibular. |Oculomotor.
Mouth. Mandibular. Mandibular. Trigeminal.
Spiracular. Hyomandibular. | Hyoid. Facial and Auditory
Ist visceral. Ist Branchial. | lst Branchial. Glossopharyngeal.
1X Cea 2nd bes 2nd Bi lst Branchof Vagus
3rd i 3rd a 3rd a 2nd Cy, _
4th a 4th i. | 4th 4 Srd; visi i
5th A 5th As 5th a Ath. ,, ¥
6th 5 6th i 6th uM A aad 5
7th 5) 7th 59 7th ss 6th! 55 3

The position of the second arch articulating with the anterior
extremity of the trabecule appears at first an insuperable
objection to this theory; but when we consider that it is
possible that concomitantly with the bending up of the trabe-
culz and the overgrowth of the cerebrum there was a carrying
forwards of this arch without any bending up, a way out of the
difficulty is presented. The fact that nerves pass out through
openings in the walls of the skull formed by the trabeculz does
not appear to be of any weight against the theory, since it is
well known that when in the process of extension cartilage
meets with a nerve or blood-vessel instead of contracting and
closing or interfering with the action of that structure, a wide
foramen is left through which it may pass freely. In favour
of the theory this may be said, that it reduces to a natural
group structures which, viewed in any other light, are ex-
ceedingly abnormal, and is what one might expect from what
occurs in the persisting ancestral forms.'

III. Tue Norocuorp and VERTEBRAL COLUMN.

In Stage a (fig. 1) the notochord in the cranium is en-
sheathed by the parachordal cartilages. Below the medulla
oblongata it bends abruptly downwards at an angle of 45°.

1 By this expression I mean Amphioxus and Ascidians, which, though pro-

bably not in the direct genealogical line, may still be recognised as types of
the direct ancestors.

OSTEOLOGY, ETC., OF SYNGNATHUS PECKIANUS. 647

Posteriorly at the tail it is somewhat bent upwards, and
tapers abruptly to a point, so that the tail-fin is at this stage
heterocercal.

In Stage B, on transverse section, a very tough fibrous
membrana limitans externa is seen, having in fact exactly
the same appearance as the membrane, in which eventually
calcic material is deposited to form the parasphenoid. No
cells were to be distinguished in it as described by Gétte.
Within this is the notochordal sheath, apparently cellular,
as if the cells of the notochord forming it had not had their
protoplasm displaced. In older stages a membrane similar to
the membrana limitans externa forms round the spinal
cord, and lateral condensations of tissue form the rudiments of
the transverse processes. In the membrana limitans ex-
terna ossific matter becomes deposited, and similar depositions
take place in the other membranes, resulting in the formation
of the vertebral centra, and spinous and transverse processes.

In the adult the vertebre are of the ordinary amphiccelous
type, about 2 mm. in length, and presenting little variation in
appearance. The anterior ones are slightly modified, chiefly
in the transverse and spinous processes. In the first the latter
structures are expanded at their extremities, having a wing-
like shape; the spinous process is ridge-like, produced into a
process posteriorly. The anterior face of the centrum is very
much enlarged to form an articulating surface for the occiput,
while the posterior is the same size as in the succeeding
vertebree. From the posterior edge of the base of each neural
arch, a zygapophysis is given off, which passes directly back-
wards to articulate with the second vertebra, and serves to
strengthen the articulation.

In the second vertebra there are both anterior and posterior
zygapophyses, and in the third posterior ones only. In the
succeeding vertebre no zygapophyses are present; the trans-
verse processes are straight, somewhat flattened, projecting out
at right angles to the hour-glass-shaped body. The spinous
processes vary slightly for some distance back, having a ridge-
like appearance, and extending the whole length of the ver-

648 J. PLAYFAIR MCMURRICH.

tebra. The canal for the spinal cord is larger than that struc-
ture; immediately above the cord a partition (fig. 14), down-
wardly pointed in the middle, projecting with the dorsal fissure
of the cord, divides the canal into two portions, the lower of
which contains the central nervous system. The notochord
persists to a large extent. There are no ribs.

A very close relation exists between the neural and pleural
processes and the dermal scales. These structures first make
their appearance as stout, tough, membranous plates, having
the saine appearance as has been described for the membrane
plates in the roof of the cranium. In these membranes ossifi-
cation takes place, apparently concomitantly with that of the
neural and transverse processes, and the relations found in the
adult commence to show themselves. In a transverse section
(fig. 14) there is to be seen, astride, as it were, of the spinous
process, a plate—the dorsal scale (D. S.), and articulating with
the transverse processes very closely; there is, on either side,
a corresponding lateral scale (L. S.).

These latter apparently take the place of ribs, acting as
protective structures to the organs within.

IV. THE PartreD AND UNPAIRED FInNs.

In Syngnathus the fins present are a fairly developed
dorsal, weak pectorals, and a caudal. In the young
stages an anal is present, which, however, does not pass
beyond the stage in which the fibrillation begins, but aborts,
and is entirely wanting in the adult.

In Stage B (fig. 2) the dorsal fin consists of cartilaginous
rays embedded in the tissue of the body, and resting directly
on the membrane surrounding the spinal cord. The portion of
the fin outside the body is yet cellular, but a distinct fibrillation
is present, which, increasing, results in the formation of the
horny rays.

In Stage p (fig. 6) these have fully developed, their supporting
rays (1.C.) still being cartilaginous. Above, these are united by
a longitudinal bar, similar to what occurs in the paired fins of

OSTEOLOGY, ETC., OF SYNGNATHUS PECKIANUS. 649

Elasmobranchs.! This is apparently a secondary formation due
to the coalescence of the extremities, since in earlier examples
it is not present. Resting on this longitudinal bar, opposite
the intervals between the cartilaginous rays, are oval cartila-
ginous nodules (B.C.), each one of whichs upports a horny ray
(H. R.). These do not present any transverse segmentation
To each of these nodules are attached two muscles, a posterior
smaller one (Dep. M.), which by its contraction depresses the
ray, and an anterior (Er. M.), attached to the other extremity
of the nodule, which acts as an erector.

The cartilaginous supports for the fins appear long before
there is any trace of ossification of the spinal column.

As was stated before, the tail fin is heterocercal at first, and
passes through nearly the same formative changes as the dorsal.
The urostylic cartilages are large, and become formed previous
to the completion of the differentiation of the horny rays: they
are comparable, to a certain extent, to the interspinous or sup-
porting rays of the dorsal.

V. THe Gitis ann ALIMENTARY CANAL.

The gills of the Lophobranchs have usually been described
as “tufted,” from the supposition that they consisted of tufts
of filaments. Dr. Ryder? has corrected this mistake for
Hippocampus. The diagram of a gill leaflet of Syngnathus,
given on Plate XLII, fig. 7, will show their real structure.
Coming off from either side of the rachis are a number of
leaflets, gradually increasing in size from below upwards until
near the extremity, when they suddenly decrease. Four such
rows (only two are represented in the diagram) are arranged
on each rachis, forming, as it were, a rectangular pyramid
affixed by its apex. As Dr. Ryder says :—“ There is therefore
nothing at all in these structures which is not represented
homologically in the fish’s gill of the ordinary type, since the
two series of vascular branchial appendages to each arch in
Hippocampus are perfectly comparable with the bifureated,
vascular, branchial appendages of such a form as Salmo.”

1 F. M. Balfour, ‘ Comparative Embryology,’ vol. ii, London, 1881.
7 Loe. cit.

650 J. PLAYFAIR- MCMURRICH.

Dr. Ryder refers the abnormal structure of the gill to the
degeneration which has taken place, evidenced by the rudi-
mentary structure of the arches, and by the fact that the
leaflets are far less numerous than in ordinary fishes. He
supposes that “ the reduction in number of these appendages
may have called for the extension of the area of the ultimate
branchial lamelle or pinne.” It seems to me that this de-
generation of the pinne and arches has been due to the almost
complete covering-in of the gill cavity, which is evidently an
old ancestral character, since, even in Stage a, the branchial
cavity is completely closed over by membrane.

The dilatation of the anterior portion of the alimentary canal,
seen in Stage A, reminds one very strongly of the pharyngeal
dilation in Amphioxus, and the respiratory sac in the
Ascidians.

No communication between the yolk-sac and the intestine
was noticeable. Von Baer states that communication exists
immediately behind the liver, and Lereboullet believes that
such a communication exists between the stomach and the
liver, and persists until the complete absorption of the yolk.
Balfour,’ however, from his observations on the Trout and the
Salmon, could not confirm these statements. I can agree
with him that ‘‘all communication between the yolk-sac and
the alimentary tract is completely obliterated very early.”
Even in Stage A the intestinal wall is seen to pass quite
unbroken over the yolk-sac, absorption taking place through
the cells of the wall of the intestine, or else entirely by the
blood-vessels.

Near the posterior extremity of the intestine in Stage p a
well-marked valve is present. In Stage a no trace of this is
noticeable, but it makes its appearance in Stage B (fig. 2),
where it appears as a constriction of the walls, which even-
tually increases, and closes off the rectum from the anterior
portion of the canal.

Guetrn. June 28th, 1883.

1 ¥. M. Balfour, ‘Comparative Embryology,’ vol. ii, p. 65.

INDEX TO JOURNAL.

VOL. XXIII, NEW SERIES.

Anilin dyes, on staining with, by Vin-
cent Harris, 292

Bacillus anthracis, a morphological
variety of, by H. Klein, 260
a re note on, by E.
Ray Lankester, 265
Bacteria, E. Klein on the relation of
pathogenic to septic, 1
Balfour, F. M., on the anatomy and
development of Peripatus capensis,
213
Blastoderm of the chick, a rare form
of, by C. O. Whitman, 376
Blomfield on recent researches on
spermatogenesis, 320
Bourne, A. G., on Haplobranchus, a
new genus of Annelids, 168
3 » review of recent re-
searches on the origin of the sexual
cells in Hydroids, 617
Ad », see Lankester.
Bower, F. O., on plasmolysis, 151
Budding in Polyzoa, by A. C. Haddon,
516

Carpenter, P. H.. notes on Echino-
derm morphology, No. VI, 597
Cells and living matter, by Louis

Elsberg, 87

Chick, a rare form of the blastoderm

of, by C. O. Whitman, 376
», development of the pelvic gir-

dle and hind limb of, by Alice
Johnson, 399

Chordata, the ancestral form of, by
A. A. W. Hubrecht, 349

Cunningham, J. T., on the renal
organs (nephridia) of Patella, 369

Dollo, M. L., on the malleus of
Lacertilia, 579

Dowdeswell on a minute point in
the structure of the spermatozoon
of the Newt, 336

Kehinoderm morphology, notes on,
No. VI, by P. H. Carpenter, 597
Elsberg on cells and living matter, 87
Eyes of Scorpio and Limulus, by

Lankester and Bourne, 177

Fasciola hepatica, life-history of, by
A, P. Thomas, 99

Gardiner, Walter, on the continuity
of protoplasm through the walls of
vegetable cells, 302

Haddon, A. C., on budding in Poly-
zoa, 516

652

Haplobranchus, anew genus of Anne-
lids, by A. G. Bourne, 168 ©

Harris on staining with anilin dyes,
292

Heape, Walter, on the development
of the Mole (Talpa europea), 412

Hickson, 8. J., on the structure and
relations of Tubipora, 556

Hubrecht, A. A. W., on the ancestral
form of the Chordata, 349

Hydroids, review of researches on the
origin of the sexual cells in, by
A. G. Bourne, 617

Johnson, Alice, on the development
of the pelvic girdle and hind limb
of the chick, 399

Klein, E., on a morphological variety
of Bacillus anthracis, 260
» onapink Torula, 268
», onthe relation of pathogenic
to septic bacteria, 1

Lacerta muralis, early development
of, by W. F. R. Weldon, 134
Lacertilla, Malleus of, by M. L. Dollo,
579
Lankester, E. Ray, and A. G. Bourne,
on the existence of Spengel’s olfac-
tory organ and of paired genital
ducts in the pearly nautilus, 340
53 As and Bourne, on
the minute structure of the lateral
and central eyes of Scorpio and
Limulus, 177
mA 4s note on Bacillus
anthracis, 265
Larva of a Crustacean, supposed to
be the larva of Limulus, by R. V.
Willemoes-Suhm, 145
Limulus supposed larva of, by R. V.
Willemoes-Suhm, 145

INDEX.

Limulus and Scorpio, eyes of, by E,
Ray Lankester and A. G. Bourne,
177

Liver-fluke, life history of, by A. P.
Thomas, 99

MacMurrich, J. P., on the osteology
and development of Syngnathus
peckianus, 623

Malleus of the Lacertilia, by M. L.
Dollo, 579

Marsupials, on the foetal membranes
of, by H. F. Osborne, 473

Membranes, feetal, of cpossum and
other marsupials, by H. F. Osborne,
473

Mole, development of, by W. Heape,
412

Nautilus, Spengel’s olfactory organ
and paired genital ducts in, by HE.
Ray Lankester and A. G. Bourne,
340

Nephridia of Patella, by J. T. Cun-

ningham, 369

Opossum, on the fcetal membranes of,
by H. F. Osborne, 473

Ornithorynchus paradoxus, tongue
of, by E. B. Poulton, 453

Osborne, H. F., on the foetal mem-
branes of the opossum and other
Marsupials, 473

Patella, renal organs of by J. T.
Cunningham, 369

Pelvic girdle of the chick, develop-
ment of, by Alice Johnson, 399

Perameles nasuta, tongue of by E.
Poulton, 69

Peripatus capensis, anatomy and de-
velopment of, by F. M. Balfour,
213

Plasmolysis, by F. O. Bower, 151

Polyzoa, on budding in, by A. C,
Haddon, 516.

INDEX.

Poulton, E. B., on the tongue of Or-

nithorhynchus pardoxus, 453
. ee on the tongue of

Perameles, 69

Protoplasm, continuity of, through
the walls of vegetable cells, by
Walter Gardiner, 302

Pythium, on the genus, by H. M.
Ward, 485

Saprolegnie, by H. M. Ward, 485

Scorpio and Limulus, eyes of, by E.
Ray Lankester and A. G. Bourne,
ivi)

Spermatogenesis, recent researches
on, by J. EH. Blomfield, 320

Spermatozoon of the Newt, structure
of, by G. F. Dowdeswell, 336

Syngnathus Peckianus, the osteology

and development of, by J. P.
MacMurrich, 623

Talpa europea, development of, by
Walter Heape, 412

653

Thomas, A. P., on the life history of
of Fasciola hepatica, 99

Tongue of Ornithorhynchus, by E. B.
Poulton, 453

Tongue of Perameles, by HE. B.
Poulton, 69

Torula, note on a pink, by HE. Klein,
268

Tubipora, the structure and relations
of, by 8S. J. Hickson, 556

Vascular system of Echinoderms, by
P. H. Carpenter, 597

Ward, H. M., on Saprolegnie, 272
a » onthe genus Pythium,
485

Weldon on early development of
Lacerta muralis, 135

Whitman, C. O., on a rare form of
the blastoderm of the chick, 376

Willemoes-Suhm on a larval Crus-
tacean, at one time supposed to he
the larva of Limulus, 145

PRINTED BY J. E, ADLARD, BARTHOLOMEW CLOSE.

iz.” as AS UEAUD

+

a cls eae fc Tr AR | suo te fayeul our ka ee «ha
| oo

4 A a > 04% J ty : fs : ad Te hy ars Z
A at i Me) PERN fi “ei [ + en it Bh oO ide Sut #0
Ute ected | TA a a ae
Fea! Bite . Ve teen T ‘viet h aK yor
aa? oe De ee Law * y
gies 4 un vell * Ae Sai i “ 5 alts yd ay
5 we s = F & 4 | aos ve vee ocd
in intig tae gitaptete 4A Smid
: OER eld R.8 velo

“A ““

a u

vow

Se cing) wisi eS Sy He dl lose

Abt SUD saiatne ee ot

aoe erp no. TY! hn Vd
Fe ae ee i.

' & OS
>
ss ia trees asa A An) Be a
= Aa , “ ‘
y) i: oh +4 a
ww ¥ 4 ]
ic: Thee Ger ee PR eh Oe Te
ra, a oY cite a) ees -
‘ a 4
marr sis Fae ceili Loe rae é
3 serd ¢ Bite Bema. 4 Sphay
: ae ase
4 ya _%
- - ,
= igs ‘
¥4 aa -
e a 4 a cad be } * oe . 4 ie ws
© At See nF ANS. : = WR CRs ABAD SO
Lane
, 22
Se Fares
F — ‘ re
7)
i.
; J Bai! ua ij
_ in
‘
.

;
ome
- aay ve
, 7
>
a ‘
; 4
f
we

iar ol eA
Mienily Ure
uit

Cen i;

SP oe higzse 4
aa

> he
oie
shir: 7 byt ¢
i
:
Be the :
On bak.
'
a
‘

careih He Saude r a ¢ ait 4
L ; ; : ; ; of
. ii. (G Sera oo 2 aa
Mitersie gill 4 iy
S aiees ied ee Mais 4 ¢ ;

PE taule was BO os -s Cie Liege e's

~

at

LL ye Zz « 91 i Sit wy
Uti eit 4 : Ci oe ir
UALS | eo ae os ed) Doe Y; rrp peor

PRAVE os ots At hs. hi ae uy

’ X

tid haart Ni eae ot
y a 4
| |

na Te at ay eae :

gm
: P ¢
bd Ay 4 # A
a ae ‘ /
i Vey -
“ay “ Sq 43 : = hon
: ‘a
‘Ss
ts
~ U
: ' ty Ut
. —
* As Mbiieioty HAM Gael
4 i]
Gal f} :
}
1”
S
*
: vis >
ov Lad wa 4 a te) fiiokt wat fi Be
- y : . Pe SR in Ba ed
i eae a , aA y= ie. Ee “=
wd

eV é
. AJ
i

4 - 7 {

- @

¥
‘ ~ “
oes
« Pe
ae We ai -
ee ae de :
tr ee i“ &
7 De ~ _

. eee by ; if
Sian yee ie 7g)

. en

: - a)

‘
on

5) \
War
ed

al

i a

\

‘7

«
rhs
i
Pee
5
,
. *
Je
ik
"wa!
f
i
1
+
'
4 t
“ue
bi #
A. ‘ ere '
ia ter
4 Lal o)
ue Au ‘
j ‘ ¥. ny y
Y hy : A 17 ey Oy 4 -
f 4" i ’
b a ¢
, : i i ; fi
¢ [? q
wal ‘ie bonny :
f | , i é 1 1
} awe Vas He ft Me
Mel ‘i iy ae f | wo Ty

CATAL

3 2044 106 275

perl sarersmbessmeneseun’

a one, op

:
.
.
-
‘

ne

a eT ee at 2 ee