* -
cas tiadelip ioe

ii)

fat
r

eC BEGIN: Bole

OF

MORPHOLOGY.

EDITED BY

C. O. WHITMAN,

With the Co-operation of

EDWARD PHELPS®* ALLIS:

MILWAUKEE.
Vor XhEb
me
ere BOSTON, U.S.A.:

yan GINN & COMPANY
Py 8 -~_
meee 2045!

mah Sate
g rar

yo

ren
7%

bse

2) ¥
le

Tet ite =

, oe
a ae fork, a
iis ah Pry cd \, sa

:
poke
oy

- ae OMT, ly
‘ es Oy

my
ay,

Dt!

ay rae’ 1” as * a |

ae vs i >

ae ney if

We OF
aac Me

it ie a

Las a, ;
3 et -

; ia Oe apn hau : Lt ay 7 ga
rides - Bite of orht AIEB a a
sata : i = baa en ieey 2, i ,
aciy te - pone 1 7 ne 8 i
: Th ea) of a a) ae *) ait
yi hate ca Wilh atte"
ne A

11) if

IV.

II.

CONTENTS] OP VOL« XII.

No. 1. — May, 1806.

KATHARINE Foor.
Yolk-Nucleus and Polar Rings

. WILLIAM PATTEN.

Variations in the Development of Limulus
Polyphemus .

W. E. RITTER.
Budding in Compound Ascidians, based on
Studies on Goodsiria and Perophora .

FRANK E. LILLIE.
On the Smallest Parts of Stentor capable
of Regeneration; a Contribution on the
Limits of Divistbility of Living Matter.

Tuos. H. MontTGomery, Jr.
Organic Variation as a Criterion of De-
velopment

No. 2. — February, 1897.

C. O. WHITMAN AND A. C. EYCLESHYMER.
The Egg of Amita and its Cleavage .

Tuos. H. MONTGOMERY, JR.
On the Modes of Development of the Meso-
derm and Mesenchym, with Reference to
the Supposed mag of the Body

Cavities .

17-148

149-238

239-249

251-308

309-354

355-306

iV CONTENTS.

No. 2. — February, 1897. — (Continued.)

PAGES
III. GWENDOLEN FouLKE ANDREWS.
Some Spinning Activities of Protoplasm in
stavjish and Echinus Egos. |. aus Of=36O
IV. A. D. MEap.
The Origin of the Egg Centrosomes. . . 391-394
V. FRANKLIN P. MALL.
Development of the Human Coelom . . . 395-453

VI. Hermon C. Bumpus.
A Contribution to the Study of Variation . 455-484

No. 3. — March, 1897.

I. Epwarp PuHEtps ALLIS.
The Cranial Muscles and Cranial and First
Spinal Nerves in Amita Calva. . . . 487-808

II. KATHARINE Foot.
The Origin of the Cleavage Centrosomes . 809-814

Che Atheneum Press,
GINN & COMPANY, BOSTON, U.S.A.

—
JOURNAL

MORPHOLQ

9)
EDITED BY Pa
—T
c. 0. WHITMAN

Head Professor of Biology, Chicago University,

WITH THE CO-OPERATION OF

EDWARD PHELPS ALLIS.
Milwaukee.

Vou. XII. May, 1896.

BOSTON, U.SvA.:
GINN & COMPANY.

AGENT FoR GREAT BRITAIN: AGENTS FOR GERMANY: AGENT FOR FRANCE:

EDWARD ARNOLD, R. FRIEDLANDER & SOHN, JULES PEELMAN,

37, Bedford Street, Strand, Berlin, N.W., 189, Boulevard Saint-
London, W.C. Caristrasse I1. Germain, Paris.

III.

IV.

CONTENTS OF No. 1, MAY, 1896.

Yolk-Nucleus and Polar Rings. KATHARINE
Foot

Variations in the Development of Limulus
Polyphemus. WILLIAM PATTEN

Budding in Compound Ascidians, based on
Studies on Goodstria and Perophora. W.
Be RITTER

On the Smallest Parts of Stentor capable of
Regeneration ; a Contribution on the Limits
of Divisibility of Soe Matter. FRANK
E. LILiiz

Organic Variation as a Criterion of Develop-
ment. T. H. MONTGOMERY, JR. .

Che Atheneum Press.
GINN & COMPANY, BOSTON, U.S.A

PAGES

I-16

17-148

149-238

239-249

251-308

Volume XII. May, 1590. Number 1.

JOURNAL

OF

MORPHOLOGY.

YOLK-NUCLEUS AND POLAR RINGS.

KATHARINE FOOT, Evanston, ILL.

PoLar rings have been observed in Clepszne by Grube,
Leuckart, Robin, and Whitman,! in Rhynchelmis, by Vej-
dovsky,? and in Allolobophora foetida® by the author.

In the present paper I hope to prove that the polar rings
and the so-called yolk-nucleus are one and the same sub-
stance, and that therefore the material which forms these
structures is by no means confined to the three forms just
mentioned. In the fall of 1894, I identified the granular
masses of cytoplasm found in the ovarian egg with the polar
rings of later stages, and traced the substance step by step
during the growth of the maturing and fertilized egg; but the
publication of this work was reserved to form part of a later
paper on the maturation and fertilization of the egg of AJlo-
lobophora foetida.

1C. O. Whitman, “The Embryology of Clepsine.” Quart. Journ. of Micr.
Scz.. vol. XVIII, 1878, p. 234.

2 F. Vejdovsky, “ Entwicklungsgeschichtliche Untersuchungen.” Prag, 1892.

8 “ Preliminary Note on the Maturation and Fertilization of the Egg of A//o-
lobophora foetida. JOURNAL OF MORPHOLOGY, vol. IX, 1894.

2 LOOT. [ VoL. UT:

In the following winter, the ovarian egg of Lumbricus was
studied by Mr. Calkins! in the laboratory of Professor E. B.
Wilson of Columbia College. His results differ so radically
from those obtained by the study of AJ/olobophora foetida
that I think it best to give a brief account of my results.
Calkins’s results are, briefly, as follows: ‘‘ The yolk-nucleus is
chromatin in the form of granules.’”’ ‘This granular mass
disintegrates and the parts form the yolk plates of the egg.”’

In Allolobophora foetida the “yolk-nucleus”’ can be sharply
differentiated from the chromatin, and in normal eggs I find
no structures answering to the “great yolk plates’’ described
and figured by Calkins for Lumbricus. In a few cases, how-

Fic. 1.— Section of a degenerating egg from Fic. 2.— Calkins’s Fig. 5 reduced
ovary of All. foe. Abbe camera. one-half.

ever, in ovaries from worms past the breeding season (clitellar
region no longer marked), I find, among eggs in various stages
of degeneration, some that exactly correspond to Calkins’s
figure 5 ; the chromatin has for the most part lost its granular
structure, forming one or more solid homogeneous masses
among the granules, and in the cytoplasm there are large
homogeneous masses, which appear to be identical with
Calkins’s “grezt yolk plates.”

Ovaries from worms without a defined clitellum show various
stages of degeneration. In some cases the eggs of one ovary
will be entirely degenerated, while those of the other will
appear normal; again, only the older eggs of one or both
ovaries will show degeneration; while in some cases both

1 Gary N. Calkins, ‘Observations on the Yolk-nucleus in the Egg of Lumérz-
cus.” Transactions New York Acad. of Sci., June, 1895.

No. 1.] YOLK-NUCLEUS AND POLAR RINGS. 3

ovaries will appear entirely normal. Iam inclined to believe
that the condition of the ovary in this respect depends upon
the length of time that has elapsed since the ceasing of its
functional activity.

METHOD.

The origin and formation of the polar rings were first traced
by various methods, none of which, however, differentiated the
substance from all other constituents of the cell. The fixa-
tives used were chromo-acetic, corrosive sublimate, corrosive
acetic, Hermann’s fluid, Hermann’s fluid followed by Merkel,
Merkel’s fluid, osmic acid of various strengths, picro-acetic,
Parenyi’s, Flemming’s, etc. The stains used were various
haematoxylins, various carmines, and numerous anilins. While
I was engaged in a systematic effort to find a stain that would
differentiate the polar-ring substance from other constituents
of the cell, the publication of Calkins’s results from his study
of Lumbricus made it still more important for me to sub-
stantiate my results by differential staining. The use of
lithium carmine and Lyon’s blue was suggested to me by
seeing some beautiful specimens of nerve tissue, prepared
(under the direction of Dr. Patten) by Miss Lewis of Harvard
University; and I am indebted to Dr. Patten for subsequent
advice as to the use of the stains. Shortly afterwards I found
that Korschelt ! in 1889, had used practically the same method
(borax carmine and Lyon’s blue) to differentiate the yolk-
nucleus (Dotterkern) of various insect eggs; and in spite of
his beautiful results (clearly differentiating Dotterkern from
chromatin), I find very few investigators who have repeated
his method.?

1 Korschelt, ‘‘ Beitrage zur Morphologie und Physiologie des Zellkernes. Zool.
Jahro., Bd. IV, Hft. 1, 1889.

2 Calkins seems to have tried Korschelt’s method, and he says of it:
“ After borax carmine and Lyon’s blue the yolk-nucleus and chromatin had the
bright red stain of the carmine.” In Allolobophora foetida the only eggs that do
not give a constant reaction by this method are eggs in some stages of
degeneration ; for example, the “ great yolk plates” (?) of Cut 1 deeply stain with
carmine, and persistently retain the red after treatment with Lyon’s blue long
enough to stain the degenerating cytoplasm deeply.

a FOOT. [Vou. XII.

The method that has proved most satisfactory for eggs of
Allolobophora foetida is to stain the sections from one to
twenty-four hours in lithium carmine, wash in acidulated
alcohol for a few seconds, and double stain with a very dilute
solution of Lyon’s blue. The length of time required for
staining depends upon the fixative used. The process must
be carefully watched under the microscope; for the lithium
carmine in both the spindle fibres and cytoplasmic network
may be replaced by the Lyon’s blue.’ If the staining be
properly modified to suit the special fixative, all the fixatives
tested give results more or less satisfactory; but corrosive
sublimate, with or without acetic acid, gives the most brilliant
and satisfactory reaction. A more detailed description of the
effect of the various fixatives upon the polar ring substance
will be given later.

Before describing the figures, I would emphasize the fact
that each is an exact representation of the preparation both in
form and color, most of the work having been done by an
expert draftsman and colorist (Mr. H. Bridgham) ; not even
in shade or tint is there any variation from the original.
None of the figures are in the least diagrammatic; they were
drawn from sections, the camera lucida being used in all cases.
Each egg (taken from a cocoon) was studied under a Zeiss
2mm. immersion lens, and drawn with the aid of an Abbe
camera, before it was imbedded for sectioning.2 The varia-
tions in size of eggs of nearly the same stages of develop-
ment are largely due to unequal shrinkage, dependent upon
the fixative used.

Fig. I represents a very young oécyte. In the cytoplasm,
in close contact with the nucleus and sharply differentiated
from the chromatin, nucleolus, and cytoplasm, is a_ blue,
granular mass,—the so-called yolk-nucleus. This substance

1 If the odgonial areas are overstained with Lyon’s blue, the nuclei appear im-
bedded in a matrix of blue; the cell boundaries, red cytoplasm, and capillaries
being obliterated. If preparations fixed by Hermann’s fluid are overstained, the
nucleoli as well as the yolk-nucleus will stain blue.

* These facts are emphasized in order to prevent a mistake similar to the one
made regarding the figures of my preliminary note. The American Naturalist,
vol. XXIX, January, 1895, speaks of my “ diagrammatic figures.”

No. 1.] YOLK-NUCLEUS AND POLAR RINGS. 5

greatly increases as the egg grows, and can be traced (by
the sharp contrast of color) step by step during the develop-
ment, fertilization, and cleavage. It is present at the chief
points of activity in the cell: it is in the spindle, it forms the
fertilization cone and the archoplasm of both sperm and egg
attraction-spheres ; and at the pro-nuclear stage a part of it
is aggregated at both poles of the egg, forming the structures
known as the polar rings. For this blue substance I shall
retain the term employed by Boveri in a far more limited
sense — the term archoplasm.

In no case where the method has been properly applied have
I found a cell in the resting stage, — that is, a cell containing
nucleus, nucleolus, and cytoplasm,— without being able to
recognize at least a trace of the archoplasm, the cell often
being so small as to make it impossible to decide whether it
belongs to the earlier or to the final generation of odgonia.
As growing oécytes are found close to the stem of the ovary,
surrounded by dividing odgonia, the locality of the cell is
not a trustworthy indication of the stage to which it
belongs.!

In Fig. 2 we have an older odcyte, with the archoplasm
increased in proportion to the rest of the cell.

1In the spermatogonia of Salamander mac., Meves finds masses of granular
substance, which he identifies with the “ yolk-nucleus ” of other authors and with the
attraction-sphere. F. Meves, “ Ueber eine Metamorphose der Attractionssphare
in den Spermatogonien von Salamandra maculosa,” pp. 143-4. Arch. f. mik.
Anat., Bd. 44, Hft. 1, 1894. Through the courtesy of Dr. Watasé, I have been able
to apply Korschelt’s method to the sperm cells of Szve, and I find that these con-
tain a granular substance similar to the archoplasm of AJlolobophora foetida.
Many of them resemble very closely my Fig. 1 (a young odcyte). These results
recall Nussbaum’s identification of the Vesenxkern of a variety of gland cells with
the Dotterkern of eggs and the (Vebenkern of the spermatocytes: “ Dagegen wird
man den Nebenkern der Driisenzellen wohl mit dem von Wittich entdeckten
Dotterkern der Eier, dem durch von la Valette St. George zuerst bekannt gewor-
denen Nebenkern der Spermatocyten, den von Leydig aus der Epidermis von
Pelobates-Larven beschriebenen Bildungen in eine Categorie bringen diirfen.”
Moritz Nussbaum, “ Ueber den Bau und die Thatigkeit der Driisen.” Arch. f.
mtk. Anat., Bd. 21, 1882, pp. 343-4. Later, Balbiani says: “Le noyau vitellin
(Dotterkern) des Araneides est homologue du Nebenkern (centrosome de
Platner) des cellules seminales et du centrosome des cellules somatiques.” E. G.

Balbiani, “ Centrosome et Dotterkern.” Journ. de 1’Anat. et dela Physiol., tome
XXIX, 1893.

6 FOOT. RVOL. XII

Fig. 3 represents a still larger oocyte, with a corresponding
growth of the archoplasm.

Figs, 4 and 5 represent two types of the distribution of the
archoplasm found in the older oocytes. In some cases the
greater part of the substance is aggregated at the periphery
of the egg; in others the distribution is more nearly equal
throughout the cytoplasm.!

Fig. 6 shows a new phase of the archoplasm.2, Here we
not only find it distributed through the cytoplasm, but at
points where the membrane of the germinal vesicle has dis-
appeared we see it entering into the germinal vesicle itself.

The presence of the archoplasm in the germinal vesicle
reminds one of Korschelt’s* and Jordan’s* observations on
the entrance of the “yolk-nucleus”’ into the germinal vesicle,
and Lavdowsky’s® claim to have seen yolk elements within the
nucleus. It also recalls the observations of some of the inves-
tigators who claim to have seen not only the nucleoli, but
also chromatin, /eaving the germinal vesicle; for this sub-
stance in earlier stages has been positively asserted to be
chromatin by several authors. In Fig. 6 an attempt has been
made to represent the less transparent appearance of the
Ovarian eggs at the free end of the ovary. A comparison
of this with the foregoing figures will show what I have
attempted to indicate. With very few exceptions (dependent
upon the fixative) this same lack of transparency is shown in
all later stages ; z.e. in eggs taken from the cocoon.

1 Fig. 4 recalls the Diffuse Dotterkern of Stuhlmann: “Die Reifung des
Arthropodeneies nach Beobachtungen an Insekten, Spinnen, Myriopoden und
Peripatus.” Ber. d. Naturf. Ges. z. Freiburg i. Br., Bd. 1, 1886. Cf His Fig.
137) elatenvildile

2 The lithographer has not sufficiently blended the three masses of archo-
plasm at the upper left side of the figure. In the original sketch, the masses
are more granular in appearance and are connected by less dense areas of the
same substance.

3 Eugen Korschelt, “ Beitrage zur Morphologie und Physiologie des Zellkernes.”
Zod. Jahrb., Ba. IV, Hft. 1, 1889.

4 Edwin O. Jordan, “The Habits and Development of the Newt.” JOURNAL
OF MorPHOLOGY, vol. VIII, 1893.

° M. Lavdowsky, “ Von der Entstehung der chromatischen und achromatischen
Substanzen in den tierischen und pflanzlichen Zellen.’ Anatomische Hefte:
Merkel und Bonnet’s Anat. und Entwicklungse., Rd. IV, Hft. 13, 1804.

No. 1.] YOLK-NUCLEUS AND POLAR RINGS. |

The study of eggs killed in any fixative containing osmic
acid inclines me to believe that this lack of transparency is
due, at least in part, to material absorbed from the body-cavity
of the worm, and later from the albuminous contents of the
cocoon ; for all these eggs contain in greater or less degree the
black spots (Deztoplasm) represented in Fig. 14, which can be
dissolved out with warm xylol or ether, their fatty nature
thus being proved. These, however, are not the only factors ;
for the degree of transparency shown by the young ovarian
egg is not restored by dissolving away the Deutoplasm.

The distribution of the archoplasm in all these eggs is
modified by the fixative; corrosive sublimate or corrosive
acetic are especially favorable for its study, as they cause it
to aggregate into masses, thus producing the sharpest contrast
between the red and the blue. In chromo-acetic preparations,
on the contrary, the distribution of the archoplasm is much
more nearly equal throughout the cytoplasm; but wth all
fixatives it 1s aggregated at the centres of activity of the cell.

I am inclined to think chromo-acetic the most reliable
fixative, for the following reasons:

First, it so fixes the eggs that the subsequent treatment
with alcohols produces scarcely perceptible shrinkage; as a
rule, eggs measured before killing and after mounting give
almost the same diameter.

Second, after chromo-acetic all the structures of the cell
may be constantly and sharply defined :—the cytoplasmic
network; the attraction-sphere, with its centrosome, archo-
plasm, and rays; the spindle fibres; the chromosomes; the
fertilization cone; the sperm, and the sperm granules.
Structures that are distorted or obliterated by many other
fixatives may be constantly and distinctly defined with
chromo-acetic.

I am inclined to believe that the apparently granular structure
of the archoplasm is largely, if not wholly, due to the fixatives ;
certainly the degree of granulation varies with the different
fixatives, and at the points of greatest activity of the cell the
fusing of the red and blue substances into a structureless
homogeneous mass suggests that (at least in some stages) both

8 FOOT. [Vor. XII.

substances may be fluid. (The polar rings in the egg of
Clepsine consist, according to Whitman, of “a transparent
fluid substance.) The granular appearance of the archoplasm
is most marked at the points of greatest aggregation; that
part aggregated at the poles as polar rings appearing more or
less granular with all fixatives.

Fig. 7 shows a typical example of the presence of the archo-
plasm within the spindle and at its poles!.

In some cases (possibly caused by the action of the fixative)
the archoplasm is not equally distributed in the spindle, but is
aggregated at its edges and around the chromosomes, leaving
areas entirely free from the substance; again, it is aggregated
in longitudinal lines, which appear as heavy blue fibres among
the red anastomosing spindle fibres. At the poles of the spindle,
we see the red cytoplasmic rays of the attraction-spheres and a
pronounced aggregation of the archoplasm. This egg (Fig. 7)
was killed in Merkel’s fluid, which thus far has proved relatively
unfavorable for chromosomes and centrosomes.”

Fig. 8 shows a cross-section through the fertilization cone
(the structure which appears to be identical with the “cone of
attraction’ of other animal forms).® The archoplasm is aggre-
gated around the sperm, and the part in contact with the
sperm appears to be blended with the red cytoplasm, forming
a homogeneous mass, suggesting a fluid condition of both
substances. This figure gives us a typical illustration of the
action of corrosive acetic upon the archoplasm, causing it to
form into masses, as mentioned above.

1 This phase of the archoplasm recalls Strasburger’s ‘‘Kinoplasm”; Flem-
ming’s “ Zelle.” Merkel und Bonnet’s Ergebn. der Anat. u. Entwickg., Bd.
III, 1893, pp. 66, 104.

* The lithographer has not indicated the faint chromosomes shown in the
original sketch. For the shape of the chromosomes of this spindle see my “ Pre-
liminary Note on the Maturation and Fertilization of the Egg of AJl/lolobophora
foetida.” JOURNAL OF MORPHOLOGY, vol. IX, p. 94.

8 For further details as to the cone and the fertilization of this egg, see my
“ Preliminary Note on the Maturation and Fertilization of the Egg of A//olobophora
foetida,” JOURNAL OF MORPHOLOGY, vol. IX, 1894. The “sperm granules ” give
a staining reaction similar to the nucleoli; with aurantia and iron haematoxylin,
the nucleoli of the pronuclei, the nucleoli outside the first maturation spindle, and
the sperm granules, all stain yellow, while the chromosomes and sperm stain blue.

No. I.] YOLK-NUCLEUS AND POLAR RINGS. 9

Fig. 9 represents a longitudinal section through one of the
fertilization cones, that part of the sperm which is still within
the cone being indicated. In this figure we again see the ten-
dency of the archoplasm to accumulate at the centres of activity ;
it is collected around the entering sperm. The extent of the
cone 1s dependent upon the length of sperm within the egg, and
the aggregation of the archoplasm (especially after fixatives that
destroy radial structure) seems to be the sole factor that gives
form to the cone. In applying the same method of staining
to the spermatozoén at various stages of its development, I
have not been able to obtain any reaction indicating the pres-
ence of archoplasm ; hence it appears that this relatively large
mass of archoplasm of the cone has not been brought into the
egg by the sperm. The question is here suggested : Is not
the archoplasm of the cone identical with the sperm archo-
plasm of other eggs? Boveri! says: “Auf diesem Stadium
nun finden wir das Archoplasm als einen dichten kugeligen
Hof um das im Centrum des Eies gelegene Spermatozoon
(Fie, ro-und; 11, Taf. 1; Fig. 26, Taf. Il): Es) stellt: sich’ an
den beweisenden Praparaten als eine betrachtliche Ansamm-
lung einer gleichmassig kornigen Substanze dar, die nach
aussen ziemlich scharf abgegrenzt ist, wahrend die iibrige
Zellsubstanz vollkommen homogen erscheint”’ (p. 65).

Fig. 10. As soon as the first polar body is constricted off
we no longer find a fertilization cone; but we find a sperm
attraction-sphere with d/ue archoplasm at the point previously
occupied by the blue archoplasm at the apex of *the cone.
At this stage (Fig. 10) the sperm head is contracted into a
relatively short, thick rod, with its attraction-sphere occupying
a position nearer the attraction-sphere of the lower pole of the
spindle than does the rod itself. The structure of the sperm
attraction-sphere is identical with that of the egg attraction-
sphere described above (Fig. 7). This figure (Fig. 10) suggests
that the archoplasm of both sperm and egg attraction-spheres
is furnished by the egg alone. In the case of the sperm
attraction-sphere we have evidence only that the middle-piece

1 Th. Boveri, ‘‘Zellen-Studien.”’ Hft. 2, Jena, 1888.

10 FOOT. [VoL. XII.

of the sperm produces the centre of activity, around which
the archoplasm (a/ready present in the egg) aggregates.1 In
the egg of Allolobophora foetida this blue archoplasmic mass ts
so pronounced that there can be no question as to its individuality ,
zt 1s as distinct as the cytoplasmic network itself, thus supporting
Bovert’s assertion of the specific character of archoplasm.

In this figure (Fig. 10), drawn from preparations fixed in
chromo-acetic, we find not only the centrosome but the red
anastomosing rays of the attraction-spheres sharply defined.

At the stage represented in Fig. 11 most of the archoplasm
(especially in chromo-acetic preparations) is aggregated at the
periphery of the egg, only a relatively small amount being
present around the very small male and female pronuclei; but
in corrosive acetic preparations the archoplasm is not limited
so nearly to the periphery, some masses being present through-
out the cytoplasm of the egg. This presence of the polar-ring
substance on the periphery of the egg, prior to its aggregation
at the poles, supports Vejdovsky’s? observations on Rhynchelmis.

Fig. 12 represents the stage at which both pronuclei are
formed, a large part of the archoplasm having aggregated at the
two poles of the egg, thus forming the polar rings. In addition
to the archoplasm at the poles, a relatively large amount has
aggregated around the pronuclei (one of which is shown in the
figure); and at the points where the membrane is breaking
down the archoplasm is entering into the pronucleus itself.
The archoplasm massed around the pronuclei has a different

1In the relatively large spermatozoa of Amphiuma (which the kindness of
Professor Conklin has enabled me to study), I have obtained a reaction to Lyon’s
blue both in the middle-piece and in the tail. The red cytoplasm and the blue
archoplasm are present in both structures, suggesting that the apparent absence
of the archoplasm in the spermatozoén of AJlolobophora foetida may be due to
faulty technique. I think, however, there can be no question that the relatively
large amount of archoplasm in the cone and attraction-spheres of the egg of
Allolobophora foetida is merely an aggregation of the archoplasm already
present in the egg. This conclusion is in accord with the observations of Wheeler
on the egg of AZyzostoma, in which he finds the archoplasm of the cleavage
attraction-spheres furnished by the egg alone. Wm. M. Wheeler, “The Behavior
of the Centrosome in the Fertilized Egg of Myzostoma glabrum Leuckart.”
JOURNAL OF MORPHOLOGY, vol. X, 1895.

2 F. Vejdovsky, “ Entwicklungsgeschichtliche Untersuchungen.” Tafs. IV und
V, Prag, 1892.

NiO: 2.i] YOLK-NUCLEUS AND POLAR RINGS. LT

shade of blue from that at the poles. It shows (though in a
less degree) the blending of the red and blue described under
Fig. 8.

In a somewhat later stage (Fig. 13) the membrane of the
pronuclei is entirely gone, the chromatin is in the form of
loops, and the archoplasm is present around the loops and in
the cytoplasm.

Fig. 14 represents one of the polar rings before staining
with Lyon’s blue. The egg was killed in Hermann’s fluid,
and the section was stained for twenty-four hours in lithium
carmine. We see here how (in Hermann’s preparations)
lithium carmine stains the cytoplasm, and does not stain the
polar ring substance. This section also represents the black
spots referred to above (Deutoplasm), which subsequently were
completely dissolved away by soaking the sections for twenty-
four hours in warm xylol.

In studying the literature on the so-called yolk-nucleus I
have found that in many cases the substance follows more or
less closely a course of development similar to that of the
polar ring substance of this egg. It appears first close to the
germinal vesicle, increases in amount as the egg grows, and in
some forms becomes distributed in granular masses near the
periphery of the egg, while in others it appears as numerous
clumps or granules throughout the cytoplasm. In many cases
the figures show aggregations of the substance which strongly
suggest polar rings. I shall, however, quote only those cases
where the author specifically states that the substance is finally
aggregated at one or both poles of the egg.!

Stuhlmann’s? results from investigations of a variety of
insects furnish the strongest evidence that the Dotterkern of
these animals and the polar rings of Al/olobophora foetida are

1 For the latest historical sketches of the ‘“ yolk-nucleus,” and for the literature
on the subject, see L. F. Henneguy, “ Le corps vitellin de Balbiani dans l’oeuf des
vertébrés,” Journ. de lanat. et de la physiol., ann. 29, 93; and H. Mertens,
“Recherches sur la signification du corps vitellin de Balbiani dans l’ovule des
mammiféres et des oiseaux,” Archives de Biologie, tome XIII, 1893.

2 Franz Stuhlmann, “ Die Reifung des Arthropodeneies nach Beobachtungen
an Insekten, Spinnen, Myriopoden und Peripatus.” Ser. d. naturf. Ges. z. Freiburg.
i. Br., Bd. I. 1886.

.

2 FOOT. [VoL. XII.

identical structures. Stuhlmann distinguishes two kinds of
Dotterkern: t.e.* Diffuse Dotterkerne,” that which is distributed
throughout the cytoplasm (cf his Fig. 137 with my Fig. 4), and
« Eigentlicher Dotterkern,” that which is aggregated at one of
the poles. For examples of polar Dotterkern see Figs. 164, 165,
at, 1X4 and examples:on Dat. XxX.

In a recent paper by J. W. Hubbard! on the “ yolk-nucleus”’
in Cymatogaster, he clearly traces the “yolk-nucleus” from the
germinal vesicle to one pole of the egg ,; and his figures show
an aggregation of these granules fully as pronounced as either
of the polar rings of Al/lolobophora foetida.

Leydig’s? description of the granules (in the egg of Argulus),
which aggregate at both poles of the egg, is strikingly suggestive
of the polar rings. “In der Substanz des Spongioplasmas
erscheinen jetzt auch wirkliche Granula oder dunkelrandige
Kornchen, welche sich zu zwet Haufen an betden Polen des Ezes
ansammeln”’ (p. 300).

I am greatly indebted to Professor Wheeler for generously
allowing me to study his AZyzostoma preparations, and provid-
ing me with adult JM/yzostoma to prepare with Korschelt’s
method. The study of these preparations has shown that in
various tissue cells the cytoplasm can be differentiated into
two distinct substances, one reacting to lithium carmine and
the other to Lyon’s blue.

It gives me pleasure to express my great obligation to Dr.
Whitman for most kind and thorough criticism of my work.

1 J. W. Hubbard, “The Yolk-nucleus in Cymatogaster aggregatus Gibbons.”
Proceed. of the Amer. Philos. Society, vol. XX XIII, No. 144, 1894.

Franz Leydig, “ Beitrige zur Kenntniss des thierischen Eies im unbefruchteten
Zustande.” Zool. Jahr., Abth. f. Anat. u. Ontog., Bd. III, Hft. 2, 1888.

No. I.] YOLK-NUCLEUS AND POLAR RINGS. ig

PAPERS REFERRED TO.

BALBIANI, E. G. Centrosome et Dotterkern. Journ. de l’anat. et de la
physiol, tome XXIX, 1893.

BoverI, TH. Zellen-Studien. Hft. 2, Jena, 1888.

Bover!, TH. Ueber das Verhalten der Centrosomen bei der Befruchtung
des Seeigel-Eies. Wurzburg, 1895.

CALKINS, GARY N. Observations on the Yolk-nucleus in the eggs of
Lumbricus. TZvransactions New York Acad. Scz., June, 1895.

FLEMMING, W. Zelle. Ergebnisse der Anat. u. Entwickg. Merkel u.
Bonnet. Bd. III, 1893.

Foot, K. Preliminary Note on the Maturation and Fertilization of the
Egg of Allolobophora foetida. /ourn. of Morph., vol. 1X, 1894.
HENNEGUY, L. F. Le corps vitellin de Balbiani dans l’oeuf des vertébrés.

Journ. de Vanat. et de la physiol., tome XXIX, 1893.

HUBBARD, JESSE W. The Yolk-nucleus in Cymatogaster aggregatus
Gibbons. Proceed. of the Amer. Philos. Society, vol. XXXIII,
No. 144, 1894.

JORDAN, Epwin O. The Habits and Development of the Newt. Journ.
of Morph., vol. VIII, 1893.

KORSCHELT, EUGEN. Beitrage zur Morphologie und Physiologie des
Zellkernes. Zool. Jahrb., Bd. 1V, Hft. 1, 1889.

Lavpowsky, M. Von der Entstehung der chromatischen und _ achro-
matischen Substanzen in den tierischen und pflanzlichen Zellen. Anat-
omische Hefte. Merkel u. Bonnet’s Anat. und Entwicklungsg.,
Bd. IV, Hft. 13, 1894.

LEYDIG, FRANZ. Beitraége zur Kenntniss des thierischen Eies im unbe-
fruchteten Zustande. Zool. Jahrb., Abth. f. Anat. u. Ontog., Bd. III.
Hit. 2, 1888.

MERTENS, H. Recherches sur la signification du corps vitellin de Balbiani
dans l’ovule des mammiféres et des oiseaux. Archives de Biologie,
tome XIII, 1893.

MEVES, F. Ueber eine Metamorphose der Attractionssphare in den
Spermatogonien von Salamandra maculosa. Arch. f. mik. Anat.,
Bd. XLIV, Hft. 1, 1894.

Nussspaum, Moritz. Ueber den Bau und die Thatigkeit der Driisen.
Arch. f. mtk. Anat., Bd. XXI, 1882.

REINKE, F. Zellenstudien, II. Theil. Avch. fi. mtk. Anat. Bd. XLIV,
Hft. 2, 1894.

STUHLMANN, FRANZ. Die Reifung des Arthropodeneies nach Beobach-
tungen an Insekten, Spinnen, Myriopoden und Peripatus. Ber. d.
Naturf. Ges. 2. Freiburg. t. Br. Bd. 1, 1886.

14 FOOT.

VeEypovsky, F. Entwicklungsgeschichtliche Untersuchungen. Prag, 1892.

WHEELER, WM. M. The Behavior of the Centrosomes in the Fertilized
Egg of Myzostoma glabium Leuckart. Journ. of Morph., vol. X,
1895.

Wuitman, C. O. The Embryology of Clepsine. Quart. Journ, Micr.
Scz., vol. XVIII, 1878.

e
‘
*
‘
.
2
+ -
5 :
-
,
¢i
"hi rn
:
’ f
-
f q '
7 . *
'
ivi
‘
f , ;
- < \ .
a , i
hy i)

16 FOOT.

EXPLANATION OF PLATEI.

All figures were drawn from sections; in all but three cases the entire figure
was drawn from a single section.

Zeiss. hom. immer., 2mm., 140 ap.

Abbe Camera.

Figs. 1-5, ocular IV.

Figs. 6-12, ocular II.

Figs. 13-14, ocular IV.

Stains, lithium carmine and Lyon’s blue.

Archoplasm (‘“ yolk-nucleus,” “archoplasm,” polar rings), blue.

Chromatin, nucleolus, and cytoplasm, red.

Fic. 1. Very young oocyte, p. 4.

Fic. 2. Older odcyte, p. 5.

Fic. 3. Later stage, p. 6.

Fic. 4. Archoplasm dispersed in irregular patches throughout the oocyte, p. 6.
Fic. 5. Archoplasm peripherally distributed, p. 6.

Fic. 6. Egg detached from free end of ovary (section through germinal

vesicle), p. 6.

Fic. 7. Egg taken from freshly-laid cocoon, showing longitudinal section
through first maturation spindle, p. 8.

Fic. 8. Fertilized egg. Transverse section through the fertilization cone, p.8.

Fic. 9. Longitudinal section through fertilization cone, p. 9.

Fic. 10. Egg after formation of first polar body (archoplasm in polar body).
Figure drawn from three sections, p. 9.

Fic. 11. Egg after second polar body has been formed (the polar bodies not
represented), p. 10.

Fic. 12. Egg in pronuclear stage; archoplasm aggregated at the poles as
polar rings, and concentrated around the pronuclei (only one of which is
represented), p. Io.

Fic. 13. Later stage; chromatin in form of loops, p. 11.

Fic. 14. Transverse section of one polar ring, egg in pronuclear stage, p. 11.

a
Tih. Anch v Werner & Wieser,

VARIATIONS IN THE DEVELOPMENT OF
TIMOLUOS, POLVFIEMG Ss

WILLIAM PATTEN, Ph.D.

PROFESSOR OF ZOOLOGY IN DarTMOUTH COLLEGE, Hanover, N.H.

CONTENTS.
Pace
PINE @ VU CTI OIN a. 5-isciiensocsensnnsnsnceotene gue ctaetarne eo aeeattee ye Seae ey Saneren ett eemeeneteme 21
ELROD SK Of TAA. ACHEIT, | SELLILL ILE NCCC mera eee ea 23
WVELHOMS Of AOU LAI EIS: NIGEL LL ame eat ae ee ee 2
COS ESTO PMV GT LALEOT ATE LILLE RTLLL noe arene ena eae 28

DESCRIPTION OF THE DIFFERENT CLASSES OF VARIATION — METHOD
OF CLASSIFICATION.

Lee LIN VAG EN-ACIO IN| (OF ZAYP IPE NID IAG ES 32.2 ecco, tenses tesa nace cee eee ee ee 30
Hivasination not due toexternal stress: co. 22-ccec.cec.ceseeneierasee teense eee 31
Bearing on lung-books of arachnids, and visceral cleffs of vertebrates.. 32

TT PABSENCE OF APPENDAGES..i:...c:.tccccescedae afl ets ec bat ros aes se eee 33
The three anterior pairs of thoracic, and the abdominal appendages,
ALE eMOS tatne qimeM tly, ASE mit essence cece eee eee nee eee 34
A. Absence of Abdominal Appendages VELY COMMOMN........-.0.--v0eeeeeveee- 34
The Telopore: relations to primitive streak, anus, apex of body,
and apical growth: comparison with vertebrates ..........0...20.0.....--..- 35
D SER 10) Bd BS) Ko) oY see eae eR Ee renee 36
Baw Asymmetry of Abdominal APPCHAGLES..--cncccr nce sercea na ravaertcecee see eeyeeee aes 36
GaerAbsenceof LkOr Acie APPCHAGRES 0.0.0 ceccxsectnssasetecaeesseseestrne eee 36
Mimetor appearance of same in mormall eg psec steers eee 36
Observations on disappearance of same in “ving abnormal eggs........ 37
al Seyi wp Wu Si. 2.222 seaecenteneeéeneskctcensscccberesnusete tac houees usec soahe eee eee een 37
Progressive reduction of thoracic appendages from before back-
WAN GS ieee ofc iencsecebansecees sccece onsnsnscsusn Giasenscadisctotcuseeuass@iec bette cetera as gee Be cena 37
Greatemvitality of last) three panes cou cs.3.s0.2<cazce acyconrnantiapeaeeeee erereeeoeee 38
Symmetrical and asymmetrical reduction of same ...........-.--..---...------- 38
III. MULTIPLICATION OF APPENDAGES AND OTHER ORGANG..............-...-. 39
Multiplication of chelicerae and corresponding sense-organs (lateral
GYES)y ANG! DEUTOMELES...-...2:.s.-c3 ayant. tvacesagertoesae eeeeeeee tee ee ee ee aac 40
IV. Fusion oF RIGHT AND LEFT HALVES OF THE EMBRYO AND AN-
MERO-ROSTERIOR: DEGENERATION -icsssscac-ore-e-c-eemercs reese eeteeeenseesee eecencee 40
Method of growth and arrangement of Aa/fmetameres...............-.-.----- 41

Fusion and degeneration of segmental organs is in reverse order of
Gein FOTIM AON +..<:..--222-650.ceseecdecrcunaweneeae ne wc eaceee tere ieee cere pence ete scans 4I

18 PATTEN. iVoL. sce

PaGE

Method of fusioniand degeneration: 2. scccscccscececcseeresncrennnz-scecersusasaecsese acs 42

Resultant froma rhomboid esos ceee ce seecses ss coceeccenesnenes cereee seaesrersnaeaecs 42
Median fusion, cranial and caudal flexures a necessary result of apical

FeO 9) Oy eee eee Be eR ee orto ee ae eee OS Sacer bee Lae 43
Contrast between arrangement of organs on a half-metamere and

the arrangement of metameres in a whole animal.........0....22..2... 43

Apical growth of half-metameres on a spherical surface necessarily
results in differential growth-tension, which determines general
shape of the body and its division into regions having different mor-
phological characters and different physiological potentialities... 44-45

Concrescenceof Margin of Mesoderm1e Axed, nese ccsstsnasassteec tsetse 45
Lateral Apical growth of each mesodermic segment from a Jateral
primitive-streak-like proliferation. ........... ...-----:csccscceecceessseeeseseeceeeeeeess 46
Mareini of mesodermicianea acest. cae se cease cease heen, eee ese teeter enne ee eee 46
Concrescence of same to form post-anal cloud...............::e::seeeseeeeees 47
Description and comparison of special CaSeS.........-...-..-0ce-cessseesecsesceees 48
Relation to apical growth, primitive streak, and to telopore................ 49
Explanation “of Spiralise pm ents: sees ceree areas ene eee ee 49

Polarity of the embryo is determined by the arrangement of half-
metameres and the conditions produced by their mode of growth.. 49

DUlus trations Of Fusion ees eee eee eee ee 52-56
Exceptions to general law of median fusion: fusion greatest at
posterior end'; or same at both vends.-22 22 scene eae ee 53

Abdomen unusually perfect and well developed in embryos whose
anterior portion is greatly reduced by median fusion and degenera-
tion.

Change in Shape of Fused and Degenerating Appendage’. .......2...0.0.0000-+ 56

Increased length and diminished diameter of newly fused appen-
dages. Irregular shape of same, etc.

V. (GENERAL "PROGRESSIVE SDEGENIER A1IO Nip ceeece: eee eee eee ee 56
Is probably due not to external conditions, but to either a general or
a local lackof formative emerge: sessee een teeet ces eecee tech meaner eee 56

Four varieties of general progressive degeneration: (1) Reduc-
tion in rate of development. (2) Reduction in size of embryo.
(3) Local reduction of organs. (4) Reduction in range of devel-
opment of the entire organism, or parts of same, followed by com-
plete retrogressive degeneration. Result is a new form of death

forhighly orzanized*animals: ie. Seer ee ee ee 57
Absence of organs does not affect character of remaining ones.......... 58
The trunk is a colony of half-metameres...c:-20201-icccofes secs recto 58

Originally absent organs are probably not restored.
The variations described due to internal conditions, not to variations

invexternal “factors vives cases es cee A A 58
Frequent absence of chelicerae. Bearing on small size of same in

allarachmids.5.20:5: 0.00 See te) Se Oona Oe oe ee 59
Weakness of anterior part of thoraxe. 0o toe a 59
Weakness of anterior margin of cephalic lobes.................-::000eee-se0uees 59

Weakness of anterior end of abdomen..3.2) 0 59

No. 1.] VARIATIONS IN LIMULUS POLVPHEMUS. 19
PaGE

Relation of these inherent lines and regions of weakness to meta-

merism, to transverse fission in annelids, and to the morphological

and physiological differentiation of the body into head, thorax,

abdomen wetcre was Weasntly Saude we eee cina id eet 61
Median fusion an indication of structural weakness and degeneration. 61
In the case of organs near the apex of the body, ze. ocelli, olfactory

organs, and lateral eyes, those on the most anterior segment show

the greatest amount of median fusion. Conversely the degree of

median fusion is an index of the original sequence of organs.......... 62
Further degeneration of embryos........................ shade de aaed encmease Meee. ease 63
Reduction to separate head and tail rudiments. .....0........-.eececseceeeeeeee-s 65
Reduction to separate invapinated ‘Sacs.........c.cc:t..205-s0ls.cseseeeeeeess 65
Completeidisappearance Gf latter: <..-c:--22:cscreceesacenctc cesta ee 66
Comparison of same with primitive cumuli and with the primitive

head and itronk of an annelid larvac 2c). .c:2.c---<c4-ccsessceaececeeseetee sent 66
Bearaneon. trochosphere. theory: ..-.s2.c.seccc te eee oe 66

Wid RSIS C0 <1 act Pes Oe BO a ne ee EE ED eee? SEL DW 66
Absence of multiple embryos in early stages of development............ 66
Transverse fission and longitudinal fission. Fundamental difference

betweemitnestwoumethods: 2.0 82)22....09 he ee ee 67
Transverse fission due to local degeneration and median fusion,

longitudinal fission to the intercallation of new organs.................... 67

BFL REID CRSEM TESSIOM 6) ac: a Sdscnnansszacteoxs kage see ee ee 67
Potential fission plane, indicated in abnormal embryos by: (a) space
between third and fourth pairs of appendages ; (4) difference in

the inclination of appendages on either side of this line; (c) con-

striction in region of fourth thoracic segment, due to median

fusion and degeneration of adjacent organs. ...............::10:c10-ee 68
Ostenorn portion) shows preatest vitality. :s:se = ace oe 70
Similar potential fissure planes in region of chelicerae and cheliceral

SEPIeN Gm eeeese Nf cesta SO Be See eae) Oe eee eon 71
Bearing on the morphology of trunk regions throughout segmented

animals) (see:also SOG: LV) \stoscccetssgeccceseses. aeececee ct oes cones ee meee 71

B. Longitudinal Fission; Double and Triple Embry u....c1ce2ecceen- 71
INotidueito:tusion of separate embryos ies ce se.scsess = ss ece eee eee neces ies
1. Double Embryos.
Due to intercallation of new halves. New and old halves always

matchweach) ‘other! exactly. .cs:f 208 oo. .scesoc-ncsssecvescusee seeceee caseeee seescearere eae 72
ew organs appear in reverse order of their disappearance by median

fuston. When new parts are added, organs morphologically the

oldest appear first, as in ordinary growth, but in the reverse direction,

— that is, from before backwards............-. dain ceieefasav cap tedznasns te scotaagasssewoustace 73
Wawa formatlon! Of NeW; OLEAN S oer cseee se sec eee este eee meena es 74
Axis of repetition and axis of differentiation at right angles to each

OLDER ene ie ate lees sebcebtbaSe conte scnetscesaee eRe ese an rama tee HCL ReE cee 74
Splitting of unpaired appendages begins at apex and extends towards

the base of the appendage: this is the exact reverse of what occurs

MM), MVE AM LUSL OM! < cn cs- ccc. csescnssoncasedasscdasnushecechseetbecccseanwes tec accnenerosstesseoe 75

20

PATTEN. [Vou. XII.

Pace

Internal differential tension due to movement of the embryo over
the yolk has no effect on the shape of organs..........-2-2...:-00-+ 76

Causes influencing complete separation from each other of double
TINE Ty, OS! Serene cera cones aaa ne ae eae cane genera rere ee oa ne hte rane ene 77

Formation of new embryos by excessive growth, followed imme-
diately, or accompanied, by degeneration = see ere eee 80-81

2. Triple Embryos.

Due to addition of four new halves to the old embryo.............2...0..... 82
Relation ‘of newihalves)to) Oldies tse nee essere cree secre eo ce 83
Mother, daughter, and granddaughter hearts. ................:::-eesceccceeceeee-os 83

Difference in age of the individuals in triplets is not primary cause
of their difference in degeneration.

Discussion of observations on defective and exuberant embryos.... 84-88

Various classes of facts and explanations to be harmonized before
we can arrive at any explanation of the phenomena observed in
triple embryos, viz. :

(1) There is variation in the duration of the growth period of the
whole, or of a part of the organism. We may assume this to be
due to variations in the intensity of growth, or differentiation,
forces; but not to the lack of formative material.

(2) There may be an entire absence of certain organs at the outset
and a subsequent failure to regenerate the same. This we may
assume is due to absence of specific formative material.

(3) General or local structural weakness of an embryo is indicated
by median fusion of its organs in the reverse order of their age
and specialization.

(4) The existence of abnormally large embryos may indicate forced
growth or an excess of formative material. Multiple embryos
may be explained in the same way. In either case excess of
material or excessive growth gives rise to the formation of new
organs in reverse order of that in which they disappeared by
median fusion and degeneration.

(5) When the formation of new organs is followed by degeneration,
we assume that the multiple parts are formed not from an excess
of formative material, but by a kind of forced growth obtained
by drawing from a fixed capital.

(6) The same explanation must apply to the cases where the indi-
viduals of triple embryos are at first very dissimilar, but finally
before their complete degeneration become equal in size and
structure.

No production of new formative material by abnormal environment,

Multiple embryos probably due to abnormally rapid, or forced, growth.

New organs in multiple embryos are formed at point of greatest
differentiation tension ............... ES OTT She ane oor mee Le LS 89

Comparison of recuperation of weaker part from stronger, in multiple
embryos, and absence of same in single embryos.........-.-.-..------0--+-++ go

Explanation of more frequent presence of asymmetry at posterior
end of, body than at anterior. .....2¢ =: -0 eee ee go

Nort) ) VARIATIONS IN LIMULUS POLVPHEMUS. 21

PaGE
Variation in the growth force is merely a convenient term to express
the results following indirectly from variations in internal condi-
tions. It is not a direct expression of variation in the quantity
of some particular kind of energized matter. Variations in the
internal conditions produce an apparent increase, diminution or
subdivision of the growth force, because those conditions deter-
mine the possibilities for a longer or shorter or different sequence
of chemical changes, that express themselves to us in change of

POTMAGPR cassia a Cate Boao Bcf ted sa decors oh eee ee eee, 90-94

VII. DEGENERATION AND DEATH OF LIMULUS EMBRYOS...........:.0:::c00000---- 94
Structure and Function of the organism a resultant of (1) Produc-

tion, (2) Specialization, (3) Longevity, (4) Death rate of cells........ 94

Death of Limulus embryos due to a gradual reduction in numbers
and in specialization of the constituent cells till only very few and
very simple cells remain. These also finally disappear.................... 95

Comparison of this kind of death with the so-called “natural death ”
due to the impossibility of adapting new and old parts to the new

conditions entailed! iby their own! growth. -o1.cccececccececaceseeeseeeecseveeeeeee 95
No necessity for assuming that there is an essential difference be-
tween body and germ cells. ‘ Mortality” is not necessarily inher-

ent in any living organism ; it is rather the result of an imperfect
organization in which growth, repair, and waste continually pro-
duce new conditions unfavorable to the continuation of the same,
and to which the organism cannot adapt itself... 96-100

INTRODUCTION.

Most of the material for the following paper was obtained
during the summer of 1891 at the U. S. Fish Commission
Laboratory at Woods Holl. The preparations and drawings
were made during the following winter. But since then, from
time to time, new material was obtained, till the number
and variety of abnormal embryos at my disposal became
very great. The principal value of the material lies in the
large number of abnormal embryos, and in their range of
variation, from nearly normal ones to those so modified as to
leave a hardly recognizable being behind. To make this range
and diversity of variation obvious, great pains were taken to
present as large a number of surface views as possible, made
from stained and mounted embryos. In nearly every case the
embryos were subsequently sectioned, and used as a guide for
the interpretation, or correction, if necessary, of the surface

22 PATTEN. [VoL. XII.

views. But the latter usually tell the whole story, so that but
few figures of sections have been deemed necessary.

The amount of good material was so great that I have in no
case used embryos for illustration if there was the least doubt
of the abnormality being a real one, and not one due to post-
mortem abrasion, shrinkage, or other causes of like nature.

The present paper is an unexpected by-product of work
along other lines. The time necessary for this digression
from a prescribed course was somewhat grudgingly given. It
is partly for that reason, and not because their value for this
purpose was unappreciated, that I have not entered into a
critical discussion of prevalent theories of heredity and devel-
opment in the light of these new facts, although I have ventured
to make a few suggestions of a theoretical character that natu-
rally arose out of their consideration.

It seemed to me that about everything of value in the way
of argument has already been said and resaid on the various
phases of epigenesis versws evolution. When the smoke from
the volleys of words discharged in the last few years has
cleared away somewhat, it will probably be found that the rival
disputants are in closer agreement than has been suspected.

But, after ail, the most convincing arguments are the plain
solid facts. They are always eloquent, and we cannot have too
many. One who has often had occasion to search the litera-
ture on a given subject for definite information must be
impressed with the uncertainties that hover thickly around the
majority of observations, and which render them practically
worthless for constructive purposes.

While it is not claimed that the following descriptions are
less open to this criticism than most others, it may be a point
in their favor that, almost to the last, the embryos were selected
and drawn with great care, merely as an illustration of isolated
cases of variation, and without any preconceived ideas as to
their meaning or mutual relations.

It was my intention originally to publish a description of the
normal development first, as I had abundance of material care-
fully prepared and studied for this purpose. The reader would
then be better able to appreciate the meaning of the abnormal

INOS i] VARIATIONS IN LIMULUS POLYPHEMUS. 23

embryos. But in view of the long delay that might attend the
publication of the first part, and among other reasons, on account
of the interest now centred in variation and abnormal develop-
ment, it seems advisable not to delay further the publication
of the present paper.

But in order to afford some ready means of comparison with
normal embryos, prepared by my methods and interpreted on
the same basis as the abnormal ones, Pl. I has been introduced,
in which are represented the more important stages in the
development of normal embryos.

Methods of Hardening, Staining, and Clearing the Embryo
for Surface Views and for Sections.

Surface views of opaque embryos are useful for some pur-
poses, certain points being brought out with special clearness
shortly after the eggs are put into the hardening fluid. But
in order to make out many important details, it is absolutely
essential to stain the egg, and clear in clove oil, balsam, or oil
of cedar. The latter often gave the best pictures and the eggs
could be kept longer in this fluid without discoloring the yolk,
than in oil of cloves.

To obtain the best surface views, the embryos should be
stained and mounted as soon as possible after hardening, or if
that is not convenient, preserved in perfectly clean alcohol in
glass-stoppered bottles, as the tannin, or other substances in cork
stoppers, is dissolved out by the alcohol and discolors the yolk.

Either picro-nitric or undiluted picro-sulphuric acid or
Perenyi’s fluid may be used for hardening. The eggs should
be immersed in ¢he cold solution from ten to twenty-four hours.
The yolk is thus made quite hard, and the membranes swell so
that they can be easily removed by fine-pointed forceps. The
membranes must always be removed before placing in alcohol ;
otherwise they shrink, and the embryos are injured or dis-
torted. Moreover, a white albuminoid substance collects under
the membranes, and is precipitated on the embryo when it is put
into alcohol, obscuring very much the beauty and clearness of
the preparations.

24 DHT AINE [VoL. XII.

After the eggs are shelled they are rinsed in the hardening
fluids and transferred to a large quantity of strong alcohol, say
94%, which is changed frequently the first few days. If not
treated in this way, the yolk is likely to swell and crack. _ Brit-
tleness of the yolk is much diminished by the prolonged treat-
ment with acids.

The most beautiful surface views are obtained by staining
the whole egg in borax carmine, or almost any haematoxylin,
for a very short time, one-half to one or two minutes, and
then washing in acid alcohol. This method gives very sharp
and luminous surface contours. It has been used on a great
variety of objects, arthropod and vertebrate embryos, etc., and
is very useful.

But this method cannot be used if there is any surface cuticle
present. The whole egg must in that case be stained through-
out, and the color subsequently drawn completely out of the
yolk by acid alcohol (10 to 15 drops strong hydrochloric acid
to 100 c.c. of 70% alcohol). The decolorizing process may
require several days, and the acid alcohol must be changed as
often as it is discolored. When this is successfully done, the
yolk has the transparent, yellow color it had before staining,
but the nuclei are bright red.

If the eggs are to be mounted, after clearing in oil of cloves,
they are split in halves with a delicate knife, made by grinding
the end of a needle down to a very thin blade.

The halves with the embryos on them are arranged like serial
sections, in shallow cells, and fastened in position with a small
drop of thick collodion and clove oil. The fixative is hardened
by washing in turpentine, and finally the eggs are mounted in
balsam.

Each embryo was numbered, drawn, and in most cases finally
taken out of the balsam and sectioned.

The embryos were studied with raised condenser and wide
open diaphragm, so that they appeared bright red on a clear
yellow field.

In such preparations the elevated surfaces and protruding
organs appear dark, the depressions light. All the drawings
have been outlined with the aid of a camera from such prepara-

No. 1.] VARIATIONS IN LIMULUS POLYPHEMUS. 25

tions, and, with few exceptions, drawn to the same scale, so that
the difference in size of embryos of the same age is very evident.

Methods of Obtaining Material.

My attention was first drawn to the presence of abnormal
embryos, by finding a nest in which about 90% of the eggs
were about ready to hatch, while the remaining ones were
apparently in very early stages of development. A few of the
latter contained double monsters. A number of different nests
were then examined, and many were found containing abnormal
eggs. The number of abnormal eggs ranged, at a rough esti-
mate, not by actual count, from none to about 10%, or perhaps
more.

In order to obtain a greater abundance and variety of abnor-
mal embryos, about 50,000 eggs, from many different nests and
in different stages of development, were collected. The sand
containing the eggs was placed in shallow dishes and stirred
about by strong currents of water till the eggs rose to the
surface, when they were poured off into other jars and washed
again till perfectly clean. Placed in hatching jars, they sink
to the bottom, and are kept constantly but gently agitated by
a current of pure sea water. As fast as the young larvae
begin to swim about, they are swept out of the top of the jar
through the escape pipe.

After from eight to ten weeks most of the larvae have
hatched. Many thousands of the larvae thus obtained have been
examined, and no abnormal embryos found among them.

In the residue left in the jar are many apparently normal
and well advanced embryos, some of which hatch from time to
time for many weeks or even six months or more after the
regular period of hatching has passed. All those that eventu-
ally leave the membranes, whether early or late, and swim about
freely, are, with very rare exceptions, normal. Among the
remaining eggs are found embryos in all stages of degenera-
tion. A very few only are dead and decomposing. The rest
appear fresh and healthy, and one would not suspect from their
general outward appearance that they were abnormal embryos.

26 PATTEN. MOL ex is

The eggs left in jars that had been running from ten to fourteen
weeks were finally treated with hardening fluids, making the
shape of the embryos very conspicuous. The whole collection
could then be examined under a dissecting microscope, and
the most important forms picked out and shelled at once.

As many embryos were already far advanced (six to nine
weeks) when placed in the hatching jar, it is evident that the
remaining abnormal forms were at least ten to fourteen weeks
old, and some may have been even more than twenty, although
as seen by the figures, many of them appear to be not more than
from ten days to two weeks old. Most of the monstrosities are
apparently the result of retrogressive development, rather than
of a slow, progressive one. That is, embryos normal in out-
ward appearance may develop up to stage #, Pl. I. They then
appear to develop less rapidly, or they may remain for a long
time practically stationary. But they finally become smaller,
and by the fusion and complete atrophy of their various organs,
dwindle to some insignificant remnant which is in turn absorbed.
Thus a once well-grown embryo may disappear completely,
leaving a healthy-looking egg behind, but one which consists
of nothing but yolk.

I was able to confirm the above statement by selecting a
dozen or more abnormal embryos and keeping them under
observation for four or five weeks. It is difficult to make out
details on living eggs, but enough could be seen to prove
beyond any doubt that in each case the characteristic abnor-
mality, such as asymmetry, fusion, or reduction of appendages,
etc., became more and more marked from day to day, until my
observations were brought to a close by the complete disap-
pearance of the embryos, due to the increasing opacity of the
chorion. On killing and staining these eggs, the embryos
could be easily distinguished again, in advanced stages of
degeneration.

In other instances I have kept large numbers of the residual
eggs in open dishes, for five or six weeks after they were taken
from the hatching jars. In such cases, the number of indi-
viduals showing extreme median fusion, or atrophy, or general
degeneration, seemed to be greater than before. The number

INO: 1.) VARIATIONS IN LIMULUS POLVYPHEMUS. 27,

of abnormal embryos seemed to be greatest in the nests taken
from sand that was blackened and foul from decaying organic
matter (decomposing eggs?); but on the other hand, they
were often found in abundance in perfectly clean sand close to
nests where nearly all the eggs were normal.

It occurred to me that the abnormal eggs were produced by
the unusual conditions in a hatching jar, such as the constant
movement and the exposure to light, or perhaps to the differ-
ence in the temperature or density of the water.

This last summer, therefore, hoping to obtain new classes of
variation, about 25,000 eggs were placed in shallow dishes.
Some were kept in the dark, others in the direct rays of the
sun. The water in some dishes was allowed to evaporate
almost to dryness, leaving a thick crust of salt in them. A
large quantity of fresh water was then added. Under this
treatment, many of the older larvae died, and their bodies were
allowed to putrefy in the dishes, so that the water became very
foul. At the close of the season, all the eggs that survived
these indignities, about two-thirds of them, were brought to
Hanover, and kept in shallow covered dishes, exposed during
some part of the day to the direct rays of the sun. Up to No-
vember, the water was turbid and filled with bacteria. Shortly
after that, algae began to grow, and now cover the sides and bot-
tom of the dish with athick, green scum. In December, it was
thought a large number of abnormal embryos might be obtained,
and about 5000 were killed and examined, including trilobite lar-
vae and unhatched eggs, but nota single abnormal larva or embryo
was found in the lot! This fact is not easily explained, because
in any case we ought to find a certain number of abnormal
forms. But it seemed probable that all those embryos origi-
nally abnormal were early exterminated by this drastic treat-
ment, and only the normal ones survived, that is, normal in
every respect except their very slow development, and this
was probably due in part at least to the increased density of
the sea water through evaporation.

28 PATTEN. [Vou. XII.

Causes of Variation.

It is assumed that all the embryos were at first normal in exter-
nal appearance, and that the condition in which they were found
was attained by a gradual fusion and atrophy of organs, according
to the methods described under the various headings. But we
have always had in mind the possibility, indeed the strong proba-
bility, that in some cases the normal form was not actually ex-
pressed, but that the embryo from the first appeared in some one
of the various stages of fusion and degeneration. For example,
when a particular appendage is completely or partially invagi-
nated, we have no means of determining whether it was first
normally formed and was subsequently invaginated, or whether
at the moment of its first appearance it gradually assumed the
condition in which we find it. The one condition may be
regarded as the resultant of the simultaneous action of normal
and abnormal factors, the other the resultant of the abnormal
factors acting after the normal ones have found expression.
No number of cases, however great or however varied, can
settle that point. But a complete series of them properly
arranged may be taken to indicate the successive stages a
fully developed normal appendage would assume when subse-
quently invaginated, and the same applies to any of the other
modifications that have been observed.

All variations of the same class are no doubt due to the
action of similar abnormal factors that tend to throw the
normal mechanism out of its beaten track. But while simi-
lar factors will tend to produce similar divergences in the
end, the nature of the preceding forms will depend on the
intensity of their action and on the period in the develop-
ment at which they begin to act. The variations first to
appear are of the greatest import, because “the ‘at | first
slight divergence becomes continually greater till it leads
to impossible combinations which may necessarily be fatal
to the future organism. Moreover, an organ on its way
toward a position of stability is more readily affected by
external agents than one that has settled down as it were
into that mature and stable condition characteristic of all older

No.1.] VARIATIONS IN LIMULUS POLYPHEMUS. c9

organs. But there is nothing to indicate with certainty
whether the initial cause of variation is due (1) to a variation
in the combination, or the quantity, or the quality of the
original constructive materials, or (2) to the variation in condi-
tions external to the ovum, or (3) to a combination of both.
But as very divergent forms appear among eggs kept under
apparently the same conditions, it would seem more than likely
that the first set of causes are the real ones.

Variations in the unfertilized ova, and in the spermatozoa
or polar globules, at once arise before the mind in their
familiar attitude, as factors in some way connected with the
phenomena of variation. And there is the whole infinitely
complex Weismannian mechanism, with its endless army of
corpuscular brownies, on whose sins of omission or commission
we may easily throw the responsibility.

But these corpuscular theories fail to explain anything. Their
agents come or go at the beck and call of him who commands
them. We are in the end left with the sterile formula, that
this or that organ is as it is because the necessary corpuscles
were there to make it so !

All the variations so far observed can be traced back to
variation in the relative rate of growth, specialization, and
degeneration in the numerous groups of cells that constitute
the embryo. The problem then is to explain why one group
of cells grows faster or slower in one embryo than in another,
or why it grows at all. But the conditions determining
differential growth are very complex, and may be different in
each particular case. No simple general statement will suffice.
We have passed that stage where it is enough to know that a
machine is made of iron and run by electricity. We must
know what electricity is, and the structure and bearing of
every pin, screw, and wheel in the whole mechanism. It is
needless to say that we are very far from having any such
knowledge even of the simplest bit of living matter.

The facts here presented enable us to catch a glimpse of
embryological processes under new conditions and from a
different standpoint, and while many of them increase rather
than diminish the existing difficulties, they will perhaps, in

30 A TMEV [VoL. XII.

some future theories of development, find their proper place
and partial explanation.

It seemed inadvisable to add further to this paper by detailed
reference to the voluminous literature bearing only indirectly
on the facts here set forth.

So little is known about variations in arthropod embryos,
and especially arachnids, and the facts we present are so differ-
ent in character from those already known, that no injustice
will be done previous workers along similar lines by not refer-
ring in detail to their publications, The only reference to an
abnormality in Limulus that I have been able to find is in an
article on “ Diploteratology, An Essay on Compound Human
Monsters,” by Geo. J. Fisher, Albany, 1868. A good figure
is there given of an adult (?) animal with a double caudal spine
and a symmetrically forked abdomen.

DESCRIPTION OF THE DIFFERENT CLASSES OF VARIATION.

I. INVAGINATION OF APPENDAGES.

This remarkable modification is of comparatively common
occurrence in forms which are in other respects more or less
abnormal. It is confined, so far as I have observed, to the
thoracic appendages, and most commonly affects the middle
ones of the series.

It may begin after the appendage is fully formed as a minute,
slit-like depression at its distal end, Figs. 10, 11, th. ap,
th. ap. The slit is always transverse to the long axis of the
body, and appears in the stained specimens as a fine line in
the middle of a clear band devoid of nuclei. When the invagi-
nation is complete, the whole appendage is carried inward, so
that in its place is an opening leading into a deep tube with a
flattened, conical lumen.

The third or fourth appendages on either or both sides may
be invaginated, Figs. 10, 11, or, as in one case —the only one
observed —all the thoracic appendages, with the exception of
the first and sixth pairs, may be invaginated, the infolding being

No. I.] VARIATIONS IN LIMULUS POLYPHEMUS. 31

greatest in the second pair, and diminishing in depth from that
point backwards.

The embryos just described are in other respects nearly
normal, the principal deviations being in the abbreviation and
atrophy of the abdominal region in Figs. 8 and 10, and the
modification of the cephalic lobes in Figs. 10 and 11. But
invaginated appendages are frequently found in other types of
embryos, as in Figs. 14, 16, 18, 33, 34, 40, 61, 65, 70, 98.

Any appendage, except perhaps those of the first pair, may
be invaginated ; the third pair appear to be most frequently
affected in this way.

Sections have been cut through several invaginated appen-
dages, and confirm completely the conclusions drawn from
surface views. They throw little further light on the sub-
ject. Pl. X, Fig. 10, represents a longitudinal, vertical section
through the embryo shown in Fig. 10. The only point of
further interest here is seen at the inner end of the appendage,
where the ectoderm seems to become continuous with the
mesoderm, as though there was an inward cell proliferation at
that point. If such were the case, a communication, through
the hollow appendage, might be established between the exterior
on one side, and the mesenteron on the other. A condition
would then result like that which obtains in the gill slits of
_ vertebrates.

As to the cause of the invagination, very little can be said
further than that it is a local, internal, rather than an external
one.

We cannot see how any of the general external conditions,
such as the density or the composition of the surrounding
medium, could produce such a purely local effect as the invagi-
nation of one out of several apparently identical appendages.

There is no evidence whatever that local pressure, such as
that which the egg membranes, or the adjacent appendages,
might exert, was the cause of the invaginations, for the mem-
branes stand far away from the embryos at this period, and
besides, an examination of most any of the figures shows
clearly that the position of the invaginated appendages is such
that the membranes could not touch them. Such cases as

32 PATTEN. [VoL. XII.

those shown in Figs. 10, 11, or 65 could not be due to pres-
sure of membranes or of adjacent appendages.

It seems to me, therefore, that the immediate cause is a
local, internal one, independent, to the same extent as the
normal growths, of external conditions, and having its source
in some remote instability of the innermost mechanism.

This class of variations may therefore be called xormal vari-
ations, to distinguish them from those due to the more imme-
diate action of the environment. They must be regarded as
necessary incidents of a particular structure and likely to occur
in a certain percentage of cases, irrespective of the immediate
environment.

In vital processes, what may be at one time or place an inci-
dental and occasional phenomenon, may become elsewhere
under other conditions a constantly recurring result. It is
therefore clear that in estimating relationships by means of
morphological characters, such variations deserve careful con-
sideration. They are as likely to throw light on phyllogenetic
problems as ontogeny. The latter indicates the established
paths connecting the present with the past; the study of
normal variations shows us possible paths leading out of the
present into the future.

These facts, then, are interesting morphologically in two ways:

(1) They may be regarded as forming an indirect confirma-
tion of the view that the lung-books of scorpions and spiders,
as claimed by Lankester and Kingsley, are invaginated gill-
bearing appendages, modified for breathing air.

(2) I have maintained in a former paper on the “Origin
of Vertebrates from Arachnids,” that the complete visceral
arches of vertebrates are homologous with the appendages of
an arachnid-like ancestor, because in Limulus and scorpions,
the number of these appendages, their innervation, the position
of their important sense organs, the structure of the meso-
blastic cavities associated with them, and the nature of the
muscles arising from these cavities, resemble as a whole the
corresponding structures in vertebrates.

Aside from other considerations, the striking difference
between arthropod appendages and the gill arches of verte-

No. 1.} VARIATIONS IN LIMULUS POLYPHEMUS. 33

brates renders any relation between them at first sight very
improbable. But the unexpected discovery that in this par-
ticular arthropod the appendages are frequently invaginated,
leaving in their place a series of slit-like gill openings, such
as those shown in Fig. 1, is a potent factor in favor of my
view.

If in Fig. 8, the thoracic appendages had been provided
with gill leaves like those on the abdominal appendages, we
would then have, in place of typical arthropod appendages, a
series of respiratory sacs, the cavities of which, by the persist-
ence of embryonic conditions might, after the manner of
vertebrate gills, communicate with the alimentary canal, either
through its nephridium and somite, or by a secondary opening
at the apex, where ectoderm and mesoderm appear to be con-
tinuous. The exact relations in Limulus of the invaginated
appendages to the mesoblastic somite, and of the latter to
endoderm have not been determined, as in the cases studied
the cavity of the somite had disappeared. But there seems to
be no reason to doubt, from what occurs in the normal embryos
of Limulus and other arthropods, that something like the
condition mentioned above might arise, for each of the assumed
conditions is known to occur.

The great difficulties involved in determining what is ecto-
dermic, mesodermic, or endodermic in the vertebrate head, we are
just beginning to realize. The use of the terms is founded on
the supposition that the embryological processes in vertebrates
present a modification of those assumed to occur in some real
or imaginary invertebrate, which is further assumed to be a
more or less remote ancestor of the vertebrates. While the
problem is still in this uncertain condition, the fact that these
‘a priori’? methods of interpretation do not harmonize with
the suggestions here made cannot be used as an argument
against them.

II. ABSENCE OF APPENDAGES.

We shall consider all embryos showing the absence of one
or more appendages under the above heading. In all cases
the reduction of an appendage seems to be accompanied by a

34 PATTEN. [Vor X1k

degeneration more or less complete of all the other organs on
the same half metamere, but as these organs are less easily seen,
I have confined my statements in the main to the appendages.
The degeneration of an appendage is an indication that degen-
eration has taken place or will take place in the other organs of
that half-metamere, but the reverse is not true, for an appendage
frequently persists long after all associated organs have disap-
peared. Absence of appendages is one of the most common
abnormalities, and like the invagination of appendages is most
likely to occur in embryos showing other indications of abnor-
mality. Zhe three anterior thoracic, first, second, and third, and
the abdominal appendages are most frequently absent. The three
posterior pairs of thoracic appendages are found so frequently
after everything else has disappeared, that one might suppose
a nauplius-like larva was a normal feature in the development
of Limulus, but a little consideration will show that such is not
the case. t)

While the appendages are frequently absent in pairs, leaving
bilaterally symmetrical embryos, this is by no means the rule.

A. Absence of Abdominal Appendages.— This is perhaps
the most common of all the modifications observed. A great
many of the abnormal embryos showed some abbreviation of
the abdomen, accompanied by the reduction or the entire ab-
sence of the appendages.

Fig. § shows the condition of the abdomen in normal
embryos at the time all the thoracic appendages have appeared.
The primitive streak is still present as a narrow, longitudinal
furrow, from the floor of which there is an’ inward.proliferation
Oficells: >

One frequently finds embryos, otherwise normal, in which
the marginal fold, f., which represents the margin of the
future thoracic and abdominal shields, sweeps across the
median line just back of the first pair of thoracic appendages.
They are so much like what occurs in Fig. 8, that it was not
necessary to represent them. In some of these cases, the

1T have unfortunately not tabulated all the cases seen, but I havea strong
impression that degeneration is more likely to affect a half metamere than a
whole one; as though the former were the unit of structure rather than the latter.

No.1.] VARIATIONS IN LIMULUS POLYPHEMUS. 35

abdomen is perhaps merely retarded and may appear later in
its normal condition. This is probably the case with Figs. 9,
10, and 15.

The TLelopore.—In the normal embryos, at a very early
stage, the posterior end of the body, or the anal plate, consists
of a broad mass of proliferating cells, representing a modifica-
tion of the posterior of the two primitive cumuli that constitute
the beginning of the embryo.

In Fig. 1, there is already formed a primitive streak-like
invagination in the anal plate, representing a specialization of
this proliferating area. Along this proliferating furrow, ecto-
derm and inner-layer cells become continuous. Without fol-
lowing out its history in further detail here, it is enough to
state that it persists, either as a solid mass of cells or as a
furrow, up to stage C, after which it gradually disappears. The
anus does not appear on the flat abdominal plate till about
stage £, Fig. 7, long after the primitive streak has disappeared.
But these normal conditions are frequently deviated from,
giving rise to a great range of variations.

[n many cases when the abdomen ts abbreviated or absent, the
entire remaining region, where the abdomen should be formed, ts
tnvaginated to form a deep depression, varying greatly in size
and general appearance. ‘This depression, or ¢elopfore, is not a
modification of the anus or of the primitive streak, for one or
the other of them may, in some cases, be found at the bottom
of the depression. It seems to be the result of the active pro-
liferation that is going on in this region during the formation
of new segments.

A common form of the telopore is shown in Figs. 9, 10, and
fiemanotieran Migs), 16, 19, 21, 22, 23) 24, 36; 50,62, 07,69.
In Figs. 20, 21, 23, 51, the whole abdominal region is so deeply
depressed that the posterior thoracic and abdominal appendages,
when they are developed, are carried into the cavity.

At the bottom of the depression, I have often found in sec-
tions the proliferating primitive streak, not markedly different
from that in the normal position.

The anus may finally appear at the bottom of the telopore,
after the primitive streak has disappeared. In some cases, it

36 PATTEN. [Vou. XII.

is difficult to distinguish the telopore from the anal invagina-
tion, or in other cases from the primitive streak, so great is
the range and number of modifications that may be observed.

The infolded abdominal region seems to straighten out in
some cases, and in the end give rise to a normally segmented
abdomen. In the majority of cases, however, its presence
indicates a general weakness that leads ultimately to complete
degeneration of the whole embryo. In some cases the defect
persists to very late stages, for trilobite larvae are not uncommon
that are perfectly formed except for the abdomen, which may
show any one of the numerous stages of degeneration.

Whether these defective larvae die, or the abdomen is
restored after successive moults, is not known. I have looked
in vain for traces of aborted abdomens and in fact for any kind
of abnormalities in individuals older than the tribolite stage.

B. Asymmetry of Abdominal Appendages has been observed
in a large number of cases. This condition is well shown in
Figs. 29 and 31, where there are three abdominal appendages
on the right but no trace of them on the left. In Fig. 29 the
abdomen is thrown round to the left by the unequal develop-
ment, and a peculiar, hood-like fold of ectoderm covers its
posterior portion.

In Fig. 30, there are four abdominal appendages on the left,
and only two on the right. In Figs. 32, 37, 38, 53, 54, there
is strongly marked asymmetry.

C. Absence of Thoracic Appendages. In the normal em-
bryos, the second, third, and fourth thoracic appendages appear
simultaneously, and shortly afterwards in the order named,
the first, fifth, and sixth.

In the abnormal forms, any one thoracic appendage, or almost
any combination of the twelve, may be absent, but, as one might
infer, from what has been said in reference to the invagination
of appendages, it is very difficult to determine in any given
case whether the appendages failed to develop, or whether
their absence is due to degeneration.

In order to obtain some light on this point, a dozen or more
living abnormal embryos were kept under observation from
the 20th of July till the 17th of August. Most of the eggs

No.1.] VARIATIONS [IN LIMULUS POLYPHEMUS. 37

either died or became so opaque that after a few days nothing
was clearly visible in them. It was, however, clearly estab-
lished, by means of careful drawings made from time to time
on such eggs as were transparent enough to allow one to follow
the changes going on within them, that a gradual atrophy of the
anterior and posterior extremities took place, while the whole
embryo showed a marked decrease in size. These facts were
enough to show that progressive degeneration did take place in
these embryos, and it is very probable that a similar gradual
atrophy would have taken place in most of the cases of
abnormal embryos here cited, if they had been allowed to live
long enough.

The simplest cases of the absence of thoracic appendages
are those where one or more appendages are absent on one
side of the body, as in Pl. IV, Fig. 30, where the right second
and third are absent. Similar cases are shown in Figs. 13, 14,
ZOO 32,33; 34, 35, 30, 37, and others:

These cases are common, but such as that in Fig. 38 are
extremely rare, only three similar ones having been seen. In
this embryo the whole of the left cephalic lobe, nerve-cord, and
mesoderm, — everything, in fact, except what appears to be the
left sixth appendage, is absent, while on the right side every-
thing is perfectly normal, except for the slight spiral curvature
to the left due to the unequal cell stress.

A very frequent form of abnormality is where there has
been a nearly bilaterally symmetrical reduction of the anterior,
or the posterior, end of the embryo, or both. The cases, in so
far as they affect the abdomen, have already been considered.
This process carried still farther would affect the reduction of
the posterior part of the thorax, but that condition seems to
be comparatively rare. There is on the contrary a tendency
to leave this region intact, reducing instead the three anterior °
segments one by one, from before backwards, producing con-
ditions like those in Figs. 16, 18, 24, 25, 26, 27. The small
embryos with but three pairs of appendages represent the last
stages of the process, and they are so abundant that I at one
time supposed they might represent a true wauplius stage. It
may be that similar embryos gave rise to the statements of

38 PATTEN. [Vou. XII.

Dohrn and Osborn to that effect. But that they are not
nauplii is obvious because they are mere fragments of embryos,
minus brain, oesophagus, and one or more of the anterior
thoracic segments. One cannot always determine with cer-
tainty just what appendages are preserved in these false
nauplii, but in most cases they appear to be the three posterior
pairs of thoracic appendages.

There are three reasons in favor of this view: (1) In some
cases that clearly belong to this category, the cephalic lobes
and the two or three following segments are quite rudimentary,
showing marked anterior degeneration, while the three pos-
terior appendages are of fair size. In Fig. 25 is a typical case.
Here the cephalic lobes are reduced to a rounded plate of cells
with the remnant of the oesophagus in the centre, and they
are separated from the thorax by a wide space along which no
organs are developed.

(2) In cases where there has been anterior degeneration,
accompanied by the fusion of right and left halves, the last
three pairs of appendages (known to be such by the presence
of the flabellum on the sixth) undergo the least modification or
degeneration, and persist long after every organ in front of them
has disappeared. Figs. 44, 48, also 94, 96, 97, 102-104.

(3) When transverse fission occurs, it divides the embryo
between the third and fourth pairs of thoracic appendages into
two parts; the posterior one usually persists for a long time
after the first part has disappeared, showing that it has the
greatest vitality. Figs. 51, 98, 103.

It is: therefore very probable thatin Pigs. 25,26, 27,-33.734,
35, 62, 65, 66 the segments present are the 4th, 5th, and 6th
thoracic. But it must be observed that in Figs. 31 and 32 the
second and third pairs of appendages are much better developed
than the posterior ones. And also in Fig. 27 the rudiments of
the oesophagus and cephalic lobes lie directly in front of the
first pair of appendages. The same position of the cephalic
lobes and oesophagus is observed in Figs: 33,.34, 45, 62,05,
66,67. It is thus obvious that entire segments may be omitted
without leaving any apparent break in the series. This is
especially clear in Fig. 27, where three thoracic segments are

No. 1.] VARIATIONS IN LIMULUS POLVPHEMUS. 39

certainly absent, because the cephalic lobes and abdominal ap-
pendages are in their proper relative positions, and yet only
three out of the six thoracic segments are present. It cannot
therefore be argued that because in Fig. 62 the oesophagus
lies just in front of the first pair of remaining appendages,
that they are the original first pair and not the fourth.

Figs. 33, 34, 35, 65, 66 probably represent modifications
of this three-legged form in which the appendages have
undergone various modifications through invagination, median
fusion, and degeneration.

In Fig. 34, the last pair have fused to form a median
appendage, and in Fig. 35 there has been such distortion and
reduction that it is hard to recognize what remains, but it
appears to consist of the three left thoracic appendages and one
right. The oesophagus lies at ve., and the telepore at &., at
the bottom of a furrow partly covered by a hood-like fold.

In most of these small three-legged forms, the median ventral
surface is deeply depressed and closely surrounded by a thick
and high marginal fold.

III. MuvLTIPLICATION OF APPENDAGES.

Multiplication of definite regions or organs of the body, not
including in this category double or triple monsters, is very
rare. JI have observed but one case. It is interesting and
important, since it proves conclusively that a half of a metamere
already laid down has the power to multiply independently of
adjacent organs. This is shown in Pl. II, Fig. 12, where the
right chelicera— which fortunately can be identified here with
absolute certainty has divided twice. The first division
apparently gave rise to a’** and a***. A subsequent division
separated completely a? from a*. But the third division
effected only a partial separation of a‘ froma’. The abnor-
mal growth has also modified the right nerve-cord at this
point, throwing the head over to the left. None of the
neuromeres are very clearly brought out in this preparation,
so it is not certain whether there is one for each of the four right
chelicerae or not. But there is a prominent enlargement of the

40 PATTEN. [VoL. XII.

right nerve-cord, s, which from its shape and position appears
more intimately related to the third left chelicera than to
any of the others. What appears to be the neuromere of
the fourth right chelicera is small and triangular, and pushed
to one side by the growth of the neuromere just in front of it.
That this growth affects the whole right half of the segment is
shown by the presence there of two rudimentary lateral eyes, Ze.

IV. Fusion oF RIGHT AND LEFT HALVES OF THE EMBRYO
AND ANTERO—POSTERIOR DEGENERATION.

This remarkable phenomenon has been observed in so many
different stages, that there can be no doubt as to the manner
in which it takes place. Median fusion and degeneration
begins at the anterior end of the embryo (except in the hour-
glass type), and gradually extends toward the posterior end.
In the typical cases, each organ in one half of the embryo
unites with its fellow of the opposite side to form a median
unpaired organ. Those nearest the median line unite first,
and then degenerate, and those lateral to them follow in the
order of their position, till the whole of the segment has
disappeared. The steps in the process are most clearly shown
by the appendages, the dorsal organs, and the nerve-cords.
The other paired organs probably fuse and degenerate in the
same manner, but owing to their indistinctness in surface
views, it is not so easy to follow in detail their successive
modifications. In this way A-shaped embryos are produced,
showing various stages in the progress of the degeneration
from the anterior toward the posterior end.

Toward the close of the process we may find either a median
row of papillae, representing two or three pairs of medianly
fused appendages, Pl. VI, Fig 50, or an exhausted mass of cells
at what was the posterior end of the embryo; and finally these
useless remnants may in their turn disappear. The principle
involved in the process is well illustrated by the adjacent Figs.
Paani. 92:

In Fig. 1, each square represents a segmental organ of some
kind. The lower half of the diagram shows how the organs

Nost.\] VARIATIONS IN LIMULUS POLYPHEMUS. 4I

are produced by the usual

it is more correctly a double rectilinear growth,

one and a lateral one at
montuaneles? to sit. vit he
longitudinal growth may
bey nepresented “by: the
lines parallel with A-4,
and the lateral growth by
the lines parallel with A-—
£. The relative age of
each organ is determined
by its position in relation
to these two sets of lines,
the uppermost @ being the
oldest and most special-
ized organ in the body, and
the lowest A the youngest.
Now it is obvious on exam-
ining the diagrams ¢hat the
median fusion and antero-
posterior degeneration of
the organs takes place in
the reverse order of their
formation, the oldest dying
first and the youngest last.
Considering for the pres-
ent only the small-lettered
part of Fig. 1, median
fusion and degeneration
such as occurs in Limulus
will gradually carry the
hypothenuses of the
shaded triangles toward
each other till they meet
in the median line, the
shaded areas themselves
gradually disappearing.
The half of the embryo

method of apical growth, or rather

a longitudinal

BB
a8
SE

Fic. 2.

Diagrams to illustrate the laws of growth of segmental
organs and their disappearance by median fusion.

42 Ad EM [VOL Ln

that dies first, therefore, consists of the two quarters 2.4.x. and
7.p.z., and with their disappearance the remaining quarters are
thrown toward the median line in such a way that the position
of each organ is shifted a step farther toward the median line
than the corresponding organ just behind it. Furthermore a
row of dissimilar, unpaired organs is formed along the median
line in the same sequence from behind forwards, as from the
median line toward the sides, Fig. 2, @, 0, c, d, 2, and A, B,
CD ae

The steps in this process are as follows. The most median
organs, a@ and a in the anterior row, Fig. 1, fuse with each other
in the median line and then disappear. Their place is imme-
diately occupied by the fused organs that were originally next
to them on the outer side, namely 4 and 4; at the same time
a and a fuse in the second row. In the third step, the fused
organs 4 and 6 of the first row disappear and the fused organs
c and c take their place ; 6 and @ take the place of a+a in
the second row, and a and a fuse in the third. The following
steps are of the same nature and are repeated till the condition
like that in the anterior part of Fig. 2 is reached. It may be
continued still farther till all the organs have disappeared, the
last fused organs to disappear being the two most lateral ones
of the, first line:!

In reality, however, no such condition as the one just
described would ever be completely realized, owing to the
presence at the posterior growing end of the body of groups of
segments in which the organs are arranged in an inverse manner
to that just described, see lower half of Fig. 2. The tendency
will be, therefore, to produce a body in which the arrangement of
the organs is determined by the interaction of growth and
degeneration, one having its greatest activity at the pos-
terior, the other at the anterior, end. The resultant form
will be therefore a more or less elongated rhomboid. The
tendency of all segmented animals to assume such a form,

1 It is not claimed that the process of median fusion and degeneration always
follows exactly the steps indicated above, as will be observed on examination of
the surface views of various embryos, but the variations from it are not of such
a nature as to invalidate the general law.

No. 1.] VARIATIONS IN LIMULUS POLYPHEMUS. 43

owing to the fusion or reduction of organs at the extremities
of the body, and the fullness and completeness of structure
shown in the intermediate regions, testify to the universality
of this double law of growth and degeneration.

What is the cause of this mode of degeneration? It seems
to be an exaggeration of the forces which have determined the
form and manner of growth in normal embryos.

Assume, for example, that the embryo of a segmented animal
consists of a double row of independent half-metameres, placed
with their oldest ends or #eads toward each other, and grow-
ing laterally like an acrogenous plant. The series of organs
thus formed, extending from the oldest or median end to
the youngest, or lateral, extremity, such as the endoderm,!
mesoderm, neuromere, nephridium, appendage sense-organs,
and various somatic structures, will then appear on the half-
somite, roughly speaking, in the order of the appearance of such
organs throughout the animal series. And they are analo-
gous to the metameres of a whole animal, or to the succession
of morphological units produced in an acrogenous plant. There
is, however, a sharp distinction to be drawn between the
arrangement of organs on a half-metamere and the succession
of metameres in a segmented animal.

In the latter case we have a succession of homologous parts
more or less modified by their position; in the former, a suc-
cession of fundamentally different parts—a logical sequence
of morphological structures in accordance with the genesis of
physiological specialization. The dorsal and ventral surfaces
are thus forever fixed as parts fundamentally different, and
less likely to be confounded through secondary changes than
the anterior and posterior extremities of the body. Broadly
speaking, therefore, the ventral surface of a segmented animal
is the oldest and most specialized, —the dorsal the youngest
and least specialized.

If we had to deal only with lateral growth, —z.e. growth
at the apex of each half-metamere, and with the formation of new
half-metameres at the posterior end, and all this taking place on

1 The endodermic and mesodermic portions of a half-metamere are enfolded
at an early period, and consequently do not appear to form a part of the series.

A4 PATTEN. [Vor. XII.

a flat, instead of aspherical surface, — we should have a succes.
sion of organs like that in the lower half of Figs. 1 and 2. But
each half-metamere not only grows in length but in width,

bc

ee
WY

we
a.

a
. ei
J

as

Fic. 4.

Diagrams to illustrate how a double series of half-metameres formed in succession from before
backwards, and growing on a spherical surface, will produce differential growth forces.
These in turn produce median concrescence and degeneration, cervical and caudal flexures,
and the modification of the trunk into regions of different morphological and physiological

potentialities.
and as this is most manifest at the lateral end, they will tend
to be triangular inform. Moreover, each succeeding metamere
will have to grow under other conditions of cell tension and
available yolk surface than the metamere that preceded it.

a

No. 1.] VARIATIONS IN LIMULUS POLYPHEMUS. 45

The whole tendency will be then to produce a form like that
in Fig. 4, where an attempt has been made to illustrate the
action of growth forces by lines corresponding with the arrange-
ment of organs. The increase in width of the lateral ends of
the metameres will produce a constantly increasing tension
which will find relief, and thus favor still further growth,
by movements forwards and backwards. The ends of the
most anterior metameres will thus be forced together along
the median line, causing the median fusion of organs in front
of the true anterior extremity of the body. Conditions like
these have probably caused the fusion of such organs as the
median ocelli and the olfactory organs of Limulus, and in insects
the upper lip, which arises, as has been repeatedly shown, from
the fusion of originally paired organs lying on the very anterior
margin of the cephalic lobes. (See “Eyes of Acilius.’’)

At the posterior end of the body, concrescence of the meso-
dermic area is the result, and the true apex of the body is shut
off from growth over the surface of the yolk. The new seg-
ments formed after this period are therefore forced to grow
vertically upward and forwards. Hence the conical, forwardly
directed tail seen so constantly in vertebrates and arthropods.

The segments formed at the apex of this conical tail lobe
will be produced under conditions very different from those
found elsewhere, and we can readily see how these conditions
might not only be the direct cause of the diminution in size and
fusion of the organs there, but also prohibit the further addi-
tion of new segments.

At the head end there is no necessity for sucha form, because
no new segments are formed there. But there isa gradual thick-
ening of all the organs along the median line, which tends to
find relief forwards as well as laterally ; but as the anterior mar-
gin of the cephalic lobes is, as it were, shut out from the median
line by the ingrowth of the mesodermic area, its only relief
isin buckling upward and forwards. Hence the cerebral flexure,
and the general S-shaped contour of the whole embryo, Fig. 3.

Concrescence of the margins of the mesodermic area occurs
in the normal embryos, as shown by the figures in Pl. I.

46 PATTEN. [Vou. XII.

But there are some very interesting phases of it seen in the
embryos that are much reduced in size, and which may be
best considered in this connection. The margin of the meso-
dermic area is there frequently enlarged so that it becomes
very conspicuous in surface views. It serves as a pretty
safe index of the grade of development and degeneration,
and also to identify the posterior end of the body in embryos
reduced to such low terms as those in Pls. VI and VII. This
structure is formed by the fusion of the peripheral margin
of the mesoblastic somites with the ectoderm and with the yolk
cells! A long primitive, streak-like thickening ts thus formed,
by the proliferation of which the germ layers are extended laterally
in exactly the same way that we are familiar with as occurring
at the posterior end of the body.

There is a great difference in the lateral extension of the
segmented portion of the mesoderm. In normal embryos it
may reach the very margin of the mesodermic area, while in
abnormal cases it may not extend beyond the appendages.
But there is universal agreement in normal and abnormal
embryos in the union at the periphery of ectoderm, mesoderm,
and yolk cells.

The lateral growth then of a metamere is comparable with
the posterior apical growth by which the body of the embryo ts
encreased in length.

In normal embryos, the lateral margins of the mesoblastic
segments, when first formed, are quite indistinct, Fig. 1, but
they soon fuse to form the conspicuous marginal thickening just
described, Figs. 2-4, m.a. The egg being nearly spherical, the
anterior and posterior portions of the margin have a shorter
distance to go in order to unite with each other in the median
line than the middle portions. For this reason and others that
we have already discussed, the margins, following the path of
least resistance, tend to form anterior and posterior loops, which
finally meet in front of and behind the embryo, PI. I, Fig. 5,
forming a rounded, mesodermic area in the centre of which the

1 We cannot for lack of space discuss the origin and history of the mesoderm
here in the detailed manner that it deserves. We can only point out briefly the
more important facts that bear on the subject matter of this paper.

No. I.] VARIATIONS IN LIMULUS POLYPHEMUS. A

embryo lies. Its thickened rim, 7. a., usually forms two con-
spicuous posteriorly directed loops, which may be found in all
stages of concrescence behind the apex of the abdomen. The
sides of the loops concresce like the closing of the arms of a \,
and a longitudinal post-anal thickening is thus formed, below
which lies a great cloud of cells, brought together at that point
by the union of the proliferating rim.

In the abnormal and degenerate forms, the margin of the
mesodermic area is usually very distinct, although every out-
ward trace of segmentation in the mesoderm, and even the
greater part of the embryo itself, may have disappeared.

The margin also shows zz many cases more clearly than in the
normal embryos the posterior, median concrescence, and it takes
place in such a manner as to bring very forcibly to mind the
similar phenomena in vertebrate embryos. One of the most
striking instances of this exaggeration of the mesodermic
margin is shown in Pl. VI, Fig.63. Unlike all the rest, this
is an opaque embryo shown by reflected light. The margin of
the mesodermic area forms a conspicuous ridge, which is
thickened posteriorly. Where concrescence has taken place, a
broad median elevation is produced.

Nothing like the thickened margin here described is known
to occur in any other arthropod, but a similar concrescence of
mesodermic segments probably occurs in all arthropods.

In the embryos very far advanced in degeneration, the margin
of the mesodermic area breaks up into isolated, irregular masses
containing closely packed nuclei, Pls. VI and VII, Figs. 60,
72, 76, 88, etc. These masses may lie deeply in the yolk,
but they still show very clearly their derivation. They fre-
quently send out pseudopodia-like streamers of nuclei, which
fade at the distal ends as though they were gradually dissolving
in the yolk.

In Figs. 67 and 72, the margins of the mesodermic area first
united a short distance behind the end of the abdomen, leaving
a pear-shaped area of yolk, covered by the blastoderm only.

In Figs. 10, 18, 24, 61, 68, the margins have fused over a
long distance, leaving along the line of fusion a cloud of yolk
and mesoderm cells. The posterior margin of the fused areas

48 PATTEN. (Vou. XII.

is usually distinctly indented in such cases, as though the
concrescence were proceeding still farther backwards.

Various modifications of the different phases of concrescence
are shown in Figs. 14, 16, 18, 20, 24, 43, 40; 50, OL, 62,:63,
GS 1075.08) 705) 71 a7 217.03 000) 2,005.

The traces of segmentation in the mesodermic area and the
very conspicuous concrescing folds are interesting features ox
Figs. 14 and 20. Observe also the curious knob-like ecto-
dermic thickening at the posterior end of the concresced area
in Fig. 16, a not infrequent occurrence. Compare also Pl. III,
Fig. 20 with Pl. V, Fig. 49, where the heart is apparently
formed for a part of its length.

The whole process of concrescence, as shown by these em-
bryos, reminds one of the concrescence of the so-called margin
of the blastopore in vertebrates. The principal difference is
that in vertebrates,the ectoderm, mesoderm, and endoderm grow
over the yolk together, there being no single layered blastoderm
covering the whole yolk before the germ layers begin to form,
as is the case with Limulus and the arthropods generally. But
the presence of this layer of cells can hardly be regarded as a
serious objection to a comparison of the processes of concres-
cence in these two great groups of animals.

In the light of this comparison, the fact that in very rare
cases, as shown by Ryder in certain fish eggs, the segmenta-
tion of the mesoderm extends to the margin of the concrescing
lips of the “blastopore”’ is very significant. It seems to me
most easily explained as a reversion to a condition like that in
arthropods.

There are some other interesting suggestions that arise from
this comparison. We must not forget that 2 Limulus at any
vate the growth backwards of this line of concrescence does not
increase the body in length. It is merely a part of the haemal
surface, thrown back temporarily upon the neural surface by the
presence of the yolk. The true increase in length of the axis of
the body comes from the proliferation of cells in a primitive
streak that les just in front of where concrescence began,
It is obvious that an application of these principles of growth
to vertebrate embryos might clear away some of the difficulties

No. I.] VARIATIONS IN LIMULUS POLVPHEMUS. 49

that have arisen in attempting to explain, first, the increase in
length at the posterior end, and second, the relation of the
primitive streak to the concresced lips of the ‘“blastopore.”
In fact, the whole manner of interpreting the formation of
germ layers in vertebrates may be advantageously reconsidered -
from this new standpoint.

I can only touch on these points here, but I shall discuss
them more fully elsewhere. The figures are too suggestive to
be passed over without comment.

That a half-metamere is to be regarded as the unit of structure
of segmented animals is shown, above all (1),as already indicated,
by their method of growth and direction of specialization, and (2),
by the manner in which half-metameres are omitted or increased
or diminished in size, rather than whole ones. On this supposi-
tion, certain imperfections in the segments of otherwise typi-
cally segmented animals can be explained ; such, for example, as
spiral segments, which may be accounted for as due in part to
the imperfect union of twin half-metameres on the dorsal side
of the body. Instead of meeting each other squarely, and thus
forming a line of equalization of forces, they have grown past
each other in opposite directions, and thus each forms a more
or less perfect spiral !

Polarity and embryonic axes thus appear in a new light.
In any segmented animal the embryonic half-metameres group
themselves in such a way that they may be represented by
axes meeting at a shifting point, as shown in Fig. 5. The line
H.. T. represents the median line and the axis of longitudinal
growth, or of repetition, and the triangle ad 7, the form that
would result if the individual half-metameres did not mutually
modify each other.

Wx. and p.x. represent the axes of specialization, along
which new organs are produced in each half segment. The
growth along these axes is at right angles to the median line,
and also parallel with it. The tendency to a concentration of
organs at the forward end of the embryo, or toward a median
fusion and antero-posterior degeneration there, may be explained

50 PATTEN. (Vo. XII.

as due to the fact that there is greatest tension along those
lines, Ax., and less growth power to resist it, because the
tissues at those points are oldest, and have already exhausted
more of the inherent powers of growth than at 7.

It may not be venturing too far to claim for this principle still
further applicability ; for example, in explaining the more pro-
found and remote physiological differences between the anterior,
middle, and posterior regions of the body. Or rather, it would
be better to say that this morphological law is the formal

Fic. 5.

expression of the fact that differential growth tension has fixed
the posterior, middle, and anterior regions of the body as the
seats, respectively, of constructive, elaborative, and destructive
physiological processes. But it will not do to press this
thought too far, certainly not without a precise statement of
the way it is to be applied.

It would also be important to determine whether there is
any relation between these laws of growth and decline and the
different powers of regeneration shown by various regions of
the body, and in this connection we would recall the difference
so manifest in this respect between the head and tail region.
If such were, indeed, the case, there might be some foundation
for the supposition that growth and regeneration are associated

No. 1.] VARIATIONS IN LIMULUS POLYPHEMUS. 51

phenomena, antithetical to the associated phenomena of degen-
eration, specialization, and lack of regenerative power.

We have thus seen how differences in the time and place of
growth will in normal embryos produce conditions that cause
the fusion and degeneration of organs.

In such cases the fusion and degeneration took place in front
of or behind the true extremities of the body. But we see no
reason why the same kind of factors should not produce the
more extensive median fusion and degeneration seen in the
abnormal forms.

This supposition becomes all the more plausible when we
consider that the lines of fusion and degeneration are coinci-
dent with the lines of greatest stress.

Again, we can see how reversing the conditions that have
brought about median fusion and degeneration, z.e. diminished
lateral cell stress at the anterior end, might permit the forma-
tion of double and triple monsters, — as shown by Fig. 7, p. 73.

And finally, if the half-metameres were very much reduced
in numbers, the tendency to increase in width at the lateral
end would have greater freedom, and more or less ovoid bodies
would result, in which the segmentation would be lost sight of
in the antero-posterior expansion of the half-metameres. The
organs would then tend to be formed along concentric circles,
somewhat as in the embryo of a cephalopod.

Having considered what we have regarded as the general
principles involved in this class of variations, we will now
examine the individual illustrations of the same.

An early stage of fusion is well shown in Pl. V, Fig. 39.
The cephalic lobes are constricted, and without character.
The chelicerae are absent. The next two appendages are
brought closely together, but the space between the appendages
of the following pairs becomes greater, till in the last thoracic
metamere they are separated from each other by the normal
distance.

A tendency in the same direction is seen in Fig. 47, as is
shown by the mere trace of cephalic lobes, the absence of

52 PATIEN: [Vo.L. XII.

the oesophagus and first two pairs of appendages, and the
drawing together of the appendages of the fourth pair.

In Fig. 46 is a more typical condition. The cephalic lobes
are entirely absent, and the anterior limbs of the thoracic
margin are converging to form the apex of an inverted V.
The chelicerae have united at ck., and the second pair of
appendages forms a single, long, coiled and slender process, af?.
The third pair has almost fused, but each appendage still retains
its characteristic shape.

In Fig. 44 the second and third pairs have united, while
every trace of the cephalic lobes and of the chelicerae has
disappeared. In Fig. 42 about the same change has taken
place, except that the degeneration of fused appendages has
progressed farther backwards, for here both the chelicerae and
second pair of appendages have disappeared. The fusion of
the appendages of the fourth pair is not quite completed.

Fusion accompanied by degeneration of the two anterior pairs
of appendages is shown in Figs. 43 and 49. In Fig. 48 fusion
and antero-posterior atrophy have given rise to a form frequently
seen in which little of the embryo but the three posterior
thoracic appendages is left. In Fig. 41 is a similar embryo in
which the oesophagus is still visible a long distance in front of
the embryo. It is a pit-like depression merging gradually into
a long cloud of cells lying below the surface, and extending
backward toward the embryo, doe. These cells represent
either the degenerating remnant of the anterior portion of the
embryo, or of the oesophagus.

If the process of fusion and degeneration progresses in the
way the various stages just described indicate, we should obtain
embryos in which all the paired organs have been fused and
subsequently absorbed, except the last pair of appendages, or
perhaps the tip of the tail. This appears to be the case with
the one in Pl. IX, Fig. 106. We should not forget that
this embryo has been developing as long as the others here
described ; and that it really is in a late stage is shown by its
size, and by the concrescence of the posterior limbs of the
germinal area back of the median appendage or lobe. There
is no trace whatever of a nervous system, unless the dark area

Nov T-] VARIATIONS IN LIMULUS POLYPHEMUS. 5

around the stomodaeum represents the remnants of the cephalic
lobes.

What appears to be a still more degenerate condition is
shown in Pl. VI, Fig. 69. Here the inner layer cells of the
margin of the germinal area are breaking up into dense, irregu-
lar masses of degenerating cells, while every trace of the head
is absent.

Degeneration of fused organs does not always proceed in
that way, as is illustrated by the remarkable embryo shown in
Pl. VI, Fig. 50. This is the only instance observed in which
three successive metameres show the same degree of fusion.
There is no indication as to what appendages are represented
in this embryo, probably the last three thoracic, as we have
seen that they have greater vitality than any other organs of
the body.

Similar conditions are shown in Figs. 52-58, but complicated
by partial transverse fission. See Section VI.

The only exceptions observed to the law of fusion and degen-
eration illustrated by the preceding figures—to which many
more might have been added — are shown in PI. V, Figs. 40 and
45. In the first, the cephalic lobes and cheliceral segments
have disappeared. The second and sixth pairs of appendages
are fused, so that the whole embryo is spindle-shaped. The
identity of the appendages is determined by the ‘dorsal
organs,” which are very plainly developed. This embryo, then,
seems to furnish a case of median fusion and degeneration at
both ends. In Fig. 45 the last pair of legs has fused. It
probably represents the fifth pair, the sixth having dis-
appeared.

Nearly all of the older multiple embryos show in one or
more of the component individuals the typical antero-posterior
fusion and degeneration. But a modification of this process
occurs there of extreme interest. In such cases the fusion is
greatest between the third and fourth thoracic metameres, and
diminishes from that point both forward and backward, pro-
ducing what I have called hourglass embryos. The con-
striction thus produced may separate the anterior and posterior
regions completely. This condition is very beautifully shown

54 PATTEN. [Vor. XII.

in Pl. VIII, Fig. 98, also in Pl. IX, Figs. 102 and 103. See
also pages 67—71.

The ordinary antero-posterior fusion is well shown in Fig. 94,
where the left-hand embryo consists merely of a tail, the nor-
mal fifth and sixth, and the fused fourth, pairs of thoracic appen-
dages. A mere trace of the anterior part of the embryo is
shown at oe., and the well-developed approximated dorsal
organs are clearly seen at do. No nervous system or other
organs are visible. In the upper embryo of Fig. 97 fusion and
complete degeneration of the anterior end of the embryo are
shown. ‘The fourth pair has fused and is much reduced, the
fifth is fused and very long and much folded, the sixth pair
is approximated, but not fused. In the opposite embryo the
process has gone on a little farther, resulting in the fusion of
the fifth and sixth pairs. The fourth pair was very small and
partly concealed under the folds of the long and much crumpled
fifth pair. The dorsal organs in the upper embryo are close
together, in the lower they are completely fused, do.

Still another condition, but similar to that in Fig. 97, is
shown in the right-hand embryo of Fig. 96. There is no mar-
ginal fold here or any dorsal organ. The tail has been thrown
outwards, making a sharp bend in the longitudinal axis, just in
front of the sixth pair of legs. The latter is nearly fused; the
fourth and fifth pairs completely so. I can form no plausible
conjecture as to what the process to the left of the left flabellum
may represent. It is entirely out of place as an appendage, so
far as we may judge from the other cases. It is important to
observe that in Fig. 96, and in both embryos in Fig. 97, the
flabella are larger than usual in embryos of this age, resem-
bling, except in position, normal thoracic appendages. The
problematical appendage of Fig. 96 may be an extra flabellum
belonging to the left fifth thoracic appendage. But no indi-
cation of such an organ has ever been seen in any other
specimen,

In the curiously aborted embryo seen in Fig. 100, the
posterior three or four (?) pairs of thoracic, and two or three
abdominal, appendages are fused along the median line. Here,
all three thoracic appendages show about the same increase in

No. I.] VARIATIONS IN LIMULUS POLYPHEMUS. 55

length and irregularity in shape, which is very unusual, indicat-
ing that they fused nearly simultaneously and to an equal degree.

In Fig. 93 the left-hand embryo, at its anterior end, is medi-
anly fused and partially degenerated. The minute pit oc.
probably represents the remnant of the fused second pair of
appendages.

In Fig. 95 fusion and degeneration result in the formation
of a slipper-shaped embryo in which nothing is left but a single
median appendage and a pit that may represent the last of an
appendage, or perhaps the fused dorsal organs.

In Fig. 98 is represented an extremely interesting condition,
due in part to transverse constriction or fission. The constric-
tion, followed by fusion, and finally by degeneration, takes place
between the third and fourth pairs of appendages, and dimin-
ishes from that point in both directions. In Fig. 103, embryo
A, the same process is carried a little farther, all the anterior
appendages being fused to form a row of four median projec-
tions which preserve approximately the relative proportions of
the three pairs of appendages from which they arose. In the
same figure, fusion and degeneration are carried further in
embryo 4, and further yet in embryo C, where nothing is left
but the fused dorsal organ at c, and the last two fused thoracic
appendages.

In the triple embryo shown in Fig. 104, all three individuals
are in essentially the same condition as embryo C of Fig. 103.

In Fig. 102 we have a triple monster in which embryo 4 is
nearly normal. JZ has undergone transverse fission, as in Fig.
go, and the anterior half has undergone degeneration till nothing
is left but a single median projection representing an imper-
fectly fused pair of appendages. Embryo C is represented by
the fused sixth pair of legs and by traces of the abdomen.

It is an interesting fact that in nearly all cases when there
has been undoubted antero-posterior fusion and degeneration,
the abdomen with its appendages is nearly normal and very
well preserved. This is somewhat surprising, as the abdomen
is usually much abbreviated and often entirely absent in em-
bryos that show a tendency to abnormality in other respects
than in multiple fission or in median fusion.

56 PATTEN. [Vor. XII.

It should also be observed that in Figs. 102 and 103, much less
clearly in Fig. 104, the embryos, beginning with the most perfect
embryo, A, show increased concrescence and degeneration as we
pass in a spiral to embryos Band C. This interesting fact will
be discussed under the head of double and triple monsters.

Change in Shape of Fused and Degenerating Appendages.

When two appendages unite, they fuse at the base first, and
the fusion extends from that point toward the apex. The
resultant appendage is at first much longer and more slender
than the unfused ones. It is also usually folded back and forth
several times, and otherwise irregular in shape. It subsequently
diminishes in length, and finally forms a minute, conical papilla
arising from the centre of a saucer-shaped depression. The
papilla then disappears, leaving a shallow depression that cannot
be readily distinguished from the oesophagus or the fused
dorsal organs.

As a general rule the order in which fusion takes place is
indicated by the length and coiling of the appendages. When
there are several fused appendages visible, they show a gradual
diminution in these characters from before backwards, as in
Figs. 40, 44, 46, and 102. In rare cases the same characters
are presented by all the fused appendages of a series as in
Figs. 50 and 100.

V. GENERAL PROGRESSIVE DEGENERATION.

A careful study and comparison of many abnormal embryos
of various ages and conditions indicates pretty clearly that
most of the abnormalities are due to either a local or a
general lack of formative energy. We picture to ourselves
two sets of factors at work, the action of one being to increase
the quantity and diversification of protoplasm, the other to
reduce it to its lowest terms. The action of the first may
temporarily prevail, but in the end the second factors are
certain to prevent the work of the creative ones.

No. I.] VARIATIONS IN LIMULUS POLVPHEMUS. 57

We assume that there are certain conditions resident in the
ovum which, under the action of normal surroundings, guide
it through a long series of changes to the expression of that
form and mode of action characteristic of what we call the
normal organism.

Among the embryos of Limulus, there are some in which
there appears to be a very slow discharge of vital processes,
producing the retarded or belated forms ; again, nearly perfect
individuals are produced within the normal period, but very
much reduced in size throughout, suggesting the small but
perfect embryos of Amphioxus that have developed from frag-
ments of segmenting ova. There are embryos constituting a
third class, in which a particular charge of formative energy
was apparently omitted, resulting in the absence of an eye, a
leg, a neuromere, or a large, definitely circumscribed area of
the embryo, but without visibly affecting the remaining organs.
Finally, there is a fourth class, where the embryo seems prop-
erly loaded and the various charges properly connected, and it
starts off well, following its normal line of flight for a while.
But through some inherent defects, the nature of which we can-
not even conjecture, progressive development ceases at a point
very far from the mark. Then follows a decline, manifested
outwardly by a general decrease in size, by fusion and com-
plete atrophy of one organ after the other, till the whole
embryo disappears. But as the animal still lives during this
decline, and in all outward appearances is sound and healthy,
showing even in the last stages the presence of karyokinetic
figures, it is obvious that we can only explain this condition
by assuming that the death rate among the cells is greater
than the birth rate. And as the last survivors are nothing but
indifferent, lymphoid cells, we must also assume that the
gradual approximation of the death to the birth period cuts
off more and more from the period necessary for cell special-
ization. The result is a new kind of death for highly organized
antutals, — one, namely, in which the component cells gradually
decrease in number and in specialization till nothing remains of a
once complex organisnt but a few indifferent cells, which in turn
themselves disappear by a continuation of the same processes.

58 PATTEN. [Vor. XII.

Almost any organ of the body, but more especially the brain,
oesophagus, nerve-cord, abdomen, or appendages, may be absent
from the start, or may be behind time, or developing normally
may quickly disappear, without in any case, so far as could be
learned, either affecting neighboring organs or being affected
themselves by conditions other than those produced by their
own growth.

The variation can be usually traced back to variations in
half-metameres. This would seem to indicate that the trunk
of the embryo is a group of integral parts arranged in a double
series like two rows of segmented animals placed head to head.
Each half-metamere seems to be endowed at the outset with a
fixed capital of formative material, which when absent in whole
or in part, or exhausted, cannot be restored. In addition to
this the growth of a half-metamere may be hindered or favored
by local mechanical conditions similar to those producing
median concrescence and degeneration.

No other supposition, it seems to me, can explain why one
leg, for example, out of twelve utterly fails to develop, while
the rest go on as usual, although all are equally surrounded by
nourishing yolk and by the same medium.

With these considerations in mind we can understand how a
weakening in the developmental forces might be indicated by
the four following classes of variation, namely: (1) slowness
of development, (2) small size, (3) absence of organs here and
there, and (4) gradual reduction of the whole body till it com-
pletely disappears.

Let us now consider these four classes in more detail.

(1) Almost every one of the embryos we have figured is behind
time, as we have already explained in stating the methods by
which the material was obtained, and we have also described
how single organs, such as the appendages, disappear or fail to
develop, and that the absence of any organ is often followed by
the complete degeneration of the rest of the metamere.

(II) One of the most characteristic features of the following
embryos zs their small size, a fact well brought out by compar-
ing the figures on Plates I, VI, and VII, all of which are
drawn to the same scale. In these cases the reduction in size

No. 1.] VARIATIONS IN LIMULUS POLYPHEMUS. 59

is permanent, and if very marked seems to lead to the final
disappearance of the embryo.

(III) Adsence of tndividual organs.

A. Atrophy of the thorax. The most frequent defect in the
thorax 1s the absence of the entire cheliceral segment, a fact of
striking significance tn connection with its diminutive size
throughout the entire group of arachnids.

Further degeneration may do away, one at a time, with the
succeeding thoracic segments in their order from before back-
wards, but usually leaving two, or more frequently, the three
posterior ones, nearly intact. This mode of degeneration may
not be preceded by the median fusion previously described. The
missing organs are either absent from the start or else degen-
erate very early. Meantime the cephalic lobes and oesophagus
may persist, but in variously modified conditions, as shown in
Pigs: 26," 27; 33, 34, 35, 62, 67. These facts show that the
anterior part of the thorax 1s the weakest part and most likely
to be absent or to degenerate, and that this weakness gradually
diminishes towards the posterior end.

B. When the cephalic lobes show partial degeneration, the
reduction seems to take place first along thetr anterior margins,
and to progress backwards independently of anterior degeneration
in the thorax. The law of degeneration of the cephalic lobes as
here stated is not so clearly shown as in the case of the thorax,
on account of the difficulty of distinguishing the parts. But
ievisecms. toy hold good in Figs. 18, 20, 21, 22, 27, 47.) Tims
perfectly certain, however, that the cephalic lobes and oesophagus
may be preserved after every trace of one or more segments behind
them has disappeared, as in Fig. 47 and in Fig. 25.

The area where degeneration of the anterior margin of the
cephalic lobes has taken place, and which has not been covered
by the contracting marginal fold, is nearly always flattened and
depressed, as shown in Figs. 10, II, 16, 20, 27, etc.

C. In the abdomen, the same law of degeneration seems to hold
good, t.e. the most anterior abdominal appendages and neuro-
meres are the first to degenerate, the posterior ones being the
most persistent. It should be borne in mind that the normal
growth of the abdominal appendages is similar to that of the

60 PATTEN. [Vou. XII.

thoracic. For example, three pairs of thoracic appendages, the
second, third, and fourth, appear first, followed by a very small
pair in front, the chelicerae, and by two pairs of appendages
behind ; the most precocious part is therefore the middle por-
tion. In the abdomen, the second and third pairs first appear,
followed by the rudimentary chelaria in front, and by the
remaining appendages behind. Compare Figs. 4, 6, and 7.

In the degenerate embryos shown in Figs. 10 and 15, the
only abdominal appendages present appear to be the second
and third pairs. It is hard to say whether the ones that
should develop back of them are merely belated or have degen-
erated, the general impression at first sight being that there
has been a shortening of the abdomen by suppression of its
posterior end. However, in more typical cases, the facts
seem to support a different conclusion. For example, a large
part of the abdomen may be absent, and in its place may be
seen either a conspicuous tail-like projection resembling the
post-abdomen in young scorpion embryos, or an ingrowth into
the yolk, which varies greatly in size and depth in different indi-
viduals. If we may use the presence of this depression as an
indication of the position of the original posterior end of the
abdomen, z¢ zs obvious that it ts the anterior metameres that
have disappeared, because we usually find this depression or
projection, as the case may be, carried forward to a point just
back of the thoracic appendages. Compare Figs. 8, 13, 16, 20,
21, 2A, etc:

In Figs. 29, 30, and 32 the same law is illustrated by a dif-
ferent class of cases. Here, also, we see that the principal loss
of material is at the anterior end of the right or the left side
of the abdomen, not at its posterior end.

The abdomen, therefore, ike an independent embryo, develops
zts metameres in a sequence similar to that in the thorax.
They degenerate from before backwards, and independently of
degeneration elsewhere in the embryo.

The cases we have just described show that there are three
separate points at which backward degeneration may begin,
namely, at the anterior end of the cephalic lobes, at the anterior

No. I.] VARIATIONS 1N LIMULUS POLYPHEMUS. 61

end of the thorax, and at the anterior end of the abdomen.
Each of the regions between these points may degenerate inde-
pendently of the other, and in the same way that the whole
embryo sometimes does, suggesting the idea that the embryo
consists of a chain of tmperfect individuals, such as we see
in annelids as the result of tmperfect fission.

A less strongly marked division occurs in Limulus, across the
middle of the thorax between the third and fourth segments.
It is shown (1) by the marked tendency of the first three seg-
ments to degenerate and of the last three to persist, (2) by
the transverse fission that occurs there, see following para-
graphs (p. 67), and (3) by the presence of an enlarged pair of
sense-organs (“dorsal organs’’) opposite the first of the last
three segments, just as the lateral eyes lie opposite the first
one of the first three segments.

When the degeneration is incomplete, it may manifest itself
by a partial median fusion and a decrease in size of the organs
along the lines separating the regions above indicated.

These cross-lines, where degeneration ts most likely to occur in
Limulus embryos, are the regions showing the least “vitality’’ ;
they correspond to the regions where a reduction in size and a
tendency to undergo median concrescence are frequently seen
in the adult individuals from various groups of arthropods.
In Limulus these lines of weakness are cleavage planes separat-
ing regions having different potentialities, and corresponding to
regions possessing different morphological characters in many
other arthropods. Broadly speaking, the salient morphological
characters seen in the various regions of the body of arthropods
are similar to those that appear as abnormal variations in Limulus
embryos ; and, as we have seen, these characters are mainly those
due to difference in size, specialization, and median fuston, and
these tn turn have to be referred back primarily to differences im
the power of growth.

The four regions of the body in Limulus embryos are
roughly shown in the adjacent Fig.6. The lateral constrictions
and the transverse lines indicate approximately the position
and amount of median fusion and degeneration.

The repetition of three segments in successive regions of

62 PATTEN. [Vov. XII.

the body is a striking fact, and it should also be observed
that each individual part of the embryo, and especially the
ends, tends to assume the shape of a single or double V,
characteristic of entire embryos
undergoing atrophy by median or
apical fusion and degeneration.

In comparing these variations in
Limulus, Fig. 6, with the normal
conditions in other arthropods, I
a will merely bring to mind the fol-
lowing facts, to which others might
be added.

(z) In the first region we have to
recall the absence, or median fusion,
or degeneration, of appendages at
the anterior margin of the cephalic
lobes, especially in insect embryos,
and the absence or fusion or de-
generation of the anterior pairs of
ocelli in arachnid embryos (scor-
pions, spiders, Limulus) and the

Ee: great development of the antennae
Diagram to\show Jines/and regions'of ana dateral eyes arising sro athe

most frequent degeneration. The

lateral constrictions indicate potential posterior margins, as in insects and
fission planes due to partial median 1
fusion. The cross-lines indicate the crustacea.

relative (weakness of: the /dittercat 7) he reductionwandenhemine-

soaks quent median fusion of the ap-
pendages just back of the mouth in insects, crustacea and
arachnids, and the much greater size of those at: the posterior
end of the thorax. The frequent absence of the chelicerae
in Limulus is to be compared with the small size of these
appendages in the vast majority of arachnids. The manifest
weakness of the first three pairs of thoracic appendages in

=a a

Bt SzSs
GP

1 The degree of median fusion may therefore be taken as an index of the
original order of serially homologous organs. Applied to vertebrates, this would
indicate that the most anterior sense-organ is the pineal eye, followed by the
olfactory organ and the lateral eyes, just as is the case in Limulus. This
harmonizes with the conclusions that I have reached on anatomical and embryo-
logical grounds of an entirely different nature.

No. I.] VARIATIONS IN LIMULUS POLYPHEMUS. 63

Limulus, as shown by their more frequent absence and median
fusion, is comparable with the small size and frequent median
fusion of the first three pairs of post-oral appendages in insects
and other arthropods.

(c) There is at the anterior end of the abdomen a region of
condensation, specialization, and degeneration, forming what I
have called the vagus region of scorpions and Limulus, and which,
in a paper on “ The Origin of Vertebrates from Arachnids,”
I showed was probably of wide distribution in the arachnida,
including the trilobites and related forms. In scorpions, it
consists of four very much condensed metameres, and in
Limulus of two or more, which in both forms are provided with
rudimentary appendages, nearly or quite fused.

(2) Finally, the very wide distribution of well-developed
terminal appendages through all groups of arthropods is a
manifestation of the same law of growth.

In conclusion, therefore, it would appear that the division of
the body of arthropods into successive regions composed of several
segments 1s not due primarily to specialization or adaptation,
either by or for any particular use, or through disuse, but to
some form determining forces that govern growth. The same
factors in all probability produce in a similar way metamertc
segmentation, transverse fission in annelids, and determine
the length of the body in a given individual.

(IV) The complete degeneration of the whole embryo is an
obscure process, and probably varies considerably in different
individuals. A continuation of the various local degenerations
previously described has, without doubt, gradually produced
the misjointed fragments of embryos seen in Pls. VI and
VII. Théy admirably illustrate Empedoclian fancies. The
embryos themselves give some indication of the various ways
in which they have degenerated. They also indicate that the
final stages of degeneration lead to a tolerably uniform con-
dition.

With the disappearance of all the appendages, the embryo
may, in the class of cases we shall now consider, be reduced to
a mere pit or sac, yet preserving certain features which show

64 PATTEN. [Vou. XII.

clearly the stage in which the whole embryo would have been,
had no degeneration taken place.

Perhaps the best illustration of this is shown in Pl. VII,
Fig. 82. The mesodermic area is relatively large, and its
posterior margins are thickened and well advanced toward con-
crescence. There is a very large, projecting tail lobe, like that
in Figs. 48, 49, and 94, and yet of the body proper (which should
have all the organs seen in Figs. 5 and 6) there ts nothing left
but a deep, thick-walled pit, with a triangular opening to the
exterior.

A similar condition is shown in Fig. 83, where the undif-
ferentiated remnant of the body forms a Y-shaped sac with an
oval opening at its posterior end.

Figs. 84, 85, 86, 87, and 88 are various modifications of the
same condition.

In Fig. 88 the thickened margin of the mesodermic area
has broken up into the star-shaped masses of degenerating
cells, so frequently seen in the later stages of degeneration.

These embryos represent what is left after invagination,
median fusion, and progressive degeneration have attacked with
varying success every part of the body.

Most of the embryos shown on Pls. V and VI would probably
have reached this condition ultimately. In some of these cases
one can distinguish here and there an appendage, or some
other organ; but the remaining parts may be so distorted or
misplaced that it is impossible to distinguish literally as well
as figuratively any head or tail to them, as in Figs. 71 and 56,
and others which we have not space to figure.

In Fig. 81 the large posterior depression appears to be the
remnant of the anal plate, and the three obscure pits in front
of it, the last of three fused pairs of invaginated appendages.
In Fig. 80 everything has disappeared except the large pit,
situated at what appears to have been the anterior end of the
embryo.

Whether the huge projection in Fig. 69, Pl. VI, and Fig.
106, Pl. IX, represents a fused and partly invaginated pair of
thoracic appendages or a tail-like projection of the anal plate
is hard to determine.

NOs 1.) VARIATIONS IN LIMULUS POLYPHEMUS. 65

The curious pits and sacs just described themselves finally
degenerate into a mere cloud of cells varying in form and
appearance, Fig. 89. Many embryos of this kind show slight
indications of a central depression, probably the last trace of
a sac like those just described.

I have found some eggs among those that had been kept alive
Jor three or four months in which no trace of cells or nuclet was
visible in surface views, yet they appeared perfectly sound and
free from decay, even when stained and cleared in clove oil.
They are probably eggs that have passed so far beyond the
stages of degeneration shown in Fig. 89, that even the last
few cells have disappeared.

There is another series of degenerated embryos that are
very interesting from their resemblance to the early stages of
normal embryos. They consist of two groups of cells like
two primitive cumuli, one corresponding to the head and the
other to the tail end of the body. This condition has already
been passed by the embryo shown in Fig. 1.

Whether or no degenerating embryos as a rule pass through
this condition with separate Az/agen for head and tail is
doubtful, nevertheless it is a variety very frequently seen. It
is well shown in Figs. 72-78.

There is no way to determine certainly whether an embryo
like that in Fig. 73 has degenerated, or whether it has been
kept, by something hindering normal development, in approxi-
mately its original condition. But the great age of the
embryo, the evidence in other embryos like this one, of con-
crescence, and the breaking up of the margin of the circular
mesodermic area, bears out the assumption that it has under-
gone profound degeneration. The anterior mass of cells and
the pit usually seen in its centre probably represent the
degenerated cephalic lobes and oesophagus, and the poste-
rior cloud, the anal plate, and telopore. Both of these discs
are in all this class of cases about the same, but the poste-
rior one is usually the larger and thicker. Sections of the
discs show various conditions, from one where there is merely
a thick, homogeneous mass of cells, flush with the surface,
but sending into the yolk pseudopodia-like masses of degen-

66 PATTEN. [Vou. XII.

erating cells, to one in which it consists of a thin layer
of cells deeply invaginated in the centre, and from the in-
vaginated part alone arises a cloud of scattered yolk nuclei.
BeOS hie.73.

Whether the depressions in Figs. 72, 79, and 81 represent
the oral and anal depressions, or the pits left after the degenera-
tion of fused and invaginated appendages, or of one or more
segments, as in Pl. VI, Fig. 61, cannot be determined.

In some cases, as in Figs. 69 and 73, where the outline of
the axis of the embryo may be faintly distinguished, there is
no cephalic cloud of cells visible.

It would thus appear that degeneration may carry old embryos
back to a state resembling that seen in very young embryos, 1.e.
one where tt consists of a cluster of proliferating cells at either
end.

These facts seem to indicate that the body of the embryo ts
not a single organic unit, such as it would be if it were an
elongated gastrula with fused lips, but rather one of a double
origin.

We may regard the head Azlage, from which arise the
cephalic lobes and oesophagus, as representing the remnants
of a trochosphere, and the posterior Anz/age as the primitive
trunk which arises from the trochosphere as a bud-like out-
growth. I have already shown elsewhere,! that the origin of
the stomodaeal nerves in Limulus supports such a view.

VI. Fission.

All the cases of fission that I have seen in Limulus affected
embryos in which the thoracic appendages were well developed,
or if they were not present it was evident that their absence was
due to degeneration.

The multiple embryos must necessarily, from the methods of
obtaining them, have been well advanced ; and any abnormality
in the early stages is easily overlooked among the hundreds of
eggs that one must examine in order to find them. Neverthe-
less it seems probable that fission does not usually begin till a

1 Morphology and Physiology of the Brain and Sense-Organs of Limulus.

No. 1.] VARIATIONS IN LIMULUS POLYPHEMUS. 67

comparatively late period. This is certainly the case with the
embryos shown in Figs. 90 and gI, the type most frequently
seen.

In the formation of multiple embryos, we may distinguish
two kinds of fission:

(1) Transverse fission, dividing the embryo into anterior
and posterior portions, the plane of fission being usually
between the third and fourth thoracic appendages.

(2) Longitudinal fission, beginning at the anterior end.
This is the most common form, and one presenting a great
number of modifications through degeneration. The process ts
essentially different from that of transverse fission, for the latter
is the result of a local transverse concrescence and degeneration,
while longitudinal fission consists in the formation of two new
halves of an embryo along the median line of one already extst-
ng. The formation of the new halves begins at the anterior
end and extends gradually backwards, one new half being a
mirror image of the other. The old halves are thus thrust
apart, and with the newly formed halves make new embryos}

This process may be repeated a second time tn one of the new
embryos, thus producing three embryos, tatl to tatl, consisting of
the two original halves plus four new ones.

A. TRANSVERSE FISSION.

There are more or less clear indications of this form of
fission in a great many embryos, but it is rarely that it is very
clearly marked. The fact that the plane of fission occurs at a
definite point is very remarkable, and indicates a break in the
morphological continuity of the embryo of considerable theo-
retical interest.

In normal embryos, neither the order of development of the
appendages, nor their size or shape, gives any indication of this
cleavage line. The great development of the lateral, segmental

1 Indications of longitudinal fission, beginning at the posterior end and extend-
ing forwards, are very rare. No dona fide case has been observed, and I doubt
whether it ever occurs in Limulus, the cases in which it appears to exist being

perhaps better explained as malformations of the posterior end of the body, rather
than as the beginning of true fission.

68 PATIL. [VoL. XII.

sense-organ, d@.0., is the first thing to suggest a break in the
apparently homogeneous series of thoracic metameres.

But we frequently see in embryos which in most respects do
not depart from the normal type, quite constant differences in
the arrangement of the thoracic appendages, which may be
regarded as indications of the cleavage plane in question. In
Pl. III, Fig. 23, the potential fissure plane 1s indicated by an
increased distance separating the third and fourth pairs of
appendages. This would be of little importance perhaps taken
alone, but such cases are frequently seen, so that it probably
has the significance attached to it.

Another class of cases illustrating the same thing consists
of embryos in which the ¢hree anterior pairs of appendages have
a markedly different direction of growth from the posterior
ones, suggesting the characteristic difference in appearance
between the mandibles and maxillae of an insect embryo and
the thoracic appendages. These cases are also comparatively
common, typical cases being shown in PI. III, Figs. 19 and 22.

None of these cases would attract attention if it were not for
the fact that the constriction about to be described occurs
between these two sets of thoracic metameres.

A most beautiful example of transverse fission, and one that
throws a good deal of light on the ones about to be considered,
is shown in Pl. IV, Fig. 29. This case is, however, compli-
cated by the partial degeneration of the left half of the thorax,
and the complete absence of the left half of the abdomen.
The constriction is due to the presence of a transverse line of
degeneration having its greatest intensity along the fourth
segment.

In Pl. VI, Fig. 51, is an obvious constriction between the
third and fourth thoracic metameres. When the constriction
is more conspicuous, it has evidently been preceded by fusion
of the right and left halves, the degree of fusion diminishing
from the point of greatest constriction toward the anterior and
the posterior end. This is beautifully shown in the lower
embryo of Pl. VIII, Fig. 98. As every pair of appendages is

present in this case, there is no question about the exact point
of constriction.

No. 1.] VARIATIONS IN LIMULUS POLYPHEMUS. 69

A similar case is shown in the single embryo in Fig.52. A
comparison of this embryo with the preceding leaves little
doubt that the constriction has taken place between the third
and fourth segments. The cephalic lobes are much degener-
ated, and the chelicerae are absent. The modification of the
remaining appendages is easily determined from the lettering.

In Figs. 54 and §5 are two other illustrations of transverse
fission accompanied by degeneration. In Fig. 55 the anterior
portion of the thorax, consisting of three pairs of appendages,
but without cephalic lobes, oesophagus, or neuromeres, is widely
separated from the posterior portion, which consists solely of a
conical projection representing either an enlarged caudal lobe,
or else the last pair of fused, thoracic appendages. One or two
small pits in front of it indicate, probably, where the other
posterior thoracic appendages have disappeared. In Fig. 54 is
a curious modification not easily explained. It appears to be
due to a transverse constriction separating the thorax, the only
part that is left, intotwo portions. The two pairs of appendages
of the anterior part have fused to form a median row, the
fusion being indicated by the great length of the two twisted
median appendages. The appendages were apparently absent
on the right side of the posterior portion, so that it is spirally
twisted.

In Fig. 53 an earlier stage of the same process is shown.
There has been a partial constriction between the third and fourth
pairs of thoracic appendages. In the anterior portion the third
pair have fused, the second are still widely separated, while
the first pair and most of the cephalic lobes are absent. In the
posterior portion the fourth pair are fused, and the right fifth
and sixth appendages, including their neuromeres, are absent.
The fifth and sixth appendages of the left side are very large,
and owing to the unequal bilateral development, thrown spirally
toward the right. The distortion of the axial line is also shown
by the position of the telopore, 72.

Still other embryos where median fusion has played a con-
spicuous part, but which still show evidences of transverse
fission, are shown in Figs. 56, 57, and 58. In Fig. 58 the ante-
rior projection probably represents the remnant of the fused

70 PATTEN. [VoL. XII,

anterior thoracic appendages; and the posterior infolding, the
invaginated remnants of the posterior thoracic appendages.
It is not improbable that Figs. 74 to 78, Pl. VII, represent
still further degeneration of this kind.

We cannot be certain whether a common form of embryo
such as that in Figs. 48 and 94 is due to gradual antero-posterior
fusion and degeneration up to the fourth thoracic appendages,
or whether there has been previous fission at that point, fol-
lowed by the degeneration of the anterior portion. But in
either case it shows a distinct difference in the vitality of the
two parts, which confirms our view of their morphological
difference.

In Fig. 103 we have the most remarkable example of con-
striction, followed by antero-posterior degeneration, that has
been met with. The least modified embryo, A, is a further
modification of the condition seen in the lower embryo in
Fig. 98. All the organs in the anterior portion have fused
leaving a median row of three long appendages, and fairly
developed cephalic lobes and oesophagus.

In embryo & the whole anterior portion has degenerated,
while the closely approximated dorsal organs appear like a pair
of eyes in front of what is left of the posterior part of the
thorax.

In embryo C the same process is carried still further, the dor-
sal organs now forming a median pit in front of the remainder
of the thorax, analogous to the median eyes of the cephalic lobes.

This condition has been reached by all three of the embryos
shown in Fig. 104.

The persistence of the posterior part of the thorax, following
transverse fission, is also well shown in the very advanced
embryos in Figs. 96,97, and 100. The point where transverse
constriction has occurred, and the degree of antero-posterior
degeneration are about the same in each.

That in cases like those just referred to there may be actual
separation of the anterior part of the thorax from the posterior
is shown by Fig. 102, in embryo 2, where in the two widely
separated portions of the thorax there is nothing left in the
anterior one except an imperfectly fused pair of appendages.

No: tr.) VARIATIONS IN LIMULUS POLYPHEMUS. ree

In conclusion, therefore, tt may be stated that transverse fission
occurs in the region of the fourth thoracic segment, dividing the
body into two parts, which show different morphological characters
and different degrees of vitality. The separation of the two parts
7s the result of a constriction brought about by the successive
median fusion and degeneration of the organs lying along that
segment, the most median, and therefore the oldest and most
specialized, fusing and degenerating first, and the others follow-
ing in the order of their position on the segment. There is a
gradual diminution of the effects of degeneration both in front
of and behind this line. A similar line of degeneration ts seen
zn the region of the cheliceral segment, just back of the cephalic
lobes, and in the region of the first abdominal or cheliceral seg-
ment, Fig.6. Further evidence of degeneration at these points is
seen in normal adult animals, in the small size of the organs on
these segments, and in their tendency to unite in the median
line. In the abnormal embryos it is shown by an exaggeration
of these conditions, resulting very frequently in the absence of
entire segments.

The inherent tendency to diminution of growth along such lines
7s what has, in all probability, led to the division of the body of
arthropods into successive regions, such as cephalic lobes, pro-
thorax, thorax, abdomen, post-abdomen, etc.

B. MEDIAN LONGITUDINAL FISSION; DOUBLE AND TRIPLE EMBRYOS.

This method of forming multiple embryos is comparatively
common. It often begins at a late period, after the full number
of normal appendages and metameres is formed.

The first steps in the process have not been observed, but
a study of Fig. 90, where fission has not progressed very far,
shows what the initiatory processes must have been like.
There can be no doubt that we have actual fission here, and
not fusion of two originally independent embryos. It is also
clearly shown by such cases as that in Fig. 98, where the left
half of the larger embryo is defective, owing to the lack of the
formative material necessary to produce an entire new half;
and especially by the fact that in all these cases the embryos

72. PATIIEWN. [VoL. XII.

match each other exactly, and always in the same way, which
could hardly be the case if two separate embryos had united with
cach other through accidental contact.

In Figs. 90 and 91, then, if we have fission instead of fusion,
one half of each new embryo must come, not from the fission
of already existing organs, but by the formation of two new
halves. Where does this new material come from, and by what
processes of growth are the new halves formed ?

In answer to the first question we may say at once that
there is not the slightest evidence of the existence of any
formative material in the shape of proliferating cells along the
median line where the new parts are forming. The old halves
are, to all appearance, separate from the new. As they are
already specialized, and sharply circumscribed, there seems no
way open to explain the origin of the new half of a segment
by lateral budding, or by regeneration, or by growth, from the
corresponding old one. We might suppose, perhaps, that the
new neuromeres in Figs. 90 and gi come from a kind of regen-
eration of the old one, but that could not possibly be the case
with any of the new organs lateral to the neuromere, such as
the appendages, sense-organs, and margin of the mesodermic
area.

We are left, then, entirely to conjecture as to the origin and
causation of the new growth. We shall return to this point
later.

The new halves are formed, however, in a very definite
manner, which we shall now proceed to explain. Jz brief, they
appear in exactly the reverse order of that by which the old ones
disappear by median fusion !

I. DOUBLE EMBRYOS.

In the formation of double embryos, two new halves
belonging ultimately to separate embryos, are produced, each
half being the mirror image of the other, to which it is
united along its /ateral margin. The new halves first appear
at the anterior end of the median line of the old embryo,
probably between the anterior median margins of the cephalic

No. I.] VARIATIONS [N LIMULUS POLVYPHEMUS. Wes

lobes. They form a triangular body which grows backwards at
the apex, and laterally in either direction, along lines parallel
with the base of the triangle. The two old halves are thus
wedged apart till, with the complete formation of the new
halves, the embryos form a straight line, tail to tail. Fig. 92.

Fic. 7.

Diagram to illustrate the law of formation of new halves in double embryos.
The new halves are shaded.

Each new organ of a metamere appears first as a single organ
common to both embryos, and having anormal position for each,
Fig. 7, Additional organs are formed in the same way, in the
order of their arrangement on the metamere. For example,
the organ nearest the median line is formed first; this then
divides into two, and the one lateral to it appears between
them as a single organ common to both embryos; this divides,
and the next one appears in the same place, till all the organs
of a given metamere are formed. The same process takes
place in the next posterior metamere, but it is always one step
behind that in the metamere in front of it.

The result is that in an embryo that has nearly completed
its division, as in the diagram (Fig. 7), we find a row of

74 PATTEN. [VoL. XII.

median, unpaired organs, which in their serial arrangement
follow the same order as in the metamere itself, namely, a, 3,
Grae,

It is thus obvious that the rate and direction of growth is
such that each new half tends to form a right-angled triangle,
the apex of which coincides with the posterior end of the
embryo, the altitude with the mid-ventral line, and the base
with the width of the oldest half-metamere.

We thus see a gradually increasing series of new organs
appear along the altitude of the triangle, or the old mid-ventral
line. They attain their full size and perfection of form for
that stage; then each divides into two (the one a mirror image
of the other), which move away from the median line, and in
their former place appears a new unpaired set, composed of
the organs that normally lie lateral to the ones just formed.

The successive ‘eruption of new series of organs along this
median line, and the manner in which they divide and move
away from it to right and left, is so entirely different from
what we have been accustomed to see that it is very impressive.
This effect is not diminished on further reflection.

Examination of the diagrammatic figure illustrating an
incomplete double embryo shows that each metamere has a
lateral growth similar to that which occurs at the posterior end
of segmented animals. Jz posterior, apical growth a number
of like parts, or segments, tncreasing in age and in differentia-
tion toward the anterior end, is produced. In the lateral growth
of a half-metamere a series of unlike parts ts produced. Lut
while there is a similar method of growth in both cases, there ts
in the second case a greater increase of specialization in passing
Jrom the growing point toward the part first formed.

If we carry the process seen in Pl. VIII, Fig. 90, back to its
beginning, we are led to conclude that fission began by the for-
mation of new organs in the median line at the very anterior
end of the body, that is, in the indentation separating the right
and left semicircular lobes of the brain. The first organs to
appear then must have been the cephalic lobes. But each half
of a cephalic lobe consists of at least two parts, a lateral one,
the optic ganglion, and a median one, the semicircular lobe and

No. I.] VARIATIONS IN LIMULUS POLYPHEMUS. 75

the cerebral hemisphere. It is not probable, therefore, that a
common median cephalic lobe developed in the same way as a
common appendage, for the optic ganglion of one side would
have to take the place of the cerebral hemisphere of the other,
or vice versa. The different parts are probably formed as
independent organs in a sequence from the median line later-
ally, as is the case with the different organs on a metamere.
That is, a common semicircular lobe is first formed, then the
cerebral hemisphere, then the optic ganglia, and finally the
lateral eyes.

In the earliest stage of a double embryo observed, Fig. go,
we see how the characteristic median row of unpaired organs is
forming. The fourth neuromere and the third appendage are
unpaired, but the tips of the second appendage are just visible
as two minute papillae, at the summit of a great bilobed
projection.

A comparison of this appendage with the third median pair
in Fig. 91 shows that cach new median appendage divides first
at the apex, the separation gradually extending toward the base.
This, tt will be observed, is the exact reverse of what occurs
when the appendages of the right and left sides unite to form a
single median one. Compare Figs. 42, 43, 48, and 49.

In the median line, at the anterior end of the double embryo
shown in Fig. go, is a small depression in a dark mass of cells.
The pit probably represents the common Ax/age of the dorso-
ventral muscles, which are seen to the right and left of each
head, reaching the surface ectoderm near the apex of the optic
ganglion.

In Fig. 91 the separation of the two heads has, by the wedge-
like ingrowth of the two new halves, been carried down to the
fifth thoracic metamere.

There is perhaps an actively growing point at the apex of the
V, which gradually works backward, thrusting the old halves
apart. But there is no indication whatever in surface views of
such a proliferation, for each new part has the appearance of
being as complete in every detail as the corresponding organs
in the old halves.

My sections of double embryos were not perfect enough to be

76 PATTEN. [Vou. XII.

of much assistance. I doubt, however, whether the most perfect
sections would indicate any materially different condition from
that seen in the surface views. It is somewhat surprising that
the tension necessary to push the old halves over the yolk does
not produce in them any other distortion than a gentle curvature
to right and left. It shows how truly each part assumes its char-
acteristic forms, dominated by its inherent structure rather than
by the mechanical stress of adjacent organs. For example, in
swinging the head of either embryo to the right or left, the
movement may coincide on one side with certain lines of
growth, but be directly opposed to them on the other. This
may be seen in the diagram, Fig. 7, where it is obvious
that the left side of the right-hand embryo is being swept
along in the direction of its own lateral growth, and it must
receive some additional impulse with which to overcome the
resistance to it. But on the right side of the same embryo
the movement is against the line of lateral growth. The
internal stress at the lateral C’s is very different from that
at corresponding points in a single embryo, and very different
from what it is at the median C, and yet the resulting organs
under these diverse conditions are the same!

The reason these varying mechanical conditions have so
little effect on the form of the organs is probably because they
are so transitory. They differ essentially from the permanent
and gradually increasing stresses that produce concrescence
and degeneration.

In Fig. 92 the separation into two embryos began at an
earlier period and is carried much farther than in the ones just
considered. The head of each embryo has been swept over
an arc of 90°. Further movement in that direction is pre-
vented by the interference of the lower margins (as the figure
stands) of the old mesodermic area. It is probable that the
tension produced by this interference would about equal the
devaricating force in the new halves, as soon as the two
embryos formed a straight line tail to tail. If this condition is
reached at an early period, the same forces, 7.e. the tendency
of the posterior margin of the mesodermic area to concresce
behind cach embryo, will ultimately force them apart. But

No. 1.] VARIATIONS IN LIMULUS POLYPHEMUS. 77

if the embryo is well developed before division takes place,
the head ends will meet each other on the side of the egg
opposite to the tails and will thus tend to prevent their further
separation.

In Fig. 93 separation of the two embryos has taken place,
as shown by the arrows, in a manner similar to that in Fig. g2.
But before the new sixth thoracic appendage and those of the
abdomen were produced, median fusion and antero-posterior
degeneration took place in the left-hand embryo in the manner
so frequently observed in single embryos. See Pl. V. The
cephalic lobes have disappeared, and the first three thoracic
metameres have fused in the median line, leaving nothing but
a minute pit to represent the second pair of appendages, and
a small, median papilla to represent the third and fourth.

In Fig. 94 the same process is carried further. In the way
the egg now stands, the common axis of the two embryos was
originally nearly vertical, as in Fig. 95. The lower one then
moved upwards past the left side of the other one to its present
position. It now occupies the free surface of the yolk between
the dorsal margins of the right-hand embryo. Median fusion
and antero-posterior degeneration then followed, whether before
or after separation cannot be determined, reducing the left-
hand embryo to the posterior part of the thorax and the
abdomen. The latter has a very conspicuous tail lobe similar
to that in Pl. V, Fig. 48.

It is seen on comparing Figs. 90, 92, and 98, that one of the
hearts and tail lobes must have belonged to the original em-
bryo; the others must be entirely new. For example, in Fig. 92
the heart and tail lobe just above y will be formed by the con-
crescence of the right and left margins of the old embryo,
while the heart appearing at x will be entirely new. The two
sides are so much alike in Fig. 97 that we cannot tell which is
the old half and which the new. Now if the tail ends of
these embryos should grow past each other, as in Figs. 94 and
96, then one embryo would carry off, according as it passed to
the right or left of the other, either a new tail and a new heart,
or the old ones. If we knew whether the heart of the original
embryo in Fig. 97 was to the right or the left, as the figure now

78 PAT ike Ve [VoL. XII.

stands, we could tell which was the new tail and which the old.
This can be done in Fig. 96, for it is obvious that both halves
of the tail lobe of embryo A, probably as far up as the last left
thoracic appendage, are those of the original embryo. Embryo
B, however, has carried off an entirely new tail lobe. The result
is that the right half of embryo 4 will consist of the right half
of the original embryo, and all its left half will be a new forma-
tion, except the abdominal part. In embryo 4, its left half is
that of the old embryo, except the posterior end, which is
new, and was probably produced by the backward regeneration
of that half of the body. Its right side is entirely new. A
similar condition must prevail in Fig. 94, except that embryo
B has here carried off the o/d abdominal lobe, and A the new
one.

These conditions are best seen in the diagrams, Figs. 8 and
9, where the shaded portions represent the old parts, and the
light ones the new. In Fig. 9, embryo J pushes past the left
side of A and carries off a new heart and tail lobe, and the
posterior portion of the abdomen on the left side.

In the well-balanced condition seen in Fig. 97, it is unlikely
that further changes in the relative positions of the two
embryos would take place. But if they should be forced to
grow past each other, a different proportional combination of
old and new elements would be produced from that in Fig. 96,
and the cause of this would be due, in part at least, to the
different periods in the formation of the new halves at which
the embryos separated from each other.

In Fig. 95 the two embryos, arising by longitudinal fission as
in the preceding ones, are still united tail to tail, but degenera-
tion of the lower one has progressed so far as to reduce it to a
slipper-shaped thickening with a depression in its centre, from
which arises a papilla, probably representing the last trace of
the fused sixth pair of thoracic appendages. A dark-rimmed
depression in front of this probably represents the remnants of
another pair.

In Fig. 98, division has produced two nearly complete em-
bryos, the new half of the abdomen of each embryo being
absent. The lower embryo has already begun to degenerate,

No. 1.] VARIATIONS IN LIMULUS POLYPHEMUS. 79

and presents a very good case of transverse fission, accom-

Fig. 8.

Diagram to illustrate the probable proportional composition of double embryos out of
old and new parts.

Fic. 9.

Diagram to illustrate the relation of old and new parts in the abdomen of separated double embryos.

panied by median fusion. There is a marked “weakness”
of the new half of the upper embryo, shown by the absence of

80 PALE. [Vou. XII.

the second, and the small size of the invaginated fourth,
appendage. There is no indication of segmentation lateral to
these appendages, and the left half of the neuromere opposite
the fourth appendage is apparently absent. The weakness of
this half is further shown by a gentle curvature of the head
toward the left, although at this stage in other double embryos
it is curved to the right.

In Fig. 97 is a very late stage (about the trilobite stage) of a
double embryo. The two embryos are tail to tail in a straight
line, and so symmetrically developed that there is no indication
whatever of the direction in which the two embryos separated
from the primitive median plane. Median fusion and antero-
posterior degeneration of each embryo has progressed as far as
the fourth thoracic segment. As indicated by the large seg-
mental sense-organ and the appendages, fusion and degenera-
tion have progressed farther in the lower embryo than in the
upper one. Interesting features of these two embryos are the
two perfect hearts and tails, extending laterally at right angles
to the continuous nerve-cords, and the large flabellae, which
look like a separate set of appendages.

In Fig. 96 is what appears to be a case of division of a dif-
ferent nature from those just described ; but a careful examina-
tion will show, I think, that it is after all the same. We can
easily and satisfactorily explain its condition by assuming that
longitudinal division gave rise to two embryos in a straight line
tail to tail, and that they separated and pushed past each other
in opposite directions. The anterior end of the right hand
embryo is turned vertically downward and has undergone me-
dian fusion and degeneration. This resulted in the fusion of
the fourth and fifth pairs of thoracic appendages, leaving the
sixth pair and the abdomen in a nearly normal condition. The
abdomen, as indicated by the dotted line, has finally been thrown
sharply to the right by the growth of the left side of the
thorax of the larger embryo. The result is an embryo much
like that in Fig. 97, only its longitudinal axis is bent at
right angles, the anterior portion of what is left being in its
primitive position, parallel with the axis of the larger
embryo.

No. I.] VARIATIONS TN LIMULUS POLYPHEMUS. 8I

In Fig. 100 is a much older embryo, with the remains of a
second rudimentary one attached to its right side. Separation
of the two embryos probably took place in the direction indi-
cated by the arrows,— the left-hand embryo undergoing
median fusion and antero-posterior degeneration. The re-
maining appendages are so twisted and obscure that their
identity could not be certainly determined. They appear to
represent the fused third, fourth, and fifth pairs of thoracic
appendages arranged in single line. The sixth pair are fused
at the base, leaving the ends free. The first two pairs of abdom-
inal appendages have also fused, something that rarely occurs
with them, to form two median, tongue-like projections, each
of which is bent almost at right angles.

On examination from the dorsal side, Fig. 1o1, the outlines
of the two embryos are distinctly seen. The median dark
streak in the smaller embryo probably represents the rem-
nants of the oesophagus, or perhaps the heart. The large semi-
lunar band of cells consists of the fibre cells that I have
described elsewhere, and represents the concresced margins of
the mesodermic area. It should lie on the anterior dorsal
surface of the thorax, but is here thrown forward towards the
ventral surface.

The only instance observed in which there seems to be a
deviation from the method of forming double embryos, just
described, is shown in Fig. 99. This embryo was accidentally
destroyed before a finished drawing of it was made, but
it had been completely outlined and carefully studied. There
is no question therefore about the correctness of the details of
structure, as far as they are given.

At first sight the right-hand embryo appears to have arisen
by longitudinal regeneration from the left side of the larger
one. Aside from the fact that no indication of such a process
has been seen elsewhere, it is difficult to imagine in detail the
method by which it was brought about. It is not necessary to
discuss these possibilities, as long as we can reduce this type
to the usual one by assuming that the form of degeneration
frequently seen in single embryos, has affected one of the
embryos almost coincident with its separation from the other.

82 PATTEN. [VoL. XII.

For example, if fission progressed in Fig. 91 down to the
abdominal region, leaving the newly formed abdominal ap-
pendages as unpaired organs, just as the fourth appendage now
is ; and if the anterior part of the right-hand embryo under-
went median fusion and antero-posterior degeneration up to
the fifth thoracic metamere, we would then have a condition
like that in Fig. 99. I see little reason to doubt that the
embryo in question was formed approximately in that way.

2. TRIPLE EMBRYOS.

Illustrations of triple embryos are seen in Figs. 102 to 104.
The steps by which they were produced were probably as
follows: It is assumed that in the beginning there was a
single, normal embryo, and that it gave rise, by longitudinal
fission, in the manner already described, to two embryos, each
one composed of a new half and an old one, Fig. 7. The
right-hand embryo then divides in the same way as the first,
by the formation of two new halves, Fig. 10. The result
is that the halves of the original embryo are now separated
from each other by an angle of about 240°. The second
embryo, &, is an entirely new formation, but embryo A con-
sists of the original right half plus a new left half, and embryo C
consists of the old left half plus a new right half.

The original line of concrescence of the posterior margins of
the mesodermic areas, 77.’ (along which the heart is formed ),
remains unchanged, except in its position on the yolk, just as
the original line of concrescence becomes the lower one in the
double embryo. The other two heart lines, //z.?2 and “73, are
entirely new formations.

If we may speak of the new halves as separate genera-
tions, their relations to each other in a triple embryo are
as follows: In embryo A the body consists on the right
side of the right half of a mother, and on the other of the
left half of a daughter. Embryo 2B is composed on the right
of the right half of a daughter, and on the left of the left
half of a granddaughter. Embryo C is composed on the left
side of the left half of a mother, and on the right side of the

NiO: 1, VARIATIONS IN LIMULUS POLYPHEMUS.

83
right half of a granddaughter. The line of concrescence of
the heart lines will have at H/7.? on one side the right half of
one daughter, and on the other the left half of another daugh-
ter. These two halves cannot be said to belong to the
same daughter without assuming some preformation jugglery
by which the molecules of one half were shifted to the
wrong side of the other. At //73 there is on one side of
the concrescence line the margin of the right mesodermic

m

So) = PES

all

fee

FIG. 10.

Diagram of a triple embryo, to show the relation of the old and new parts.
1+1'3 second generation, 2+ 2'; third generation, 3 + 3’.

Original halves,
area of a granddaughter, and on the other the margin of an-

other granddaughter. At //7.7 the concrescence line has, on

either side, in their original relations, the right and left mar-
gins of the mesodermic area of the mother. We may there-
fore call the hearts, or other organs developed along these three
lines, respectively the mother, daughter, and granddaughter
hearts, etc.

In explaining the condition of the triple embryos in Figs.

102 to 104, we shall assume that the original embryo divided

84 (PAIL: [VoOE.Zon:

lengthwise in the manner already described, producing A and
BC, and that BC divided, giving rise to Band C. In Fig. 102
A remains practically normal ; 4 has undergone median fusion
and transverse fission across the line of the fourth segment.
The abdomen and last two thoracic appendages are practically
normal, while the anterior part of the thorax and cephalic lobes
have disappeared, except one pair of fused appendages.

Passing around to embryo C, we find median fusion and degen-
eration have obliterated everything but the abdomen and the
last pair of fused thoracic appendages.

In Fig. 103, 4 has undergone median fusion and degenera-
tion, forming a pretty good example of an hour-glass embryo.
The same process has affected 4, obliterating entirely the
cephalic lobes and anterior portion of the thorax. The dorsal
organs, however, are not quite fused in the median line. But
this has taken place:in C, and in other respects the degenera-
tion is carried farther than in &. In both these triple embryos,
then, the path of increasing degeneration ts that of a spiral
from A to C.}

In Fig. 104 all three embryos are reduced so nearly to the
same level that it is hard to determine which is the most de-
generate. They are reduced to the last two fused thoracic
appendages and a remnant of the abdomen. Embryo C has the
smallest appendages and may be taken to be the most degener-
ate. But between A and BZ there is so little distinction that one
cannot determine whether the line of degeneration follows a
right or a left handed spiral. We shall return to this point later.

Discussion of Observations on Defective and Exuberant Em-
bryos. — There is little hope in the present condition of our
knowledge of finding anything like a satisfactory explanation
of the phenomena of either defective or exuberant embryos, be-
cause the solution of the problem is bound up in that of vitality.
While the futility of seeking final explanations of vital pheno-
mena is fully recognized, we have ventured, supported by the
facts on variation here described, to approach some of the out-
posts of the subject. These facts are numerous and in some

1 Possibly, from C to 4, as more recent evidence indicates.

No. I.] VARIATIONS [IN LIMULUS POLYPHEMUS. 85

instances novel, and it will probably be admitted that they are
somewhat conducive to speculation. I shall make them alone
the basis of the argument which follows. There is an obvious
advantage in treating the subject in this way, for the chances
of misconception due to confounding unlike results or condi-
tions with one another are thereby reduced to a minimum.

We may assume that the three embryos of triple monsters
are endowed at the outset with equal potentialities and that it
is merely a question of time that will bring them all to the
same condition of degeneration. But one embryo must be
older than the other two, both of which are of the same age.
If the age of the embryo, that is the time that has elapsed since
it became an independent embryo, determines the amount of
degeneration, then C, which is the most degenerate, ought to
be the oldest ; and 4 and C, being of the same age, ought to
show the same degree of degeneration. But as this is not the
case we must assume that some other factor than the time each
has had to degenerate determines the degree of degeneration.
We may also dismiss as possible factors the environment, for as
we have seen that was the same for all classes, yet, in spite of
that, defective, exuberant, multiple, and normal embryos were
produced ; and also that when the conditions of development
were made excessively abnormal, no abnormal embryos were to
be found.

In discussing the phenomena of multiple embryos as well as
the other variations described, we must bear in mind the follow-
ing facts, which although apparently contradictory in some
cases, must nevertheless be made to harmonize and be mutually
confirmatory before any approach to an explanation is possible.
These facts are :

(1) Whatever variations are here considered are probably
due primarily to structural variations resident in the ovum, and
not to differences in the environment.

1 It is obvious that we should find at least as many abnormalities under the
abnormal condition as under the normal, especially if the primary cause of the
variation is to be sought in the eggs themselves. The absence of abnormalities
under the former condition is probably due to the fact that under the prolonged

drastic treatment to which they were subjected, only the normal healthy ones
survived.

86 PATTEN. Vou EXD;

(2) There is a great difference in the growth period under
apparently the same conditions.

(3) There is a great difference in the size of different
embryos, some being much larger than the normal and others
smaller.

(4) Certain organs or regions of the body may be entirely
absent and are not subsequently restored.

(5) When organs once formed disappear in certain regions,
it is usually by median fusion and degeneration in the reverse
order of their age and specialization.

(6) Multiple embryos are due to the formation of new parts,
which appear in the reverse order of that in which old organs
disappear by median fusion and antero-posterior degeneration.

(7) Multiple embryos thus formed quickly disappear again
by median fusion and antero-posterior degeneration.

(8) Individuals of triple embryos recently formed differ
greatly in size and in the amount of degeneration.

(9) In old triple embryos the individuals are more nearly alike.

Taking up each set of facts, except the first, it would seem
from a consideration of the variations of the second class that
every individual and every part of it has a definite vate and range
of growth, a rise and a decline like a time clock that has been
set to go a definite period at a definite rate. The time for the
whole embryo may be reduced apparently to almost any frac-
tion of the normal one, as in those that die a natural death
before reaching stage C’; or individual organs may die and dis-
appear by median fusion and degeneration before the other
organs have completed their development ; and defective parts
either remain defective, or dwindle and disappear in the midst of
plenty, side by side with healthy flourishing organs. Embryos
six or eight months old are found in the same stage, to all
appearances, as normal ones only six or eight weeks old.
Both these kinds of variation are apparently best explained
by assuming a primary variation in the quantity or quality
of growth material, or of both. Either variation may ex-
press itself as a variation in the intensity of the “growth
force” that may be measured in terms of the range of develop-
ment, that is, the number of intermediate stages produced, or

INOS 1.] VARIATIONS IN LIMULUS POLYPHEMUS. 87

in terms of the rate. The original variation may be local or
general, but in either case, the deficiency, if there be one, is
not restored by nutrition, or by drafts on a general supply.

(3 and 4). Variation in the size of separate organs or of the
whole embryo, or the entire absence of organs, seems to be
due to a variation in the amount of formative material of that
particular kind out of which the variable organs are formed.

The variation in the amount of formative material expresses
itself in these cases in a variation in the size of the parts, not
in their rate or range of development. The formative material
when diminished or absent does not appear to be restored or re-
plenished, because regeneration of defective or absent parts
does not take place, although the organs that are present find
plenty of material with which to continue their own growth.

The thing then that is lacking in this case seems to have a
definite location and is not distributed throughout the embryo.
The deficiency is due apparently to the absence of some kind
of formative material, and not merely to a diminution of its
formative powers or to a variation in the quality.

(5). Organs once formed frequently disappear by median
fusion and degeneration, zz the reverse order of their age
and specialization. In other words, on a given segment
the right and left organs telescope into each other at the
median line, and disappear one after the other in the order
of their position and original formation. This mode of
degeneration must be due to some inherent structural con-
ditions, and not alone to mechanical stress or tension, because
it always occurs in the same way at various but well-determined
places, where the mechanical stresses due to growth of the sur-
rounding parts must be quite different. We may assume that
the reason the median end of a half segment disappears first is
because it is most specialized or most highly developed, and
therefore most likely to feel the effects of diminished vitality and
increased tension. The other parts follow in order for the
same reason, but the reason the latter move bodily toward the
median line is because a path of least resistance is constantly
reéstablished there by the degeneration of the organs nearest
that point.

88 PATTEN. iVous Xie

(6). Multiple embryos are produced by the formation of new
parts which appear in a definite sequence and in a definite
place, and in a manner the reverse of that by which old organs
disappear by median fusion and degeneration.

This result may be attributed either to the presence of an
excess of formative material, or to forced drafts on the reserve
brought about by some unknown internal conditions: If the
former supposition were correct, there is no obvious reason why
the additional parts should not be permanent acquisitions. We
cannot assume that such an arrangement of parts as we see in
multiple embryos is necessarily fatal to their codrdination, be-
cause the large numbers of double embryos of all kinds that
have reached advanced stages of development, as is well known,
would testify to the contrary. On the other hand, our next
fact (7), that multiple embryos almost immediately degenerate
by median fusion arid antero-posterior degeneration, shows, it
seems to me, that ¢here zs no production of new “ formative
material,’ but a misdirection of that already existing. There
7s in Some way the dividing up of the sum total of formative
material so that it crystallizes out along two or three lines
znstead of one.

The fact that these new centres start out all right but soon
begin to degenerate in the same way that poorly endowed
single embryos do, shows that there was not enough formative
material to go round ; that multiplying the formative centres
simply cuts off specialization and longevity at the other
end.

This supposition is still further supported by the next two
facts, namely (8) :

The individuals of triple embryos recently formed differ
greatly in size and in the amount of degeneration, and (9)
In old triple embryos the individuals are more nearly alike.
Before we consider these two points further, let us assume for
a moment that in the formation of multiple embryos there
has been, figuratively speaking, a forced growth producing
material faster than it can be differentiated, and that the point
of greatest differentiation tension is at the oldest point, where
the most growth and specialization had already taken place.

OO Es

No. I.] VARIATIONS IN LIMULUS POLYPHEMUS. 89

This point, as we have already shown, is the median part of
the most anterior end of the body. The tension there would
be most likely to be relieved by the production of a new organ
at that point, exactly duplicating in size and differentiation the
one side of which it was formed. If the differentiation tension,
if we may use the term, is still too great, it will be relieved by
the production of more new organs at the next point of greatest
differentiation, namely, at a point lateral to the two newly
formed organs, and on the median side of the corresponding
organ in the next posterior segment. The process might
go on till two entirely new halves were produced in that
way, or it might stop at any time that the extra tension was
relieved.

This supposition may in a measure account for the remark-
able fact that the newly formed organs do not pass through
the various stages of development the other organs did, but
assume at once whatever characters the old organs may have
at that time. For example, the tip of the new unpaired append-
age in Figs. 90 and gI is just as perfect in form and character
for that age as the other appendages on the old halves, and for
every successive part of it that appears, it is the same. It
looks as though the perfect appendage had been previously
concealed in the yolk, and gradually rose to the surface point
first, till completely exposed.

If two entirely separate embryos are produced by this forcing
process, the growth force, if it is an exhaustible quantity, will
be irretrievably subdivided between two embryos. Each will
have half the energy of the parent, provided there has been an
equal distribution between the new and old halves. But on the
formation of a triple embryo, the undivided one will have twice
the energy of either of the other two.

On the other hand, if there is at first no recuperation of
embryo B from A, or C from &, the latter must contain the least
amount, and C less than 4, provided there is recuperation of
the weaker half of each embryo from the stronger during the
process of division. This is clearly in accordance with their
degrees of degeneration, as shown in Fig. 102. But if we
suppose that there is subsequently a slow diffusion of formative

gO PATTEN. [Vor. XET.

energy throughout the three embryos, so that each one gets an
equal share, there will be a tendency to bring all three embryos
down to the same grade of degeneration. We may thus explain
the diminished difference between the three embryos in Fig. 103
and those in 102; and finally in Fig. 104, where the three embryos
have evidently been formed a long time, all three are reduced
to nearly the same condition. At first sight these conclusions
seem to be in direct opposition to the fact that in single
asymmetrical embryos ¢here 7s no indication whatever that the
stronger half possesses any power by virtue of which missing
parts on the opposite side of the median line can be restored.
If one embryo can produce two new halves by drawing on
its bank account, as in the formation of double and triple
embryos, why do not defective single embryos restore an
absent half or quarter? We can only say that in such
cases the right half of the embryo, for example, was absent
because the formative material for that part was absent
from the start. Theoretically there is no reason why the
lost part should not be restored by a forced draft on the
corresponding organs of the whole side. Aut there is no
particular reason for assuming that forced growth would be
likely to occur in an embryo that was already defective in
growth material.

On the other hand, it is a point to be borne in mind that in
all the defective embryos shown in Pl. IV very few show a
defect at the anterior end of the body, unless, as in Fig. 38,
the whole left half is absent, or in Fig. 37, which is probably
a double embryo. The parts most frequently absent are the
posterior ones, or ones across the middle of the thorax. That
as, they are the parts least likely to be restored from the opposite
side, provided their restoration took place in the same order that
the organs of new halves are formed in double and triple
embryos.

According to Hertwig and others, the stimulus of a modified
environment alone is sufficient to call forth the new embryo. But
as in the cases we are considering the environment was nearly the
same for all, it can only mean that out of many thousands of eggs
the few that produced double or triple monsters were different at

No. I.] VARIATIONS [IN LIMULUS POLYPHEMUS. gI

the outset, and for that reason responded ina different way to the
same environment. Again, if itis the environment alone that pro-
duced the excessive growth manifested in the formation of two
or three embryos out of one, we should expect that under these
favorable conditions the new embryos would be large and
vigorous, but as a matter of fact they are not, for no sooner
are they once formed than they begin to degenerate, and
finally all of them may be reduced to mere remnants.

Hertwig’s criticism of Weismann’s explanation of polymor-
phism in ants and bees, as well as of multiple embryos, is, it
seems to me, a valid one. For, as he points out, we have no
right to assume that the embryo is provided with several sets
of germ plasm destined to develop, under proper stimuli, into
new embryos or organs, unless the necessary stimuli are likely
to occur in nature. And it might also be added that this con-
dition could not have been fixed by natural selection, if its
realization brought death with it.

On the other hand, the assumption that Hertwig makes, that
the stimulus of a new environment is alone sufficient to pro-
duce new formative material, is not a necessary one. Indeed,
the facts seem to me to point to an opposite conclusion.

The factors producing forced growth of this kind, it seems
fair to assume, are the reverse of those that cause median
concrescence, because the sequence of events is nearly the
reverse. That is, excessive tension causes organs to fuse and
disappear along the median line. Is it not probable then that
abnormal reduction of this tension would facilitate the forma-
tion of new organs there, which will then reverse the tension
conditions and cause the organs to disappear ?

The formation of multiple embryos may be the result of
some inherent defect in the ovum, and the stimulus of a modi-
fied environment may exaggerate these defects, and increase
the percentage of multiple embryos; but there is no evidence
to show among the higher, segmented animals, at least, that
double or triple embryos may be produced by the sole action
of any definite environment.

The fact that double and triple monsters appear under abnor-
mal conditions, is not a valid argument against the mosaic

92 PATTEN. [Vou. XII.

theory, or, at least, not in its widest sense, but only against
the rigid limitations placed on it by Weismann and others.

It is well known that an embryo may, under stress of new
environment, divide and produce two or more new ones ; and
perfectly formed embryos are produced from fragments of
segmenting eggs, but these embryos have suffered either a
diminution tn size, or in vitality, or both. No one, so far as I
know, has succeeded in raising a fraction of the original ovum
into a mature individual showing the same longevity and
vitality as those raised from whole ova. The same is true of
double or triple embryos produced by excessive growth.

If we assume with Weismann that the material for two or
more embryos is present in each egg that is capable of producing
multiple embryos, and that the stimulus of the environment has
called them all into activity, it will be difficult to explain how
it happens that the newly formed parts disappear so soon after
formation, and also—a difficulty that, so far as I know, has not
been foreseen before — why the new material should be in tso-
lated halves of different individuals instead of in one or more
entirely new individuals distinct from the old, or why these
halves should develop and unite with the old ones in the way
they do.

It is probable that the same method of forming the new
parts in multiple embryos of Limulus is followed in other
animals, only it has not been recognized there, because the
embryological processes are more obscure.

It should also be observed in this connection that there is
no evidence that the new organs come from some indifferent
reserved cells. In addition to the fact that, under very favor-
able circumstances, no trace of such cells can be seen, it would
be difficult to explain why they should always first manifest
themselves at the point of greatest specialization, that is, at
the very anterior median line, and grow backwards instead of
forwards. We should, on the contrary, naturally look for
them at the posterior end, and expect that they would produce
a new embryo there in the usual manner.

It seems to me that we must assume that in all normal ova
there is a definite quality and quantity of formative material

NOs i] VARIATIONS IN LIMULUS POLYPHEMUS. 93

that will under normal conditions produce at the end of a
certain time an animal of a given form, capable of performing
a number of activities. The capital each ovum starts with
determines the result, if the conditions are normal. A change
in environment may retard or accelerate the mechanism ; it may
throw certain parts into new tracks ; it may influence the dis-
tribution of formative material as a whole, but it cannot, within
the period of one generation, change the nature of the formative
material.

However the machine may vary, we recognize it as the same
machine, — incomplete perhaps, but if so, we recognize the
vacant places. In very rare cases (only one observed) a Limu-
lus embryo may have more than twelve thoracic appendages,
but the extra appendage is exactly like one already existing.
In no case is there found a new organ or part different
in kind from those already existing, and in no case is an
organ, out) of place in reference to others. The chelicerae
always come back of the cephalic lobes; and a flabellum,
if it is present at all, always occurs on the outer margin
of the sixth pair of appendages. We can only attribute
the original absence of an organ, or of any part of the
embryo, and the subsequent failure to reproduce that organ,
to the original absence or diversion into other channels of
the material out of which that organ was to have been
formed.

We can explain the formation of double and triple embryos,
not on the assumption that the original formative material has
been increased, but that it has been divided and diverted into
separate channels, and the consequent diminution in the quan-
tity available for the new embryos has been one of the causes
of their subsequent degeneration.

There are in normal embryos inherent lines of weakness
along which there is certain to be diminished growth. They
mark off the different regions of the body from each other; and
when through the action of the environment or through con-
genital conditions there is a diminution of vitality, it is shown
by the diminished size of these regions and by a tendency to
fuse along the median line.

94 PATTEN. [VoL. XII.

Increased vitality is shown by increased size, persistency of
form, and by the production of new organs in the reverse way
of that in which they disappear. If the increase is due to the
original presence of an increased amount of formative material
in the ovum, the increase in size, vitality, and number of organs
will be a permanent characteristic of the individual ; but if it is
due to the accelerating action of the environment, there will
be a corresponding loss at some other time or place. We may
further conclude that :

(1) In multiple embryos of Limulus the new organs are
formed by forced drafts on the old material.

(2) That the new halves are weaker in formative power than
the old.

(3) That equality is established by the interchange of ma-
terial from the stronger to the weaker halves.

(4) That at first, the sum total of formative energy in
both halves of embryo C is less than in Z, and in B than
in A.

(5) That equality is finally established between all these
embryos by interchange of material, so that in the end all
three are reduced to the same grade of degeneration.

(6) In defective single embryos, the absent parts are absent
because their specific formative material was absent. Theo-
retically the absent organs might be restored by forced growth
of the other side. But there is no reason to expect forced
growth will occur in an embryo already defective in formative
material. If it does occur, it will more likely be at the an-
terior end, and of course the restoration will not be detected.
If it does occur in that way, it explains why unilateral defects
are more frequently seen at the posterior than at the anterior
end.

VII. DEGENERATION AND DEATH OF LimuLUS EMBRYOS.

The causes of degeneration and death of the embryos de-
scribed in the preceding sections are due to abnormal limita-
tions in either the power of division, specialization, or longevity
of the cells, or to various combinations of the same.

No. I.] VARIATIONS IN LIMULUS POLYPHEMUS. 95

The structure of a fully developed adult animal, it seems to
me, must depend on the relation that exists between (1) the
rate of production of new cells, (2) the rate and the amount of
specialization of these cells, (3) the longevity of the completely
specialized cells, and (4) the death rate of the cells, or their
rate of decline toward a simpler, less specialized condition.
The interrelations of these factors must be extremely complex
and to a certain extent independent of each other ; for repro-
duction, specialization, and decay may, apparently, take place
simultaneously in any part of any organ, or in the entire
organism, in almost any conceivable proportion.

An essential feature of organic death in the higher animals
is the cessation of normal activity in many different organs,
because other organs, by accident, or by inherent conditions,
cease to perform some particular work, perhaps insignificant
in itself, upon which all the others depend.

The nerve cells, for example, may be performing faithfully
and well their particular work, and yet may perish for lack
of proper nutrition, or owing to the presence of poisonous
substances that should have been eliminated from the body.
We do not know how long they might have continued to act
under favorable conditions.

Every living thing has a more or less definite size, form,
life period, and kind of activity. It is apparently assumed by
some writers that these manifestations are not merely pre-
determined by the organization of the ovum, acting under the
guidance of a changing environment, but that there is an actual
deposit of formed material corresponding in some way to every
one of the almost infinitely numerous parts of the future
organism. That there are definite potentialities in every ovum
cannot be denied, but these potentialities must not be con-
founded with what actually exists as specific matter, and which
forms the actual physical basis of the ovum at a given time.

The future organism is the resultant of the action on an
initial, specifically constructed mass of forces outside of and
independent of the mass (1) of such forces acting on the newly
formed material added to the old, and (2) on the continued
interaction of these masses and forces on one another, under

96 PALTIEN. [Von. XIT-

constantly changing conditions brought about by these inter-
actions. To affirm that the resultants of such complex inter-
actions can be in any sense preformed is grossly inaccurate
and misleading.

To use the term “preformation” to designate in any way
the remote antecedents of a bit of living protoplasm is like
asserting that the germs of a banner cloud are preformed in
the south wind, or that the germs of a watch are preformed
in the iron and coal used in making it.

Embryological processes are not to be interpreted merely
as necessary preliminaries to a higher condition. There is no
beginning or end. Each phase and part of a living thing
should be treated as it actually is, and be accorded its full
value as a perfected, completed thing, — not as it is going to
be, ignoring the present to catch some imaginary glimpses of
the future.

In the higher animals, death comes to the organism in most
cases, it would seem, as a result of the increasingly complex
interrelation of cells. Non-living compounds accumulate in
the tissues with age and destroy their elasticity or permeability ;
the calibre of conducting tubes is diminished and they fail to
furnish the necessary supplies, or excessive demands lead to
rupture, lack of codrdination between volume and surface ex-
posure, excess of capillary friction accompanying increased
size of organs, etc. Such slight defects may place insuperable
barriers to further growth and specialization. The same may
be true of the individual cells themselves. We need not
assume that they cease their activities at the end of a certain
period because a certain amount of inherited energy has been
liberated, as in the uncoiling of a watch-spring, but because
they cannot clean themselves or adjust the necessary repairs to
the new conditions. Environment, within and without, winds
up and liberates the springs, not the ancestors. As long as the
environment does this, the vital mechanism will continue to go
till stopped by the products of its own activity, and the mechani-
cal impossibility of adjusting old parts to new requirements.

The more complex the relation of parts and the more per-
fect their mutual adaptation, the more sure and complete will

)

No. 1.] VARIATIONS IN LIMULUS POLYPHEMUS. 97

be the collapse of the whole, when the working of one part
is interfered with. Each organism by itself is a world of
individual cells where heredity, use, disuse, and struggle
for existence are the determining factors of form and func-
tion, just as in the larger world organisms as a whole are
the resultants of these factors. In both cases, sudden and
great changes in the environment result in a rapid col-
lapse or death, owing to the impossibility of the cells on the
one hand, or the organism on the other, adapting themselves
to the new conditions ; and the result is death of the indi-
vidual organism in one case, or extermination of the species
in the other. If the change is a slow one and one calling
for less specialization, in fewer directions, degeneration, or
reversion to a simpler plan of structure, may follow till perhaps
something resembling the original starting-point has been
reached.

It 1s an exactly analagous process to this that occurs in some
of the degenerating embryos of Limulus. It is a new kind of
death for an individual organism, or, at least, not like the one
with which we are familar. The embryo, a community of
thousands of different kinds of cells, does not die like a nation
swept by a pestilence, or like the starving inmates of a
helpless vessel, but like a flourishing community in the midst
of plenty, where some of the infinite niceties of adjustment are
such as insensibly to reduce the birth rate below the death
rate, to reduce the complex interrelation of individuals, until
after many generations the last survivor, reduced to the lowest
terms, disappears.

The point we wish especially to emphasize is that in the
degenerating embryos of Limulus, cell reproduction, cell
specialization, and cell decay are progressing side by side in
every part of the body. Karyokinetic figures and the frag-
ments of decaying nuclei are found side by side. Whether
the animal develops or degenerates depends on the relative
intensity of these three factors. The embryos dwindle in size
because the death rate of the cells is greater than the birth
rate. The embryo loses nerve centres, sense-organs, and ap-
pendages because the amount of specialization of individual

98 PATTEN. [Vou. XII.

cells is cut down more and more, till only the simplest kinds
remain, or because the new ones die before they become
specialized. The lack of specialization may affect different
parts unequally, as when the surface details of structure fail
to appear on the otherwise well-developed appendages, as in
Fig. 21, or the nerve-cords, the appendages, or sense-organs
are lacking.

As we might expect, the process is never exactly the
same, but it invariably tends to carry the organism back, in
the main, over the old lines of progressive development, till
it reaches its primitive condition, namely, a small community
of untrained, insubordinate individuals, which in turn die
one by one, till the death of the last survivor exterminates
the race.

This may be called the true natural death of an organism,— all
others are more or less catastrophic, due to the increasing lack
of codrdination and adjustment to the new conditions produced
by itsown growth and specialization. A natural death like this
is only possible where the degree of specialization has been com-
paratively small, and where every cell may receive the stimulus
and material necessary to the continuance of its activities and
the discharge of its waste or noxious products.

If this be true, then there is no such distinction to be made
as “mortal and immortal” protoplasm. All protoplasm is “ im-
mortal,” in the same sense that chemical compounds, or mix-
tures of the same will continue to be formed, manifest their
specific properties, and disappear so long as the proper environ-
ment is maintained.

Cessation of vital activity, then, is due solely to inadequate
environment, whether we are dealing with highly organized
individuals, or with bacteria, amoebae, or human ganglion cells,
or with any part of these.

The growth of the smallest protoplasmic part of a cell differs
from the growth of a crystal in the fact that, having grown, the
conditions of growth and persistency are more materially
altered in the former than in the later case.

The metabolic changes that take place in and around the cell
are to a certain extent processes of filtration. The vital pro-

No. I.] VARIATIONS IN LIMULUS POLYPHEMUS. 99

cesses cease, not because some hidden spring within needs to
be renewed, or because, like a spent rocket, it needs to be
recharged, but because the still perfect mechanism is choked
with dust and weighed down with uneliminatable material.
Before the final collapse some virgin fragment of the original
material escapes to begin the same process anew.

We need not assume that this new material, or germ, is es-
sentially different from the living parts of the old machine.
We certainly need not look on it as a wonderful piece of pyro-
technics, with countless preéxisting fuses, percussion caps, and
hidden chambers, adjusted to discharge themselves at the right
instant and in the proper sequence —a something which needs
but the spark and afflatus of proper environment to start on
its heedless career with infinite splutterings and explosions, to
cease only when it becomes an empty shell!

There is in our opinion nothing to be gained by such a view
of developmental processes. Nevertheless, in discussing such
problems the use of figurative language cannot be avoided, for
nothing more than a vague conception can be formed of the
infinitely complex mechanism at work within the embryos even
of the very lowly organized animals. To follow its normal
course of action or development to the end, there must be in
the ovum infinite niceties in the qualitative and quantitative ad-
justment of the new parts to the old ; the chemical time-locks,
that are to fix or release, must be set with exactness to the time
and place, and the whole mechanism made adjustable to a
rather wide variation in its surroundings. This being the case,
who will venture to say that the entire absence of the third
thoracic appendage was due to the fact that a certain particle,
destined to grow into that particular appendage, was accident-
ally omitted from the chromatin of the segmentation nucleus ?

Is it necessary to suppose because the third left thoracic
appendage in Limulus was invaginated and its mate projected
freely from the surface, that the former was derived from a
peculiarly constructed molecule, or biophor, or whatever you
please to call it, that was bound to develop into an invaginated
appendage and not one projecting in the normal way? As well
assume that a river inherits two kinds of sand grains, one of

100 PAT TEIN.

such a peculiar structure that they give rise to sink holes, and
the other to sand banks !

The facts of variation are of value to the morphologist as
well as to the biological metaphysician. If, for example, invagi-
nated appendages are formed, and are formed frequently, it
shows that in that direction is, in its broadest sense, a path
of low resistance likely to be followed again and again. The
same is true of the deviations affecting the margin of the
mesodermic area, and in fact any of the deviations that occur
often and in a definite way.

In the inevitable shifting of internal relations that occur
in all living organisms, this path is as likely to become a per-
manent path of least resistance as the other is to remain as
it is. In other words, the normal and abnormal exchange
places.

In conclusion, there seems to me no evidence in the varia-
tions here described to support the theories which attempt to
explain heredity by assumptions that leave no room for the idea
that the ovum is an organism — theories which make the ovum
a mere receptacle to hold job lots of ancestral organs, to be
shuffled together and dealt out again during segmentation by
some archoplasmatic prestidigitator! It makes little difference
whether the germules, plastidules, biophores, or whatever we
may choose to call these corpuscular ‘‘ brownies,” come from
immediate producers of the ovum, or from ancestors ten thou-
sand generations back; they are things, it seems to me, which
exist only as mere names that help to bring before the mind a
few of the factors in an extremely complex process.

i ste re can
; ™ i
ea dee Aah
1% 7 ee ae a
Depa faa wy: Rie
a RTS Ars ava
eee br ;

ie wey by Voi

¢ ; wine ile hy 7 ‘ y

Wi erin!) Ae

Pann et i

oe 7) Oh _ Das hare . | a
nee Me i‘ : as
| | '¥ } ie
nya) Oda ee tae
| Py Qleey.
| ee) ov Coane
| | . ; Tae fata et yi
pe 0 iene il D:
WAeranh hae bl aarh i; tpn
‘eae er ner
yen anne eet eh |

102 PATTEN. [Vou. XII.

EXPLANATION OF PLATE II.

Nearly all the figures were outlined with a camera and drawn to the same scale. They were
made from mounted and cleared specimens, viewed with a raised condenser and wide open
diaphragm. In this way they appear bright red on a clear yellow field ; the elevations showing dark,
the depressions light.

The embryos 4 to Z represent normal stages introduced here for comparison with the abnormal
ones. Figs. 1 and 2 belong to an earlier series, where a different method of designating the stages

has been adopted.

Fic. 1, X 50. Surface view of a normal embryo, stained and cleared in oil of
cedar. The cephalic lobes form a clearly outlined, semicircular plate of ectoderm,
thickened on each side. There is no trace of an oesophagus, but very faint indi-
cations of the cheliceral segment may be seen on the posterior margin of the lobes.
The next three thoracic metameres are fully formed. In reflected light they
appear, in surface views of opaque preparations, as gentle undulations of the
surface, forming a series of ridges and valleys. The lateral ends of the low
ridges are bent backwards and unite with each other near the margin of the
mesodermic area. When the eggs are cleared in oil, the mesodermic segments
are seen beneath, and about coextensive with the surface ridges. The fifth seg-
ment is just separating off from the anal plate. Along the median line of the
latter is a slit-like invagination to form the Ze/opfore, from which a sheet of inner
layer cells extends laterally and forwards, while beneath the invagination a line of
cells extends into the yolk.

Fic. 2, X 50. Surface view of an embryo about 10 days old. The drawing was
made from studies of the opaque embryo in alcohol, and from studies of the same
embryo cleared in clove oil and stained in borax carmine. The sixth segment has
appeared on the anterior margin of the anal plate. The lateral ends of the four
preceding mesodermic somites have become confluent, and the mesodermic area
thus formed is provided with a distinct thickening along its lateral margins, m. a.,
forming what I have called the “ thickened rim of the mesodermic area.”

The cheliceral segment is now quite distinct, and a faint, dark band is seen in
front of the cheliceral segment connecting the two sides of the cephalic lobes.

Fic. 3, X 55. Surface view of embryo about twelve days old, stained in borax
carmine and cleared in oil of cedar.

All six thoracic segments are now formed. The sixth, as was the case at a
corresponding age with each of the preceding segments except the chelicerae, is
very plainly marked by its dark color, and by the sharp furrows that lie on either
side of it. ,

The second, third, and fourth pairs of appendages have appeared, and on either
side of them are faint spots, that are perhaps the beginnings of the sense-organs,
seen in this region more clearly at a later stage.

Fic. 4, X 50. Surface view of a normal embryo 14 days old, stained in borax
carmine and cleared in oil of cedar. The three pairs of appendages are more
conspicuously developed, the oesophagus has appeared at the anterior margin of
the cephalic lobes, and the brain and ventral nerve-cords are now easily distin-
guished by the characteristic mottlings, due to the presence of many minute pits,
precisely similar to those described and figured by me in scorpions. The appear-
ance is due to the presence in the nerve-cord and cephalic lobes of numerous
independent, bud-like thickenings of the ectoderm, which in histological special+

Nas Fo] VARIATIONS IN LIMULUS POLYPHEMUS. 103

zation and in general appearance resemble sense-organs. The suggestion that
the appearance is due to a folding to increase the surface, or that they represent
“neuroblasts” (Wheeler) is inadequate, and both suggestions are based on mis-
apprehension of the facts. We shall treat of them more fully elsewhere.

The rim of the mesodermic area is considerably enlarged, and is growing
towards the median line back of the anal plate.

The telopore in this particular individual has disappeared, but the median
streak of inner layer cells, derived from it, is still visible.

Fic. 5, X 60. Surface view of a normal embryo 18 (?) days old. All the
thoracic appendages except the first pair are now elongated processes having the
characteristic shape of thoracic appendages. The sixth pair are still transversely
elongated, and resemble in form the early stages in the development of the
abdominal appendages. The most striking feature of this stage is, however, the
series of large, shallow depressions on the outer side of the marginal fold, m./f.
It is only in exceptional cases that the depressions are visible. They are most
readily seen by reflected light in shelled ova, shortly after treatment with the
hardening reagent, picro-nitric acid. The second and third, and the fifth and
sixth, quickly disappear, leaving no trace behind. The first is a little smaller
than the rest and lies a little in front of the chelicerae, but appears to belong to
that segment. It develops into the lateral eyes. The fourth, d.0., lies exactly
opposite the fourth thoracic appendage, and growing rapidly, gives rise to the con-
spicuous organ, undoubtedly of a sensory nature, found in nearly the same position
as late as the time of casting the first larval shell. The lateral eyes gradually
move backwards till they lie on the dorsal side of, and posterior to, the so-called
dorsal organ, d.o. The rim of the mesodermic area is decidedly thickened, and
consists entirely of the remarkable cells containing a coiled fibre. Along the
whole extent of this thickening, the inner and outer layers of cells are con-
tinuous; elsewhere, except along the median line, they are sharply separated.
Sections show clearly that the inner layer cells along the entire length of the
thickening are receiving extensive additions from an inward proliferation of the
overlying ectoderm. The posterior margin of the mesodermic area is drawn out
into two backwardly directed lobes, which frequently show traces of segmentation
comparable with that seen in the early stages of the older somites (Fig. 1).

Fic. 6, X 41. Surface view of entire ovum about 21 days old. Picro-nitric
acid, borax carmine, balsam. The stained embryo is shown as a transparent
object.

The operculum and first pair of gills are well developed, but the chelaria are
as yet barely visible. The cephalic lobes have become specialized into the semi-
circular lobes, s./., and the optic ganglia of the lateral eyes, o.g.; the remaining
portions constitute the brain proper, 47. The ectodermic thickening to form the
corneagen of the median eyes is shown at m.e. The openings of the tubular
invaginations of the median eyes and their nerves are shown at /.e.¢. The fused
distal ends of these tubes will form the median eye vesicle ; the remaining distal
third of the tubes is converted directly into the corresponding portion of the
median eye nerve. The proximal portion of the nerves is formed by the separa-
tion of the nerve fibres from the peripheral surface of the tubes, leaving a collapsed
epithelial tube behind, which persists for a long time after hatching as a function-
less remnant. The separation of nerve fibres from the proximal ends of the eye
tubes takes place before these parts of the tubes fuse with each other, so that the

104 PATTEN.

distal end of the median eye nerve is unpaired, while the proximal end is paired.
The roots of the ,-shaped nerve terminate in the semicircular lobes.

The lateral eyes have moved backward, away from their original position
to the dorsal surface opposite the second and third thoracic appendages, Ze.
The “dorsal organ” has retained its original position on the fourth segment. The
dorso-ventral muscles, dividing the yolk into segments, and passing between
the primary liver lobes, have appeared.

Fic. 7, X 40. Surface view of a normal embryo about four weeks old. Picro-
nitric acid, borax carmine, clove oil. The point that will interest us here is the
further specialization of the cephalic lobes. The median eye tubes are still
unfused, and open by separate pores, #.¢e.¢. The proximal portion of the median
eye nerves is seen at p.m.n., connecting the median eye tubes, m.e.2., with the
semicircular lobes, s.2. The latter, which are formed by invagination of the
anterior margin of the cephalic lobes, have grown inwards and backwards, so that
they now lie on the dorsal surface of the same. They are seen through that part
of the brain which lies over them, and which will soon give rise to the cerebral
hemispheres.

The olfactory organs, o./., are now visible for the first time as oval ectodermic
thickenings on the anterior margin of the cephalic lobes, at the point where the
future cerebral hemispheres, the optic ganglion, and the semicircular lobes meet.
The coxal sense-organs are visible as elongated transverse ridges at the median
margin of the base of the appendages. From these thickenings arise the enor-
mous ganglia of the pedal nerves and the coxal sense-organs. See my paper on
the “ Brain and Sense-Organs of Limulus.”

i
ve

Pa

VARIATIONS IN LIMULUS POLYPHEMUS. 105

EXPLANATION OF PLATE III.

Fic. 8, X 60. This is a well-developed and nearly full-sized embryo of stage
C-D.

It is remarkable on account of the invagination of all the thoracic appendages
except the first and last pairs. The first pair are small for this stage, and hardly
recognizable even in sections. The last pair are well developed, and project freely
from the surface in the normal way.

The invagination of the remaining thoracic appendages is greatest on the left
side. The series of low oval elevations represents the uninvaginated bases of the
appendages, and the transverse slits at their summits, w.af., the openings leading
into deep cavities formed by the invaginations of the outer two-thirds of the
same. The dark ring around the slits represents the optical sections of the
ectoderm nuclei that form the wall of the invagination. Sections of these
appendages show that they do not differ materially from those shown in Pl. XI,
Fig. I.

The abdominal plate is very short and devoid of appendages. A small, deep
pit is present in it, which in sections resembles the proctodaeum. Comparison
with stages C and D indicates that the embryo is too young to be normally pro-
vided with a proctodaeum. It probably represents the last stages of the se/opore.
The cephalic lobes are abnormal in shape and structure. They contain two
central depressed areas of thickened ectoderm, appearing as light round spots in
surface views, 47.zv, and in section, Pl. XI, Fig. 8. The optic ganglia, of.g., are
partly concealed by a broad fold of ectoderm, whose free edge is directed diagonally
backward, g.f, Pl. XI, Figs. 8 and 8%. This fold does not extend across the
median line in front of the oesophagus.

The lateral eyes are conspicuous, but the sense-organs opposite the third
thoracic appendage are not visible in surface views, although they can easily be
detected in sections.

Fic. 9, X 60. This embryo is between stages C and D. The anterior end of
the body is narrowed, and there is a slight lateral constriction opposite the fourth
pair of appendages. The latter are completely invaginated into the yolk, leaving
two slit-like openings on the surface of the embryo. The boundaries of the inner
ends of the invaginated appendages are not sharply defined, and in the yolk
around them is a halo of degenerating nuclei apparently formed by the disintegra-
tion of its walls. The thoracic portion of the marginal fold is very faint, except at
the posterior end. The whole embryo lies flush with the surrounding surfaces,
instead of being deeply depressed as in Figs. 8, Io, 12.

The cephalic lobes are clearly outlined. A series of transverse sections (Pl. XI,
Fig. 9) shows that the dark area bounded by the lines ec. and gf is covered by an
amnion-like, ectodermic fold (although the middle layer is very obscure), extending
backward and medianly over the cephalic lobes. Over the light inclosed area, 6r.,
there is no superficial ectoderm, and the very light spots, 47.zv., are the same pit-
like depressions of the brain seen in Figs.8-12. The ectoderm over these depres-
sions is covered with a layer of vertically striated, cuticular substance, looking like
a layer of cilia, and similar to that seen over the surface of most developing sense-
organs.

106 PATTEN. EVOL, xaT

Fic. 10, X 60. A large embryo, a little older than stage C-D. All the thoracic
appendages are large. The third pair are completely invaginated, and project
their whole length into the yolk. The pockets thus formed open to the exterior
by two sharply circumscribed transverse slits, 7.

The apex of the fourth appendage of the left side, a/., is invaginated, and the
base is expanded and folded as though vertically compressed. A section of the
third pair is shown in Pl. XI, Fig. 10%. The ectoderm is sharply defined except at
the tip, x, where the arrangement of the nuclei indicates a proliferation of ectoderm
cells into the yolk. A similar condition has been observed in other embryos.
The cephalic lobes are modified in appearance by the extensive overgrowth of an
amnion-like, ganglionic fold along their anterior margin. Longitudinal sections in
the planes indicated by the lines 1, 2, and 3 are shown in Pl. XI, Figs. 1o!, 10%,
and 10%. Abnormal depressions in the region of the future cerebral hemispheres
are seen at 4y.zv., in Fig. 10, and in sections in Fig. 107.

Near the median line the brain is covered with a layer of ectoderm cells, Fig.
103, not developed in the lateral areas, Figs. 107 and 107.

The lateral eyes have moved back to a point opposite the second pair of
appendages. The large thoracic sense-organs are very feebly developed, if not
entirely absent.

There is a conspicuous mass of cells lying in the yolk back of the tail lobe,
pac. It is formed by the concrescence of the thickened rim of the mesodermic
area.

Fic. 11, X 60. This embryo is similar in general appearance to that shown in
Fig. 10. The thoracic appendages are, however, longer, and there are four pairs
of abdominal appendages present. The distal half of the third thoracic appendage
on the left side is invaginated, ¢4.af. The fourth and fifth appendages are
directed forwards instead of backwards. The brain-pits, 47.2v., are here smaller
than in Fig. to, and pushed farther forward beneath the ganglionic fold. The
whole brain and optic ganglia, of.g., are depressed below the level of the ventral
cord and the margin of the ganglionic fold. The ganglionic invagination, 2.2v., is
about the same as in Fig. 10. The light area, which indicates the mouth of this
slit-like depression, gradually runs into the rounded brain-pits on the median side,
brav. The ganglionic fold is continued as a thickened band, without infolding,
across the median anterior border of the cephalic lobes.

The lateral eyes are unusually conspicuous, and lie very nearly on a line with
the chelicerae.

A large peripheral vesicle, #v.ve., extends inwards almost to the centre of the
egg. It is a conical cavity in the yolk, lined with a thick layer of cells, derived
from an abnormal growth of the thickened margin of the mesodermic area. A
well-marked depression is present in the abdominal plate that in surface views
appears to mark the beginning of the proctodaeal invagination.

FIG. 12, X 60, stage C.D.

The thoracic appendages have the character of those seen in the early
phases of stage D, while the abdomen has the characters seen in stage C. The
apex of the third left leg is reduced in size and deeply invaginated.

The cephalic lobes, which have run together to form a _ broad, unpaired
thickening, are thrown to the left by a remarkable growth of the right half of the
cheliceral segment. The original right chelicera has apparently divided, produc-
ing appendages a3-4 and a’*. A second division followed, producing a3 and a4,

No. I.] VARIATIONS IN LIMULUS POLYPHEMUS. lO7

and a third has imperfectly divided a’-*. The resemblance of the extra append-
ages to the normal chelicerae, and the fact that all the remaining appendages are
in their proper positions, remove all doubts as to their identity.

It should be observed that the fourth chelicera, if this explanation is correct,
is the oldest, as indeed its size and isolation suggest. This condition, however, is
the reverse of what usually obtains in segmented animals. There is no distinct
segmentation of the right cord corresponding to the extra appendages. But if
present it could not be easily seen, as the segmentation of the nerve-cord in other
parts of the embryo is very indistinct. But there is a triangular swelling of the
nerve-cord, .S, which widens forwards, and seems to include the first three append-
ages of the right side. A small, triangular thickening, posterior and lateral to the
one just described, lies opposite the fourth chelicera.

The lateral eye of the left side is normal. But on the right are two dark areas
which probably represent two right lateral eyes. It would thus appear that the
right half of the cheliceral metamere has given rise by division to four more or
less complete half-metameres.

Fic. 13, X 60. A large embryo in stage C.

The first two thoracic appendages are absent on the right, causing a spiral
curvature of the head toward that side. The anal plate forms a conspicuous oval
protuberance, suggesting the early stages in the formation of the tail in insects and
scorpions.

The most singular feature, and one not seen in any other abnormal form, was
the fact that the cephalic lobes formed a continuous thickened mass of ectoderm,
without distinction into right and left halves. In the median line was an enor-
mous wedge-shaped ectodermic thickening, reinforced by an underlying layer of
mesoderm, extending backward as far as the third thoracic metamere. This
median thickening seems to be, in part, an exaggeration of a median post-
oesophageal proliferation frequently seen in normal embryos of this age. On the
right side of the cephalic lobes mesoderm and ectoderm are indistinctly separated
from each other, and the former consisted of a mass of the characteristic, degen-
erating nuclei. In place of the first two appendages on the right, are irregular,
flattened masses of loose cells containing degenerating nuclei. Nearly all the
mesoderm of the right side, and especially the great masses in the appendages,
showed the same kind of degenerating nuclei. None were seen on the left
side.

A large area in the yolk, beneath the posterior, thoracic region, and nearly the
whole surface of the embryo, especially on the right side, was strewn with
innumerable, intensely red dots, that look like bacteria. But these dots are also
similar in size and color to the chromatin granules in the yolk nuclei, and also to
those in the nuclei of certain ectoderm cells. After a careful study of them it
seems probable that the red dots will turn out to be bacteria, but we must not lose
sight of the possibility that they may be chromatic granules, liberated by the
rupture of degenerating nuclei. The granules appear to lead an independent
existence, at least for a limited period.

Fic. 14, X 60. A small embryo in stage C.D.

The right chelicera is absent, and the last two thoracic appendages on the
right side are completely invaginated. There is a difference in direction of the
first three and the last three pairs of thoracic appendages, similar to that in
Fig. 22.

108 PATTEN.

The abdominal plate forms an elongated, thick-walled lobe, the posterior half
of which projects freely from the surface of the egg, a not infrequent abnormality.

The cephalic lobes are well defined, and in shape and in the numerous small,
pit-like depressions of the surface suggest the early stages in the cephalic lobes of
scorpions and spiders.

The thoracic portion of the marginal fold is very faint except at the anterior
and posterior ends.

The posterior margins of the mesodermic area form conspicuous, triangular
bands, approaching each other toward the median line. Just beneath them in the
yolk are many degenerating nuclei, especially numerous just beneath the posterior
median margin, where they form two elongated, parallel cords of yolk nuclei.

Fic. 15, X 60. This is a remarkably large and well-developed embryo in stage
D. It represents a type very frequently seen. It is quite normal except in its
size, in the abbreviation of the abdominal plate, and in the absence of the in-
vagination of the cephalic lobes. The whole embryo is deeply imbedded in the
yolk, and with a very prominent marginal fold. There is a series of rounded
ectodermic thickenings on the outer side of the base of each of the first four
thoracic appendages, resembling the flabellum of the sixth pair, but being in reality
the ventral ends of the dorso-ventral muscles.

.

of Morphology Vol. XiL

ap?

i 7 -
* .
5
i ;
a
4
f .
,
x
«
rn
7
ts ca
~.
;
J =’
‘
2
2
1
‘
| a i
>
j
; 5
7 i
1 .
e
rom 1
1
j 7

VARIATIONS IN LIMULUS POLYPHEMUS. 109

EXPLANATION OF PLATE IV.

Figs. 16, 17, (18), 19, 20, 21, (22), 23, 24, 26, 27 represent very common types
of about stage C. They are often laterally compressed, as well as diminished in
length by the antero-posterior compression of the head region, and by the
absence of the abdomen. Except in Fig. 19, the marginal fold, g.f, is very con-
spicuous, extending nearly around the embryo. All the parts included in this
fold, z.e. the nervous system and appendages, are deeply depressed below the
surface of the ovum. In Fig. 19, there is a difference in direction between the
appendages of the anterior and posterior thoracic regions, and a wide space
between the third and fourth pairs, in Fig. 23. In Fig. 16, the second and third
pairs of existing appendages are partly invaginated. There is no invagination of
the abdominal plate in Figs. 17 and 19, but a very conspicuous one in Figs. 16,
20, 21, and 23. The full number of thoracic appendages are present in Figs. 17,
19, 20, 22, and 23. In Fig. 16, two metameres are absent, probably the first two,
although there is no positive evidence in this case as to which ones. In all the
figures the cephalic lobes, and the cheliceral segment when present, are much
more deeply depressed than the rest of the nervous system. This is a very
common feature in this type of embryo, and is especially well shown in Fig. 16,
where there is a very abrupt descent from the level of the ventral nerve-cord to
that of the cephalic lobes. Compare also Figs. 62-67, etc.

Two extreme modifications of the ganglionic folds of the cephalic lobes are to
be seen. One is shown in Fig. 19, where the cephalic lobes are nearly flat and
naked, or as in this embryo, covered by a very thin single layer of cells, ¢.f, ad-
vancing posteriorly and medianly over the lobes and adhering very closely to
their outer surface. The other type is seen in Fig. 21, where there is a deep
invagination on either side of the cephalic lobes, overarched by a very prominent,
amnion-like ganglionic fold. In Figs. 17 and 23, only a’small part of them is so
concealed, while in Fig. 16 they are entirely exposed. In Pl. IV, Fig. 34, the
whole cephalic lobes are covered by this fold.

In nearly all of these figures, the lateral eyes are very well developed for this»
stage, and are easily seen in surface views about opposite the chelicerae, or if the
chelicerae are absent (as in Fig. 16) about opposite the position the chelicerae
would have occupied had they been present. On the other hand, the so-called
“dorsal organs” are either entirely absent, or so faintly developed that they can-
not be detected in sections or surface views. They are shown in Fig. 22 only.

In these compressed embryos, the rim of the mesodermic area is generally very
thick and distinct. In Fig. 20, the outlines of the posterior mesoblastic somite
are visible up to the margin, and the manner in which they grow toward the
median line, and unite there to form the post-anal cloud of cells, is clearly shown.

In most embryos of this type, the outlines of the somites are not preserved, but
the posterior portion of the thickened rim of the mesodermic area, and the part
that has concresced in the posterior median line are all the more distinct. One
of these forms, seen as an opaque object and illustrating the extreme development
of this character, is shown in Pl. VI, Fig. 63. An important point to be observed
here is that the rim of the mesodermic area is so abnormally large, that it appears
as a white and prominent ridge, resembling in shape and position the so-called
concrescing margin of the blastopore in sharks and reptiles.

110 PATTEN. [Vou. XII.

The prominent mesoblastic rims of this type of embryo are subject to local
vesicular enlargements, which are usually filled with a clear fluid and lined with
several layers of rounded cells derived from the mesoderm. The round cells
present all stages of degeneration.

The individual characteristics of these embryos are as follows :

Fic. 16, X 60, sectioned. Two thoracic metameres are absent, probably the
first two. Last three pairs more or less invaginated at apex. Embryo somewhat
elevated as a whole, but depressed in centre, and surrounded by a high, thick,
marginal fold. Abdomen absent. Cephalic lobes, sharply depressed below level
of rest of nervous system, forming a steep descent between the first two appen-
dages. ‘Two deep oval invaginations on the lateral margins of the cephalic lobes.

There is a slight depression at the summit of the elevation between the first
two thoracic appendages, from which arises by inward proliferation, a great mass
of degenerating cells, which give this area its dark color. The region of most
active proliferation seems to extend a short distance along the median line, the
crowd of cells thus produced mingling anteriorly with those about the oesophagus,
and diminishing posteriorly till they disappear about opposite the anterior margin
of the second pair of appendages.

The mesodermic area is extensive anteriorly, but without a conspicuous rim,
which was too remote from the body of embryo to represent in the figure. Two
broad masses of mesoderm radiate from the head of embryo to the rim of the
mesodermic area. Posteriorly the mesodermic rim is well marked and notched in
the median line, where there is a small ectodermic elevation, /#.a.c., continuous
with a great mass of underlying yolk cells. The latter are also continuous with
the great cloud of cells arising from the deep oval invagination, ¢./., in the anal
plate. There isa small marginal vesicle on the right side, #.v. No trace of
segmentation in the mesodermic area.

Fic. 17, X 60, sectioned. Embryo short and rather broad. Abdomen absent.
Marked difference in size of the first, the following three, and the last two, pairs
of appendages. Cephalic lobes broad and disproportionately large. Large oval
depressions, 67. 7v., on sides of lobes partly covered by an overhanging ganglionic
‘fold. Marginal fold conspicuous posteriorly, where it extends across the median
line, forming a very prominent spindle-shaped enlargement. No marked con-
crescence of the posterior margin of the mesodermic area.

Fic. 18, X 60, sectioned. The body of this embryo lies well below the sur-
face and is much compressed laterally, so that the nerve-cord forms a single,
median, roof-like ridge, not adequately represented in surface views. The
cephalic lobes are absent, but there is a small median invagination, probably
representing the oesophagus.

There is a very large tail lobe, projecting upward and forward, and entirely
separated from the body except at its posterior end. Its interior is filled with
yolk. On either side are seen the median portions of two great marginal vesicles,
m.v. The outer margins of the vesicles coincide with the peripheral margin of
the mesodermic area. In the depressions beneath the legs and the tail lobe, on
the nerve-cord and elsewhere, the surface of the embryo was covered with clusters
of the intensely red dots that look so much like bacteria.

Beneath the base of the tail lobe is a conical invagination best seen in sections,
directed diagonally backward into the yolk. From its posterior end, streamers of
closely packed cells extend into the yolk. Around these streamers are many free
nuclei resembling yolk cells.

No. 1.] VARIATIONS IN LIMULUS POLYPHEMUS. EEL

Back of the tail lobe is the usual ectodermic thickening, f.a.c., and beneath it
is the usual post-anal cloud of mesoderm, formed by the concrescence of the
margin of the mesodermic areas.

FIG. 19, X 60, not sectioned. Embryo not depressed. Thoracic appendages
show a common, but abnormal mode of growth. Abdomen absent. Cephalic
lobes slightly convex, with very thin ectodermic fold, probably consisting of a
single layer of cells growing over their lateral portions. Marginal fold knotted
and conspicuous posteriorly, but not continuous across the median line. No in-
vagination of anal plate. Mouth very long and narrow. Mesodermic area, com-
paratively small, circular, and with conspicuous rim, m.a., especially at sides. No
conspicuous mass of cells due to concrescence of mesodermic rim back of anal
plate.

FIG. 20, X 60, sectioned. Embryo short and narrow and deeply depressed.
Marginal fold sharply defined. Cephalic lobes shortened and partly concealed by
a broad overhanging ganglionic fold.

Abdominal plate without appendages and sharply infolded to form a very deep
oval invagination, from the walls of which arises a cloud of degenerating yolk
cells.

The mesodermic area is extensive. It rim is well defined, and posteriorly
shows very beautifully its mode of concrescing. In this case the outlines of the
posterior, mesoblastic segments are very clearly shown.

FIG. 21, X 60, sectioned. Chelicerae and abdomen absent. Large open
depression on margin of cephalic lobes, bounded on the sides by a prominent,
overhanging lip. Mouth very large, with rostrum-like projection in front. Body
of embryo depressed and surrounded by prominent, knotted, marginal fold, which
is discontinuous posteriorly.

_A deep and broad invagination of the anal plate has carried the last pair of
thoracic appendages inwards till they project from its sides. A conspicuous
elevation back of anal plate, Z.a.c., due to the concrescence of the posterior
margin of the mesodermic area. The lateral and anterior portions of the rim
are not represented.

Fic. 22, X 60. A very common form of embryo in stage Y. Not sectioned.
Differs from the normal in its shortened, compact shape, and in being deeply
sunken in the yolk. It was outlined to show a rather frequent modification of
the direction in which the thoracic appendages grow, the anterior pairs pointing
backwards and laterally and the posterior ones forward and inwards. It is
interesting because it occurs frequently, but especially because it is between the
third and fourth thoracic metameres that transverse division sometimes takes
place. The difference in the disposition of the appendages recalls a very similar
one that obtains between the mouth parts and the walking appendages of insect
embryos.

Fic. 23, X 60, not sectioned. Embryo narrow and depressed. Appendages
normal except in the separation of third and fourth pairs, a rather frequent occur-
rence, that seems to have some connection with the transverse fission that often
occurs here, and with the difference in direction between the appendages in front
of and behind this space. Two pairs of abdominal appendages present, but they
are carried inwards by the invagination of the anal plate, so that they project
from the sides of the cavity. Marginal fold prominent, especially posteriorly, and
continuous anteriorly with a well-defined fold overhanging the invagination on
the sides of the cephalic lobes. Mesodermic area not conspicuous or well defined.

112 PAW TEV:

FIG. 24, X 60, sectioned. In this embryo the cephalic lobes and first two tho-
racic segments have disappeared. The marginal folds have closed in front of the
third pair of appendages, and the nerve-cord terminates abruptly just back of
them. The abdomen is absent, and the marginal folds are contracted so as to
extend across the median line just behind the sixth pair of thoracic appendages.
In the median line, the fold is interrupted by a deep, thick-walled invagination
with an oblong lumen.

The invagination dips deeply into the yolk, and from its thick walls arise
numerous nuclei which are seen scattered about in the neighboring yolk. The
yolk nuclei are most numerous back of the invagination. An ectodermic thick-
ening forms the broad, dark, post-anal band seen in surface views, 7.a.c. At its
posterior end, the band becomes continuous with the thickened rim of the meso-
blastic area. The post-anal cloud of yolk cells is formed in part by cells that
have migrated from the walls of the invagination, and in part by those arising
from the thickened mesodermic rims, which have concresced along that line.

A large marginal vesicle, m.v., is seen in the right anterior margin of the
mesodermic area.

Fic. 25, X 60, sectioned. A remarkable embryo in stage C, in which the
cephalic lobes are reduced to a flat circular plate, slightly depressed, so that it is
surrounded on all sides by a vertical wall. Near the centre of the disc, which is
separated by a considerable distance from the remainder of the embryo, is a small
pit representing the oesophagus, and in the yolk below the disc is a great, irregular
mass of cells, with numerous pseudopodia-like streamers of cells, extending still
deeper into the yolk.

The body of the embryo, which consists of three appendage-bearing segments,
is bent into the yolk at the posterior end. Back of this abbreviated trunk is a
broad depression bounded on either side by steep walls, which gradually shallow
posteriorly to the surface of the ovum.

Fic. 26, X 60, sectioned. A very compact embryo of stage Cand YD. All the
ectodermic layers are very thick, and the mesodermic area is constricted so that
its peripheral margin lies close around the body of the embryo. The greater part
of the mesoderm forms two thick bands on either side of the body, close to the
median line.

The cephalic lobes are completely covered by a hood-like fold extending back
almost to the second pair of thoracic appendages. At the bottom of the cephalic
cavity is a small pit, the histological character of whose walls indicates that it is
the oesophagus.

Surface of embryo covered with bacteria.

FIG. 27, X 60, sectioned. Very much shortened embryo in stage C.

The rudimentary cephalic lobes are covered by a backwardly directed fold, and
only three pairs of appendages, probably the fourth, fifth, and sixth, are represented.
What looked in surface views like an anus was present, but its nature could not
be determined in sections. Such embryos as those in Figs. 26 and 27 are rather
common, and possibly they were seen by Dohrn and Osborn, and gave rise to the
statement that Limulus passes through a nauplius stage.

Fic. 28, X 33, not sectioned. An embryo in stage C.

It consists of an abdominal plate and four metameres, probably representing
the last three thoracic and first abdominal ones. The appendages are separated
by a very wide space, over which the neuromeres (?) and mesoblastic somites
extend as long, narrow bands.

- = iy
oe 2
c.
i
-
ef
~ =
os =
laa
_ a ae
7 -
\ x a
, Pig
- €
-
I
2 7 a ey

>

fo phology Vol. NIL

cma

W Pacer il

VARIATIONS IN LIMULUS POLYPHEMUS. EE?

EXPLANATION OF PLATE V.

Fic. 29, X 60. Embryo in about stage £

The most striking features are : (1) a diffuse atrophy of the left side, resulting
in the complete disappearance of the left abdominal appendages and neuromeres,
and the reduction in size and absence of surface details in the left thoracic appen-
dages; and (2) the hour-glass form due to a diffuse, transverse atrophy of both
halves along the fourth thoracic segment. When examined more carefully, it is
seen that the three anterior thoracic appendages on the right side are nearly
normal ; the second and third being perhaps a little stouter than usual, but still
showing all the characteristic details of this stage. The fourth right appendage
is reduced to a broad, low elevation, probably representing the basal joints of
the same, the rest of the appendage being reduced to a very small, medianly
directed protuberance springing from an oval depression. The next appendage is
larger than the fourth, but smaller and less perfect than the sixth; and that is
smaller and less perfect than it should be. The last two look as though made of
wax that had been warmed on the surface, thus melting off sharp angles and
other details.

The right side of the abdomen is apparently perfect.

On the left half of the thorax the chelicera is thrown outwards, but is otherwise
normal; but the second and third appendages are strikingly smooth and rounded
on the tips, as though the surface details had been melted off. The fourth is a
minute papilla, and the fifth and sixth are small, rounded, unjointed elevations.

The whole left half of the abdomen is absent, except a trace of the marginal
fold.

The cephalic lobes are somewhat distorted, and the semicircular lobe is
apparently interrupted in the median line.

The median ocellus and the nerves extending to it are very distinct.

The ventral nerve-cord is not clearly divisible into right and left halves. Back
of the third pair of thoracic appendages, it narrows, and finally the right half only
extends beyond the rudimentary fourth pair of appendages into the abdomen. In
another specimen similar to this, it is very clear that the right half of the cord
only is represented on the posterior part of the thorax and abdomen.

The marginal fold is deeply constricted opposite the fourth thoracic segment.
The right lateral eye and “dorsal organ” lie beyond the limits of the figure, thus
preserving their normal position in reference to the body of the embryo, but not
in reference to the marginal fold. The left eye and dorsal organ could not be
seen, and were probably much reduced, if not entirely absent.

Fic. 30, X 60, not sectioned. Embryo with two thoracic appendages and
the corresponding neuromeres of the left side absent. The missing appendages
appear to be the second and third, as the next three correspond to the last three
of the opposite side. It should be observed that the second and third appendages
of the right side are unusually large, as is the case with the appendages on the
corresponding segments in Figs. 31 and 32. There appeared to be four abdominal
appendages on the left and only two ontheright. There is a very large peripheral
vesicle on the right side, /.v.

FIG. 31, X 60, not sectioned. Embryo with the last three thoracic appen-
dages on the left side absent. The left half of the abdomen is also absent. The

114 PATTEN. [Vox. XII.

remaining half, which is provided with two distinct appendages, is thrown sharply
to the left, so that its long axis is at right angles to the rest of the body.

All the thoracic appendages are present on the right ; but the second and third
are very large, and the fifth and sixth correspondingly small. On the left, three
appendages are absent, probably the last three of the series, as the most anterior
one seems to represent the chelicera.

FIG. 32, X 60, not sectioned. On the left side the second and third thoracic
appendages, as in the preceding figures, are unusually large, and the last three
thoracic appendages are abnormally small. On the right the chelicera is absent,
and the last three appendages, of which one is entirely absent, are very small.
The right half of the abdomen is absent, but on the left three appendages and
their corresponding neuromeres are present. What is left of the abdomen is
thrown to the right, and its end covered with a large hood-like overgrowth of the
marginal fold.

FIG. 33, X 60. Small embryo in stage C.

Cephalic lobes are very small and partly concealed by the ganglionic fold,
which nearly reaches to the mouth. Three well-formed appendages, fourth, fifth,
and sixth (?), are present on the left. On the right the corresponding appendages
are invaginated, forming three shallow pits, 7.2. The abdomen and the anterior
part of the thorax are absent.

The outline of the mesodermic area is visible, showing its thickened rim, wa.
Its concresced posterior portions form the usual post-anal cloud of cells,
p-4.c.

Fic. 34, X 60, not sectioned. A much reduced embryo in stage C.

The head is turned toward the left, although that side is better developed
than the right. The cephalic lobes are covered by a distinct, amnion-like fold,
that extends back to the first remaining appendage. Through the fold may be
seen the reduced oesophagus. Two appendages are present on the left, and
corresponding to them on the right, a shallow pit and small papilla. Back of
these appendages is a large unpaired one, probably formed by the fusion of the
appendages of the sixth pair. Back of this is a deep furrow, flanked on either
side by a rounded elevation.

On either side, just within the limits of the mesodermic area, are two large
marginal vesicles.

FIG. 35, X 60, not sectioned. A much distorted and abbreviated embryo.

The cephalic lobes are broad, and contain a good-sized oesophagus. The
remainder of the embryo is bent sharply to the right. This is due to the presence
of three distinct appendages on the left side, and only one (the second?) on the
right. The tail end of the embryo is invaginated, and covered by a small, over-
growing fold. The margin of the mesodermic area is clearly defined, and plainly
shows the notch, due to concrescence of its posterior margin. At the anterior
end of the embryo is a dark band of yolk cells (?) extending forward to the anterior
border of the mesodermic area.

Fic. 36, X 42. Embryo in somewhat older stage than Fig. 29.

The whole right half is practically normal. The most characteristic feature of
the embryo is the absence of most of the left half of the thorax ; although the left
half of the cephalic lobes, and of the abdomen, is nearly normal.

The whole embryo is somewhat shortened. The appendages of the right side,
and the right ganglionic fold, extends backwards over the outer surface of the

No: 1.]] VARIATIONS IN LIMULUS POLYPHEMUS. II5

brain and optic ganglion farther than usual. Otherwise the right side appears to
be normal.

The left sixth thoracic appendage is a small, three-jointed organ, showing
distinctly the characteristic flabellum near its lateral margin. The fifth is a mere
papilla, and the third and fourth are entirely absent. The second is relatively
large, and seems to be partially invaginated at the tip. The left chelicera is
absent. In the left half of the abdomen the appendages are partly fused, forming
a great three-lobed mass.

It is, therefore, plain that while the left half of the thorax is greatly reduced, as
a whole, it shows in addition a diffuse transverse atrophy with its line of greatest
intensity between the third and fourth appendages, and diminishing gradually in
front and back of this line. We thus have a case of hour-glass atrophy, confined
to the left half of the body. (Compare Fig. 29.)

The nervous system is not sharply outlined. The parts of the cephalic lobes
are run together, and a great dark fold extends diagonally backwards over the
right half. The left is much smaller, and has been carried backward by the
contraction of the left side. The oesophagus and mouth are small and inconspic-
uous. It is not clear from surface views whether a part of the left nerve-cord is
absent or not.

The lateral eyes and dorsal organs were in their normal positions on the right
side, but could not be seen on the left.

The marginal fold is normal on the right. On the left it disappears near the
optic ganglion. There is a large, laterally directed fold opposite the left sixth
appendage that consists, in part at least, of the marginalfold. In the abdomen the
left marginal fold is normal, and it meets the fold just described at a sharp angle,
but does not appear to be continuous with it. Between the two folds is a very
deep triangular depression. This condition of the marginal fold is the usual one
when there has been atrophy of the anterior quarter of the same side.

In the anal region the marginal fold is greatly thickened, forming a conspicuous
U-shaped boundary to the posterior part of a deep depression.

Fic. 37, X 40. This embryo is, in part, in an advanced stage of develop-
ment, corresponding to that seen in embryos about ready to hatch. There are
only a very few markedly abnormal embryos found in this late stage.

The right side is nearly perfect and normal. The ganglionic fold over the
right half of the brain is large, as in the preceding figure. The right nerve-
cord, which was very well developed and plainly outlined, is bent sharply at
the junction of the thorax and abdomen. This bend, as seen in the drawing,
is due, to a small extent only, to the foreshortening produced by looking down
on the upturned abdomen. Almost the entire left half of the cephalic lobes
are absent, but the left nerve-cord, as far as the beginning of the abdomen,
is present in its normal condition. All the left thoracic appendages have
disappeared completely, leaving no trace behind except in the third segment,
where there is a shallow depression, seen edgewise in the figure, with a small papilla
projecting from its centre. This papilla probably represents the last trace of the
third thoracic appendage. The left nerve-cord seems to terminate abruptly at
the posterior end of the abdomen, without uniting with its abdominal part. The
anterior end of the latter turns off sharply to the left, toward a large, dark,
conical elevation. To the left of it is a short, conical projection, with its apex
directed forwards.

116 PATTEN.

The embryo is probably in a state of incomplete longitudinal fission, or a
double embryo, similar to those in Pl. IX, Fig. 98. The two problematical
appendages, x and y, would then represent the medianly fused, posterior thoracic
appendages of a second embryo, whose axis is in the line 4. The two embryos
possess an abdominal nerve-cord in common. The abdominal appendages on the
left side of embryo & are absent (its right abdominal half was not formed at all
according to this supposition), and in embryo 4 the thoracic appendages failed to
develop after the longitudinal fission of the embryo.

This explanation, requires us to assume nothing more than what we know
takes place in the double embryos described in plates IX and X. If it is
correct, this embryo, although resembling that in Fig. 36, owes its present condition
to a totally different sequence of events.

Fic. 38, X 33, unsectioned. A well-advanced embryo, the right half of which
is complete and perfectly normal, except in its slight curvature to the left. Of
the left half nothing remains but a dark band of inner-layer cells and a small
posterior appendage situated in a rather deep depression. The long axis of the
mouth is rotated nearly 45°, so that the apex of the rostrum is thrown toward the

left.

pmn

al

ap®

ma

|

+

“,
i

- PAC.

i

VARIATIONS IN LIMULUS POLYPHEMUS. iy

EXPLANATION OF PLATE VI.

Fics. 39-49 illustrate the more important phases in the formation of the
inverted V-shaped embryos. These embryos are formed by the fusion in the
median line of the corresponding right and left organs of each metamere. The
organs nearest the median line are the first to unite, forming in that way an
unpaired organ, having the characteristic features of each member of the pair.
The unpaired organ thus formed then decreases in size, and finally disappears. In
its place the organs next to it, on the same metamere, unite, and in turn degener-
ate; and so on till the whole metamere has disappeared. The process seems to
begin in every case in the most anterior metameres ; and in the most typical cases,
as soon as the first unpaired organ, formed in, say, the first thoracic metamere,
has disappeared, the same organ is found unpaired in the second metamere; and
by the time that has disappeared the unpaired condition of that organ obtains in
the next following metamere, and so on, till every paired organ has become median
and unpaired, and then disappeared. In the last phase of the process, if realized
in full, there would be nothing left of the embryo but a single unpaired organ,
situated at what was the posterior end of the body, and formed by the median
fusion of the most laterally situated paired organ of the last metamere.

Such a condition has not been observed, the nearest approach to it being an
embryo of which nothing remained but the mesodermic area and a posterior
unpaired process, representing either the last thoracic appendage or the tail lobe.

In very rare cases one of the posterior pair of appendages may fuse in the
median line, while there is no indication of fusion in front of that point. But in
such cases there is no evidence of a progressive median fusion and degeneration
extending toward the anterior end.

There is another exception to the median fusion and progressive antero-
posterior degeneration seen in the hour-glass embryos shown in Pl. IV.

FIG. 39, X 60, not sectioned. This embryo is instructive, as it is apparently
in the early stages of median fusion. The cephalic lobes are reduced to a thin,
circular disc, showing no trace of separate optic ganglia and cerebral hemispheres.
The oesophagus is a shallow pit in the centre of the disc, and the cheliceral
segment has disappeared entirely. The appendages of the fifth and sixth thoracic
segments are nearly normal in position, but in passing forward from this point the
appendages decrease in size and approach more and more the median line, till in
the second segment they have nearly united with each other. The nerve-cords
terminate back of either the third or the fourth pair of appendages, but as this
embryo was not sectioned that could not be determined with certainty.

The end of the left third appendage is invaginated. The abdomen is normal
except for the rather prominent anal plate.

FIG. 40, x 60. This isa very unusual form, and the only one of its kind observed.
Median fusion has taken place at both ends. The dorsal organs are very con-
spicuous, and as they always lie opposite the fourth pair of appendages we can see
that the following changes have taken place :

The cephalic lobes, oesophagus, and first thoracic neuromere have disappeared.
The appendages of the second pair have fused completely, and those of the third
pair have approached each other, preparatory to the same change. The fourth
pair is nearly normal. ‘The fifth appendage on the right is invaginated for half its

118 PATTEN. fViows cite

length. The appendages of the sixth pair are short and thick, and have fused
with each other except at their tips. Just in front of them is a large, deep pit, z.v.
I do not know of any explanation of the presence of such an invagination at this
point, unless it may be regarded as an invaginated tail lobe or a telopore, carried
forward to its present position before the fusion of the last pair of thoracic
appendages.

Every trace of the abdomen is absent, something that is very unusual, for it is
a noteworthy fact that in this class of embryos the abdomen usually remains
nearly normal, however profound may be the changes that have affected the
anterior part of the embryo.

Fic. 41, X 60, sectioned. In this embryo the cephalic lobes and first two
thoracic segments have disappeared, apparently after median fusion. The primi-
tive position of the cephalic lobes is indicated by a conical mass of degenerating
cells projecting into the yolk, and seen in profile on the edge of the egg. From
the apex of the mass, an irregular train of the same kind of cells, lying deep in the
yolk, extends backwards to the surface at the anterior end of the embryo. These
cells probably represent the last remnant of the degenerating oesophagus and
anterior portion of the embryo.

The right appendage of the third thoracic segment is absent, and the fifth pair
is invaginated. Three depressions along the median line are indicated by white
areas.

All the parts of the embryo appear very dark, owing to the unusual thickness
of the cell layers, and the large amount of staining fluid the cells have absorbed.

FIG. 42, X 60, sectioned. This is a rather common form. ‘The cephalic lobes,
oesophagus, and first two thoracic metameres are absent. The appendages of the
first pair are fused nearly to their tips, those of the fourth pair are fused at the
base only. The two nerve-cords fuse with each other just in front of the fifth pair
of appendages, and the single median cord thus produced terminates abruptly just
back of the base of the fourth pair. Within the fused bases of the appendages of
the fourth segment is a large mass of cells that looks like the remnant of the
neuromere of this segment, but forced to assume a spherical form by the fusion of
the appendages. It is connected posteriorly with the rest of the nerve-cord by a
narrow chain of cells. In front of the fused appendage is merely a thin layer of
ectoderm and mesoderm, every trace of the nerve-cord being absent.

Immediately in front of the fused appendages of the third thoracic segment is
a short, flattened tube, directed diagonally forward into the yolk. It is surrounded
by a thin layer of mesoderm, and opens outwards by a transverse opening between
the marginal fold and the anterior edge of the unpaired appendage. It has all
the appearance of an oesophagus, and probably is one, but of course it is entirely
out of place here.

There is a slight asymmetry of the abdomen, due to the absence of one appen-
dage on the left side, but otherwise the embryo back of the fifth pair of thoracic
appendages is quite normal.

FIG. 43, X 60. This embryo is similar to that in Fig. 49. The original position
of the cephalic lobes is indicated by a shallow, saucer-shaped, ectodermic thicken-
ing, c./. In the figure, owing to the curvature of the surface of the ovum, it is
seen edgewise, and only the thickened posterior rim is shown. It is connected
with the remaining part of the embryo by a broad train of yolk ced/s, lying close
to the surface. They appear in the figure as a broad, faint band, y.c. The third

Now t:] VATAATIONS TIN VEIMOLUS POLVPHEMUS. IIg

pair of appendages are fused at their bases, and back of them terminates the
ventral nerve-cord.

The abdomen is smooth and flat, showing no trace of appendages.

The limits of the mesodermic area are clearly defined by the usual thickened
rim. The mesodermic area back of the tail lobe is darker, and its posterior
margin, owing to incomplete concrescence, is deeply notched.

Fic. 44, X 60. In this embryo the cephalic lobes and the cheliceral segment
have disappeared, leaving some distance in front of the embryo a large irregular
patch of cells, continuous right and left with the mesodermic rim. In the figure
it is seen in profile on the edge of the egg, c./.

The appendages of the second post-oral segment have fused with each other,
and a similar fusion of appendages has taken place in the third segment. The
smaller size of the anterior median appendages indicates that it was formed
previous to that of the following segment, and has already undergone some
degeneration.

The marginal fold, m.f, extends anteriorly between the fused appendages of
the second and those of the third segment. It usually appears to contract with
the atrophy of the anterior end of the body, so that it closely encircles all the
remaining appendages. If that really occurred here, the fold must have in some
way passed around the second pair of appendages, a process difficult to explain.
The only alternative is to suppose that a new fold was formed back of the second
pair of appendages, and that the old one has disappeared or was never formed.

The nerve-cord terminates as a blunt, unpaired process, a little in front of the
fourth pair of appendages. The posterior part of the thorax and the abdomen
are normal, except that the apex of the latter projects freely away from the ovum,
and the whole abdomen hangs over a great blister-like vesicle filled with fluid and
enclosed between the blastoderm above and a thickening of the mesoderm below.
There is nothing in the preparation to suggest that this condition is due to
shrinkage, etc.

The outlines of two similar vesicles_(or marginal vesicles) are seen in the
mesodermic area on either side of the h..u region, m.v.

Fic. 45, X 60, not sectioned. This embryo shows a slight reduction of the
cephalic lobes and of the third and fourth thoracic appendages of the right side.
In place of the fifth and sixth appendages and the abdomen, is a large median
conical protuberance. Whether the latter represents the fused fifth and sixth
appendages or the tail lobe could not be determined.

Fic. 46, X 60, not sectioned. The cephalic lobes have disappeared, or at
least together with the oesophagus are reduced to an invaginated, conical layer of
cells, constituting the dark mass at the apex of the V-shaped marginal fold, d.c./.

The chelicerae have fused to form a short median process, which lies in a
depression that in front is partly overarched by a semicircular fold of the
ectoderm.

The appendages of the second pair have fused to form a very long, slender,
corkscrew-like filament. It is attached to the embryo just back of the apex of
the fused chelicerae. The appendages of the third pair are fused at their bases
and somewhat diminished in size. The nerve-cord terminates abruptly just back
of them.

The parts of the embryo lying back of the third thoracic segment are prac-
tically normal.

120 PATTEN.

Fic. 47, X 60, not sectioned. The first and second thoracic segments are
absent. The cephalic lobes, however, persist as a faint disc of cells without
character. This case differs somewhat from the preceding ones, in that the effects
of degeneration do not increase gradually from before backward, for degeneration
has been greater in the first two post-oral segments than in the cephalic lobes.

The median space between the appendages of the remaining pairs gradually
increases toward the posterior end.

A still further difference between this embryo and the preceding ones is seen
in the fusion across the median line of the abdominal appendages, the last
appendage being the smaller one.

Fic. 48, X 60, sectioned. In this embryo, the sixth pair of thoracic appendages
are identified by the well-developed flabella. The long median appendage is
formed by the fusion of the fourth pair of the thoracic appendages, the unfused
tips of the two original appendages being still visible. The first three pairs
of appendages and the entire cephalic lobes have disappeared. The nerve-
cord is normal and well developed from the root of the tail lobe to the base
of the unpaired appendage of the fourth segment. The abdomen is small, but
possesses two pairs of normal appendages. There is a very conspicuous tail lobe,
the size and shape of which suggest the idea that with the diminution in size of
the embryo, the marginal fold became too large, and the excess accumulated at
the posterior end to form the tail lobe. Back of the tail lobes ts a small conical
projection, also visible in the sections, having all the appearance of an unpaired
thoracic appendage. The presence of an appendage in this place is very remark-
able, and I am unable to offer any explanation for its occurrence there.

The body of the embryo forms the floor of a deep depression bounded by the
marginal fold. The embryo stained very deeply, as it is composed of dense
tissue containing a large amount of chromatin. Anteriorly, the rim of the
mesodermic area, which was not visible, or perhaps overlooked in the surface
views, is easily seen in the sections, apparently preserving its normal size and
extension for this stage. In the sections that pass a long distance in front of the
present anterior end of the embryo, the mesodermic rims are seen as widely
separated as in the normal embryo of this stage. In the line midway between
them, instead of the nerve-cords and appendages, there is merely a thin, undif-
ferentiated layer of ectoderm with a few scattering mesoderm cells beneath it.

Fic. 49, X 60, sectioned. The cephalic lobes and first two thoracic segments
are absent. No trace whatever of these organs is to be seen in sections, although
there is a dark area, due to an accumulation of loose cells, where the anterior end
of the embryo should be.

The marginal folds, m.f, are distinct and well developed, and extend across
the median line just in front of the second pair of thoracic appendages. The
latter are fused with each other at their bases, and just back of them the nerve-
cord teminates abruptly in a blunt, unpaired process.

The abdomen is very well developed, and terminates in a broad tail lobe that
projects upwards and forwards, thus overarching the posterior part of the
abdomen.

Such a prominent tail lobe, although occasionally seen, is not a common form
of abnormality It is suggestive of the prominent tail fold in insects and
crustacea, but in this peculiar case recalls that seen in scorpion embryos.

=y

Pe

1 of Morphology Vol. XU

Mor cl.

tH.

L

122 PATTEN. [Vou. XII.

EXPLANATION OF PLATE VII.

FIG. 50, X 60. This embryo evidently represents an extreme case of median
fusion, but differs from those on the preceding plate in that there are no indica-
tions of that V-shaped arrangement of paired organs usually seen when a progressive
antero-posterior degeneration has followed the median fusion. The cephalic
lobes, oesophagus, and nervous system are entirely absent. Nothing remains of
the body but an oblong elevation of thickened ectoderm with three median
appendages arranged along its summit. Behind the third median appendage was
a partly invaginated plug of cells, ¢.¢., that may be the remnant of the telopore.
Besides this there was nothing to indicate to what segment these fused appendages
belonged, or which was the anterior and which the posterior end of the body. The
mesodermic area was slightly raised and was pentagonal in outline, with a
thickened rim.

Fic. 51, X 60. A large embryo in stage C showing a distinct transverse con-
striction between the third and fourth thoracic segments.

The anterior part of the embryo is normal, except in the absence of the cheli-
cerae. ;

The posterior portion is infolded between the appendages of the sixth thoracic
segment to form a deep, circular cavity. The sixth pair of legs, identified by the
flabella on their outer margin, have been carried inwards by the infolding, till
they project toward each other from its lateral walls.

FIG. 52, X 60, sectioned. An hour-glass embryo in stage C.

The cephalic lobes are deeply depressed, and nearly covered by two lateral
folds. The chelicerae are absent, the next two appendages are brought closely
together, and the third pair completely fused to form a thick median appendage,
ap, extending backwards.

These three appendages project from a circular depression, bounded on all
sides by a thickened margin, which anteriorly forms the folds overlapping the
cephalic lobes.

The fourth pair of appendages are also fused, af‘, and the fifth are either
absent, or fused and invaginated to form the pit just back of the base of the
preceding pair, ap°.

The sixth pair are small and devoid of flabella, but are nearly normal in shape
and position.

The posterior part of the thorax and a part of the abdomen are nearly sur-
rounded by a marginal fold, well developed posteriorly, but not extending across
the median line in front of the fused appendages of the fourth segment.

FIG. 53, X 60, sectioned. A remarkable embryo in stage C that has under-
gone extensive reduction and fusion. As nearly as one can determine by a study
of surface views and cross-sections, the following changes have taken place. The
cephalic lobes, oesophagus, and chelicerae have disappeared, leaving hardly any
recognizable traces behind. The second thoracic appendage on the left is nearly
normal, that on the right, a low, irregular papilla. Between the two, in sections,
traces of a double nerve-cord may be seen. Back of this point, the nerve-cords
fuse and finally disappear as such, near the large, irregular, median protuberance
that probably represents the fused appendages of the third segment, ap >.

No. 1.] VARIATIONS IN LIMULUS POLYPHEMUS. 123

Following the fused appendages is a rather broad expanse, covered by a thin
(single?) layer of cells.

The next thickening, af‘, is elongated transversely, and probably represents
the fused appendages of the fourth segment.

The next two appendages of the left side are of full size and normal, but owing
to the absence of the corresponding opposite appendages and nerve-cord they
have been twisted spirally toward the right.

The abdomen is absent. At what represents the posterior end of the embryo
is a deep, tubular invagination, 4/., with thick walls, from which have arisen
innumerable cells that form a dense cloud in the surrounding yolk. The embryo
is therefore to be regarded as an hour-glass embryo, still further modified by the
absence of the right posterior half of the thorax. It should be compared with
the following one.

The surface of the embryo is covered with bacteria.

FIG. 54, X 60, sectioned. This remarkable embryo in stage D-Z belongs to
the hour-glass type, but it is modified by the median fusion of the anterior end,
and by the absence of the right half of the posterior end of the thorax. It seems
to be the result of still further progress along the lines followed by the embryo
shown in Fig, 53.

In such cases as this it is very difficult to determine just what changes have
taken place, especially as regards the amount of nerve tissue, if any, that is left,
and the segments to which the remaining appendages belong.

As near as can be determined, the changes have been as follows : The cephalic
lobes are absent, and in their place is a broad, triangular, ectodermic plate,
beneath which is an unusually large number of yolk cells.

The marginal fold has contracted anteriorly to form a thick rim around an
oval depression. At the anterior end of this depression is seen in surface views a
dark spot, which is the optical section of a very long oesophagus-like tube extend-
ing vertically into the yolk, about as far as the adjacent appendage is long. The
tube is not closed at its inner end, and its walls are folded longitudinally, as in a
true oesophagus.

Back of this tube is a long irregular appendage arising at first directly upward
from the bottom of the depression, and then bending over to the side, so as to
lie flat on the surface of the egg. Back of this is another appendage; it is bent
double, and in such a way that its distal end points toward the median line,
toward a point a little in advance of its proximal end.

The lateral growth of these two appendages at first seemed to me to indicate
that the whole right half of this portion of the embryo had disappeared, and that
the left half had rotated on its own longitudinal axis 180° toward the right. But
after a careful examination of the sections, I failed to find any traces of either
nerve-cord ; and, on the whole, the sections seemed to indicate that each appen-
dage had been formed by the fusion of a pair, and, as is so frequently the case in
such embryos, the nerve-cord had entirely disappeared.

The posterior end of the embryo, which is separated from the anterior one by
thin, characterless layers of ectoderm and mesoderm, consists of a deep, thick-
walled and irregular sac, from the posterior wall of which arise two long-pointed
appendages, extending upward and forward. There are some parts of the wall of
the sac that look histologically like parts of a nerve-cord, but the whole layer is
so folded and crowded together that nothing in regard to this point could be

124 PATTEN. [Vou. XII.

determined with certainty. There is, however, little doubt in my mind that the
two appendages in question are the fifth and sixth thoracic appendages of the left
side. The sixth has been moved in the plane of the paper 45° toward the
embryo’s right side. The cause of the rotation is to be sought, as in Fig. 53, in
the absence of the posterior right half of the thorax.

The mesodermic area is much contracted, and in place of the continuous
thickened rim are isolated masses of closely packed nuclei, some of which lie
quite deeply in the yolk; others are continuous with the surface mesoderm. The
masses are irregular in shape, but are usually provided with radiating streamers
containing many nuclei.

Fic. 55, X 60, sectioned. This is also an hour-glass embryo, but one in which
the two parts are completely separated. The cephalic lobes are absent. The
second pair (chelicerae ?) of thoracic appendages have fused in the median line.
The next pair are widely separated, and following them is a second pair of fused
appendages. Between the unfused appendages is a well-developed double nerve-
cord, which gradually narrows at either end to an unpaired cord. One end lies
just back of the anterior median appendage and the other in front of the posterior
one. 2
Back of these four appendages is an area devoid of external features. In
sections it is seen to be composed of a slightly thickened homogeneous layer
of ectoderm with a similar underlying layer of mesoderm. At the posterior end
of this area is a deep, tubular invagination or telopore, ¢.f., directed vertically into
the yolk, and surrounded by the usual cloud of migrating cells. In front of the
telopore are two small depressions, /.a., that may represent invaginated appendages.

Projecting forwards over the telopore is a broad conical process. It probably
represents a tail lobe such as is seen in Pl. VI, Figs. 48 and 49.

Back of this lobe is a broad convex area, composed of slightly thickened
ectoderm and mesoderm, and evidently formed by the concrescence of the pos-
terior margins of the mesodermic area. The rest of the mesodermic area is
roughly A-shaped, the two backwardly directed lobes showing notches produced
by partial concrescence.

Fic. 56, X 30, not sectioned. This embryo is in about stage C.

The remnants of the right and left sides have undergone complete median
fusion. There is now nothing left but an anterior, oesophagus-like tube, oe.,
opening beneath a backwardly directed fold, and two median appendages. The
posterior one is the larger, and projects backwards a short distance over the floor
of the depression in which both lie. This depression is roughly boat-shaped, but
wider and considerably deeper at the posterior end.

The nervous system is entirely absent, or, at any rate, the ectoderm over the
bottom of the furrow is not thicker than, or in any way distinguishable from, that
covering other parts of the embryo. The mesoderm is much thickened under
what may be considered the body of the embryo, as shown by the dark rim in
the figure. Buta thin layer of mesoderm extends much farther than this, and
along its periphery may be seen, in surface views, irregular star-shaped masses of
cells, deeply imbedded in the yolk, but probably derived from the disintegrated,
thickened margin of the mesodermic area.

As seen in sections, the tissues appear to be perfectly normal and healthy.

There is an indication of a transverse constriction between the two appendages.
As similar constrictions occur most frequently between the third and fourth

No. I.] VARIATIONS IN LIMULUS POLYPHEMUS. 125

segments, it is possible that the appendages on each side of this constriction
belong to the third and fourth segments.

FIG. 57, X 60, not sectioned. This is an embryo in stage C that has under-
gone partial median fusion, accompanied by transverse constriction. It could not
be determined which was the anterior end of the embryo. It has been placed in
its present position on account of the hood-like fold at what is now the anterior
end, as it resembles a fold that sometimes covers the cephalic lobes of other
embryos. At the anterior end of the embryo are two pairs of appendages nearly
fused. Back of them is a thin layer of undifferentiated ectoderm with a thick
mass of underlying mesoderm, the latter causing the dark color of the prepara-
tion. There is a longitudinal fold of ectoderm, probably the displaced marginal
fold, on the left side of this area, and in about the middle of the same is a conical
invagination, 2.a.

At the posterior end, the fold extends around a second depression, from which
arises a large protuberance, evidently formed by the fusion of a pair of appen-
dages. There is no trace of a nervous system in this embryo, or any distinguish-
able lateral boundaries to the mesodermic area.

Fic. 58, X 60, sectioned. Here we have a thick, oval mesodermic area, from
whose inner surface several pseudopodia-like bands of nuclei extend vertically
into the yolk. There are no appendages or nerve-cords, but there is an axial,
ectodermic thickening, which, at what we may call the anterior end, forms a large
pyramidal elevation, containing a large number of degenerating nuclei. This
elevation probably represents one or more pairs of fused appendages. At the
posterior end, the axial thickening gradually culminates in a tub-shaped elevation
with a slit-like depression on its upper surface. At the bottom of the slit is a small
circular depression. Underlying the whole is a great mass of mesoderm cells.

FIG. 59, X 60, sectioned. In this embryo the cephalic lobes are hardly distin-
guishable. There are three small appendages on the right and four on the left
side. The identification of the appendages can be approximately determined by
the presence on the right of the dorsal organ. There is a great depression across
the posterior thoracic and abdominal region, and below its thickened ectodermic
floor, and apparently arising from it, are a great many free-yolk cells.

The mesodermic area is nearly circular. Along the whole extent of its
thickened margin arises a cloud of isolated cells that extends deeply into the
yolk.

The concrescing posterior margins of the mesodermic areas are very beautifully
shown. The median limbs, c.m.a, are the most conspicuous in surface views.
Sections show that this is due to the presence of a ridge-like thickening of the
ectoderm, continuous with a compact band of underlylng mesoderm.

On the peripheral margin of the mesodermic area the ectoderm is thin, and the
underlying mesoderm is composed of isolated, lymphoid cells. The dark spot on
the right, where the third appendage should be, is due in part to a small protuber-
ance there, but in the main to the presence, below the surface, of an oval sac with
clear-cut walls. It has the appearance of an enlarged mesoblastic somite.

Fic. 60, X 60, sectioned. This is probably a very old embryo, for there are
thick layers of chiten over the tips of the remaining appendages, such as is only
seen in embryos ready to hatch. It at first seemed probable that the small
depression, #/., was the remnant of the cephalic lobes. But on sectioning the
embryo, the thickenings at what is now the upper end seemed to be, without much

126 PATTEN. [Vou. XII.

doubt, the brain and optic ganglia. The deviations from the normal, then,
in this embryo are as follows: The whole right optic ganglion and a part
of the right cerebral hemisphere are absent. On the left side the cephalic
lobe is perfect, and one can distinguish in sections the minute pit that develops
into the lateral eye, Ze. The chelicerae are probably absent, the two pairs
of appendages now present belonging to the second and third thoracic
segments. The appendages of the fourth segment have fused in the median
line, and the marginal fold has closed in behind them. From this point in
the marginal fold a faint cloud of yolk cells extends backward a consider-
able distance to a small thick-walled depression in the ectoderm, 4. From the
posterior wall of the shallow depression arises a minute papilla, which may be the
remnant of the tail lobe. (Compare Figs. 27,43, and 64.) It seems too far back to
represent the fused appendages of the last thoracic segment. The anterior part
of the depression is still further invaginated to form a short, conical pocket,
directed diagonally forward into the yolk. The inner end of the pocket is con-
tinuous with a great mass of cells which seem to be wandering into the yolk
there to degenerate and disappear. It would appear, therefore, that transverse
fission has taken place, and that contrary to every other case that has come under
my notice, except, perhaps, Fig. 45, the subsequent fusion and degeneration have
been greatest in the posterior half of the thorax.

FIG. 61, X 60, sectioned. The mesoderm and ectoderm are much thickened
and concentrated along the axial portion of the embryo. There is no distinction
into head, thorax, and abdomen. The body of the embryo forms a deep, elon-
gated furrow, constricted laterally to form a chain of four or five clearly marked
dilatations, each one of which represents a metamere. At the bottom of two
of the dilatations are shallow, slit-like infoldings that represent the remnants of
invaginated thoracic appendages, za. In sections there is no recognizable remnant
of the nerve-cords to be seen.

The mesodermic area is sharply defined on its periphery, where there is the
usual thickened rim, but it contains very few mesoderm cells, except beneath the
axial portion, where they form a very thick, compact layer ten to fifteen cells deep.
It is this deposit of cells which forms the dark ring seen in surface views.

This embryo is also remarkable for the fact that the surface ectoderm of the
entire mesodermic area is covered with a very thick deposit of a soft, vitreous
exudation, resembling chiten. Over the axial furrowit has accumulated in great
botroidal masses, among which in each section may be seen four or five rounded,
amoeboid cells that look like those seen in the peripheral vesicles. There is no
indication as to where these isolated cells, which are separated by considerable
distance from the embryo, come from, or what their function may be. Their
occurrence at this place suggests a rudimentary amnion and serosa, but otherwise
there is nothing to indicate that they develop in the manner characteristic of these
organs.

FIG. 62, X 60, not sectioned. Embryo with three pairs of appendages, per-
haps the fourth to sixth pairs. The cephalic lobes are deeply depressed, and
separated from the thorax by a sharp, transverse fold. The rim of the mesodermic
area is very distinct, and shows clearly the posterior concrescing margins, and the
post-anal cloud of yolk cells between, #.a.c.

The abdomen is deeply depressed, forming a cavity with nearly vertical walls
surrounding a triangular opening.

No. 1.] VARIATIONS [IN LIMULUS POLYPHEMUS. 127

Fic. 63, X 50, not sectioned. This embryo is shown as an opaque object by
reflected light. It was drawn shortly after it was killed in picronitric acid, and
while still in the fluid.

The embryo is small, but has six pairs of thoracic appendages and well-
developed cephalic lobes. It shows in a very striking manner a conspicuous ridge
extending completely around the embryo. It is broadened and thickened poste-
riorly, and is continuous, with a thick, broad elevation extending forwards to the
posterior portion of the body. ‘This ridge is the very well-developed marginal
thickening of the mesodermic area, such as has been shown in many of the preced-
ing figures. The post-anal band is formed by the concrescence of the posterior
margin of the mesodermic area. When seen in this way, as an opaque object, its
resemblance to the blastodisc of a vertebrate embryo is obvious (shark, lizard, or
chick).

Fic. 64, X 60, not sectioned. A peculiar embryo without appendages, but
showing three pairs of distinct, hollow, mesoblastic somites. In shape it resembles
the embryos of scorpions or spiders more than those of Limulus.

Fic. 65, X 60, sectioned. A small, abbreviated embryo in stage C. The
rudimentary cephalic lobes are depressed and rotated to the embryo’s right.
Their lateral margins are still further invaginated to form two oval depressions,
nearly concealed by distinct folds that have grown over them from the sides, 4.7v.
The oesophagus is comparatively large, and the slit-like opening is also rotated
toward the left.

The chelicerae are probably absent. On the left are the fourth (?) and sixth (?)
appendages, but the fifth is invaginated. On the right the fourth (?) and fifth (?)
are invaginated, and the sixth (?) persists.

A telopore is present behind the last pair of appendages. Back of the embryo
is seen a long, faint band of yolk cells, .¢.c., formed by the concrescence of the
margin of the mesodermic area.

Fic. 66, X 60, not sectioned. A small embryo in stage C.

The cephalic lobes are not distinguishable as such, but the oesophagus is
distinct. There are three appendages (fourth, fifth, sixth[?]) on the left, and
two on the right. In place of the posterior part of the thorax is a broad
depression.

FIG. 67, X 60, not sectioned. A narrow embryo, depressed along the median
line, and surrounded by high and sharp marginal folds. The cephalic lobes are
depressed below the level of the thoracic nerve-cord. The second(?) and third(?)
pairs of thoracic appendages are present with one more appendage on the
right. In the abdominal region is a deep, pyramidal depression, directed nearly
vertically downwards.

The dark area along the marginal fold of the embryo is produced by unusually
thick layers of mesoderm. The margin of the mesodermic area is not clearly
defined ; it lies, in part, at least, beyond the limits of the figure. On the left side
of the embryo, in the thickened mesodermic margin, is a large, marginal vesicle.
The two dark areas back of the abdomen are formed by longitudinal, ridge-like
thickenings of the ectoderm and mesoderm. Between the two ridges is a depressed
area covered by a thin layer of mesoderm and ectoderm. It appears, therefore,
that the posterior median margins of the mesodermic area have not completely
united, except at their very posterior ends. In front of this point is a rather
large rhomboidal space not covered by the mesodermic area.

128 PATTEN.

Fic. 68, X 60, sectioned. In this embryo the cephalic lobes and cheliceral
segment have disappeared, except for a cloud of degenerating cells in the underlying
yolk. These cells lie where the cephalic lobes ought to be, and perhaps they repre-
sent the remnants of these organs ; but, on the other hand, they have more the
histological character of loose mesoderm cells, and are produced, in part, at least,
by the concrescence of the thickened anterior margins of the mesodermic area ;
they thus resemble the post-anal band formed by the concrescence of the posterior
margins of the mesodermic area. This anterior band of cells gradually becomes
more dense posteriorly, till it finally merges into a flattened, oesophagus-like
invagination, which opens just over the anterior margin of the fused second pair
of appendages, ve.

A remarkable feature of this embryo is the curious invagination extend-
ing along the whole left margin of the mesodermic area, 7.m.a. This very
narrow-mouthed infolding converts the mesodermic rim into a thin shelf of
cells, projecting deeply into the yolk. In most sections the cavity of the
invagination is very plain, and it is then seen to be lined with a continuation of
the thin surface cuticula. (See sections, Pl. XI, Fig. 68.)

Along this furrow, and in the margin of the aborted cephalic lobes, the surface
of the ovum is covered with masses of minute, intensely red dots that look very
much like bacteria. The dots are also found just below the surface, and then it
often appears as if some of the nuclei had ruptured, allowing a swarm of chromatin
granules to escape.

The second pair of appendages have fused, forming a broad, biloped plate back
of which the nerve-cord terminates. The fifth and sixth appendages of the left
side are absent. The abdomen is absent ; in its place is a very deep, thick-walled
depression. The marginal fold is very thick posteriorly, and terminates abruptly
in a foot-shaped enlargement that dips down into the depression, and forms a part
of its lateral wall. The posterior end of the right marginal fold is weak, and is
not connected with the fold on the opposite side.

The dark area back of the abdomen, /.a.c., represent a great cloud of cells lying
in the yolk, and formed by the concrescence of the posterior portions of the
mesodermic rim.

Fic. 69, X 60. Embryo with nearly circular mesodermic area. The nerve-cord
appendages and cephalic lobes have disappeared.

At the anterior end is a mass of mesoderm with a conical cloud of yolk cells
beneath, so faint as to be hardly visible in surface views. From this point the
longitudinal, ecto-mesodermic band, gradually increases in distinctness toward the
posterior end, where the two layers that compose it are sharply separated. The
outer layer is here raised to form the median protuberance, which probably
represents an enlarged tail lobe. The outline of the mesodermic area is clearly
defined, and along its lateral and posterior margin are numerous irregular masses
of cells that extend deeply into the yolk.

The tissues appear healthy and normal, but there is no trace of organs other
than the median appendage.

Lith. Anst.v. Werner &Winter, Frankfurt $M.

NMA,

61.

muy

~—— .

69.

mh.

va

ip

Tih Instn Werner aMinter, Frankfurt,

—

mr Pt ae
rd et nae
s) ane
5 |

af

Mane

Ls!

i
}
7

130 IBS GIVE [VoL x1

EXPLANATION OF PLATE VIII.

Fic. 70, X 60, not sectioned. A very much reduced embryo probably in
stage C. There are faint indications of the nerve-cord and cephalic lobes and
oesophagus. Two appendages are present on the right; on the left the corre-
sponding appendages are invaginated.

Back of the embryo is a long band of inner layer cells formed by the concresced
margin of the mesodermic area. The length of this band affords some indication
of the advanced stage of development of the embryo, and the amount of reduction
it has undergone.

The mesodermic area is nearly circular and very large, only the posterior por-
tion being shown in the figure. Its extent may be approximately determined by
continuing the curved lines, c.m.a.

Fic. 71, X 60, sectioned. This embryo is so flattened and distorted that its
parts are hard to identify. Beginning at the anterior end of the embryo and
passing backward, we find first a deep depression, at the bottom of which is a
well-defined nerve-cord. From the right wall of the depression arises a large
appendage, af", that projects over and conceals the nerve-cord. Its mate on the
opposite side is absent.

Back of the depression, the surface of the embryo rises to nearly the level of
the ovum, and here is a nearly normal neuromere, 7.#., with an invaginated
appendage on either side of it, a*. The ectoderm then thins out considerably,
but thickens again to form an indistinct nerve-cord, .m *, with a large invaginated
appendage on the left, af °. Following this are two swellings that may represent
appendages ; and finally this decidedly mixed-up collection of organs terminates
in a thickened ectodermic disc, which appears to represent the anal plate. In
both sections and surface views indications of the thickened mesodermic margin
are seen on the side. Posteriorly the margin is very broad, and, owing to
concrescence, deeply notched.

This embryo is in the last stages of degeneration that show recognizable organs.
In the following figures, even these rudiments disappear, and nothing remains
but simple layers of mesoderm and ectoderm, complicated by more or less irregular
foldings.

In the following figures, 72-78, so far as one may judge from what little
remains of the embryos, there has been median fusion, and antero-posterior
degeneration. The embryos were probably laid down throughout their whole
extent, and then underwent a gradual median fusion and degeneration, coz/ized
principally, however, to the axial organs, and to the region midway between the two
extremities. In none of this series is there any recognizable trace of nerve-cord,
appendages, or other axial organs; although the peripheral mesoderm retains in
many cases very nearly its normal character. It is assumed in each case that the
present condition was preceded by one having nerve-cords, appendages, cephalic
lobes, and sense-organs, and that these have gradually faded away, leaving only
the mesodermic area in a recognizable condition. Within this area are two (sel-
dom but one) great masses of mesodermic (?) or yolk (?) cells, either with or
without an overlying ectodermic invagination. They represent the remnants of
the head and tail.

No.1.] VARIATIONS IN LIMULUS POLYPHEMUS. E3t

It is very interesting to note that these two cell masses resemble in position
and general appearance the two proliferating areas, “ primitive cumuli,” which at
a very early period give rise to the head and trunk of normal embryos.

No trace of primitive cumuli is visible in normal embryos after stage C, and
as these embryos are certainly as far advanced as that stage, they either must
have retained to this late period a very early embryonic character, or else the
processes of degeneration have carried them back a second time to their original
condition. In either case, the facts seem to emphasize the morphological impor-
tance of these two proliferating centres.

There is no very reliable indication as to what is the head or tail Ax/age,
except perhaps in Fig. 73. The embryos have been oriented in each case with
the smaller An/age at the anterior end, and arranged and numbered as far as
possible in accordance with the degree of degeneration.

It is obvious that such a plan of arrangement cannot be followed with any
accuracy, since each embryo has its own way of degenerating, begins to degenerate
at different stages of development, and may have been defective at the outset.
What finally happens to these degenerated embryos cannot be determined, since
further degeneration of either head or tail 4z/age would make it impossible to
determine whether it belonged to this series or the next. But I see no reason to
doubt that degeneration goes on till even the head and tail Ax/agen are ab-
sorbed and disappear completely.

While Fig. 73 is the first one of the series here represented, numerous inter-
mediate stages between it and normal embryos of stages C and D were found.

Just how far normal] development may proceed before this general degeneration
begins could not be determined. I do not recollect seeing any embryo undergoing
this phase of degeneration, which in whole or in any part had reached a higher
stage of development than that characteristic of stage D. But it must be borne
in mind that all these embryos had been developing or degenerating in the same
jars with normal eggs, for from fifteen to eighteen weeks. The normal ones
passed through their whole development and escaped from the jars as trilobite
larvae, in but little more than half this time.

FIG. 72, X 60, sectioned. In this embryo, every trace of cephalic lobes,
oesophagus, nerve-cords and appendages has disappeared. The mesodermic area
is very large and irregular. Anteriorly the margin of the same is pretty close to
the axis of the embryo. It there consists of numerous star-shaped masses of cells,
densely crowded together, m.a. The posterior margins are much thickened
to form two great diverging arms, which are united at their distal extremities, but
are separated medianly by a triangular area devoid of mesoderm cells (compare
Fig. 67). These thickened arms unquestionably represent the posterior margins
of the mesodermic area which has not completely concresced along the median
line. It is important to recognize this, because it shows approximately to what a
late stage of development this embryo belongs. Such embryos as this and those
that follow on this plate, cannot for a moment be considered as belated normal
embryos, in early stages of development. Their great age (two or three months),
the concrescence of the mesodermic margins, the histological character of the
tissues, and a comparison with the early stages of normal embryos, is sufficient to
prove this beyond any question.

A thin layer of ectoderm covers the greater part of the mesodermic area, but
it is thickened and elevated in the axial region, over what probably represents the

132 PATTEN. [Vov. XII.

body of the embryo. The elevation forms a broad, figure-8-shaped ridge
(marginal folds ?), with a slit-like depression in about the centre of each loop.
There is a thickened layer of mesoderm beneath the whole of this axial portion.

Beneath the anterior end is a very large solid mass of mesoderm cells, ¢.77.5.,
from which pseudopodia-like streamers of cells radiate deeply into the yolk.

Fic. 73, X 60, sectioned. In this embryo the tissues are beautifully clear,
and sharply differentiated, resembling histologically the late stages in the forma-
tion of the blastoderm of normal embryos. The ectoderm, instead of containing
flattened cells with little protoplasm, is composed of high columnar cells with
nuclei at their outer ends, their inner ends being filled with masses of yolk gran-
ules. The whole layer is sharply separated from the underlying yolk and
mesoderm.

There is no cluster of yolk cells at the anterior end. The dark median band
consists of columnar ectoderm cells, like those described above, but higher, and
below them is a clear band of mesoderm. Both layers are entirely undifferentiated.

At what seems to be the posterior end, is a deep, oblong depression with two
infoldings of its anterior wall, like steps leading down into it. Numerous de-
generating nuclei arise from the inner surface of the infolded layer, and lie scattered
about in the neighboring yolk. The mesodermic area is circular, with a faint
thickening of its anterior and lateral margins. There is no trace of a posterior
concrescence.

FIG. 74, X 60, sectioned. Embryo similar to the preceding. It consists of
two disc-shaped ectodermic thickenings, connected by a longitudinal band of the
same nature. Beneath the thickened ectoderm are corresponding mesodermic
thickenings, but especially enlarged at either ena.

There is a shallow depression in the centre of each ectodermic disc. In both
Figs. 74 and 76 the ectoderm of the posterior thickening shows a higher grade of
histological differentiation than those of the anterior one.

The peripheral outline of the mesodermic area was not visible.

Fic. 75, X 60, sectioned. An embryo with a circular, mesodermic area, and
a “blastopore ’’-like invagination at either end.

The invaginations are similar to those already described, except that com-
paratively few yolk cells surround the invaginations. The outer surface of the
invagination is covered with a thick layer of bacteria (?). The ectoderm is a
uniformly thin layer of cells, showing no trace whatever of a median thickening
connecting the two invaginations. The mesoderm is remarkable in that it is of
nearly uniform thickness throughout its whole extent, except at the margin where
it is decidedly thickened, 7z.a.

Most of the marginal mesoderm is composed of the characteristic, striated cells
seen in stage D, and later.

Fic. 76, X 60, sectioned. This embryo consists of a circular, mesodermic
area containing scattered, star-shaped masses of cells, either extending from the
mesoderm into the underlying yolk, or lying freely in it. There is.a deep, circular,
ectodermic invagination at the anterior end with a thick layer of mesoderm sur-
rounding it, and a similar, but larger, deeper, and more irregular cavity, at the
posterior end. The ectoderm lining this invagination is considerably thicker than
that of the anterior one, and the thickening extends along the surface for some
distance back of the infolding, f.a.c. This thickening is probably the result of
the concrescence of the posterior margins of the mesodermic area. There are

Nosiay | VARTA TIONS IN EIMCLUS: POLYPHEMUTS. 133

many degenerating nuclei in the ectoderm and mesoderm of the posterior thicken-
ing, but comparatively few in the anterior one.

There is a gently elevated, thickened band of ectoderm, connecting the two
invaginations, but it does not show any histological differentiation, and the
slightly thickened layer of mesoderm underlying it is not segmented or otherwise
specialized.

FIG. 77, X 60, sectioned. In this embryo everything has disappeared except
two masses of cells, one at either end of what was the body of the embryo.
There is a small cluster of mesodermic cells at the anterior end, and a similar one
at the posterior extremity. Over the latter is a thickened ectodermic area with a
rather deep, vertical invagination in its centre. There is no indication of any
ectodermic or mesodermic thickening connecting these two masses of cells.

The outlines of the mesodermic area are not well defined.

Fic. 78, X 60, sectioned. Embryo with well-defined, circular, mesodermic
area. There is an axial band of thickened ectoderm, slightly enlarged at either
end, and beneath it is a broader band of mesoderm, which becomes enlarged at
either end to form two great irregular masses of cells, c.¢. The latter project
deeply into the yolk, and from them radiate pseudopodia-like processes, composed
of dense masses of cells. Over the centre of these masses, ectoderm and meso-
derm are continuous, as though the inner cells arose by proliferation from the
outer ones. The outer layers at these two points, however, contain no karyokinetic
figures, but instead, numerous degenerating nuclei.

The ectoderm over the posterior mesodermic enlargement is composed of a
very thick layer of columnar cells, but shows no noteworthy histological differen-
tiation.

No bacteria (?) were visible on the surface of the ovum, and the tissues
appeared healthy and normal.

Fics. 79-89. In this series, the processes of degeneration have followed
somewhat different lines than in the one just described (Figs. 72-75). But we do
not assume that there is any sharp distinction between the two series.

We have here in some instances undoubted evidence that the embryos in
question represent mere remnants of highly developed individuals, or to put it
more accurately, some of these embryos may consist of a single recognizable
organ only, which may be in the condition seen in normal embryos of stage D,
while every other organ characteristic of this stage has either failed to develop or
has completely degenerated.

This is especially true of the embryo shown in Fig. 82, where the large tail
lobe and the concrescing mesodermic margins suggest the conditions seen in stage
A and D (see Figs. 5 and 6), while the remainder of the body of the embryo is
reduced to a mere sac, invaginated into the yolk.

In the cases that seem to lead up to this condition, the body of the embryo is
usually pinched and narrowed and much depressed, with high, conspicuous,
marginal folds that approximate each other toward the median line, Fig. 70.
The appendages may or may not be present, but if present, they are usually
invaginated.

Practically all conditions, from normal embryos, and ones like those in Figs.
61, 62, and 67, to those shown in Figs. 81 to 89, have been observed.

From the first condition towards the last there is a gradual reduction of sense-
organs, nervous system and appendages, till nothing is left of the embryo, but a

134 PATTEN. [VoL. XII.

more or less clearly defined mesodermic area, with a central thick-walled sac,
representing the remnant of the axial portion of the body. This sac may be a
single one, regular in outline, or it may be more or less irregular, as though com-
posed of two or more sacs arranged one behind the other along the median line
(Fig. 81).

These sacs may be of a different nature in each case. In Fig. 61, the dilata-
tions of the common furrow represent single metameres, from which the ap-
pendages and nerve-cords have nearly disappeared. In Figs. 72 and 79, each sac
probably represents the remnants of several segments. These embryos may
belong to the hour-glass series, where there is a transverse constriction between
the third and fourth thoracic segments dividing the embryo into two distinct
parts. Subsequent degeneration of the appendages and other organs would
reduce each part to a separate sac.

When the sac is comparatively large, and communicates with the exterior by a
small opening only, as in Figs. 82, 83, 84, 85, and 86, a study of the sections
through the sac reveals an unexpected complexity of structure, the walls being
thickened or variously curved, so as to suggest that the cephalic lobes, with their
various ganglia, and the anterior part of the nerve-cord, still persisted, while the
appendages had disappeared. I have not been able to satisfy myself that such
was really the case, for the various parts are thrown so much out of position, that
it is impossible to identify them. But the whole histological character of the
thickened walls of these sacs resembles nerve tissue more than anything else.
For this reason, and owing also to the shape of the sacs, cross-sections of them
resemble cross-sections of embryonic vertebrate brains. This is true of Figs. 83
and 84, especially the latter, where the tubular outgrowths from the anterior part
of the sac recall the similar outgrowths to form the lateral eyes in vertebrates.

These sacs themselves degenerate. Their walls collapse and lose their
character as distinct cell layers, till there is left merely a central mass of cells ina
much contracted mesodermic area. All distinction between these two parts, z.e.
ectoderm and mesoderm, is gradually obliterated, and what then remains seems
to disappear completely by absorption. It would be a very difficult matter to prove
that this complete absorption really does occur, for one cannot follow any single
individual through the process, but there is no difficulty in finding among these
degenerating eggs a few apparently healthy ones in which there is no trace of
nuclei or cell clusters to be found. There is also no difficulty in finding ova, showing
various stages, from embryos composed of a barely recognizable cluster of cells,
to ones like that in Fig. 89. These ova without embryos may be three or four
months old, and if they had failed to develop in the first place and died, it is
difficult to understand how they could be preserved so long without putrefaction.

In all these embryos, 79-89, as well as those in the preceding series, the
tissues at first sight appear perfectly normal and healthy. They stain well,
and one may find some nuclei in karyokinesis. But in one part or another
may be seen, among the perfectly normal ones, numerous nuclei, in which
the chromatin is crowded at one or both poles, leaving the centre perfectly
clear. In those apparently just entering on the process, the chromatin stains
very deeply, as in nuclei undergoing karyokinesis, but in the later phases
the color becomes less intense. The nuclear wall disappears, and finally nothing
is visible but a few faint, isolated chromatin granules, which in turn disappear
also. Immense masses of these nuclei are sometimes seen at the anterior and

Nott) “VARIATIONS LNOEIMOLGS: POLYPHEM GS. £35

posterior ends of the degenerating embryos. They are also seen along the
longitudinal, median line, in the nerve-cords, and in the appendages.

I have described and figured just such nuclei as these in the ectoderm about
the eyes of Acilius, and have seen them in the degenerating masses of endoderm
cells, or yolk nuclei, at the anterior and the posterior ends of healthy and normal
scorpion, Limulus, and insect embryos.

Similar nuclei are found, and they undergo the same changes, in the degener-
ating amnion and serosa of insects, where these membranes, after forming a
“dorsal organ,” are invaginated into the yolk.

In none of these cases are the degenerating nuclei ingested by more vigorous
amoeboid cells; on the contrary they disintegrate and then fade out of sight,
probably by forming solutions that diffuse gradually throughout the yolk or other
tissues.

FIG. 79, X 60, sectioned. The mesodermic area is ovoid. There is a well-
marked marginal thickening; also a post-anal one, /.a.c., showing where con-
crescence has taken place. Along the axis of the embryo the thickened ectoderm
is elevated and contains two distinct, elongated invaginations. At the posterior
end of the second ore is a small and still deeper invagination. The lateral walls
of the second invagination are constricted at a point that probably marks the line
of separation between two adjacent segments; compare Figs. 61 and 81. The
invaginated ectoderm, here, as in the preceding figures, probably represents the
remnant of the whole axial ectoderm of normal embryos, z.e. cephalic lobes, nerve-
cords and appendages, as far laterally as the marginal fold.

This embryo may represent the last stages of an hour-glass embryo, the
anterior sac being the remnant of the cephalic lobes and first three pairs of
appendages, and the posterior sac the remnants of the last three thoracic
appendages and the abdomen. Compare Figs. 52, 53, 54, 55, and 57.

Fic. 80, X 60, sectioned. This is similar to the preceding, only the con-
crescence of the posterior margins of the mesodermic area is much more clearly
shown, /.a.c. There is hardly any trace of the axial portion of the embryo, other
than an elongated, ectodermic thickening with a large diamond-shaped depression
extending along it. In surface views, there are no indications of nerve-cords or
appendages, but in sections there are slight differentiations of the thick wall of
the furrow, that may represent the last traces of the nerve-cords and appendages.

Fic. 81, X 60, sectioned. Over a heart-shaped, mesodermic area lies a thin layer
of ectoderm, elevated considerably and thickened along the median line. Along the
summit of the ridge is a series of depressions increasing in size from before back-
wards. The third depression is so deep as to be almost tubular. The last is
broad and shallow, but has a rather deep depression in its centre. Each depres-
sion probably represents a segment from which nerve-cord, appendages, etc., have
disappeared. The ectoderm has the appearance of old specialized tissue, although
one cannot identify any particular kind of tissue or organ in it.

Fic. 82, X 60, sectioned. This extremely instructive specimen shows clearly
the real stage of development, and the nature of the changes this class of
embryos has undergone. The mesodermic area is kidney-shaped, and precisely
defined by a thickened margin. The latter forms two distinct lobes that are
manifestly growing toward each other in order to concresce along the posterior
median line. On examination of the normal embryos it will be seen that this
phase of concrescence is not due till stage C-D. The abdomen of this embryo

136 PATTEN. [Vou. XII.

has apparently persisted as a large, conical lobe, projecting almost vertically
upwards from the surface of the yolk. It is evidently the same kind of a lobe
seen in Figs. 13, 48, and 49. The body of the embryo, as in the three preceding
figures, is reduced to a mere sac. The opening of the sac is triangular, and almost
as wide as the sac is broad. A little to one side, on the floor at the anterior end,
is a small, tubular invagination, oesophagus (?), directed vertically downwards.

Fic. 83, X 60, sectioned. This is another important embryo. The meso-
dermic area is relatively small, and not very sharply defined anteriorly, owing
to the small number of mesoderm cells contained in it. Posteriorly, and
this is a point of value in determining the poles of the embryo, the meso-
dermic margin is decidedly thickened, and shows very clearly the effects of
concrescence, f.a.c. The axial ectoderm shows the usual thickening, and is
invaginated to form a complicated, thick-walled sac, opening by an oval aperture.
The sac projects forwards beneath the ectoderm, and at its anterior end expands
into two transverse tubes. If the nerve-cord and the cephalic lobes are still
present, although they cannot be recognized as such, it is obvious that the floor
and sides of the sac will consist of these organs, while the lateral eyes would be
situated somewhere near the end of the lateral diverticula. This condition is
what normally obtains in vertebrates, and is of obvious significance in view of
the other resemblances which we have pointed out elsewhere, between the normal
brain and eyes of Limulus and those of vertebrates. Such embryos are extremely
rare, and may have no morphological value. But the possible significance of such
variations should not be overlooked.

Beneath the aperture of the sac, the floor of the same suddenly slopes down-
ward, and at the bottom of this depression is still another smaller and tubular
one that projects downward and forward into the yolk. The character of the
tissues, which seem to be perfectly normal, and further structural details, are
shown in the three transverse sections (Pl. X, Figs. 83',”,%).

Fic. 84, X 60, sectioned. Of the mesodermic area, which is nearly circular,
only the posterior portion is represented. The whole area consists of a thick layer
of rounded, and often isolated, lymphoid cells, lying in a space between the yolk
and a thin layer of ectoderm. ‘There is a prominent marginal thickening extend-
ing completely around the mesodermic area. In the centre of the latter the
ectoderm is thickened and deeply invaginated to form a roughly cubical cavity,
partly closed over in front by a backwardly directed lip, or fold, and behind by a
similar one directed forwards.

The thick walls of this cavity probably represent that part of the embryo from
which the nerve-cords and appendages are developed, the whole complex mass
being reduced to a continuous, undifferentiated layer.

From the anterior wall of the cavity projects a short, tubular outgrowth that
looks very much like an embryonic stomodaeum, the resemblance being increased
by the presence of a pair of muscle strands, which extend laterally from the tip of
the tube to the surface ectoderm. The presence of this specialized tissue offers a
sharp contrast to the very early embryonic character of the rest of the embryo.

This fact supports my view, if further evidence were necessary, that these are
very old embryos whose organs have been reduced, with this exception, to the
simplest embryonic tissues.

Fic. 85, X 60, sectioned. The mesodermic area here is small, circular, and
poorly defined. The margin, however, at the anterior border is greatly enlarged

No.1.] VARIATIONS IN LIMULUS POLYPHEMUS. 137

to form a marginal vesicle, #.v. In the centre is a kite-shaped sac opening to the
surface by an elongated figure-8-shaped opening,c. The anterior part of the floor
of the sac protrudes upwards and forwards (cephalic lobes[?]), and from this
inclined wall projects forwards a short tube or stomodaeum(?), 6. Both structures
are seen in sections in Pl. X, Fig. 85',?.. The sac extends backwards some distance
beyond the aperture, c, as a cylindrical tube which ends blindly deep in the yolk.

There is a shield-shaped thickening of the ectoderm, lying over and a little to
the right of the sac. It is represented by the shaded area, d. A section through
this thickening and the posterior end of the sac is shown in Fig. 857.

Fic. 86, X 60. This embryo is similar to the preceding one. Unfortunately, it
was lost or injured before it could be sectioned. In the centre of the mesodermic
area is a thick-walled sac, which communicates with the exterior by a diamond-
shaped opening. The anterior end of the sac is very broad and spacious, and the
posterior part a deep, nearly closed furrow terminating in a tube that dips vertically
into the yolk. The relation of these parts, so far as they could be determined
in surface views, is shown in the accompanying median, longitudinal, optical
section. Fig. 86, 2. .

Fic. 87, X 60, not sectioned. The mesodermic area appears to be nearly
circular, but its outline is very indistinct. In its centre is a conical, thick-walled
projection, with an opening at the summit leading into a large, nearly spherical
sac. Specimens resembling this one in all but unimportant details are comparatively
common. They are evidently further modifications of the conditions seen in the
five preceding figures. This specimen was not sectioned, as it did not appear to
present any noteworthy features.

Fic. 88, X 60, sectioned. In this case a sac like that seen in the preceding
figures is beginning to break up. The mesodermic area is distorted, and on the
left ill defined; on the right the rim of the mesodermic area shows the character-
istic, star-shaped masses of cells that are spreading out into the yolk, preparatory
to their final dissolution.

In sections of other specimens, apparently in a similar condition to this one,
the walls of the sac have lost their sharp outlines, as though they were gradually
falling apart ; and the cells composing them have the character of closely packed,
lymphoid cells, rather than that of the columnar ones seen in the preceding figures.
Still one would hardly suspect, from a mere inspection of the sections, that these
specimens were anything else than very early stages of normal embryos.

Fic. 89, X 60, sectioned. In this remarkable specimen we have a good
example of the last stages in the degeneration of the embryo. In all the other
cases I have studied, where the thick walls of the sac had broken up into a form-
less mass of cells, the latter were usually rounded, lymphoid ones, showing no
further differentiation. But in this instance the entire mass, representing all that
is left of the embryo, is composed of those peculiar cells, containing coiled filaments
that are only found at a late embryonic period in the thickened margin of the
embryonic area. I have already referred to these cells in one of my earlier papers
(“Origin of Vertebrates from Arachnids,” p. 375), and at that time supposed they
were purely embryonic, and gave rise in some instances to the future muscle cells.
Since then, however, I have found them as free, amoeboid cells, in great
numbers in the tissues of the adult animals. They differ somewhat in histological
characters from the embryonic ones of the mesodermic margin, but there can be
no question whatever about their identity. In the adult these cells are very

138 PATTEN.

abundant all through the tissues about the median and lateral eyes and the
olfactory organs, and I see no reason to doubt that they will be found as
abundantly elsewhere. Kingsley failed to confirm the existence of these cells in
the embryo, which is somewhat surprising, for in stage they form as conspicuous
an object in sections and surface views as any other organ of the body.

I expect to describe in more detail the history of these cells in another
paper. I refer to them here to make clear the remarkable fact that in these
degenerate embryos there is almost nothing left but a mass of these highly
specialized and peculiar cells. The only thing remaining besides them are the
ordinary yolk cells and the blastoderm, and even this disappears, as such, over
the mass of cells under consideration. It is as though a mammalian embryo
should gradually degenerate into a mass of cells consisting solely of osteoblasts,
or muscle cells, or any other specialized form of tissue. However, it must
not be assumed that other kinds of cells are during degeneration converted
into the fibre cells. The latter merely persist as such after the others have
disappeared.

On careful study of one of the sections we see the fibre cells usually in ill-
defined groups, with intermediate areas containing some nuclei, unquestionably in
karyokinesis. Other nuclei, however, seem to have made preparations for division,
but instead of doing so they break into numerous deeply stained globules, which
become disseminated through the yolk, and growing fainter and fainter, finally
disappear.

A portion of a section through the embryo is shown in Pl. X, Fig. 89. Below
the disc is a clear area containing a few yolk globules, and many degenerating
nuclei and cells, apparently derived from the mass of cells above.

im
T
L : Ss
a -
'
— -
'
‘
'
.

‘a4

s

34,

&2Z

1 i

35.

cma. >
87.

Lith Anstr Werner Winter, Prankfirt #0.

Tor, Aeon
ee ty
OTT. cet

140 PATTIEN, . [Vorcxt:

EXPLANATION OF PLATE IX.

In this plate, double monsters are shown in various stages of formation. In
all the cases I have seen in Limulus, double embryos are formed by fission, or
more correctly by the gradual intercallation, beginning at the anterior end, of two
new halves between the old. If there are five paired organs on each segment,
and a is the most median one, and e the most lateral, then a will be the first
new organ to appear, and it will appear in the median line of the first segment, as
an unpaired organ, having the same appearance as each member of the paired
organ. It divides, and in its place in the same segment will be found an unpaired
organ like organ 4. But at the same time a new, unpaired organ, like a, will be
formed in the median line of segment number two. At the next division, organ
a will be produced in the median line of the third segment, 4 in the second, and
c in the first; organs a and 6 being now completely formed in pairs in the first
segment, and organs 4 in the second. This process goes on till two complete
new halves are wedged in between the old, and two new individuals are produced,
each individual consisting of an old and a new half. The old halves were
produced by normal apical growth from behind forward ; the new half was also
produced by apical growth, but from before backwards. Each new half is a
mirror image of the other, and at the same time the mirror image of the old half
next to it; but the new halves are united by their lateral margins instead of the
median ones.

The very first steps in this process have not been seen. Hence the manner in
which the new halves of the cephalic lobes and the oesophagus are produced, can
only be inferred from the manner in which new organs are formed in the post-oral
segments. The process of forming two new organs by the division of the
unpaired ones, takes place in exactly the reverse order that unpaired organs are
formed by the fusion of paired ones.

FIG. 90, X 33. The entering wedge formed by the newly produced organs has
reached the fifth thoracic segment. At the end of the wedge is an unpaired
neuromere ; in the next in front of it are two new halves of a neuromere and an
unpaired appendage. In the next segment in front of that are two complete and
newly formed chelicerae.

FIG. 91, X 33. Here the process has progressed still further. The division of
the new appendage in the third segment is almost completed, the separation
extending from the tip nearly to the base. In the fourth segment the new
appendage is perfect, but shows no trace of division.

FIG. 92, X 30. The growth of the new halves is practically completed, forming
two distinct embryos, which have, however, an abdomen in common. The new
halves lie on the upper side, as the egg is placed in the figure, the old ones being
below. Each of the old halves has been rotated on its tail end 90°, one to the
right, the other to the left. They have been forced apart in this way, at first by
the wedge-like ingrowth of the new nerve-cords. But as the organs lateral to the
two nerve-cords develop, they push the previously formed parts still farther right
and left. This goes on till the new embryos form a straight line tail to tail.
Further rotation of the old halves will be stopped by the interference of the
lower margins of the mesodermic areas at y.

No.1.] VARIATIONS IN LIMULUS POLYPHEMUS. I4I

FIG. 93, X 33. The process seen in the preceding figure has evidently begun
here. But before the two new halves were completely formed, the left-hand
embryo began to concresce along its own median line. The cephalic lobes and
first two segments have disappeared in that way, exactly as in single embryos
(see Pl. IV). The sixth pair of appendages has not been formed in the new
halves.

FIG. 94, X 33. Here the process seen in the preceding figure has been carried
further. Fission probably occurred at first, as in Fig. 92; degeneration of one
embryo then followed, as in Fig. 93; and finally the two embryos separated,
moving tail first in opposite directions to the position they now occupy. The
right-hand embryo is apparently normal; the left-hand one has been reduced by
antero-posterior fusion and degeneration, till the last two pairs only of thoracic
appendages remain. The fourth pair have united in the median line. The
“dorsal organs,” d.o., are still separate and very distinct. Between them is a
small mass of cells, the remnant probably of an unpaired appendage. This
embryo, with its five legs and large tail lobe, is very similar to that in Pl. V, Fig. 48.

Fic. 95, X 33. Thisis a double embryo which, at one stage of its development,
was probably like that in Fig. 92. The normal upper embryo still occupies its
former position tail to tail with the lower one, which median concrescence and
antero-posterior degeneration has reduced to a small, median appendage, project-
ing from an oval depression. The abdomen is a narrow oblong thickening, and
at the opposite extremity of the body is a median depression that may represent
the fused dorsal organs (compare Fig. 97), or the remnants of an oesophagus.

On either side of this much-reduced embryo is an irregular, dark band, formed
by the concrescence of the margins of the mesodermic areas of the two embryos,
and probably representing two rudimentary hearts.

Fic. 96, X 33. This is a much older embryo than any of the preceding. It
has passed successively through approximately the same stages seen in Figs. g2-
94. Thé exact sequence of events, of course, cannot be determined ; but we can
understand the position of appendages in the right-hand embryo, by assuming
that after a tail-to-tail condition was reached median fusion and antero-posterior
degeneration of the lower embryo followed, somewhat as in Fig. 93. Then the
tail of the lower embryo pushed past the tail of the upper, going to the right of it
instead of to the left, as in Fig. 94. But meantime the margin of the thorax of
the larger embryo had grown so large that it had practically preémpted the
territory in that region ; the tail of the smaller embryo was thus forced to bend
to the right, through an angle of about 90°. The curvature thus produced of the
axis of the smaller embryo is shown by the arrow 2, and by the dotted line which
runs along the median line from the fourth thoracic appendage, af 4, to the abdo-
men. The latter is perfectly normal and well developed. The cephalic lobes and
first three neuromeres have disappeared, through median fusion and antero-
posterior degeneration. The appendages of the fourth and fifth segments have
fused along the curved median line, and are very much reduced in size. The
sixth pair, af, are still quite large, and fused only at their base. The flabella,
ji., are very large, and each is separated by a relatively wide space from the base
of its respective appendage.

Just under the end of the left sixth appendage is a small projection that
cannot be accounted for, as its position is such that it cannot be brought into
line with any of the other appendages. It may be the left flabellum of embryo 4.

142 PATTEN.

FIG. 97, X 33. This embryo is in an older stage than the preceding one. It
is an almost perfectly symmetrical double embryo, tail to tail as in Fig. 92.
Median fusion and antero-posterior degeneration have affected both embryos,
but the lower one more than the upper. In the latter, everything as far as the
fourth segment has disappeared. The dorsal organs are very large, and are
supplied on one margin with considerable black pigment, a condition that has not
been seen in the normal embryos. They have approached the median line, but
have not fused with each other. The appendages of that segment, ze. fourth,
have fused, and the single appendage thus produced is reduced to a small conical
projection, af4. The fifth pair have also fused, forming a long zigzag appendage.
The sixth pair have fused at the base to form a large oval vesicle, from the
summit of which project the separate ends of the reduced appendages. The
chelaria and abdominal appendages are normal.

In the lower embryo, 2, the ‘dorsal organs” on the fourth segment have
fused. Everything anterior to that has disappeared, but the appendages of the
fifth and sixth segments remain as elongated, crumpled, unpaired organs.

The nerve-cords extend without interruption or modification from one animal
to the other. There are two perfectly normal hearts and two abdominal lobes,
both organs being shared in common by the two embryos.

Fic. 98, X 33. This is a very interesting case, as it shows clearly three entirely
separate phenomena, z.e. (1) longitudinal division; (2) median concrescence;
(3) transverse fission. The two latter conditions have been seen in the single
embryos previously described. The two embryos now form almost a straight
line. The abdomen of the original embryo is intact, and would have become,
in all probability, the abdomen of embryo A. A remarkable fact is the obvious
“weakness” of the new half of embryo 4, as compared with its old, right
side. This is shown by the absence of its second and fourth appendages, and
the fourth neuromere. In place of the fourth appendage is a minute pore that
may be the invaginated remnant of the same. Embryo Z& is a beautiful example
of the hour-glass type, the constriction occurring in the usual place, between the
third and fourth segments. The appendages of the third and fourth pairs are
fused in the median line. In front of the third pair, and back of the fourth, is
a gradual diminution of the concrescence. The large size of the chelicerae
in embryo Z is surprising, considering the otherwise reduced condition of the
embryo.

FIG. 99, X 35. This is a very rare form, and the only one of the kind I have
seen. It at first sight seems to belong to a different class from the preceding,
and to have been produced in a different manner. However, it is easily explained
by assuming that during fission, like that in Fig. 91, median fusion and antero-
posterior degeneration destroyed the anterior part of embryo Z& as fast as it was
formed. Two new nerve-cords extended to the tip of the abdomen, and a row of
unpaired appendages, extending from the fourth abdominal to the fifth thoracic
segment, have been formed.

a4

=e

144 PATTEN. [Vou. XII.

EXPLANATION OF PLATE X.

FIG. 100, X 25. This is one of the oldest double monsters seen. It should be
compared with Figs. 96 and 98. The development of the new halves had appar-
ently not separated the oid ones much more than in Fig. 98. The main embryo has
the abdomen of the original embryo.

In the smaller embryo are two unpaired, tongue-like, abdominal appendages.
The appendages of the sixth segment are fused at their base, and in front of
them is a row of three long, crumpled filaments, representing the medianly fused
appendages of the third, fourth, and fifth segments. All the segments in front
of the third have disappeared.

Fic. 101, X 30. Same embryo as in the preceding figure, seen from above. The
dorsal surface of the thorax of the smaller embryo is spread out on that of the
larger. Below the dark mass of cells that represent the remnants of the thickened
margin of the mesodermic area is an elongated cloud of cells, probably the remnant
of the oesophagus, ge.

There is a well-developed heart in the abdominal region of the larger embryo,
but none in that of the smaller one.

FIG. 102, X 30. There are three embryos on this egg. Embryo 4 is normal
and perfect in everything except the abdomen. JZ has undergone median fusion
and degeneration, and transverse fission. The cephalic lobes and first four segments
have disappeared, except two incompletely fused appendages. The abdomen and
the posterior part of the thorax persists. The latter is bounded in front of the
fifth pair of appendages by a great fold that extends completely across the median
line. The nerve-cord in this posterior remnant of an embryo forms a conspicuous,
unpaired ridge.

Embryo C has undergone such fusion and antero-posterior degeneration that
nothing remains but the fused appendages of the sixth segment, and a rudimentary
abdomen.

It is probable that the original embryo divided lengthwise, giving rise to 4 and
B&C, and the latter then divided, giving rise to B and C.

FIG. 103, X 30. In this triple embryo the individuals, 4, 4, and C, were proba-
bly produced in the same way as in the preceding. As all the embryos could
not be seen at once, each embryo was drawn separately with the aid of a camera,
and finally all three united and represented as though spread out on a flat surface.

Embryo 4 has undergone median fusion and transverse fission. The fused
appendages of the first four segments are arranged in a single row; it is very
rarely that one sees as many unpaired appendages as this.

The cephalic lobes are narrowed, and covered by a hood-like fold of ectoderm,
through which one sees the oesophagus.

The marginal fold has grown across the median line in front of the fourth pair
of appendages, as in the extreme forms of hour-glass embryos. In front of this
fold, and near the median line, are the dorsal organs.

The fifth pair of appendages have not fused entirely. Embryo & has degen-
erated completely in front of the fused fifth pair of appendages, with the exception
of the dorsal organs, which have almost reached the median line.

The sixth pair of appendages have nearly fused at the base, but are distinct at
the tips.

No.1.] VARIATIONS IN LIMULUS POLYPHEMUS. 145

In embryo C the same kind of fusion and degeneration as in embryo Z has
occurred, but it has progressed still farther, for the dorsal organs have fused, and
also the six pairs of appendages. At the central ends of all the embryos are paired
and unpaired ridges, representing abdominal appendages. The dorso-ventral
muscles are well developed, and may be seen radiating from the triangular,
marginal fold to all three embryos.

Fic. 104, X 33. This extraordinary embryo is a triple one like the preceding.
By median fusion and antero-posterior degeneration, each embryo is reduced to a
condition like that in embryo C, Fig. 103. The dorsal organs and the fifth and
sixth thoracic appendages have fused in the median line. Everything anterior
to the fourth or fifth segment has disappeared. There is a dark, central area,
composed of thickened layers of mesoderm and ectoderm, that represents the
anal plates of all three embryos. A series of two or three concentric ridges
encircle it, representing the abdominal segments. Between embryos 4 and B, and
A and C, are seen the concrescing margins of the mesodermic areas, divided into
distinct mesoblastic segments. Nothing of the kind can be seen between embryos
B and C.

The three following embryos belong on the preceding plate, but could not con-
veniently be placed there.

FIG. 105, X 60, not sectioned. A very flat embryo with inconspicuous, mar-
ginal folds. The right chelicera is absent. The embryo is remarkable for the
large size of the mouth, and especially for the series of organs arranged along the
median line behind the mouth, a-d. The first two are transverse depressions, and
in the yolk beneath the second depression, as though arising from it, is a collec-
tion of cells. The third depression is similar to the second, but with a thick lip,
projecting forwards from its posterior border. The fourth is a conical, ectodermic
elevation. The abdominal plate is thrown toward the right. The embryo con-
veys the impression (more strongly than appears in the figure) that it is provided
with a series of five mouth-like openings, arranged along the median line.

Fic. 106, X 60, sectioned. This embryo is in an advanced stage of develop-
ment, but extensive degeneration has taken place, leaving but little of the original
embryo behind. The cephalic lobes are represented by a dark patch of cells,
with a still darker portion, the remnant of the oesophagus, in its centre. In
sections, it appears as a slightly thickened layer of ectoderm, with a much thicker
mass of mesoderm beneath it. The remainder of the embryo, except the terminal
lobe, consists of a single layer of flattened ectoderm cells, with a thin underlying
layer of mesoderm. There is no indication of a nervous system, or other
ectodermic structures. The margins of the mesodermic area, especially on the
side, are thickened, #.a., to form a conspicuous band, composed of several layers
of cells. At the posterior end of the body of the embryo is a median, conical
projection, probably representing the tip of the abdomen, but possibly a fused
pair of thoracic appendages. The appendage rises out of a depression, the
anterior wall of which is thrown forward into a shallow pocket.

Back of this papilla, the margins of the mesodermic area have concresced in
the typical manner for stage D, and in the yolk beneath the concresced margins
is an oval cloud of degenerating cells, Z.a.c. The condition of the posterior end
of the embryo shows clearly, in spite of its apparently simple condition, the ad-
vanced stage of development it has reached. While degeneration seems to be

146 PATTEN.

going on rapidly and extensively in this embryo, there are numerous karyokinetic
figures in both the ectoderm and mesoderm at the posterior end of the body.

Fic. 107, X 60. A very small, circular embryo, with a bare trace of cephalic
lobes and oesophagus, the latter appearing as a white spot with faint bands on
either side, which probably represent the anterior extensions of the nerve-cords.

Two pairs of appendages are present (second and third thoracic ?) and back of
them are traces of other appendages. The abdominal invagination is small, but

quite deep.

2
ioe u
: - :
“6
:
s — :
_ 7 )

-*

«

- 7

PLY

106.

omar

107.

ve

ip

M1.

i Of er i
a) ee Go
. a

S| i i! a Ae

148 PATEEN,

EXPLANATION OF PLATE XI.

The number opposite each figure corresponds to the number of the embryo
that was sectioned. The exponent indicates the position and number of the sec-
tion, and agrees with the number opposite the dotted section line, s, in the surface
views.

The sections are not intended to show histological details, but merely as a
help in the interpretation of the surface views.

Fic. 81,2, X 200. Two longitudinal sections of the brain, optic ganglion, and
invaginated appendage.

Fic. 9°, 5,7. Cross-sections of the cephalic lobes, to show the nature of the
fold that grows over the brain.

Fic. 101,2,8. Longitudinal sections showing the ganglionic fold over the
optic ganglion, and the brain pits.

Fic. 10 4, X 200. Longitudinal, vertical section of an invaginated appendage.
Same section as 101, but shows a part farther back, and more deeply invaginated.

Fic. 111, X 400. Section of a marginal vesicle, showing fatty degeneration of
the mesoderm cells.

Fic. 201, 2,3,4, X roo. Cross-sections, showing absence of nervous system at
the anterior end, and the medianly fused appendages.

Fic. 68 1,2, X 100. Cross-sections, showing invaginated, marginal thickening.
A. Same more highly magnified, to show the bacteria-like dots on the surface.

Fic. 731,2, X 100. Section of a late stage of a degenerate embryo, showing
absence of organs and histological specialization.

Fic. 771,2, X 100. Section of embryo similar to that in Fig. 66. It shows a
remarkably large, horseshoe-shaped body cavity, possibly formed by the fusion of
several somites. The cells lining the floor of the cavity are flattened and mingled
with fibres, so it has the character of connective tissue seen in very old stages.
The somatic mesoderm is on the contrary composed of lymphoid cells, which are
here and there wedged in between the inner ends of the columnar, ectodermic
ones, as though they arose from the latter by inward proliferation.

On each side of the cavity is a thickened cord of mesoderm cells, which seems
to represent the thickened rim of the mesodermic area. Such a relatively
enormous body cavity (?) has not been seen in any other embryos.

Fic. 831,2,8, X 100. Cross-sections of an invaginated embryo, showing ab-
sence of recognizable nerve-cords and appendages.

Fic. 84, X 100. Cross-section of an invaginated embryo, showing the last
traces of the nerve-cords, appendages, and eyes.

Fic. 85 1,2, X too. Cross-sections of invaginated embryo, showing the relation
of the foldings.

Fic. 89. Section of a very old embryo, reduced to a formless mass of
cells. Of the latter, some are multiplying by karyokinesis, others degenerating ;
some are lymphoid, others finally are the very remarkable amoeboid cells, con-
taining coiled fibres. These cells are found in great numbers in the marginal
thickenings of the mesodermic area of very old embryos, and are also found in
great numbers in the adult. &. Two fibre cells enlarged, showing the fibre
lengthwise and in optical cross-section.

ee

ss : ‘

a . 7 -
¥ 7 « os . a i td
7
. A
t
\ ae
i ‘ , +
“ a
7 ~ < diy
t
t 1 & -
e
‘ .*
we i ®
eS ' 4 é
\ = ’ .
‘
- 7
: ot
oan -
* &
'
- \ > uJ » a
= xs
_ 1
2 a 7
*
| : . ia
y

v
- he 7
7 '
® +
; ,
Y
> *
# £ ' '
1 ¥
— a
— . - ee
F °
‘ -
: —
¢ 7
“~.
: c
,
%
i
= a 7
i '
eed
= =
—
“
we
t
aoa od

q

POS eoore :

so]

sii

~axdc

Tah Anshx Werner shiner Frankfurt 9H.

BUDDING IN COMPOUND ASCIDIANS, BASED ON
STUDIES. ON GOODSIRIA, AND PEROPHORA.

W. E. RITTER.

CONTENTS.
PAGE
cA cee ©) @) 1) STMRSTVAS ST) UREA © 72.02) 55 79ers ee ee ae meer tages ee Pree ae ga 149
Tere NUATHERD ATC @ RC EUNTOU irc core see reese ene en ek ee eet 150
2:2 DESCRIPTION OF THE! SPECURS ese ee ae ee ee eee ISI
Bx LOOMS IN) (REE (COL ON West cece sateen cree eco cess eee 158
4. DEVELOPMENT OF THE UNDIFFERENTIATED BUD.........-2..20:::-200--- 163
IK. QOEVELOPMENT OF THE ORGANS 2. oc 22020stjoc2.ceshastteccenmetasiacnagne stesso 165
a. Lhe Branchial and Peribranchzal SACS pen. 165
bs) tke, Disesieve: Uract and is Appendage sues sa ee 171
Cle LIE MP CUACAT ALUNL MAILE: SLL EL TL Ue crere te Deee ree ee 174
Lhe Ly pophy sts ard UG Aig TtO 1 res eae eee eee eee ee 176
(a)) Lhe Gangelo-hypophiyseal Duchenne 176
(BY OLDE GOZO oes ces esos ee 178
Cnr Lee: GEretal Sy SCCM <8 eh ewen teens etiag Suet aes soe eee ee 184
Ba ER © PEO ReAG JASN INGE © GIN Soe eee eeyrecse ae besten eee eee 187
f.) INTRODUCTORY. MIATERDAIS TR CHINTQUWeteccteccrerteree ee eeseareee ee LOT,
25 RELATION OF THE BUD) TO) GHEE) STOLON cesses) eee eee 190
3: (PERICARDIUM AND) PLEAR@):... fisccs.ccsccs.cnctacestes.00ecuseetsaersneeeces See ee 195
Ant VELWPO REY SIS IA NID GAN GILIO Nit eseie: sotaercbeeee eerie teens eee eee OD
C. GENERAL CONCLUSIONS AND REFLECTIONS =...........4..-2+. 209
I. APPLICATION OF GENERAL DEVELOPMENTAL PRINCIPLES TO
ZAIN) INDE RIP RURAD TOI (OB eT) HB) ACTS asa oases cee eee 209
2. ON THE QUESTION OF DIFFERENT TYPES OF BUDDING AMONG
ARUN T GATES 4225 22 15 steed dese sessdesite pon test a cacen sane tere ee ee 213

3. THE RELATIONSHIP BETWEEN THE POLYSTYELLIDAE AND THE
BOTRYLLIDAE (see 6, under Summary of Results in General).

PD SUMIME AGRO) | OUR REBT SU ME TS Seesaw east ec cenc ape pec eeeence eee ase ne 220

INTRODUCTION.

THE material that has served as the basis of the paper here
presented was collected by myself at various points on the
California Coast during the years 1892 and 1893. A con-
siderable part of the investigation has been carried on under

150 RITTER. [Vor. XII.

circumstances that have placed me under obligations to several
persons and institutions. These I would here acknowledge,
and that most gratefully: first of all to Mr. Alexander
Agassiz, through whose generosity it was that I was enabled
to occupy a table at the Zodlogical Station at Naples for
some months during the fall and winter of 1894; next, to
Prof. Dohrn and all those associated with him in making
the Naples Station the realized ideal of what such an institu-
tion should be; and last, but by no means least, to Prof.
F. E. Schulze, who so kindly and generously placed the
excellent facilities of the Zodlogical Institution of the Uni-
versity of Berlin at my service for some months.

A. GOODSIRIA DURA NOV. SP.
I. MATERIAL. TECHNIQUE.

My specimens were all collected by myself at Santa Barbara,
California, during a brief visit there in the last days of Decem-
ber, 1892.

They were all found upon the beach where they had been
thrown by the waves; I saw no colonies in their original posi-
tions. As they were found in abundance and in a perfectly
fresh condition, I conclude the species must be plentiful at
this point, and that it lives on the sea bottom not far from
the shore. This latter supposition is likewise supported by
the fact that most of the colonies are attached either to
shore-inhabiting seaweeds, or to Szyela rubra, an Ascidian very
common at Santa Barbara on the piles of the wharf, and on the
rocks in shallow water. It is almost certain that the dredge,
when brought into use here, will bring it up from its natural
abode in quantities. On account of its obvious abundance at
this point, I have been somewhat surprised at not finding it
elsewhere on our coast; but there is little doubt that further
work with dredge and tackle will bring it to light at other
places.

As the only killing reagents with which I was provided on
my visit to Santa Barbara were picro-sulphuric mixture and

No.1.] BUDDING IN GOODSIRIA AND PEROPHORA. I51

alcohol, I was limited to these for the preservation of my
specimens. Luckily, however, the former proved to be a
favorable medium for the purpose. Some of my material has
proved to be excellently well preserved, better than any I have
been able to get of some other Ascidians by using a large
variety of reagents. Some specimens preserved in absolute
alcohol were found to be valuable as collateral material.

No opportunity has been afforded me for studying living
specimens with any detail, but they are much too dense and
opaque to permit of being very satisfactorily studied in this
condition. Likewise, since the species belongs to that cate-
gory of Ascidians in which the mantle clings closely to the
test, it is impossible to free the zooids from the colony in
preserved specimens so as to study them whole with much
satisfaction. One is consequently obliged to depend on the
study of small colonies, or small pieces of colonies cleared in
oil; on dissections; and on thin sections. The first mentioned
method is particularly valuable.

Of the numerous stains employed I have found Paul Mayer’s
Hzemalum, and Grenacher’s Alum Carmine to give the most
satisfactory results.

2. DESCRIPTION OF THE SPECIES.

As we are dealing with a new form, it will be necessary to
describe it; and this is the more incumbent because it belongs
to a group of Ascidians not very well known to science.

As mentioned in my preliminary paper ('94), it belongs to the
Polystyelidae, a family founded by Herdman (86). The author
has given a good historical résumé of our knowledge of the
group, and I may consequently touch this phase of the subject
lightly. The six genera of which the family is at present
composed were, with one exception, all established between
1843 and the time of Herdman’s description of the Challenger
collection in 1886. This one exception was added by the
author himself. They were described by different writers and
assigned by them to different larger groups already known;
Carus (43), for example, regarding the genus described by him

1.52 RITTER. [Vo.t. XII.

as being allied to Clavelina; while Giard ('74) considered the
two founded by him as related to the Cynthizdae. Most of
the authors believed the particular forms which they studied
were true compound Ascidians, though none of them were
able to produce the crucial test of this, vzz., evidence of the
occurrence of multiplication by gemmation. This evidence
remained wanting till the publication of my recent pre-
liminary.

By personally studying as many of the known species as
were accessible to him, and by critical examination of the
descriptions and figures given by other writers, Herdman has
amply justified his uniting the genera into a single new family.
He has also contributed many valuable observations on the
species described by him.

Since the work of this author appeared, there has been, so
far as I know, but one other contribution to our knowledge of
the group. This consists in the addition by Gottschaldt (94)
of a new species, Goodsiria borealis. ‘To this, further reference
will have to be made because of its close resemblance to the
species now under consideration.

The following is the diagnosis of the new species as my
present knowledge enables me to give it:

General Character of the Colonies. Predominating form flat
and encrusting. Occurs most frequently on various sea-weeds,
and on Styela rubra, and often so completely covers these that
the outlines of the colonies are determined by those of the
particular substrata. But in addition to the flat and encrusting
condition, colonies not infrequently occur with fleshy knobs
composed of a large mass of test material containing zooids in
the entire surface layer but none in the centre.

Colonies from icm. to § or 6 cm. in diameter, apparently
tending to expand equally in all directions when permitted to
do so by form of substratum.

Color in living state, dull red; light brown in preserved
specimens.

Zootds. Fully imbedded in the test, not projecting from sur-
face of colony; not arranged in systems, no common atrial
orifices ; for the most part evenly distributed in surface layer

No.1.] BUDDING IN GOODSIRIA AND PEROPHORA. 153

of test and rather close together. Size of zooids, length 3 to 5
mm., width 2 to 3 mm.

Test. Rather dense, opaque, cells small and not numerous,
no bladder cells, slightly fibrillated. Vessels numerous, much
branched and anastamosing, terminating, particularly in mar-
gins of colony, in many large pear-shaped ampullae; mostly
occupying a deeper position in test than the zooids.

Musculature. Not highly developed, fibres not arranged in
distinct bands.

Lranchtal Apparatus. Position of branchial and atrial orifices
vary with the character of the colony; when the colony is thin
and encrusting they are both placed more dorsally, 2.2. opposite
the endostyle; when the colony is fleshy and massive, the
orifices are more anterior and nearer together. No distinct
siphons; orifices not lobed, at least not discoverably so in
preserved specimens; in some specimens orifices obscurely
quadrilateral.

Branchial tentacles simple, usually twenty long and strong
ones, and about same number of smaller ones alternating with
them. Atrial tentacles present, about twenty in number,
much smaller than the branchial. Branchial sac without folds;
internal longitudinal vessels rather prominent, 5 on each side,
the two dorsal ones on each side nearer together than the
others. Small intermediate transverse vessels frequently pres-
ent. About 12 series of stigmata; 5 or 6 stigmata between
each two longitudinal vessels excepting between the two on
each side which are closer together, where there are only two
or three. Dorsal lamina a plane membrane tending to roll up.

Endocarps. Present in the form of large globular structures
attached to the parietal wall of the peribranchial sac, from
which they project prominently into the peribranchial space.

Digestive Tract. Situated on left side of branchial sac,
distinctly divided into oesophagus, stomach, and intestine.
Oesophagus nearly as broad as intestine, and approaching the
stomach in length. Stomach somewhat pear-shaped, about
8 deep folds extending lengthwise of the organ, parallel with
one another, not converging toward the point of entrance of
the oesophagus. Course of intestine, first ventralward from

154 RITTER. [Vou. XII.

stomach for a short distance, then changing by a rather wide
curve to a nearly dorsal direction which is followed for a
distance about equal to the combined length of oesophagus,
stomach, and ventrally directed limb of intestine. Anus
directed somewhat anteriorly, nearly as broad as any part of
intestine, provided with a thick lip. A peculiar thickened
band in the wall of the intestine, beginning in the loop, extends
in an oblique direction half way to the anus. Lacteal system
consists of a prominent coecum projecting from the stomach
near its pyloric orifice, into one side of which opens a much
smaller tube, the common stem of the greatly branched in-
testinal part of the system.

Sexual Organs. Both ovaries and testes in the form of “ poly-
carps” attached to the mantle on each side of the endostyle,
and projecting into the peribranchial chamber. These few in
number and small in size so far as known. Uncertain whether
the same zooids produce both ova and sperm or not.

Hypophysis. A simple duct, no branched glandular part,
the mouth a simple elliptical opening.

Ganglion. Situated ventrally to the hypophyseal duct, con-
sisting of the usual outer layer of multi- and uni-polar ganglion
cells, and an inner cell-less core of nerve fibres.

Budding. Pallial, z.e. from the parietal wall of the peri-
branchial sac.

I have hesitated much in deciding in which of the two genera,
Goodsiria or Synstyela, this species ought to be placed. The
only well-marked distinction between them appears to be in
the character of the colony, this being designated as massive
in the first, and thin and encrusting in the second. Certainly
the latter characterization applies to the greater number of
colonies of G. dura which I have seen, but at the same time it
does not apply to all in all their parts. It will surely be allowed
by all zodlogists familiar with the compound Ascidians that
the form of the colony alone is a rather frail peg on which to
hang a genus. It appears to me that the difference between a
pedunculated and a massive colony is at least as great as that
between a massive and an incrusting one. If the latter differ-
ence is worthy of being used to separate species into genera,

No.1.] BUDDING IN GOODSIRIA AND PEROPHORA. 155

the former ought to be likewise; but two of the Challenger
species of Goodszria described by Herdman, vzz., G. pedunculata
and G. placenta, are decidedly pedunculated, while this char-
racter is as markedly absent from the other known species.
As we shall see presently, if the structure of the zzdvidual
zootds and not the form of the colonies is made the basis of
comparison, the closest ally to our present species is certainly
found in the genus Goodszria, instead of in the genus Syxstyela.

I have consequently reached the conclusion that while, so
far as morphological agreements and disagreements are con-
cerned, my species might be placed with about equal propriety
in either genus, since Goodsiria is the older of the two (it
having been founded by Cunningham in 1872, while Syxstyela
was founded by Giard in 1874) it is more entitled to receive
the new comer.

The specific name dura I have chosen as applying to the
firmness of the colonies due to the density of the testicular
matter.

There are three other known species to which this one is
closely allied, and with which, consequently it must be com-
pared in some detail in order that its differentiating characters
may be brought out.

These are: Goodsiria coccinea, Cunningham, G. borealis,
Gottschaldt, and Syzstyela incrustans, Herdman. With the
first mentioned species it agrees not only in the numerous
points of structure common to all species of the family, but
also in the form and size of the zooids, points of considerable
determinative value for species in this group; but most im-
portant of all, 2 the absence of folds in the branchial sac. So
exceptional is this condition among the representatives of the
family that its occurrence in two species rather closely related
in many other particulars may be regarded as evidence of
their very close relationship. But that these two forms are
specifically distinct there can be nodoubt. The differentiating
characters are the following. The colonies are much larger
and much more distinctly massive in G. coccinea than in G.
dura. According to Herdman (86), on whose description of
the former I base my comparison, the colonies of this species

156 RITTER. [Vou. XII.

sometimes reach a length of 46 cm., while I have never seen a
colony of G. dura more than 10 cm. in length,

The branchial and atrial apertures of cocc¢nea are conspicuous
and are irregularly four-lobed; in dura they are not conspic-
uous, and in preserved specimens show no trace of lobes. In G.
coccinea the “‘meshes”’ of the branchial sac, z.e. the areas of the
sac between the internal longitudinal vessels, contain eight
stigmata. In G.dura the number is not the same in all the
meshes, but at the most is less than eight; at the least it is
three. But probably the best distinction between the two
species is in the character of the stomach. In G. coccznea this
organ, as described by Herdman, is globular in form, and con-
cerning its folds he says: ‘There are usually six well-marked
folds upon the right side of the stomach. A transverse section
shows in addition a single large fold, which projects far into
the centre, nearly dividing it into two cavities.” Reference to
my Figs. 4 and 6, Pl. XII, shows at once that the stomach of
G. dura is quite different from this. In the first place it is
not globular. It is rather schizaster-shaped, if I may be per-
mitted to suppose that the form of this echinoid is any more
familiar than that of the stomach which I am comparing with
it; but I can think of no other object which it so closely
resembles in form as it does some species of this genus. In
the G.dura stomach there is no one fold that exceeds the
others in the extent to which it projects into the chamber.

Another distinction between the two species, that would
appear to be constant, consists in the presence in G. coccinea
of a vessel which, ‘enclosed in a prolongation of the mantle,
leaves the posterior end of each Ascidiozooid and runs for a
longer or shorter distance through the test before ending in a
dilated bulb.”’ No such vessel occurs in G. dura.

Although it is evident from Gottschaldt’s description of
G. borealis that this species and G. dura are closely related, it
is unfortunately impossible to make as exact a comparison
between them as is desirable, owing to the incompleteness of
the author’s description at one or two critical points. In the
form of the colony they appear to agree more closely than
either agrees with G. coccinea, for the figure representing a

a

No.1.] BUDDING IN GOODSIRFA AND PEROPHORA. 157

colony of G. borealis, shows it to be quite “ flat and encrusting,”
though the author speaks of it as being fleshy. The points in
the structure of G. borealis which preclude the placing of our
California form in that species are these: The four-lobed
branchial and atrial orifices; the eight series of stigmata
(G. dura has twelve); and the presence of folds in the branchial
sac. It is stated by Gottschaldt that “zarte Muskelfibrillen”’
occur in the test; but almost certainly this is an error, the
“ Fibrillen’’ which he supposes to be muscle fibres, being
similar to those found in the test of many other Ascidians, but
which are in all probability correctly considered as not con-
tractile, but mere filamentous differentiations from the matrix
of the test. The author also says that the “liver gland” is
well developed; but he has in all probability mistaken the
lacteal system for a liver, since nothing corresponding to what
is generally understood to be the liver in many simple Ascid-
ians is present in at least three closely related species which I
have examined with special reference to this point. These
species are Goodsiria coccinea, G.dura, and an undescribed
Australian species of Chorizocormus kindly furnished me by
Prof. Herdman.

I have regarded Syustyela incrustans and Goodsiria dura as
closely related, more on account of resemblances in the form
and character of the entire colonies than from similarities in
the structure of the zooids. In the former particular, they
undoubtedly agree rather closely, but in the latter they are, on
the whole, as previously remarked, less closely alike than are
G. coccinea and G. dura. The colonies of Syzstyela tncrustans
are said by Herdman to reach a length of 20cm. in some
cases, while it will be remembered that no colony of G. dura
has been found more than half that size. The individual
zooids are also much larger in the former than in the latter
species, their length being as great as 8 mm. in the one, and
never more than 5 mm. in the other.

The branchial and atrial apertures of S.zucrustans are
described as ‘conspicuous, but not distinctly lobed.” We
have seen that in G. dura they are not conspicuous, nor are
the lobes recognizable at all in preserved specimens, so that in

158 LEI OIM Ee [Vor. XII.

this particular the last named species agrees about as well
and about as little with G. coccinea as with S.zncrustans. But
it is when we regard the branchial sacs of the two species now
being compared that we find their most important differences.
The sac of Syzstyela incrustans has a rudimentary fold on each
side, while it will be remembered that no folds are present in
G. dura. Unfortunately, Herdman does not tell us the number
of internal longitudinal vessels in the sac of either G. coccinea
or S.zucrustans; but he has shown eight in a figure of a por-
tion of the sac of the latter, so we are certain that the number
here exceeds the number in G. dura by at least three; of
course the excess may be greater, since it is not certain that
the eight shown in the figure referred to is the entire number
present; in fact it is more likely not to be, since in making
such preparations of the sacs as that from which this figure
was drawn one does not usually obtain intact the entire width
of one side.

In S.zucrustans there are intermediate vessels, ‘normally
three in number,” crossing the meshes of the branchial sac.
In G.duva I have never seen more than one of these, and
frequently this one is absent.

There also appears to be a difference between the two
species in the arrangement of the polycarps. In the former
they are figured by Herdman as scattered promiscuously over
a portion, at least, of the wall of the peribranchial cavity, while
in the latter, so far as I have found them present, they are
confined to a single row on each side of the endostyle. But
great weight cannot be attached to this distinction, since the
sexual organs are too poorly developed in all my specimens of
G. dura to enable one to decide what their arrangement might
be in a better-developed condition.

3. THE Zooips IN THE COLONY.

This subject is so intimately connected with that of bud-
development that it must be treated somewhat more fully than
have been the other points of adult structure. I cannot dis-
cover any constant arrangement of the Ascidiozooids in the

No.1.] BUDDING IN GOODSIRIA AND PEROPHORA. 159

colonies. Fig. 1, Pl. XII, represents a colony, natural size, grow-
ing on a laminaria leaf. As here seen, the zooids are, as com-
pared with many compound Ascidians, rather distant from one
another. They are very regularly distributed, there being no
suggestion of systems. Neither have the zooids any constant
relation either to the colony as a whole or to one another, with
reference to their antero-posterior axes; though in general
where the colonies are narrow and elongated the longer axes
of the zooids correspond to the longer axes of the colonies.
The arrangement of the adult zooids is of course determined
by the relations which the developing buds hold to their
parents, and in this there appears to be no constancy beyond
the fact that the buds are generally confined to the borders of
the colonies. I say generally this is the case; but it is not
without exception, for in several instances I have found young
blastozooids so situated that older ones intervened between
them and the edge of the colony. (Fig. 2, Pl. XII.)

The question of the relation of the buds to the older zooids
is important because it obviously involves the questions whether
all the zooids of a colony are capable of producing buds, and
at what age in the life of the zooids their buds are produced.
And this last question leads back again to the fundamental
one of whether or not the cells that initiate the bud-develop-
ment are derived as unmodified or embryonal cells from the
parent, the grandparent, and so on to the sexually produced
embryo that was the common ancestor of the colony. What
I have to say on this point will be better reserved until I
speak in detail of the Axlage of the bud, my purpose here
being to describe merely the form and composition of the
colony as a whole. In a few instances (Fig. 2, buds a and 8),
for example, there is a suggestion that at the borders of the
colonies the younger buds occupy positions alternating with
the next older ones, but a little in advance of them toward the
edge of the colonies. If, however, such a law of arrange-
ment exists, it prevails less frequently than do the exceptions
to it.

To Metchnikoff’s (69) denial that buds are produced by the
vessels in Botryllus, Giard ('72) raises the objection that if this

160 LATTER. [VoL. XII.

were true it would be impossible to explain ‘la production
d’etoiles multiples et distantes dans le cormus d’un Botryllien.”
The remoteness of the young buds from any older zooids in
Goodsiria has likewise frequently proved a stumbling-block to
me in seeing how they could in such cases have been produced
in the usual way, z.c. from the wall of the peribranchial sac.
But I have given much attention to the point, and am quite
convinced that in reality this is their only source. Herdman
(g6) expresses the opinion that the ampullae of the vessels will
be found to give origin to the buds, but such is apparently
not the case here any more than it is in Botryllus.

The frequent remoteness of the buds from their parents
must be due to their having grown away from the latter before
they become fully severed from them. From the firmness of
the test and the character of the young buds I can hardly
believe that they have any power of independent migration
throughout the test.

This conjecture, that the remoteness of the buds from their
parents is due to the growth of the bud before it becomes
severed, harmonizes with the view that pallial budding as it
takes place in Botryllus and Goodsiria is not of necessity
fundamentally different from the stolonic method of budding.
The comparison between these two methods becomes still
more interesting from the discovery that in Pevophora the
buds are connected with the septum of the stolon by their
peribranchial instead of their branchial sacs. But I shall
discuss this point further after having described in detail the
development of the bud in each species.

There is certainly no septum in the vessels of the test,
(Bigs. 3, 5.7, ete. oki. XU ec:ves.); and at budsawerestojbe
produced in connection with them, it would have to be by a
method essentially different from any of the generally recog-
nized types of gemmation among Tunicates. Since the blood
of the zooids passes freely into the testicular vessels, and since
the young sexual cells float in the blood, it is not beyond the
range of possibility that these cells may pass into the ampullae
of the vessels, and there form themselves into the “inner
vesicle,” the all important precursor of the blastozooid.

|
'
‘
|

No.1.) BUDDING IN GOODSIRIA AND PEROPHORA. 161

It is an interesting fact that Herdman ('86, p. 90 e¢ seq.) has
observed a process in Colella pedunculata which would appear
to be a realization of that here imagined. As the author
himself says, his observations are rather fragmentary, and
consequently his account is much less full than we might wish
it £6; be:

That the buds arise in this anomalous manner in this
genus, he, however, seems to be convinced, and his descrip-
tion and figures undoubtedly furnish good ground for his
conviction. It is certainly very much to be hoped that oppor-
tunity will before long be afforded some zodlogist to study
the subject more fully. In a few instances I have found a
massing of cells within the ampullae that is strongly sugges-
tive of the process described by Herdman. Minute examination
of these aggregates has, however, failed to furnish any evidence
that they produce buds. Two such cases are shown in Fig. 7,
Pl. XII. They are quite conspicuous when cleared in oil and
examined with a low magnification.

The vessels generally occupy a deeper position in the test
than do the zooids, so that the test surrounding the zooids is
less thickly penetrated with them than are its deeper portions.
This is shown in Fig. 5, Pl. XII.

It has been mentioned above that the buds become fully
severed from the parent zooids at an early stage in development.
This fact raises the question of the extent to which the zooids
in this species are independent of one another. It is certain
that many of them are in connection with the vessels of the
test for at least a portion of their lives, and are consequently
in vital connection with one another. But this connection is
almost certainly secondary, and, I believe, not essential to the
development of the zooids.

As the young buds become larger, they press more and more
closely against the vessels that were in contact with them, or
nearly so, at the beginning, and by this pressure a fusion of
the vessel wall with the outer or ectodermal wall of the bud
is produced, and then later a perforation of the fused walls
occurs, and the lumen of the vessel is thus brought into com-
munication with the body space of the zooid.

162 RITTER. [Vou. XII.

Figs. 37 and 38, Pl. XV, taken from two different zooids,
represent the points at which the vessels open into the body
space of the zooids. I would call particular attention to
Fig. 37, since this illustrates a much more common appearance
than that shown in Fig. 38. The projections, ec.ves', into the
cavity I understand to be due to the vessel’s having pushed
itself therein, either by its own growth or by its passive resist-
ance to the expansion of the developing zooid. In many cases
it appears that the wall of the zooid has been pushed in toa
considerable extent by the vessel ; and by an interfolding of the
vessel wall and the zooid wall, the character of the connection
has become quite complicated. In some instances these pro-
jections extend entirely across the body space to the wall of
the peribranchial sac.

I said above that it 1s almost certain that these communica-
tions are secondary. ‘This qualification was made because one
might ask if the vessels do not sprout out from each of the
blastozooids, as it would seem they must have done from the
sexually produced ancestor of the colony.

That this process never occurs I cannot positively affirm,
but I have no evidence that it does, and two considerations
incline me to the belief that it does not. In the first place,
the peculiar character of the connection, as described and
shown in Fig. 37, seems to me to speak against such a process ;
and in the second place, one would suppose that if it takes
place at all it would do so at a rather early stage in the devel-
opment of the zooid, before the ectodermal cells in the region
from which the vessels would have to be produced had undergone
modification. But I have searched in vain for cases of such
formation.

By carefully examining all the sections of numerous buds at
various stages in development, one can determine that there
are no connections whatever between them and the vessels.
This is the fact that makes me believe that the connection of
the buds to the vessels is not essential to the development and
life of the zooids.

No.1.] BUDDING IN GOODSIRIA AND PEROPHORA. 163

4. DEVELOPMENT OF THE UNDIFFERENTIATED Bup.

The earliest stage which I have observed in the development
of a bud is represented by Fig. 11, Pl. XII. Here the bud Axlage
is clearly marked, dd¢.a., both by the character of the cells and
by the slight evagination already seen. But as yet the ecto-
derm shows no indication of having been affected by the
process which has begun in the wall of the peribranchial sac
beneath it. It is true its cells are higher immediately over
the evagination than they are at any point more ventralward ;
but they are not higher than they are from this point upward
toward the dorsal side of the zooid, as the figure shows.
That is to say, the ectoderm cells are not different over the
bud from what they are in a corresponding position of a zooid
in which no bud is developing.

The figure shows approximately the stage of development of
the parent zooid. The endostyle, ezd., is not yet fully differen-
tiated, histologically, and the thickened places, s¢.a., in the wall
of the peribranchial sac opposite the bud show where branchial
stigmata are going to form. The string of cells, ev.c., near the
bud, is a fragment of one of the ‘“endocarps”’ that has been
broken and displaced somewhat in the section cutting. It
consequently has no significance.

To the important question of whether the bud Ax/age arises
from a “budding zone” on the peribranchial wall in which
the cells are “embryonal,” and are endowed from their very
origin with peculiar bud-producing powers, my observations do
not enable me to give a wholly satisfactory answer. However,
certain facts are suggestive in this connection. I have never
found anything that appears like a “ budding zone”’ similar to
that described by Oka, for example, in Botryllus, and I find
that the walls of the peribranchial sacs in the developing bud
are, after the very first stages of their growth, very thin through-
out, the cells being considerably flattened. Their development
is, of necessity, accompanied by cell division, but I am unable
to discover that this is more noticeable in one region than in
another; or that the cells have a different appearance or
distribution in one locality from what they have in another.

164 RITTER. [ViGL.X ie

One of two conclusions seems to be inevitable from the facts
observed: Ezther there are no budding zones containing specially
endowed cells; or a large majority of zootds are incapable of
producing buds.

It is true that a comparatively small number of zooids with
buds actually in connection with them have been found; but I
suppose this to be due generally to the early stage at which
the buds are severed from their parents, rather than to the
incapacity of the zooids to produce buds. But I would by no
means assert that all zooids are capable of asexual reproduc-
tion. Later stages in the development of buds, though before
severance from their parents, are shown in Figs. gand 10, Pl. XII.
and Figs. 12-15, Pl. XIII. Fig. 9, Pl. XII, represents a zooid
with its bud, seen as a transparent object, but not sufficiently
cleared to render distinctly visible the several organs and
layers. On the other hand, Fig. 10 was drawn from a speci-
men well cleared in cedar oil, and consequently so transparent
as to make the optical section at the level represented almost
as distinct in its various parts as an actual section in the same
plane would be. These figures explain themselves sufficiently.
The position of the buds far forward on the zooids will be
noticed in all these cases. This appears to be their usual
position. A transverse section of a bud of a stage about
corresponding to that shown in Fig. 10 is presented in Fig.
14. This represents the tenth section from the tip of the
bud. The thickness and irregularity of the inner vesicle are
here conspicuous. The sections of the same series near the
parent zooid show the inner vesicle to be here considerably
thinner than it is in the section figured. The difference in
thickness in the two regions is, I suppose, due to the fact that
growth is taking place chiefly toward the end of the bud. The
thinner, basal part is where, a little later, the endoderm will
become disorganized in course of the cutting off of the bud
from the parent.

The question may arise here if the unequal thickness of the
endoderm at various points, as shown in Fig. 10, is not an
indication that the differentiation of the organs has begun at
this stage.

SE ee eee ee

i

No.1.] BUDDING IN GOODSIRIA AND PEROPHORA. 165

This does not seem to be the case. At any rate, as we shall
presently see, at a still more advanced stage, z.e. after the bud
has become fully severed, the wall of the inner vesicle is, in
some buds at least, still entirely undifferentiated.

Fig. 15, Pl. XIII, shows a section of a bud that is very nearly
severed from its parent; in fact, the severance is practically
complete, there being a mere neck of scattered cells, v.c., mark-
ing the former connection between bud and parent. It is here
seen that the ectoderm of the zooid is not fully closed together
at the point where the bud was cut away. Here the inner
vesicle is wholly undifferentiated, it presenting in every section
the same appearance as that shown in the one figured ; and it
should be said that buds occur in which every trace of the
connecting strand has disappeared, but in which the inner
vesicle still retains its simple, unmodified condition. However,
it does not appear to remain long in this state.

In describing the further growth of the bud it will be best
to follow the course of development of each organ separately,
but before beginning the description I wish to call attention to
the interesting fact that 2” different buds the order of appear-
ance and stages in development of the several organs are subject
to considerable variation with reference to one another. I will
point out specific instances of this as I proceed with the
description.

5. DEVELOPMENT OF THE ORGANS.

a. The Branchial and Peribranchial Sacs.

I have been unable to make my observations on entire buds
and on sections agree exactly regarding the initial step in
the formation of these structures. By examining whole buds
the impression is gained that the process is begun by the
growth of two folds which will ultimately form partitions
separating the primitive simple vesicle into three portions — two
lateral ones, the peribranchial sacs, and a middle one, the
branchial sac. “These folds are seen:at pf, Fig: 17, Pl. XIII,
in which their formation is already advanced to a considerable
extent; but in Tig. 16 the slight angles marked by the same

166 RITTER. [Vox. XII.

letters are the very beginnings of the folds. In neither of
these figures, nor in fact in any of the numerous similar ones
that might be given, does it appear as if the peribranchial sacs
begin as evaginations from the primitive vesicle. It does not
appear here as if the peribranchial sacs are initiated by the
active growth of the portions of the primitive vesicle walls
which are to enter into them, but, as already said, by the
growth of the partitioning folds. On the other hand, sections
of a very early stage in the development of the sacs show that
in one instance, at least, there takes place an active growth of
the Anlage of the sacs themselves, resulting in well-defined
evaginations. (See Figs. 20 and 21, 67s.a., Pl. XIII.) These
figures are drawn from sections of a bud presenting an earlier
stage in the development of the sacs than that represented by
Fig. 16, and so it might be assumed that the sacs of the latter
individual had their beginnings in such evaginations as those
shown in Figs. 20 and 21, even though all indications of them
are by this stage completely lost.?

But if I am right in supposing the angles #.f. in Fig. 16 are
the beginnings of the folds, it is not quite clear how such a
condition as that presented by this bud would be related to
one in which the well-defined evaginations of Figs. 20 and 21
were present. The only way of harmonizing the two condi-
tions, so far as I can see, is to suppose that the sacs begin
with the evaginations, and that these grow rapidly in width,
but very little in depth, till they extend over nearly the whole
of the sides of the primitive vesicle. Posteriorly they extend
ultimately so far as to be separated only by the angle in which
the intestine (zz¢., Fig. 16) develops, and anteriorly the angles
are almost obliterated, but the sacs do not approach each other
so closely as behind, but remain separated by the thickened
area (end., Figs. 16 and 17), which will give rise to the endostyle.
Thus we should have the condition presented by Fig. 16 asa
more advanced stage of development derived directly from the
earlier distinctly evaginated stage. And the fact that the
hypophyseal duct and intestine are both begun in the bud

1] have since seen in a whole bud an evagination from the inner vesicle, at
least on one side, that in all probability corresponds to 47.5.4. in Figs. 20 and 21.

>.

Novi) (bCDDING IN GOODSIRIA’ AND PEROPHORA. 167

represented by Fig. 16, while neither is yet indicated in the
bud with the peribranchial evaginations, also suggests that the
former is older than the latter. But the discrepancy between
the two conditions may be explained in another way; it may
be due to individual variation ; in any case, however, it cannot
be a matter of fundamental importance.

The initial steps in the formation of the peribranchial sacs
accomplished, the remainder of their development is followed
without difficulty. The sickle-shaped partitioning folds, .f,
Fig. 17, extend their arms, one on the dorsal, the other on the
ventral side of the inner vesicle, farther and farther backward.
But of the two arms of each fold, the ventral one takes by far
the more important part in effecting the ultimate complete
separation of the peribranchial from the branchial sacs.

The process is clearly illustrated by Figs. 16-19, Pl. XIII,
representing dorsal views of whole buds examined as transpar-
ent objects; and by Figs. 22-24, Pl. XIII, and Figs. 26-20, PI.
XIV, inclusive. The series of transverse sections represented
by Figs. 22-24 are from a bud in a stage of development
slightly more advanced than that shown in Fig. 17, Pl. XIII.
Fig. 22 is the most anterior of the three sections drawn, and,
as will be seen, passes through the point at which the bran-
chial siphon, 67.52%., is being formed. Fig. 23 is fourteen
sections farther back. It shows both the dorsal and ventral
arms of the sickle-shaped folds, @.f. and v.f., those of the right
side (left side of the figure) very nearly meeting each other,
and so making the separation of the right branchial sac at this
point nearly complete. Fig. 24 represents a section farther
back, and shows at this position merely a trace of the folds,
the three sacs of the anterior region being here merged into
one common cavity representing the unmodified remnant of the
primitive inner vesicle. Figs. 26—29, Pl. XIV, represent similar
sections of a still older bud. That shown by Fig. 26 is most
anterior, and passes through the opening of the hypophyseal
duct into the branchial sac, Zy.m., and hence is slightly farther
back than the section shown in Fig. 22 of the preceding series.
Fig. 28 represents a section sixteen sections farther back. It
is seen that the dorsal folds take no part in the formation of

168 RITTER. [Vox. XII.

the peribranchial sacs in this region. Eight sections farther
back, Fig. 29, the ventral folds no longer appear, and we have
again an unpartitioned portion of the primitive vesicle.

From the figures of the whole buds, and also from those of
the series of sections, it will be seen that the folds do not
extend back parallel to each other, neither do they occupy a
vertical position. They converge both posteriorly and ven-
trally, so that the middle or branchial sac is made cone-shaped,
the apex of the cone being directed backwards. This becomes
still more distinct at a later stage in development (see Figs.
18 and 19, Pl. XIII). An inclination toward the left side of
the posterior extremity of the cone is now noticeable, and this
becomes more pronounced at a later time (compare Figures
mentioned). In some buds it is much more prominent than
in others.

We may now pass to the practically completed condition of
the branchial apparatus. A dorsal view of an entire bud in
such a stage is shown in Fig. 19. The only changes that will
take place between this and the fully adult condition will be
some unimportant ones, so far as our present purpose is con-
cerned, in the relations of some of the parts to one another,
and a great increase in size of all the parts. From the prece-
ding stage the most important changes have perhaps been those
taking place in the posterior region of the animal.

Recurrence to Figs. 23 and 24, Pl. XIII, will recall the fact
that at this stage the three sacs opened widely into a common
chamber in the posterior part of the bud. In the stage we are
now considering such is no longer the case. The two ventral
folds by growing up to and fusing with the dorsal wall of the
primitive vesicle for some distance, and then by fusing with each
other but xot with the dorsal wall, for the rest of the way back
to their extremities have effected the complete separation of
the branchial sac from the peribranchial sacs. These relations
will be easily understood by comparing Fig. 24 of the preceding
stage with Fig. 28 of the present one, and then Figs. 30 and
31, Pl. XIV, with Fig. 28. Figs. 30 and 31 are from a bud
considerably more advanced than the one represented by
Figs. 26-29. Fig. 31 shows at drs. the posterior tip of the

.
{
{
:
q
;

No.1.] BUDDING IN GOODSIRIA AND PEROPHORA. 169

branchial sac. One or two sections farther back this cavity
disappears, and to the very last is wholly shut off from the
large single surrounding cavity marked a7. The section shown
in Fig. 31 is thirteen sections farther back than the one shown
in Fig. 30.

It thus appears that while the peribranchial sacs become
wholly separated from the branchial sac, they do not become
separated from each other, but remain in very wide communi-
cation, this communication occurring both behind the posterior
extremity of the branchial sac and over its dorsal part. (Com-
pare Fig. 31.) It is this common cavity into which the two
peribranchial sacs open, and which may be regarded as an
unmodified remnant of the primitive inner vesicle that has
been called the cloaca ; but it is important to recognize that in
this species, as in many others, it is neither morphologically
nor physiologically wholly distinct from the peribranchial sacs.
By this time numerous branchial stigmata have formed, par-
ticularly in the anterior portion of the branchial sac. The
peripharyngeal band and the branchial tentacles are already
partly developed, and the dorsal lamina, @/, Fig. 30, has
become quite a conspicuous object on the sections, though it
has not yet become rolled over at its edge as it always appears
in the adult (Fig. 6, Pl. XII). The endostyle is also distinctly
marked out, though it is not histologically fully differentiated,
(end., Fig. 30). Likewise the internal longitudinal bars (2.2.0.,
I, 2.26., 2, etc., Fig. 6) have begun to form.

The formation of the branchial and atrial openings seems
deserving of a little special attention. The development is
initiated by an evagination from the inner sac, for each orifice
(Fig. 35, a¢.szp., Pl. XIV). A little later an invagination forms
immediately over this in the ectoderm ; the two meet, fuse, the
fused wall becomes broken down, and by gradual but later-
accomplished perforation of the overlying test, communication
is established between the outside world and the respective
sacs (Fig. 36, Pl. XV). The fact to which I wish to call atten-
tion appears in the course of later development. It is illus-
trated by Fig. 36, drawn from a section through a branchial
siphon in which the opening is completed so far as the cellular

170 LEG RS TITS, [VoL. XII.

walls taking part in its formation are concerned. The over-
lying test is not yet perforated, however. The great growth
of the epithelium lining the newly formed siphon, resulting in
its remarkable thickness, and the wide, circular double folds,
$:., Sf, and sjf.1, 1s the fact to swhich I refer. Study) of
various stages in this growth shows that all the thickened part
of the epithelium is derived from the ectodermal invagination,
and that the formation of the folds is due to growth of this
ectodermal epithelium. The secondary fold, s.f.1, represents
the evagination from the inner layer, it remaining passive,
while the other portion grows rapidly, and produces, as the
figure shows, a well-marked valve. As development goes on
the whole siphon expands, and this fold or valve becomes
narrower and narrower, and finally in the adult it is wholly
obliterated, at least in the normal uncontracted condition
of the animal. The formation of this fold is probably to
permit of the protrusion of the siphon above the general level
of the colony in its fully developed state. As shown by
this figure, at this stage, z.e. previous to the perforation of the
test, there is no such protrusion. This explanation hardly
accounts, however, for the great height of the entire lining
epithelium during these stages of development.

A very different explanation suggests itself, though one too
imaginative to be deserving of more than a mere mention.
One might conjecture that as incident to the abbreviated
development of the Ascidian blastozooid as compared with the
embryonic development, the earliest, paired stage in the origin
of the atrial opening, so well known in the larval history of
many Ascidians, had been entirely lost in the development of
the bud, and that the peculiar conditions above described rep-
resent the stage in the formation of that opening immediately
after the fusion of the two primitive paired embryonic orifices.
Van Beneden et Julin (84), pp. 625-627, to whom we are
indebted for the facts in the embryonal development above
referred to, quite insist on the importance of distinguish-
ing the cloaca proper from the peribranchial cavities. They
point out that in Phallusia scabroides the cloaca is wholly
lined by an ingrowth of ectoderm from the dorsal side

No.1.] BUDDING IN GOODSIRIA AND PEROPHORA. ev Au

of the larva between the two atrial orifices ; while, on the
other hand, “les cavités péribranchiales sont . . . delimitées
en dehors seulement par l’épiblaste,’’ and even this partial
ectodermal lining arises independently, both in time and
position, of that which lines the cloaca.

If there is any value in the suggestion here made that the
ectodermal fold, s.f, Fig. 36, is comparable to the true cloaca
of the above-quoted authors, it would not be lessened by the
rejection of their distinction between cloaca and peribranchial
cavity in so far as this distinction rests upon their view that a
portion of the peribranchial epithelium is of endodermal origin.

But, as already said, I do not regard the suggestion, with
the only facts that I now have to base it on, as deserving more
than a mere mention.

b. The Digestive Tract and its Appendages.

The Ax/age of the intestine is established at a very early
stage in the development of the bud. It appears at a point on
the primitive inner vesicle that is ventral, posterior, and slightly
tothe lett:

In the earliest condition seen it is a mere short, simple pro-
jection growing from the wall of the vesicle, situated in a
notch, or indentation of its side in the region mentioned.
Fig. 16, Pl. XIII, representing a dorsal view of a whole bud,
illustrates the above statements, but it must be borne in mind
that the bud is seen as a transparent object, and that the
hypophyseal and intestinal Az/agen would not, as represented,
be seen at the same level. The notch above mentioned is
clearly seen in this figure as far as the posterior side of the
vesicle is concerned ; but sections from similar buds enable
one to see that it is quite as well marked on the ventral side
as it here appears on the posterior side. Concerning this
notch, or rather the parts of the vesicle adjacent to it, I shall
speak further in dealing with the development of the heart ; I
shall, therefore, leave the subject for the present after mention-
ing that at o time in the life of the zootd does the digestive tract

172 RITTER, (VoL. XII.

extend so far posteriorly as do the backward extensions of the
vesicle,

During its early stages of development the intestine projects
downward and to the left from the vesicle to which it is
attached, so as to lie for the greater part in the wide body
space that surrounds the vesicle (Fig. 29, zv¢., Pl. XIV). As
development advances, however, it becomes pushed into the
vesicle, so that ultimately it lies wholly within it, z.e. within
the atrium in the adult, and by carrying the wall before it as it
enters the vesicle, it comes ultimately to be wholly enveloped
by a thin layer of epithelium, which is attached through a thin
double-layered mesentery to the epithelial lining of the atrium.
Figs. 29, 30, 32-34, Pl. XIV, illustrate these statements.
Fig. 29 is from a section that passes through the oesophagus,
oe., and at the same time cuts the intestine, zzz., in its widest
part at this stage.

At this time the intestine lies, as mentioned above, in a
notch in the ventral side of the vesicle, but projects largely
into the body space. It has as yet grown very little in length,
but its lip is already directed forward where it appears in the
fourth section in front of the one here figured. The distal
part is considerably smaller in diameter than the proximal, the
stomach and intestine proper being thus distinguished from
each other even at this early period.

The organ grows both forward and outward with advancing
development, and very soon begins to take on the curved form
so characteristic of the Ascidian digestive tract. The curva-
ture is produced by the distal end becoming directed at first
outward and dorsalward, and then a little later backward. The
plane of the loop stands at first at an angle of about 45° to the
horizontal plane. Owing to the backward direction of the tip,
a section tangent to the loop first appears in a series cut from
before backward ; and then a few sections farther back the
loop is wholly passed and a double section of the organ is
made. This condition is shown in Fig. 32, ves. being the
rectum, and s¢. the stomach. The long pouch-like appendage
from the stomach, /. coe., is the beginning of the lacteal coecum
and duct. It is noteworthy that at this stage it is almost as large

No. 1.] BUDDING IN GOODSIRIA AND PEROPHORA. 173

as the intestine proper. By this time the pushing of the organ
into the vesicle, as mentioned above, has advanced considerably
(Fig. 32). With further growth the rectum becomes directed
inward toward the branchial sac, as well as backward and dor-
salward, so that, in the nearly adult state, transverse sections
of the animal, which pass through the anal opening, already
formed at a considerably earlier time, cut the rectum length-
wise for some distance, at the same time that they pass trans-
versely through the stomach (Fig. 6, az., vec., and s¢., Pl. XII).
It will be observed in this figure how fully the whole digestive
tract has now come to lie in the atrium and left peribranchial
sac. The rectum reaches across the chamber, and comes in
contact with the wall of the branchial sac, and actually forms
a secondary attachment to it, so that a rectal mesentery is
produced (Piss 34), wes.zec,, Fl XIV). In the, section this
mesentery does not appear, but it does in sections a little far-
ther back, none of which, however, would illustrate some other
points as well as does this one; consequently, I have chosen
this for the figure, and marked the position of the mesentery
diagrammatically. The original mesentery (Figs. 33 and 34, mes.
gas., Pl. XIV) is in the adult attached to the parietal portion of
the atrial epithelium well out on its side, z.e. quite remote from
the median ventral line, and extends across as a rather narrow
band to the digestive tract in the region of the stomach and
lacteal coecum (Fig. 34). The differentiation of the lacteal
system into the coecum, the duct, and the ramifying portions,
as we have seen them to exist in the adult animal, has been
fully accomplished by the time the stage we are now consider-
ing is reached. The same is true of the formation of the
longitudinal folds of the stomach, and the peculiarly modified
longitudinal band in the wall of the intestine.

A secondary attachment of the rectum to the wall of the
peribranchial sac is said by Hjort (93), p. 596, to take place
also in Botryllus.

174 tel Pete Te. [VoL. XII.

c. The Pericardium and Heart.

The origin of the common Az/age of these two structures
has been the subject of almost as much diversity of statement
for different groups of Tunicates as has been the origin of the
nerve ganglion. It is, therefore, particularly satisfactory to
find that in this species its method of origin is so clear as
to leave no doubt about what it is.

At the time of writing my preliminary communication I had
not succeeded in finding the earliest stage in the development
of the structure ; further search, however, since then has borne
the desired fruit.

It arises from the postero-ventral wall of the inner vesicle,
but does not become fully severed until the ventral folds which
play the major part in separating the branchial from the peri-
branchial sacs have extended nearly back to the intestinal
Anlage. These folds, consequently, enable us to locate more
precisely the point of its origin.

Figs. 39-41, Pl. XV, are from a series of transverse sections
of a bud in a stage of development considerably younger than
that shown in Fig. 18, Pl. XIII. Fig. 39 represents the tenth
section in front of the one from which Fig. 40, passing through
the pericardial An/age, pc.a., is drawn. The ventrally directed
pouch-like fold, 70.s"., seen in Fig. 39, is a portion of the
right peribranchial sac, and it is at the posterior extremity of
this that the pericardial Ax/age originates. As shown most
clearly in Fig. 41, it is an evagination from the inner vesicle,
but, as just stated, is at the point reached at this stage by the
posterior extremity of the right peribranchial sac. vf, in
Figs. 40 and 41, indicates the right ventral partitioning fold.
The Anlage of the pericardium appears to become separated
from the vesicle at its anterior end first, then later at its
posterior end. A later stage, slightly after its complete separ-
ation, but while it is still a simple small vesicle, is shown in
Fig. 42, pc.v.

According to Oka (92), p. 535, the pericardium in Botryllus
arises as “eine solide Wucherung des inneren Blattes der

No.1.] BUDDING IN GOODSITRIA AND PEROPHORA. E75

urspriinglichen Knospe ; aber da die betreffende Stelle gerade
im Winkel zwischen der mittleren Blase und der Anlage des
linkern Peribranchialsackes liegt, ist es schwer zu entscheiden,
ob sie speciell der ersteren oder der letzteren entstammt.”’

Hjort also (93), p. 602, found the pericardium of the Botry/-
fus bud to appear first as a solid cell mass ; but concerning
the origin of this mass, whether from the “endoderm” or
from mesoderm cells he was not able to decide.

Pizon ('93), p. 44, on the other hand, describes it as arising
in this genus as a small diverticulum, and neither from the
branchial nor the peribranchial sacs, but from the primitive
vesicle.

But whether it begins as a solid cell mass or as a diverticu-
lum signifies very little ; and of almost as little significance is
the question whether it arises from the primitive vesicle, from
the branchial sac, or from the peribranchial sac, since its point
of origin is at this stage of development an indifferent point as
regards the three structures mentioned. It is not at all impos-
sible that absolute exactness in observation would find it to
arise from the branchial sac in some individuals, from the
peribranchial sac in others, and from the primitive vesicle in
others. Indeed, the observed facts in relation to variations
in the time and position of appearance of different organs
make such a conjecture rather probable.

The only fundamental question raised is as to the possibility
of a mesodérmal instead of an endodermal origin, This
question, and also the question of the epicardium, I shall
discuss after having presented my observations on the devel-
opment of the organ in Pevophora. The formation of the
heart by an invagination of the pericardial vesicle is here
so entirely similar to what is well known in many other
Ascidians that a detailed description of the process would be
superfluous. A mere reference to Figs. 29 and 30, Pl. XIV,
in which two stages in the development are incidentally shown,
will suffice.

176 RITTER. PVor. Shi.

d. The Hypophysis and Ganglion.

a The Ganglio-hypophyseal Duct.— The hypophysis always
arises very early, in some buds its origin being the first inter-
ruption of the simple spherical form of the primitive inner
vesicle. This is the case, for example, in the bud a section of
which is represented in Fig. 43, Pl. XV. Its method of origin
can be very well seen by examining with a low magnify-
ing power a whole bud from its dorsal side, the bud having
been first cleared in oil. A drawing of such a bud is shown in
Fig. 16, Pl. XIII, where Zy.a. is the Anlage of the organ now
under consideration. By observing it under a constantly
changing focus, one finds that it is a groove-like evagination
from the dorsal wall of the inner vesicle. A transverse section
of a bud in the same stage of development, and cutting the
evagination, is presented in Fig. 43. The groove is very simple
and is quite uniform throughout its length.

The same section is shown at éd., in Fig. 5, Pl. XII, where the
clearly seen orientation of the bud in the colony, and the posi-
tion of the evagination on the dorsal side of the bud, leave no
room for doubt that the evagination is the beginning of the
hypophysis. The Az/age remains in this groove-like condition
only a short time ; for in a stage of development of the bud
only a little more advanced than that shown in the two figures
to which attention has just been directed, it appears as a tube
wholly separated from the vesicle, except at its anterior end,
where its lumen communicates with the cavity of the vesicle.
This communication remains throughout the life of the animal
as the mouth of the hypophysis. As the tube becomes con-
stricted off from the wall of the vesicle, it terminates posteriorly
as a simple blind pouch, and does not at any stage communi-
cate with either of the peribranchial sacs, as is agreed by Oka,
Hjort, and Pizon to be the case in Botryllus. Hjort (95) also
shows such a communication in Glossophorum sabulosum. But
as remarked by Hjort and Bonnevie ('95), p. 392, this commu-
nication can have very little significance. They base this
remark on the fact that it does not take place in the buds of
Distaplia magnilarva, studied by them, while it does in the

No.1.] BUDDING IN GOODSIRIA AND PEROPHORA. i eel)

buds of genera no farther removed from this genus than are
Botryllus and Polyclinum. J agree that the fact has very little
significance, but would point to a different reason for so regard-
ing it. The posterior communication is always transient and
without physiological importance. The hypophyseal duct
forms from the dorsal wall of the inner vesicle at an early
stage of development of the bud, before the limits of the peri-
branchial sacs are sharply established ; and since their limits
on the dorsum of the vesicle are always very close to the
posterior extremity of the duct, a slight variation in the relative
position of the point at which the duct terminates, and the
course of the partitioning folds of the peribranchial sacs; or of
the time of closure of the duct opening; or the formation
of the folds, would determine whether the posterior opening of
the duct should be into the primitive inner vesicle, the branchial
sac, or one or the other of the peribranchial sacs, or into the
atrium. I shall show presently that there is a brief period in
the history of the duct in Goodszvia when it also has both an
anterior and a posterior opening. The posterior opening
closes, however, at a stage so early in the development of the
peribranchial sacs, that we can only regard it as pertaining to
the primitive vesicle itself.

As regards the later history of the duct as such, little more
need be said. In the nearly adult state its dorsal wall, Fig. 50,
Pl. XV, is exceedingly thin, it being but one layer of cells
thick, and the cells of this one layer are considerably flattened,
and their nuclei are far apart. The ventral wall, on the con-
trary, remains, at least in the oldest stages of which I have
made sections, several cells thick. Attention may here be
called to the entire absence of the hypophyseal gland.

The formation of the duct as I have here described it differs
in some unimportant particulars from its formation in Bozryllus,
where, according to Hjort and Pizon, it is not at its begin-
ning a groove, but an evagination of about equal diameters
transversely and longitudinally. This evagination then grows
out into a forwardly directed pouch, the blind tip of which
fuses later with the wall of the branchial sac, and its lumen
gains communication with this sac by a perforation at the

178 RITTER. [Vou. XII.

point of fusion. The permanent hypophysis mouth is there-
fore in this genus secondary, and not primary as in Goodsiria.
The groove-like Az/age of the duct in the latter genus is appar-
ently very similar to that which Kowalevsky (74a), p. 450, has
described in Didemnium styliferum, the bud-development of
which resembles that of Goodszria in some other respects.

B The Ganglion.—The three papers of Oka, Hjort, and Pizon,
on budding in Botryllus, reference to which has already been
made, were written so nearly simultaneously that neither author
was acquainted with the result of the others’ work while prose-
cuting his own. Jo two of these investigators agree concerning
the origin of the ganglion, though each was duly impressed with
the theoretical importance of the question, and was also familiar
with the discordance in the previous results obtained by several
investigators who had studied the blastogenesis in various other
Ascidians. This fact testifies convincingly to the difficulties
involved in the subject. To these difficulties I too can bear
witness, and while doing so may be permitted to say that I
have striven hard to search out every fact that might bear
on the question, and to draw my conclusions uninfluenced
by any bias in favor of one or another of the several
prevailing views. Four different origins have been found for
the ganglion in the buds of different compound Ascidians
by different students. 1. It has been derived directly from
the central nervous system of the parent zooid. 2. It has
been derived from the free so-called mesoderm cells of
the body space. 3. It has been derived from the ecto-
dermic, or outer vesicle of the bud. 4. It has been derived
from the inner primitive vesicle, or, more precisely, but what
is the same thing, from the hypophyseal duct. The first was
advanced by Pizon (93), for Botryllus, evidently suggested by
the method of origin of the ganglion in the buds of salpa. But
as the author does not pretend to have proven such a process
to take place here, and as the facts on which he bases his belief
are exceedingly meagre, this may be dismissed without further
remark.

The suggestion of a mesodermal origin was, so far as I
know, first made by Seeliger in his studies on Clavelina. It

.

No.1.] BUDDING [IN GOODSIRIA AND PEROPHORA. 179

was, however, based on indirect and theoretical grounds, viz. :
on the facts that in its earliest stages the constituent cells
much resemble those of the mesoderm by which it is sur-
rounded ; that no other source for them was observed which
appeared more probable ; and that in the embryonal develop-
ment a portion of the dorsal nerve undergoes dissolution, its
cells becoming transformed into the free mesoderm cells. It
was conjectured that these latter might reassemble again to
form the ganglion of the bud. But this view has, I believe, been
given up by the author. The belief in a common mesodermal
origin of hypophysis and ganglion, as described by Seeliger
(89), in the buds of Pyrosoma, appears to rest on a much
securer foundation. Very recently Lefevre (95) has main-
tained such an origin for the ganglion in Perophora.

In Goodsiria, the very early stage at which the ganglion is
found to be fully separated from the hypophyseal duct, together
with the conditions which I have observed. in Perophora, have
induced me to look for evidence of its mesodermal origin here.
The close resemblance of its cells in its early stages to certain
of the surrounding cells in the body space, z.e. mesenchyme cells,
and the difficulty of observing its origin from any other source,
are the only facts that lend any countenance to such a view —
facts certainly of little weight, particularly when opposed to
direct evidence of a contrary kind. What I shall have to
say on this point when treating the gangleo-hypophyseal
development in Pevophora will be of greater moment.

The ectodermal origin of the ganglion in the bud has
been asserted by Van Beneden et Julin (87), for Clavelina ;
by Oka (92), for Lotryllus ; by Salensky (92), for Pyrosoma ,;
and by Brooks (93), for Sada. Of these authors’ works
the most important for us in the present connection is
that of Oka. This is most important because of the obvi-
ously close relation between Botryllus and Goodsiria ; because
of the positiveness of the author; and because the work
is very recent, and consequently was done by methods in
common use at the present time, and in the light of pre-
vailing theoretical views. It must, therefore, be examined
critically. The author first describes and figures an irregular

180 RITTER, [Vot. XII.

accumulation of cells in the blood space between the already
formed hypophyseal tube and the walls of the branchial
and peribranchial sacs. This becomes more compact and
regular in form; in short, it develops into the ganglion.
Concerning the origin of the cells composing the mass, the
author’s essential words are: ‘“ Was die Herkunft dieser Zellen
betrifft, so entstammen sie der Korperwand und zwar gehen
sie durch Proliferation des dorsalen, oberhalb des hypophysaren
Rohres gelegnen Epithels hervor (Fig. 33). Dzeser Vorgang
... kann nur in giinstigen Fallen beobachtet werden, denn die
Zellen verlieren sehr rasch den Zusammenhang mit den Epithel-
zellen, von denen sie sich abspalten, und wandern einzeln oder
gruppenweise rechts und links um das Rohr herum und sammeln
sich an der unteren Seite desselben” (p.540). The italics are
mine, employed because these are to me the most significant
words of the description. No one who has had experience in
studying such a process as the author here so briefly describes
(for the quotation contains all he has to say on the point) will
be convinced by the evidence which he has produced that he
has not fallen into error. Only a single figure illustrates his
statements, and while in this two cells are shown which
undoubtedly appear as though they might be in the act of
leaving the ectoderm to enter the blood space, still in the
absence of further illustration, or more precise statement, this
instance cannot appeal to one as being necessarily more than a
deceptive appearance.

The inadequate support produced by Oka for his view,
together with the fact that he is opposed by all other observers
(Della Valle, Hjort, and Pizon, who have studied the point in
Botryllus), and likewise by my own results in Goodstria, compel
me to reject his conclusions.

Concerning the ectodermal origin of the ganglion in Clavelina,
it is worth while to point out that Von Beneden et Julin have
not given the subject special attention in their paper. Under
the topic, “Development du coeur etc.” they incidentally
mention the ganglion several times, and it is shown in numerous
figures ; in most of them in close contact with the ectoderm,
but also in most quite distinct from it. However, figures

No.1.] BUDDING IN GOODSIRIA AND PEROPHORA. 181

representing sections of the ganglion well toward its posterior
end show it apparently in organic connection with the ecto-
derm. Their most pronounced statement concerning its origin
which I can find is the following, occurring on page 311:
“Dans les dernieres coupes il n’est pas possible dé voir la
limite entre l’ébauche neurale et l’épiderme; elle parait étre
un simple épaississement de |’épiblaste (Fig. 10, 11, et 12).”
The authors’ conclusions certainly need confirmation before
they are entitled to unqualified acceptance.

The ectodermal origin of the ganglion asserted by Salensky
and Brooks for Pyrosoma and Salpa, respectively, appears to
be well supported in both cases, though it does not stand
unchallenged, Seeliger ('89) claiming a mesodermal origin for
it in both these genera. I have had no opportunity for personal
observations on the point, and consequently shall express no
opinion upon it further than this, that to my thinking even
Pyrosoma is sufficiently remote in its relationship to the com-
pound Ascidians to make possible an ectodermal origin of the
ganglion in its ascidiozooids, while the same organ arises from
the endoderm in the buds of the compound Ascidians. Seeing,
as we do, the central nervous system arising from the ectoderm
in the embryozootds, and from the “endoderm” in the blasto-
zootds of the same species, its origin from either of these
sources, or even from the mesoderm, in Pyvosoma, or still more
in Sal/pa, ought not to cause great astonishment. That, how-
ever, there are such radical differences within the same genus
as results indicate cannot be accepted till more evidence is at
hand than we yet have.

The last-mentioned source of the ganglion, z.e. from the
inner or ‘“endodermic”’ vesicle, was asserted by several inves-
tigators whose observations were made a number of years ago,
before the exacter methods of section-making had come into
use in morphology, and before morphology had gone so greatly
under bondage to the germ-layer theory as it has done more
recently. And so it happened, as is not unfrequently the case
in science, that the greater intellectual freedom enjoyed by
the earlier workers more than offsets their cruder technique,
and they were enabled to reach conclusions more nearly true

182 RITTER. iV Or ei

than are those reached by some later students under conditions
reversed as to technique and intellectual freedom.

Among the earlier works to which reference is above made
those of Giard (72), and Kowalevsky (74a and '74b), are
particularly to be mentioned.

The results of my study of both Goodsirza and Perophora
agree essentially with this last view; but since in neither
genus were they reached without special perplexities, I must
present the facts with considerable detail. In Goodsiria the
formation of the ganglion is by no means a complicated process,
but the difficulty in finding just how it occurs is due to the
apparent quickness with which it becomes fully separated from
its source, which is, I may say in a word, the ventral wall of
the hypophyseal duct. At the time of writing my preliminary
paper I was still in some doubt about its source, but more
careful study since leaves, as I hope the following will show to
the satisfaction of every reader, no room for question. Fig. 40,
Pl. XVI, presents a transverse section of the duct in the earliest
stage that I have found after its complete separation from the
primitive inner vesicle. That the stage is one very soon after the
separation takes place is certain from the state of development
of the other organs of the bud as compared with the stage in
which the duct is still merely a groove-like evagination. The
peribranchial sacs are very slightly developed, the dorsal parti-
tioning fold of the right side being barely indicated in this posi-
tion, while the left does not appear at all in the section. Yet
it is seen that the ganglion, ¢/., though very small, is already
distinctly separated both from the wall of the primitive vesicle
beneath it, and from the duct above it. The condition here
shown, and even a slightly earlier one, but with the separation
still complete, occurs not infrequently ; but the critical stage,
the one between this and that with the duct in the groove con-
dition, escaped me for a long time, and made it seem quite prob-
able that the cells entering into the ganglion were derived from
some other source than the duct or primitive vesicle, z.e. either
from the ectoderm or from the surrounding mesenchyme cells.

In only three or four buds have I found the decisive stages,
but two of these, those from which Figs. 44-47, Pl. XV,

No.1.] BUDDING IN GOODSIRIA AND PEROPHORA. 183

and Fig. 48, Pl. XVI, are drawn, are particularly to the point,
and these I have consequently selected to describe. Figs. 45—
47 represent transverse sections, in order from before back-
ward, of a bud slightly less advanced than the one in nearly
longitudinal section shown in Fig. 48. The sections are
considerably oblique, so that the one, Fig. 44, passing through
the mouth of the duct makes it appear to be inclined to
one side. The ganglion is not touched by this section, my
purpose in figuring it being to show that the hypophysis mouth
exists at this time, though very small. This section, with the
following one, not reproduced, leaves no doubt on the point.
Figs. 45 and 46 are from sections a little farther back. They
need very little explanation. They show the thickened ventral
wall of the duct, which is (particularly in Fig. 46, six sections
farther back toward the place where the duct is still unsevered
from the vesicle) not distinguishably separated from the wall
of the vesicle. In this thickened region, Fig. 46, cell division
is more frequently seen than in other parts of the tissues.
Fig. 47, one section farther back, shows that the duct still
communicates by a wide opening with the primitive vesicle.
Three more sections take us beyond the duct entirely, so it is
obvious that the communication shown in Fig. 47 is at the
posterior part of the duct. If now we turn to the longitudinal
sections, one of which is shown in Fig. 48, Pl. XVI, we find
that the lumen of the duct no longer communicates with the
vesicle posteriorly, but that there exists a trace, ~e¢.0., of the
opening by which this communication was earlier effected.
But the most important fact to be learned from this section is
that anteriorly the ganglion, g/., has now become separated
both from the ventral wall of the duct and from the wall of the
vesicle ; but that posteriorly, where the duct has not yet fully
severed its connection with the vesicle, the ganglion is still
confluent with the large cell mass that forms at the same
time the ventral wall of the duct and the dorsal wall of the
vesicle.

We are thus led to recognize that szmultaneously with the
closing off from the tnner vesicle from before backward of the
hypophyseal duct, the ganglion becomes differentiated in the same

184 RITTER. [Vou. XII.

order from the cell mass that forms the last connection between
the duct and the vesicle.

e. The Genital System.

The results obtained from the study on this system are far
from complete. It is very imperfectly developed in all my
specimens — so much so, in fact, that I cannot regard as final
any of the observations made. As already said, my material was
all procured in one locality, and at the same time, vzz., December
30—midwinter. It is not at all impossible that when oppor-
tunity is afforded to examine specimens taken from other
places, and particularly at other seasons of the year, the sexual
organs will be found to tell a different story from that intimated
by the fragmentary facts presented by the colonies so far studied.

The character of the sexual organs in the adult, already
described, and the manner of origin of the sexual cells in the
Botryllus bud, lead us to expect that here also we shall find the
youngest ova and sperm cells floating free in the body space,
having been brought hither by the blood directly from the
parent zooid. That such is the case seems quite certain from
the facts observed. In several instances I have found ova in the
blood (Fig. 52, Pl. XVI), but it is noteworthy that in all these
cases they were in duds far advanced in development. Although
I have given particular attention to the point I have not yet
succeeded in finding an undoubted case of sexual cells in a very
young bud. This seems the more remarkable when it is re-
membered that the bud becomes fully severed from the parent
at a very early stage in its development. But the young buds
examined are far too few to justify the negative conclusion
that recognizable ova never enter them before their complete
severance from the mother zooids. Nevertheless the facts
observed are of a kind and sufficiently numerous to make very
interesting the several alternative questions which they raise.
Direct observation shows that in some colonies in which some
of the blastozooids contain ova and are consequently presum-
ably capable of sexual reproduction, certain other zooids are
developed without having received recognizable ova directly

No.1.] BUDDING IN GOODSIRIA AND PEROPHORA. 185

from their parents. Do these latter zooids remain throughout
their lives incapable of sexual reproduction ; or do they receive
their ova from the blood vessels with which, as previously
shown, the zooids become secondarily connected; or have some
of the cells contained in the blood the capability of being
transformed into sexual cells; or finally, are the sexual cells
really brought into the bud from the parent, which again re-
ceived them in the same way from its parent, and so on till
their ultimate origin was the sexually produced common an-
cestor of the entire colony? The last is, I suppose, the alter-
native that would appear most probable to the great majority
of embryologists, since it is the one most in conformity with
prevailing theoretical views concerning the origin of the repro-
ductive elements. It is also in keeping with the early appear-
ance of the sex cells in the Sa/pa stolon, as made known by
Kowalensky, Salensky, Brooks, Seeliger, and others; also with
what occurs in the budding of Pyrosoma.

But most of all it is supported by the conditions presented
by the buds of Botryllus. The large ova in the young buds of
this species, known since the time of Savigny (16), have by
many writers been supposed not to appear in any of the buds
before the third or fourth generation. Pizon (93), has, how-
ever, shown that they really originate in the sexually produced
embryo, and migrate, as he expresses it, through the first
generations of buds, leaving these to develop into sterile zooids,
only to become permanently fixed as ovaries in later genera-
tions.

The fact that I have not been able to distinguish sexual
cells among the blood cells in the young buds of Goodstria
may not be regarded as proof that they do not exist there.
They may be in so early a stage of differentiation that, with
the methods of preservation and staining employed, their
distinguishing characters were not brought out, as they might
have been by some other treatment.

It is true that I have occasionally found large polynuclear
cells (Fig. 55, Pl. XVI) that have seemed to me to be possibly
the forerunner of the cell aggregates. These I have again
conjectured to be a stage in the formation of the polycarps.

186 RITTER. (VoL. XII.

But I have been unable to get decisive evidence of such a
relation between these elements; and the structure of the
youngest undoubted female polycarps found does not strengthen
the conjecture, for in these the ova are in different stages of
development, but even the smallest show the characteristics of
ova. Fig. 51, Pl. XV, represents such a polycarp. It has not
yet advanced sufficiently to have pushed its way from the body
space into the peribranchial chamber. Certain clusters of a
few cells larger, more deeply stained, and more granular than
the other blood cells, Fig. 54, Pl. XVI, I suppose to be the
male sexual cells; but this too is not beyond question.

As has been pointed out in describing the species, the poly-
carps have a rather regular position on the ventral side of the
zooids, on each side of the endostyle and not far from it. In
other words, they occupy undoubted determinate positions.
With the sex cells at first apparently wholly subject to the
movements of the blood, one would much like to know how it
comes about that their final points of attachment, where they
develop into the polycarps, are always nearly the same. That
they are in the ventral part of the zooids suggests that the
greater weight of the cells, while still swimming in the blood,
prevents their being carried into the higher parts to become
attached; but that indicates nothing as to why they should
become arranged so regularly on each side of the endostyle.
Unless we may assume determinants for these particular cells
of a considerably higher grade of intelligence than is generally
attributed to these convenient phantoms, it would seem that
there must be some physical or chemical condition in the
membrane in the region where the polycarps are situated that
attracts to it the sex cells from the blood when they happen
to be carried near it.

With the young sex cells entirely subject to the movements
of the blood, and as a consequence liable to be carried to
various parts of the colony, —into one zooid or another, — it is
interesting to notice how completely, from the sexual point of
view, at least, the colony, and not the ascidiozooids composing 11,
is the individual.

No.1.] BUDDING IN GOODSIRIA AND PEROPHORA. 187

B. PEROPHORA ANNECTENS, RITTER.
1. INTRODUCTORY. MATERIAL. ‘TECHNIQUE.

I have described this species in detail in a former paper ('94),
and consequently am relieved from doing so here. One point
only concerning it I must speak of briefly. In the description
referred to, I have said in substance that the species includes
at one extreme colonies in which the zooids are quite as
distinct from one another as they are in P. Lzsteri, for example,
they being nearer together, merely, on the stolons; and at the
other extreme colonies in which the zooids ‘are as completely
enveloped by a common test as they are in Botryllus or
Goodsiria.”

I wish here to reaffirm and emphasize this statement in its
fullness. I have several times reéxamined my material with a
view to finding some constant structural difference that would
enable me to separate specifically the one extreme from the
other; but each effort has resulted in strengthening the convic-
tion that such a separation is impossible. It may be worth
while to add that I have reéxamined the case since having had
opportunity to study several other species of Pevophora, new
and old, P. Lzstert included. Last summer I had the pleasure
of submitting my specimens to the experienced judgment of
Prof. Herdman, and he agrees with me fully that they are all
one species. I reassert, then, that we have here within the
limits of the variability of a single spectes a complete transition
Srom the Simple or Social Ascidians (depending on what author s
system of classification be adopted) to the Compound Ascidians;
or in other words, a complete transition within a single species
between two groups that have been considered as distinct
families, Bronn, Gegenbaur; or as distinct suborders, Herd-
man; or even as distinct orders, Haeckel, Claus. For a
summary of the various systems of classification that have
been proposed from time to time, see Seeliger (94). Lahille
(90) has proposed a wholly new classification and nomenclature
of the groups above genera, in which the simple or compound
conditions are entirely ignored as differentiating characters for

188 RILTER. PVioL. xa:

any groups higher than species. To some extent this radical
departure receives support from Perophora annectens. But I
fully agree with Herdman and Seeliger in thinking that the
system adopted by Lahille, at any rate as regards his families
and suborders, comes no nearer being a natural one than the
other systems which it is intended to supplant. But it would
be outside the purpose of the present paper to enter upon a
general discussion of tunicate classification.

In view of the apparent hesitation of some students of the
Tunicata, with whom I have had personal conversations, to
accept my conclusion that P.aunectens really does include
such widely varying forms, I had thought to present in the
present publication a more detailed and fully illustrated de-
scription of all the transitional forms than is contained in my
former paper. However, further reflection has made it seem
to me that such a discussion will belong more appropriately to
a comprehensive treatment of all our Pacific Coast species of
the genus. Such a treatment will be contained in a mono-
graph of the California Tunicata which is now in course of
preparation by myself and one of my students, Mr. F. W.
Bancroft.

As stated in my preliminary, my results on the budding of
Perophora have nearly all been obtained from the study of P.
annectens. From the abundance of this species in Monterey
Bay, Cal., and from the compactness of the colonies, it is
very easy to get unlimited material, and whenever I have
collected it, the buds in all stages of development have been
present in great numbers. /. Lzstevz, on the other hand, at
least in the Bay of Naples, I found to present almost insup-
erable difficulties in the way of procuring buds in sufficient
quantities to enable me to get the necessary stages, and those
in such numbers as to make a satisfactory study of their
development. Both the zooids and the stolons are very delicate,
and they cling so closely to the stones on which alone I found
them that the young buds are removed with very great diffi-
culty. I succeeded, however, in securing specimens enough
to-enable' me to make sure that there is no ditference “of
moment between the bud-development in the two species.

No.1.] BUDDING IN GOODSIRIA AND PEROPHORA. 189

For the killing and preservation of material I have used the
various methods that have been recommended by the numerous
investigators who have been occupied in recent years with
Ascidian morphology and development, but in spite of the
variety of treatment I have not been able to get preparations
of this genus quite as satisfactory as those obtained from
Goodsiria. On the whole, as in Goodsiria, the specimens
fixed by the weaker solution of the picro-sulphuric mixture of
Kleinenberg has given the best results; but the glacial acetic
acid method and the chromic-acetic acid solution have also
been found useful.

Several zodlogists have studied the budding of Perophora.
Giard (72), Kowalevsky (74), Van Beneden et Julin (g7), and
Pizon (93), have published their observations on the subject,
but of these by far the most important work is that of
Kowalevsky. It was this investigator who first described in
detail the entire development of the blastozooids, and his results
were so nearly complete that the points which he left in doubt,
with perhaps a single exception, remained in that condition up
to the present time, nothwithstanding the studies that have
since been bestowed upon the subject. The points to which I
refer are the origin of the ganglion, the heart, the sexual
organs, and the precise relation of the zooids to the stolon; or
more exactly, to the cloison, or septum of the stolon. The first,
second, and fourth of these, as the sequel will show, have been
rather obscured than clarified by the more recent works.
Concerning the third it appears that the results of Van Beneden
et Julin are more exact than those of Kowalevsky; but of
this I am unable to judge from personal observations. In
none of the specimens studied by me have the genital organs
been sufficiently developed to permit a satisfactory study of
them.

Because of Kowalevsky’s paper a redescription of the general
development of the buds would be quite superfluous. Except-
ing the genital system, I shall confine myself, therefore, to the
points mentioned above as having been left doubtful by him.

I9O0 RITTER: VoL. Xi:

2. RELATION OF THE BUD TO THE STOLON.

The point to be made in this connection may be briefly
stated by quoting what I have already said in my preliminary
paper: ‘When the differentiation of the ‘endoderm’ into the
branchial and two peribranchial sacs takes place, it does so in
such a way that the developing blastozooid is connected with
the double-walled partition of the stolon, not by the dvanchzal
sac, as has been hitherto supposed, but by the left perzbvanchial
Sat.

Figs. 57-64, and 65, 67, and 70, Pl. XVI, will suffice to
illustrate the facts. Of these Figs. 67 and 68 are from the
youngest stage. At this time the inner vesicle is still almost
a perfect sphere, no differentiation of organs having yet begun
excepting that a mere trace, fc.a., of the heart is present.
The connection of the inner vesicle to the stolonic septum,
cl’n., does not show in his section, but it does in the third
section of the series preceding this. The bud at this time
does not stand out at a right angle to the stolon, but is
inclined somewhat toward its tip. The series passes from the ,
posterior, or point of attachment of the bud, toward its ante-
rior, z.e. toward the tip of the stolon. It is for this reason
that we find the section of the stolonic septum in several sec-
tions before we come to its place of attachment to the inner
vesicle.

The process by which the vesicle becomes divided up into
the branchial and two peribranchial chambers differs so little
from the same process in Goodszria and various other Ascidian
buds, that I have thought it would be sufficient to pass over
the various stages excepting those that pertain especially to
the condition about to be described. It is, however, necessary
to refer to the account given by Lefevre of the first stages in
the differentiation of the vesicle. He states that in P. vzrzdis
the inner vesicle rotates to the right through 90° during the
initial steps in the formation of the peribranchial sacs. Accord-
ing to him the heart Ax/age, the first organ to appear, is located
at the beginning go° farther to the dorsal and left of the
vesicle than the position that it ultimately occupies in the

No.1.] BUDDING [N GOODSIRIA AND PEROPHORA. I9QI

adult zooid. The rotation is not so great in P. annectens, as
may be determined by considering Fig.67. The heart Anlage
is certainly somewhat higher than it will be at a later time, as its
position with reference to the stolonic septum, in this section
shows. I have seen no specimens in which the Az/age was at a
higher point, and I think that this particular individual furnishes
evidence that no appreciable rotation could have taken place in
the earlier stages of development. This evidence consists in
the uniform thickness of the wall of the vesicle in all its parts,
thus indicating that there has been no considerable inequality
of growth, to which Lefevre probably rightly attributes the
rotation; and to the fact that at this time the connection of
the septum to the vesicle is not yet thrust off to one side as it
later comes to be (Fig. 70) when the rotation is complete. It
is undoubtedly to this rotation that the attachment of the
septum to the left peribranchial sac is due.

The greater or less extent of the rotation in various species
probably has no particular significance beyond the general
significance that attaches to variations in development.

From this early stage we may now pass to the examination
of one much farther advanced. Such a one is shown by the
series of Figs. 57-64, all from the same bud. The sections
pass from the anterior toward the posterior of the animal, and
are seen on their anterior surfaces, thus causing an apparent
reversal of right and left. In Fig. 57 the peribranchial sacs
are entirely distinct from the branchial sacs, and this is the
condition for a considerably greater portion of the anterior
part of the bud. Fig. 58 shows the point at which the three
cavities become confluent. From here the ventral partitioning
folds, v.f, rapidly fall away; compare Fig. 59, seven sections
farther back. A few sections farther back (Fig. 60) they have
almost wholly disappeared, and the three cavities have become
practically one. This section, with those going before it, shows
clearly, however, that the bay, /.f0.s.,1s the posterior tip of the
left peribranchial sac. But this is the point at which attach-
ment to the stolonic septum, c/’z., occurs. It will be seen that
the ventral side of what will later be the branchial sac is far
below this, it being marked by the already forming endostyle, end.

192 LO IOI BY se [VoL. XII.

Series of figures similar to these, representing both younger
and older stages, might be multiplied indefinitely were it neces-
sary; but the main facts, vzz., those concerning the relation
of the zooid to the stolonic septum, are so clear that no further
illustration is required. And this is the more so because
Lefevre (95) fully confirms my results in this particular. In
Fig. 65 I have shown a single instance from many that
might be given of the same relations in a bud of Pevophora
Lister.

The connection between the septum and the peribranchial
sac becomes entirely lost at a stage only a little later than that
represented in the series 57-64. Lefevre states that the
severance does not occur until a somewhat later time in
P. viridis, This difference in the two species may be corre-
lated with the fact that the zooids of the P. annectens colony
are closely bound together by the test, and hence depend less
on the stolon for maintaining their colonial character than do
those of P. vzvidis. It is certain that in some cases, at least in
P. annectens, the buds become wholly severed from the stolon,
the ectodermal as well as the endodermal connection being cut
away. I have shown an instance of this in Fig. 66. This is
the section of a complete series in which the stolon, s¢o., with
its septum approaches most nearly to the wall of the zooid; yet
there is a distinct absence of connection between them. I am
unable to say with what frequency this complete separation takes
place, but so far as I have been able to determine it appears
that it is rather exceptional. Certain it is that the stolons of
a colony are always in communication with many of the zooids
because the blood is always in motion in the living colonies,
and there is, of course, no other propelling power than the
hearts of the zooids.

I have stated that the connection is always to the left pert-
branchial sac. This has been true for every undoubted instance
observed, but it does not seem in the least improbable that
exceptions might be found were one to examine a sufficiently
large number of cases; though that it is well-nigh the invaria-
ble rule is certain from the large number of instances of its
occurrence that have been observed.

No.1.] BUDDING IN GOODSIRIA AND PEROPHORA. 193

That Kowalevsky failed to make out the precise relations
between the bud and the septum of the stolon is not surprising,
since he studied whole buds almost exclusively ; at least he
could have had no complete series of sections. The septum is
very delicate, and is entirely surrounded by tissues denser than
itself, and for these reasons it is almost, if not wholly, impossi-
ble to trace the actual condition of things in the living bud.
But the failure of Wan’ Beneden et. Julin, and of Pizom)'I
cannot account for so easily, since these authors made use of
thin sections in their studies. It seems probable that they
either missed the stages in which the facts are most easily
seen, or that their series of sections were imperfect. Under
other circumstances it appears hardly possible that they would
have overlooked the facts, patent as they are.

It would appear that Van Beneden et Julin, having care-
fully followed through its entire development the bud of
Clavelina, and having studied the bud of Perophora sufficiently
to see that in most particulars it agreed with C/avelzna, thought
themselves safe in assuming the same agreement would hold in
all points. So many common characters do the two genera
present, not only in adult structure but also in development,
both of embryozooids and blastozooids, that such an assump-
tion might very naturally seem warranted. The case furnishes
only another illustration of the dangers to which morphologists
are subject when they assume certain things to be true of one
animal because they are known to be so for another closely
related one. The fatality lies most frequently, I suspect, in
the supposed, and not real, close relationship between the
animals.

These authors practically, though unconsciously, say that their
observations on this particular point are defective for Pevophora.
They say (p. 307), “nous n’avons pas pu nous édifier complete-
ment sur l’histoire du coeur et du systéme nerveux,” etc.
Now, as will appear from my account of the development of
these organs, it would have been well-nigh impossible for them
to fail of a clear understanding of the development and relations
of structures mentioned without likewise missing the true rela-
tions of the bud to the stolonic septum.

194 RITTER. [Vot. XII.

That they did suppose the same conditions to hold for both
genera is evident, although they say so rather indirectly.
Thus, in summing up their results on the development of the
heart “et de ses dependances chez la Claveline,’’ they say, p.
317, “Chez la Claveline, comme chez la Pérophore, la vesicule
interne du burgeon,” etc. ; and then follows a brief restatement
of the relations between the primitive inner vesicle, the bran-
chial and peribranchial sacs, the epicardium, the pericardium
and heart, and the cloison, as they have been described in
detail for C/avelina. In this connection I shall merely say
that the description which they give for Clavelina does not
apply in a single essential particular, excepting as to the forma-
tion of the heart from the pericardium, to what actually takes
place in Perophora. In describing the development of the
heart I shall point out in detail the difference between the two.

Pizon’s paper is entitled ‘ Histoire de la blastogénése chez
les Botryllides,” and the author does not claim to have made
so exhaustive a study of the numerous other genera which
received his attention as he has of Botryllus and Botryllozds.
As regards Perophora he seems to have examined it just enough
to make himself feel safe in adopting the errors on this point,
at least, into which Van Beneden et Julin had fallen. Thus,
he says, p. 105, ‘‘Disons d’abord que chez des bourgeons
a’ Amaroncium proliferum, de Circinalium concrescens (Giard),
de Perophora Listert, de Clavelina Rissoane, je suis arrivé
identiquement aux mémes résultats que Kowalevsky (Am. pro-
liferum) et que Van Beneden et Julin (Clavelina Rissoana) a
propos de lorigine du tube epicardique.” Then follows a
brief description of the epicardium and its relation to the
branchial sac and pericardium, which is the same as that already
referred to as having been given by Van Beneden et Julin.
But this writer’s statements on this point I shall also have to
consider further after having described the development of the
heart.

= eee =

a 6 oe

Se ae

No.1.] BUDDING IN GOODSIRIA AND PEROPHORA. 195

3. PERICARDIUM AND HEART.

It appears to be the rule that the pericardium is the very
first organ to be founded in the Pevophora bud. At least, I
have found a structure present in several buds at a time when
the primitive inner vesicle is still wholly unmodified, which I
suppose to be the Axdage of this organ. One of these is
“epresented im) Figs.; 67 and 68, Pl. XVI.

My reason for thinking this to be the beginning of the peri-
cardium is this: the sections of this bud are cut from the
posterior toward the anterior end of the future zooid, conse-
quently they are seen in their natural position as regards right
and left. This, as I have pointed out a few pages earlier,
makes this structure in the position, the slight lateral rotation
of the vesicle to the right having been performed, that the
heart occupies in the adult bud. The only question that could
be raised against this identification would rest on the possi-
bility, first, of error concerning the anterior and posterior ends
of the bud; and second, that in this instance we have an
exception to the rule that the septum is connected with the
left peribranchial sac. As regards the first, the possibility of
error is remote because another older bud in the immediate
vicinity is clearly cut from behind forward, and the two are so
close together that a reversal of their anterior ends is hardly
possible. As regards the second, nothing can be said beyond
what was asserted in discussing the relation of the bud to the
stolon ; vzz., that in the large number of buds observed, no
exception to the rule has been found. If this is an exception,
it is the only one hit upon.

We may then regard it as certain that the cell mass under
consideration is the Az/age of the pericardium. The origin of
it is a difficult matter to decide. The numerous and widely
separated groups of Tunicates in which the pericardium has
been satisfactorily shown to arise from the endoderm is a
circumstance that in itself makes its similar origin here
probable ; and my observations on the whole point in the same
direction, though I have been unable to remove from my mind
some traces of doubt on the subject. As shown in Fig. 68,

196 RITTER: [Vou. XII.

there seems to be in the section here represented a passing of
cells from the wall of the vesicle into the Ax/age, pc.a., but the
evidence of such a process is more convincing in this section
than in either of the other three sections of the series in which
the An/age appears ; and even here, as the figure shows, the
line of separation between the wall and Ax/age is distinct in
places.1 At the same time the cells of the Az/age are some-
what more deeply stained than are the immediately adjacent
ones in the wall of the vesicle, a fact which I have attempted
to bring out in the figure by making the former slightly darker
than the latter. Again, as seen by this figure, the cells of the
Anlage are not closely massed together as one would suppose
they would be had they been derived from the vesicle at the
single point where, as shown by the figure, they are passing
from the latter into the former.

These two last-mentioned conditions raise some doubt about
the origin of the Az/age from the vesicle, and at the same time
they, together with the close resemblance of the cells compos-
ing it to some of the surrounding mesenchyme cells, suggest
these latter as being its source. I have not seen an earlier
stage than this in the formation of the Av/age, but have found
about the same and slightly older stages not infrequently, and
in all cases the conditions presented are very similar. In Figs.
69 and 71, Pl. XVI, are represented sections from different buds
both considerably more advanced in development than that
just described, though of the two the one shown by Fig. 71
is somewhat older. Here the cavity of the pericardium is
already established, though it is quite small, particularly in
the younger bud. The portion of the wall of the organ in
contact with the wall of the inner vesicle is thicker than else-
where, and the two appear to be in organic connection ; but
any one who has had experience in determining whether two
cell masses in contact with each other are really organically
connected or not, knows well how extremely easy it is to
be deceived. I have carefully examined with an oil-immer-

The interruption of the line separating the 47/age from the wall is at a point
toward which the index line /c.a., Fig. 68, points. The lithographer has failed to
accurately reproduce my drawing here.

No.1.} BUDDING IN GOODSIRIA AND PEROPHORA. 197

sion lens all the sections of these two, and numerous other
cases, and believe that the connection is real and not ap-
parent ; and since at a later stage the separation certainly
becomes complete, the connection cannot be supposed to be
secondary. I therefore conclude ¢hat the pericardium originates
Srom the wall of the primitive inner vesicle.

At the same time, however, I must expressly state what is
already obvious from the description, that the conclusion rests
upon a preponderance of evidence; there is certainly some
evidence against it, and that is indicative of a mesenchymal
origin of the cells.

(Since the above account of the origin of the heart was writ-
ten, Lefevre’s short paper has reached me, in which he affirms
that the mesenchyme is the source of the heart. I have
reéxamined my preparations with much care, and must say
that, although, as my words above indicate, Lefevre’s state-
ments found my mind in a condition of equilibrium, almost, on
the point, I am still of the opinion that my conclusion is cor-
rect, at least for P. annectens. In fact, my more recent study
has discovered some additional evidence in support of my
earlier conclusion. For example, Fig. 69 represents a section
in which, az the focus here shown, as seen under the oil-immer-
sion lens, it is certain that no interrupting line is present, and
at 8 is one cell, at least, that is undoubtedly about to divide,
though I do not stake much on this fact. A circumstance
connected with another section of this same series is I think
quite indicative that the Az/age is in organic connection with
the vesicle. It is this. The outer thin wall of the pericardial
vesicle has been, in the section referred to, by some means
artificially broken away from the rest of the vesicle ; yet the
thickened side next the endodermic vesicle is still as indistin-
guishably confluent with the latter as in the section figured.

It seems to me quite likely that a force considerable enough
to break this wall would have moved the whole pericardial
vesicle from its contact with the endoderm were its relation
merely a contact.

I have found no sections in which at some focuses I cannot
see the separating line to be uninterrupted ; but, very distinctly

198 RITTER. [Vou. XII.

in many sections, and less distinctly in many others, I am sure
that the line is interrupted.

And, as already said, the fusion can hardly be regarded as
secondary, because at a little later time the pericardium is
clearly wholly distinct from the “endoderm.” Now what is the
explanation of the facts that have left traces of doubt in my
mind, and have led Lefevre to believe that the cells under con-
sideration come from the mesenchyme ? I believe it to be this:
That the mesenchyme cells themselves are being constantly pro-
duced from various parts of the endodermic vesicle for a consider-
able time during the early stages in the development of the bud.

My evidence for this is not as conclusive as we would wish
it to be, but at the same time I have observed a considerable
number of cases in which I believe cells are being set loose
from the “endoderm,” and are passing into the blood. Figs.
77 and 78, Pl. XVII, represent cases of this kind, d.c.', being
the cells referred to. These are both remote from the position
at which either the ganglion or heart will form.

And, on theoretical grounds, is not such an origin of the
mesenchyme cells highly probable? Certainly new ones must
be forming rapidly somewhere, for the newly developing zooids
must make large demands on them for their blood and muscles,
both of which undoubtedly come immediately from the mesen-
chyme. To suppose that all the muscle, blood, and mesenchyme
cells of a colony have been derived directly from the mesen-
chyme cells of the embryozooid does, it seems to me, rather
more violence to general developmental principles than to
suppose that each new bud is capable of producing such cells
for itself. And what part of the bud is more likely to do this
than the “endoderm”? If my belief on this point is correct,
then it is not at all impossible that the heart, or even the
ganglion, may be formed, at least in part, from mesenchyme
cells, for the mesenchyme cells would themselves have the
same source, and when first severed from the ‘endoderm”’
would be of the same character as the cells which certainly
produce the heart and ganglion in the buds of some other
species. It would be merely a question of what position on
the primitive inner vesicle the cells should be given off. And

No.1.] BUDDING [N GOODSIRIA AND PEROPHORA. 199

in this connection it is significant that the resemblance of the
Anlage cells to the mesenchyme cells does not apply to a// the
mesenchyme, or blood cells by any means. Asa matter of fact,
it is to a comparatively small number of these latter that the
former have such a resemblance.)

A detailed account of the development of the heart from the
pericardial vesicle is just as little necessary here as in the case
of Goodstria, where it was stated that such a description would
be superfluous, so entirely does the process correspond with
what has been many times described for other Tunicates. Fig.
61, Az., Pl. XVI, incidentally shows the heart beginning to
form by an invagination of the large pericardial vesicle.

The matter of chief interest in connection with the develop-
ment of the heart in the buds of Pevophora remains to be
considered. I refer to its azrect origin from the primitive
inner vesicle, and its consequent independence of the stolonic
septum. The facts in relation to the subject were partly pre-
sented, as will be recognized, in describing the manner of
attachment of the blastozooid to the septum; and the whole
question might have been very properly discussed there.
But, since it has been more commonly treated in connection
with the heart by other authors, it seemed best to adopt the
same plan in this instance. It has already been pointed out
that Van Beneden et Julin supposed Clavelina and Perophora to
agree in this respect, as they do in many others. In order to
make it clear that they do not, it will be necessary to point
out what, according to these authors, are the conditions in
Clavelina. An understanding of them in all their details can
hardly be reached, however, without aid of the numerous fig-
ures by which the authors have illustrated the subject. These
are all the figures of plate XI; figures 1 to 7, plate XV; and
figures 3%, 35, 3°, 34, and 3¢ of plate XVI of their well-known
memoir ('87).

The parts essential to an adequate understanding of the
conditions described by them are: the primitive inner vesicle,
the branchial sac, the “tube épicardique,” the “sac épicar-
dique,”’ and the “cloison stoloniale.”’

To make the subject clear, and at the same time to do so as

200 RITTER. [VoL. XII.

briefly as possible, I will quote their words in part, and in part
give my own description condensed from theirs. On page 317,
previously quoted in part, they say “Chez le Claveline, comme
chez la Pérophore, la vésicule interne du bourgeon résulte de
l’écartement des deux lames cellulaires adjacentes de la cloison
stoloniale. La vésicule, allongée dans le sens de l’axe du
bourgeon se continue en arricre dans la cavité virtuelle de la
cloison stoloniale. Cette véstcule se divise transversalement en
une portion terminale et une portion basilaire. La portion ter-
minale de la vésicule donne naissance au sac branchial et au tube
Epicardique.... La portion bastlaire engendre le sac péricar-
dique.... Puis, par une sorte ad étranglement progressif, les
deux portions de la vésicule interne primttive se séparent lune
de l'autre.’ The italics are mine, and are inserted to call
attention to the points most important for the present purpose.
One might understand from the first part of the quotation that
the “portion terminale”’ is fully separated from the “portion
basilaire” before any farther differentiation; but such is
evidently not the meaning, since the lines omitted between
the last and next to the last parts of the quotation give the
character of the connection between them and the relation of
the whole to the developing heart. I should perhaps have
included in the first part of the quotation another line or two
which state that the “portion terminale”’ gives rise also to the
peribranchial sacs and the intestine.

On pages 315, 316 the following is much to the point. “Il
résulte de l’examen de la série des coupes successives faites 4
travers le bourgeon partiellement représenté pl. XVI, figure
32 a 3°, que le péricarde, l épicarde, et le sac branchial sont des
parties, incomplétement séparées l’une de l’autre de la vésicule
interne du bourgeon.... Cependant, le sac péricardique a
commencé a se séparer de l’épicarde, et l’étude de ce bourgeon
nous permet de nous faire une idée trés exacte de la maniére
dont s’accomplit cette séparation.”” Then follows a description
of the way in which the separation takes place, and still further
on the statement: ‘Plus tard ces communications cessent
d’exister et dés lors la vésicule interne primitive du bourgeon
s'est subdivisée en deux parties distinctes: sac branchial et

es oe aod

No.1.] BUDDING IN GOODSIRIA AND PEROPHORA. 201

épicarde, en avant péricarde et cloison stoloniale en arriére.”’
And on another page they show that the cavity of the fev7-
cardium ts directly continuous into the virtual lumen of the
clotson, and that this is the condition retained in the adult
zooids.

The essential facts here set forth, summed up in the fewest
words possible, are these: The epicardium is derived from the
branchial sac, or more precisely from the part of the primitive
inner vesicle that later forms the branchial sac. The peri-
cardium is derived from the epicardium. The epicardium
remains in connection with the branchial sac, but becomes
fully separated from the pericardium. The pericardium re-
mains in connection with the cloison.

That the epicardium is a well-defined structure in Clavelina,
is obvious from the description and figures of it by the authors.
Thus it communicates with the branchial sac by two openings
called by them “orifice épicardique,” these being in reality
two short tubes passing between the branchial sac and the
“tube épicardique”’ proper, which is a single wide cavity
terminating posteriorly in two “cul-de-sacs épicardiques.”

That the course of things in Perophora is very different from
this is clearly seen by an examination of the series of Figs. 60,
61, and 62, Pl. XVI. The most striking difference is in the fact
that in Perophora the pericardium nezther at its origin, nor at
any later time, has any connection whatever with the cloison of the
stolon; see particularly Fig. 70, Pl. XVI, from another some-
what younger bud. From these it is seen that it arises directly
from the primitive inner vesicle on its rzgh¢ side, consequently
remote from the point of attachment of the cloison to the
inner vesicle, which point is on its 4f¢ side, corresponding in
this stage of development to the posterior extremity of the
left peribranchial sac. This difference implies the further one
that there ts no epicardium in Perophora, at any rate holding
such a relation to the pericardium as this structure does in
Clavelina. It may be asked if the portion of the primitive
vesicle from which the pericardium is derived may not be
regarded as representing the epicardium of Clavelina. It
certainly does hold the same relation to the inner vesicle; z.e.

202 Y § 6 bal GI PNG [VoL. XII.

it is at its posterior middle part, so that later it becomes
a posteriorly extended portion of the branchial sac; compare
Figs. 62, 63, and 64, in which 67s’. indicates the portion of
the vesicle referred to. But it must be noticed that this is
nothing more than the somewhat narrowed rear end of the
branchial sac leading to the oesophagus (0e., Fig. 62), and ex-
tends back only a very short distance behind its opening.
Fig. 64 represents a section only three sections, 7%p
thick, farther back than the one shown in Fig. 62, passing
through the opening of the oesophagus; and in this it will
be noticed that the cavity 67s’. does not appear, though the
stomach does. This is practically the condition found in the
adult zooids. However, the fact that this coecum, as it may
be called, is the vestibule to the oesophagus need not, in view
of its relations to the branchial sac and pericardium, stand
seriously in the way of regarding it as representing the
epicardium. But to make it correspond with this structure
in Clavelina, its relation to the septum would require such
a radical transformation of things as makes it very difficult to
believe that the two structures are genetically related to each
other. To shift the connection of the septum from the ferv-
branchial sac on the /eft stde of the body, to the pericardium on
the zzght stde, would be a rather serious matter, it seems to me.

Having thus shown a radical difference between Clavelina
and Perophora in the relations of the pericardium to the stolonic
septum, and of the septum to the branchial apparatus, I leave
the subject for the present to resume it again later on from a
more general point of view.

4. THE HyYpopHysIS AND GANGLION.

It was pointed out in my preliminary communication that
while in both Goodsiria and Perophora the investigation of
this subject encounters special difficulties, it fortunately hap-
pens that the difficulty is not at the same point in the two
genera.

In Goodsiria, as we have seen, the origin of the common
Anlage of duct and ganglion from the zuner vesicle is observed

No.1.] BUDDING [N GOODSIRIA AND PEROPHORA. 203

with the greatest ease and certainty; the difficulty is found in
ascertaining the source of the ganglion. On the other hand, as
we shall presently see, in Perophora the origin of the ganglion
from the common Anlage is seen with perfect distinctness,
while the source of this common Anlage is ascertained with
considerable difficulty.

It is a suggestive fact that in Pervophora the difficulty en-
countered in making out the origin of the ganglio-hypophyseal
Anlage is precisely the same as that which we have already
seen in the way of ascertaining the source of the pericardium.
In Goodsizta it will be remembered that we found both
hypophyseal duct and pericardium to arise as_ well-defined
evaginations from the primitive vesicle. In Perophora, on the
contrary, we have already seen the pericardium arise, almost
certainly from the inner vesicle, but, instead of by an evagina-
tion, by a cell egression so difficult to observe that a trace of
doubt might be entertained as to whether the cells have in
reality come from the vesicle or from the surrounding blood,
or mesenchyme cells. We shall now see the ganglio-hypo-
physeal Ax/age to arise in the same way as the pericardium,
and hence in the same way its source is not as absolutely
beyond question as might be desired. In both genera it
appears as though the influences which have determined the
character of the development of one of these structures have
also determined that of the other.

Since in Perophora the cloudy point is not the origin of the
ganglion alone, as in Goodsiria, but of the common Ax/age of
duct and ganglion, my reference to the four different alleged
sources of the ganglion, made when treating the development
of that organ in the latter species, should be recalled here
and considered in connection with the origin of the ganglio-
hypophyseal Ax/age.

The supposition that it originates directly from the nerve
ganglion of the parent zooid would, in this instance, be so
obviously unjustifiable that Pizon himself would hardly venture
to make it.

In my preliminary I have declared it to be “absolutely
certain” that the Ax/age does not arise from the ectoderm.

204 RITTER. (Vor. XII.

This declaration was made in the face of a full appreciation of
the general danger there is in unqualified affirmations that
certain developmental processes do zot¢ take place ; but in this
instance I believe such positiveness is justified. I must give
my reasons for believing so with particular care and detail,
because of the slight uncertainty of my conclusion that the
inner vesicle is the source of the Az/age, and furthermore
because of the supposition by Van Beneden et Julin (87) that
the ganglion originates from the ectoderm in the C/lavelina
bud.

It is in the character of the ectoderm and the relation of the
Anlage to it that my conviction finds its justification. The
ectoderm is composed of a single layer of cuboid cells, so large,
regularly placed, and distinctly set off from one another that
they are quite diagrammatic in their clearness in most prepara-
tions. The nuclei are round, and generally sharply contrasted
with the cell-body by their distinct membrane and their less
deeply stained ground substance. They are as a rule situated
somewhat nearer the inner side of the cells.

The inner surfaces of the cells are remarkably even and
clear cut, and the layer of protoplasm forming them appears to
be denser than the rest of the cell-body; at least, it is generally
stained more deeply (Figs. 68, Pl. XVI, and 72 and 74, Pl. XVII).
From this character of the individual cells the inner surface of
the layer which they compose appears in sections as a very sharp
line; and as this line would have to be broken were cells to
enter the body space from the layer, either by cell division, or
by migration, I cannot believe the process could escape all the
search I have devoted to the point. Further than this the dis-
tinctness of the cells from those adjacent to them in the body
space and those composing the Az/age is evidence to the same
end. Their nuclei are in general smaller ; but the most impor-
tant difference is in their behavior toward reagents. By some
methods of treatment, most markedly apparent perhaps in
some specimens preserved in Perenyi’s fluid and stained with
Kleinenberg’s Haematoxylin, the protoplasm of the ecto-
derm cells, after having been decolorized, shows a dark dirty
greenish tint that is entirely characteristic of them, not only

No: ti, BODDING IN GOODSTRIA AND PEROPHORA. 205

as compared with the blood and Ax/age cells, but also as com-
pared with any other cells whatever of the animal.

And with almost all the methods of preparation used the
ectoderm cells stain considerably more deeply, particularly on
their inner sides, than do the other cells with which we are con-
cerned. In some instances where the blood cells are particu-
larly numerous between the ectoderm and the inner vesicle, it
is wholly impossible to decide whether the “mesenchyme”
cells are being given off from the inner vesicle or not, so much
do they resemble the cells of the latter, and so imperfect is the
separating line between them. But in these cases there is
not the least possibility of deception about the distinctness of
the body-space cells from the ectoderm. The difference in
staining and the clear boundary line, as described above, pre-
clude it. Now of course, a critical reader, particularly if he be
inclined to be skeptical, might reply that for a particular
instance it may be true that the case against the ectoderm is as
clear as here presented, and that in this instance no cells are
being given off from it; but that this is very far from proving
that at some other stage in development, or in some other region
this process does not take place. I fully appreciate the weight
of this rejoinder, but in this instance think it wholly over-
balanced. As to the second part of it, I would say that the
descriptions and figures all apply to the ectoderm on the dorsal
side of the bud immediately over the Az/age, or, for stages
before this has appeared, over the region where it will appear,
consequently in the region where we should expect the nervous
system to arise, and the region where, according to Oka and
Van Beneden et Julin, in Botryllus and Clavelina it does arise.
It is also the region in which more than elsewhere the char-
acter of the cells is such that they might most easily be sup-
posed capable of giving origin to newcells. As already shown,
the cells are here cuboid in form and have round nuclei and a
considerable quantity of cell protoplasm. In all other regions,
on the contrary, the cells are flattened, as are their nuclei also,
and their protoplasm is relatively less in quantity. The first
part of the objection is more weighty than the second, but even
this must, I believe, yield to the facts. We shall allow that it

206 RITTER. iVor. XT

is not sufficient to prove that the ectoderm is not giving off
cells into the body space at the tome when the Anlage ts being
formed. It must also be shown that the process does not
take place at any earlier time. But when it is considered that
the An/age is formed at a very early stage in the life of the
bud, it will be seen not to be a matter of great difficulty to
examine a complete series of stages from the very inception of
the bud up to the time when the Az/age is fully formed. This
I have done repeatedly, that is, on numerous series of sections,
and the description of the ectoderm already given applies as
well to one stage as to another. To emphasize this fact I would
again call attention to Figs. 68 and 74, ec., the first from a bud
in which the Az/age has not yet appeared, the second from
one in which it is forming. I thus hope to have successfully
assumed the risk of posztively denying that the nerve ganglion
comes from the ectoderm in the blastozootds of Perophora, even
though I cannot be so positive as to what its source really is.
Concerning the difficulties in the way of deciding whether the
inner vesicle or the “‘mesenchyme’”’ cells are the real source of
the Ax/age, I need only refer the reader to what has already
been said about the difficulties involving the study of the origin
of the pericardium. The well-nigh universal distinctness of
the line of separation between the Az/age and the wall, the
lack of compactness of the Az/age in its early stages, and the
close resemblance of its cells to many of the surrounding blood
cells, here, in the same degree as there, suggest the latter as
being the source of the Anzlage. Fig. 72, Pl. XVII, affords a
typical illustration of this. It is drawn from the section in
which, if in any one of the series, the separating line is inter-
rupted by cells passing from the wall to the Axlage. At
the middle point of the separating line there seems to be such
an interruption.

By far the most convincing evidence that I have of the origin
of the Anxlage from the vesicle consists of the occurrence in a
single specimen of what is almost certainly az zmperfect evagt-
nation by which itis formed. Fig. 74, Pl. XVII, represents the
section in which this is most clearly seen. There is no
doubt that the separating line is interrupted here at the point

No.1.] BUDDING IN GOODSIRIA AND PEROPHORA. 207

gl.ev., which I believe to be the evagination ; and there is also
no doubt that the cells of the Az/age are continuous with those
of the vesicle wall. While it is true that, as shown by the
figure, the evagination is not clearly defined either as to its
cellular wall or as to its cavity, I am unable to see any valid
objection against regarding it as such; the chief difficulty, so
far as I see, lies in the fact that the instance is an isolated one.
It is impossible that such a condition can be general, since, as
already said, I have seen it in a single bud only among the
large number examined of practically the same developmental
stage. It may be asked if we have not here a rather late stage
of development in which the mouth of the hypophysis is already
formed ; and if the opening here figured does not represent it.
But when the entire series of sections is examined and com-
pared with those from younger and older buds, there is really
no ground for question on this point. The hypophysis mouth
is not formed till a considerably later time, and it is always a
well-defined opening, but smaller than this. It is my strong
belief that originally the ganglio-hypophyseal Anlage arose from
the inner vesicle by an evagination in Perophora, just as it does
in Goodstria, Botryllus, Glossophorum, and Distaplia. For
some unaccountable reason the process has undergone second-
ary modification, till in most cases, in P. annectens at least, the
evagination has been wholly replaced by a cell proliferation.
Occasionally, however, the earlier evaginate process is reverted
to; and such a case is presented in the individual shown in
Fig. 74. The individual variation here seems to be somewhat
similar to that which occurs in the gastrulation of Aurela
flavidula, where it has been recently shown, Miss Hyde (94),
that in some instances the endoderm is formed by a regular
invagination, while in others it is formed, in part at least, by
delamination and by inwandering of the blastosphere cells.
The completion of the hypophyseal duct and the differentia-
tion of the ganglion from its dorsal wall takes place relatively
later here than in Goodsiria. It will be remembered that in
the latter species the duct is well formed, and the ganglion
entirely separated from it at a stage when the peribranchial
sacs are but just begun and the endostyle is still scarcely

208 RITTER. (Vou. XII.

indicated. In Pevophora, on the other hand, the ganglion is
not fully separted from the duct till the peribranchial sacs are
well developed, and the stigmata have begun to form. Like-
wise the endostyle is well advanced to its final form.

The first differentiation that occurs in the solid, irregular
cell mass, that at first constitutes the Avz/age, consists in the
modeling of this into a quite regular cylinder, in which there
soon appears alumen. The wall of the tube thus formed is
several cells thick, and is of about equal thickness on all sides
throughout its length. Before the lumen is wholly completed
a fusion between the walls of the anterior end of the tube and
of the branchial sac takes place, the fused area becomes per-
forated, and thus the hypophysis mouth is produced. The
formation of the ganglion begins by a rapid proliferation of
cells in the dorsal wall of the duct. Nearly the entire length
of the duct participates in the process, although in the nearly
adult condition the ganglion does not extend entirely to either
end of the duct. Fig. 76 represents a transverse section of a
duct at an early stage in the development of the ganglion, and
Fig. 75 a longitudinal section of a much later stage —a stage,
in fact, when the ganglion is almost completely separated from
the duct.

The intervening stages I have not thought it necessary to
figure. They occur in many of my sections, however, and the
whole process is very clear, and easily observed. The separa-
tion takes place considerably earlier in some buds than in
others, and in a majority of cases it is completed before a
stage so late as that shown by Fig. 75.

I may, perhaps, here refer again to the fact that the ganglion
in this species develops on the dorsal side of the duct, while it
develops on its ventral side in Goodszria.

It is unnecessary to follow the development further. In
a short time the ganglion reaches a diameter considerably
exceeding that of the duct, and it acquires the characteristic
mantle of ganglion cells.

A single developmental point has been omitted that ought,
perhaps, to be mentioned. In a few individuals I have noticed
a distinct thickening in the ventral wall of the tube. Where

No.1.]} BUDDING IN GOODSIRIA AND PEROPHORA. 209

this has occurred most conspicuously it has been before the
appearance of the ganglionic thickening in the dorsal wall. It
is possible that this may be an embryonic trace of the glandular
portion of the hypophysis that is so well developed in some
Ascidians. There is no particular evidence for this, but the
fact that this structure is known to undergo degeneration in
some Ascidians gives the suggestion some probability.

In my description of Perophora annectens (94) I have
called attention to the fact that the posterior ends of the
ganglion and duct are not situated in the median plane of the
body, but are considerably to one side. This condition is
assumed early in the development of the organs, and is almost
always quite pronounced, but I do not see that it has any par-
ticular significance.

C. GENERAL CONCLUSIONS AND REFLECTIONS.

1. APPLICATION OF GENERAL DEVELOPMENTAL PRINCIPLES
TO AN INTERPRETATION OF THE FACTs.

It being now, as I believe, fully established that the nervous
system of the Ascidian bud has a different origin from that of
the embryo, we must seek for an explanation for so anomalous
a fact, —for the processes of evolution are of more importance
from a scientific point of view than are their particular products.

The explanation, which I think to be the true one, has already
been pointed out, first by Seeliger (85), and more recently by
Hjort (95). The latter, in particular, has dwelt quite fully
upon the point. It is, however, a matter of such importance
that I believe it deserves to be emphasized still further. This
is an instance where nature herself has performed an experi-
ment in modifying the ordinary course of development. The
revolutionary and comparatively crude, but still, in consider-
able degree, successful efforts of experimental embryologists to
artificially divert cells that would normally become ectoderm
into entoderm, or structures ordinarily of endodermic origin,
have been deservedly held to be of the highest moment for
solving the fundamental problems of animal development. Cer-

210 TITTLE RR. [VoL. XII.

tainly, then, if cases can be found where nature, with her
conservative and infinitely delicate methods, has entered upon
the same general line of experimentation, and has carried her
efforts through to complete success, such cases cannot be too
carefully studied.

Let us observe with attention the state of things in the
Tunicate bud. The ectoderm is part and parcel of the ectoderm
of the parent (this is strictly true in forms like Goodsiria and
Botryllus, where no stolon is present ; and is also essentially
true when the budding is stolonic, since here the ectoderm of
the stolon is only a prolongation of that of the parent). This
is equivalent to saying that the ectoderm of the bud is not an
embryonic structure at all. It is, on the contrary, a differentiated
organ. Its function in the parent is to secrete the cellulose
test, and in the bud from the very earliest stage it has the
same function. The specialization of this secretory function
must be deep-seated, for, as Hjort has pointed out, the cellu-
lose character of the test necessitates this. Furthermore, not
only is the specialization deep-seated, but also there must be a
great and constant activity of the cells; for not only is the test
considerable in quantity, but it must be perpetually produced
through the whole life of the zooids to replace the continual
waste that is taking place from the external surface. One not
infrequently finds great quantities of diatomes embedded in the
surface layers of test, and it is well known that many species
of animals, particularly small Crustacea, work their way into
the test of Ascidians and there lead a semi-parasitic life. Even
where no foreign organisms were present, I have often observed
the surface test in various species to be eroded and ragged.

The ectoderm, then, has a well-established physiological
function to perform in the bud from its very earliest stage of
development. How is it with the endoderm? It is scarcely
possible to see how a structure could be more favorably situated
for retaining, so far as its functional relation to the organism
as a whole is concerned, an undifferentiated character than is the

‘endoderm ” in the early bud. Not only is it wholly protected
from contact with the external world, it being enclosed in the

ectodermic vesicle, but it has nothing to do in the preparation

No.1.] BUDDING [IN GOODSIRIA AND PEROPHORA. 211

of its own food, for it is entirely bathed in the maternal blood.
It is relieved from all offices except to assimilate already
digested food, and to grow.

Why should not the production of structures which in the
embryonic development belongs to the ectoderm be here trans-
ferred to the endoderm? And soitis. This conclusion is the
more justified when it is considered how different are the con-
ditions under which the two layers develop in the embryo.
Here the neural canal is formed while yet the ectoderm is
strictly an embryonic structure, and before the endoderm
(particularly in the compound Ascidians, e.g. Amaroeciui,
Maurice et Schulgin, 'g4) is differentiated from the richly laden
yolk cells which ultimately give it origin.

We have here an instance where physiological requirements
have run counter to the course in which, through heredity
alone, development would proceed; and the former have
proved more powerful. To my mind the chief difficulty in
such a case is not that developmental conditions can so pro-
foundly modify the usual course of things, but that by such
different courses precisely the same result should be reached.
In Perophora, for example, no one has detected any difference
between the adult embryozooid and the adult blastozooid. The
same is true of C/avelina, and this case is more important,
perhaps, because the embryology of this genus has been very
thoroughly studied up to the practically adult state.

It appears certain that heredity has here an w/¢¢mate aim, as
we may call it, and that it is able to reach this even though
it be thrown considerably off its regular course by special func-
tional requirements ; in other words, that heredity is prospective.

In this connection one may refer to the fact that in some
compound Ascidians, e.g. Botryllus, there is no such thing as
an adult embryozooid, and the suggestion may be made that
the much abbreviated life of the embryozooid in such instances
is in some way correlated with the profound modification
undergone in the development of the blastozooid as compared
with the embryozooid.

I have already pointed out on another occasion ('95a) that at
least one other instance of anomalous bud origin may receive

212 PIT TER. iVon. X11,

a physiological explanation somewhat similar to that above
supported as the cause of the course of things in the Ascidian
bud. I refer to the ectodermal origin of the bud in Rathkea
octopunctata, as recently described by Chun (95). The buds of
this medusa develop on the manubrium, and on that portion of
it in which the endoderm cells, as is clearly shown by the
author’s description and figures, are highly modified for the
digestive function. On the other hand, the ectoderm cells are
as clearly much less highly modified. When we look for a
reason why the ectoderm retains to so considerable a degree
its undifferentiated character, we find a sufficient one in the
fact that it is largely relieved of the protective function that
usually belongs to this layer in adult animals by its being itself
well protected by the deep, close sub-umbrella of this medusa.
And in the two facts that the buds arise in a region where the
endoderm is highly modified for the digestive function, and
that the ectoderm remains comparatively unmodified, appear,
as I believe, an adequate reason why the ectoderm alone
contributes to the formation of the buds.

Braem (95), in puzzling over this case, suggests that the
most probable explanation of the anomalous condition is to be
found in the fact that the buds arise from the same layer, and
not remote from the position where the sexual cells are pro-
duced.. _But—Chun states, page 32> that “Mann trifft keine
Sporen oder parthenogenetischen Eizellen an, welche durch
Grosse und abweichendes Verhalten des Inhaltes sich von den
ibrigen Ektodermzellen abheben.”’

Of course, if my explanation of this case is correct, we may
expect to find other instances where similar conditions and
causes will have produced like results ; as, for example, in the
budding medusa of Bougainvillia Niobe recently described by
Mayer (94), which is likewise a deep-belled form, and in which
the buds “spring from the gastric region manubrium.” But I
do not think that if in a particular instance, where the condi-
tions are right to produce the results, they still do not appear,
we should on that account be justified in concluding their
inefficiency to produce them. It is highly probable that, in
groups of animals which reproduce by budding, the faculty is

No.1.] BUDDING IN GOODSIRIA AND PEROPHORA. 213

acquired independently by different species and at different
times.

Now it is certain, both on theoretical and on observational
grounds, that there would always be a tendency, and a strong
one, for all the germ layers to participate in the production of
the bud ; and should they be found to do so in particular cases
where the conditions are such as to cause, according to my
idea, one or the other of the layers to be excluded from the
process, this might be due to the fact that such exclusion had
not yet been fully effected because of the comparatively recent
acquisition of the property of budding.

2. ON THE QUESTION OF DIFFERENT TyYpPEs OF BUDDING
AMONG TUNICATES.

Having now before us the facts concerning the budding in
these two genera, we must compare them in order to see
whether the differences must be regarded as fundamental, or
whether the mode of development in the two cases is more
probably a modification of a common type.

The first point to be considered is the relation of the bud to
its parent. Various authors have expressed more or less posi-
tively the view that the two modes of bud origin represented
by Perophora and Goodsiria, t.e. the stolonic and the pallial,
have been derived from a common primitive type. Garstang
(95) is the most recent of these, and he has adopted this inter-
pretation apparently with considerable confidence.

It is interesting to notice that the relation between the
blastozooids of a colony in these two genera may be viewed in
such a way as to give the appearance of a rather close resem-
blance.

If we disregard the embryozooid from which the colony has
sprung in each case, and fix our attention upon the adult
blastozooids only, we may reduce the entire colony to a series
of individuals connected with one another by their peribran-
chial sacs through the stolonic septum, or cloison.

This is evident enough in Perophora, where it is essentially
the actual state of things ; the only modification of it being

214 RITTER. [Vou XII.

that in some instances at least, as already shown, the connec-
tion between the zooid and the stolonic septum becomes
severed.

To see how the same relations would be produced in Gooa-
sirta, it is only necessary to imagine the bud to retain the
connection with its parent, and for the connecting neck, shown
almost severed in Fig. 15, Pl. XII, to become more elongated ;
in other words, to form a stolon.

Text figures 1 and 2 illustrate the scheme that is here
described, 1 representing Goodsivia and 2 Perophora ; d indi-

CAGAGAGAG-

Fic. 1.

Be nhs prs

¢ Ph S: d ¢ g f

Fic: 2.

cates daughter and / parent ; 4.s. branchial sac, #0.s. peribran-
chial sac, and s¢. stolon. The stolons in Goodsiria are not
connected to the daughter zooids, since we do not know
whether the connection would be to the branchial or peribran-
chial sacs.

But when we come to examine the subject more closely we
find that the likeness of the two conditions is much less strong
than it at first appears. In the first place we see that the
blastozooids of the Goodszrza colony would stand in the relation
to one another of mother, daughter, granddaughter, etc., while
in Perophora this is not so. Here the blastozooids are rather
all to be looked upon as sisters ; that is, they are all produced
by the stolon, and we have no facts to indicate that the stolon
ever arises from any zooid excepting the original embryozooid
of the colony. Another distinction that results directly from

No.1.] BUDDING [N GOODSIRIA AND PEROPHORA. 215

the one just pointed out is this: in Goodszria it would always
be the mother ascidiozooid that would be attached to the
stolonic septum by its peribranchial sac, while in Perophora it
is the daughter zooid that is so attached. Or, to express the
same thing in another way, we do not know how any given
blastozooid of Goodstria would be attached to the supposed
stolon from which it would arise. We only know how it would
be attached to the stolon that would arise from 7¢.

In reality, then, there is a rather deep-seated difference
between the processes in the two genera. I am, however, far
from affirming that the difference is fundamental. It is not at
all impossible that they may be modifications, though rather
profound ones, of a common original process. How is the
question to be decided? The answer I wish to give with
emphasis. If it is ever to be correctly answered, it must be
by considering the evidence afforded by the d/astogenesis in
connection with the evidence to be derived from embryogenesis
and from the comparative anatomy of the adults.

It was my original plan to make the discussion of the ques-
tion as full as the data at hand would permit ; and with this in
view I have already prepared the manuscript for the anatomical
comparison, and in fact have presented it in brief on another
occasion. Further reflection has, however, convinced me that
it will be better not to attempt a full description until the
embryological data for both genera, now almost wholly wanting,
have been furnished. We must know how the peribranchial
sacs arise in the embryo in each case, and also how the first
bud arises in each case, before we have a basis on which to
profitably speculate.

It will be worth while to point out briefly how unsatisfactory
such speculation would of necessity be in view of our imperfect
knowledge of the ontogeny in these genera.

We have no observations on the origin of the peribranchial
sacs in the embryo of either Perophora or Goodsiria. Neither
do we know how the first bud arises in either case. We might
infer that in these particulars Pevophora is like Clavelina ; in
fact, such an assumption is generally made. But as I have
shown the relation of the blastozooids to the stolon to be very

216 RITTER. (VoL. ATT:

different in the two cases, it seems rather probable, at least
not at all improbable, that there may be important differences
between them in the two points mentioned, vzz., in the origin
of the peribranchial sacs and the first buds.

As regards Goodsiria we may assume that the embryonic
development is essentially like that in Botryllus. For my own
part I fully expect that this will prove to be the case. How-
ever, we are by no means certain of it, and even if we were we
should be far from clear sailing as regards the origin of the
peribranchial sacs of the embryo, since Pizon insists that they
arise from the endoderm of Sotry/lus, and Garstang is inclined
to accept his conclusion. If the peribranchial sacs of Perophora
arise from the ectoderm as they do in Clavelina, and 7f the
stolonic septum of the first bud of Perophora arises from the
pharynx as it does in Clavelina, and 7f the first bud of Goodstria
arises like the first bud of Botryl/us, and zf the peribranchial
sacs of Botryl/us arise from the ectoderm, then weighing all
the evidence together, anatomical as well as developmental, it
seems to me that the budding in the one genus is genetically
independent of the budding in the other.

This view I have practically expressed before ('95a), in con-
tending that the Compound Ascidians have had a double
origin from the Simple ones.

But, as already said, I regard it as necessary that the above
formidable array of ifs should be gotten out of the way before
the discussion can profitably be carried farther.

And now concerning the “epicardiac tubes”’ that are of so
much importance in connection with Tunicate budding. I have
elsewhere (95) said, regarding the origin of the pericardial
vesicle in Goodsiria, that an epicardium is here present “as in
Botryllus.’ As will be readily seen by comparing Figs. 18 and
19, Pl. XIII, with, for example, the woodcut given by Pizon
(93), page 30, of Botryllus, there can be no doubt that the
posterior extremities of the peribranchial sacs in my figures of
Goodsiria are the same as those marked g.v. by Pizon. The
only difference in the two cases is in the size. The two
pouches are much longer in Botry//us than in Goodsiria. Both
Pizon ('93) and Garstang (94) regard the structures in Botryllus

No.1.] BUDDING IN GOODSIRIA AND PEROPHORA. 217

as homologous with the epicardium of Clavelina, Distaplia,
Fragrotdes, and other species.

Thus the last-named author says: ‘In spite, therefore, of
the final difference of position between the epicardial (or peri-
visceral) sacs of the Botryllidae and the epicardial tubes of
Distaplia or Clavelina, there can be no doubt, as Pizon has
maintained, that there is an exact homology between the two
structures.’ Nevertheless, there is, in my mind, a very grave
doubt that the structures in Goodstria, which are certainly the
same as those called epicardial sacs in Botryllus, have anything
whatever, either morphologically or physiologically, to do with
the epicardial tubes of Clavelina.

It is well known from the investigations of Seeliger and
Van Beneden et Julin that the epicardial tubes of C/avelina
arise from the branchial sac, and they arise in the same way in
Glossophorum (Hjort, '95) and Fragrotdes (Maurice, ’gs). As
Pizon and Garstang contend, the mere fact that the structures
arise from the branchial sac of some species, and from the peri-
branchial sacs in others, would not in itself present any diffi-
culty against regarding them as genetically homologous (and I
take this to be the kind of homology that these authors mean,
for I do not suppose they recognize any other kind). But
this view would have to regard it as proven that the branchial
and peribranchial sacs both arise from the endoderm in the
embryozooid ; and this can certainly not be granted as our
knowledge now stands.

As previously said, Pizon claims that such is their origin in
the Botryllus embryo; and Garstang is inclined to support the
same view. Should it finally be established that this is in
reality the case, then, so far as this much of the evidence is
concerned, the structures may be homologous in Botryllus and
Clavelina, and of course in Goodsiria, also, if its embryonic
development here is similar to what it is held by Pizon to be
for Botryllus ; but this is still to be shown.

But now it being conceded that, so far as concerns the
evidence yet in hand relating to the origin of these structures,
they may fosszbly be homologous, we must still consider what
the evidence of their destiny is.

218 RITTER. [Vou XI1.

In Clavelina, Distaplia, Glossophorum, and Fragroides the
epicardium gives origin either directly or through the stolonic
septum to the inner vesicle, or endoderm, of the bud. In
Botryllus and Goodsiria the so-called epicardial sacs have noth-
ing whatever to do with the budding, and there is not the
slightest evidence that they ever had anything to do with the
process. The last part of this statement is at least true for
Goodsiria. Here the buds arise, as reference to Figs. 9 and Io,
Pl. XII, will show, far forward on the parent zooids, while the
epicardial pouches are at the extreme posterior end of the
animal. All they are is this: When the intestine begins to
form at the postero-ventral side of the primitive inner vesicle
it produces, as one might say, an obstacle in the way of the
further expansion posteriorward of the inner vesicle, which, of
course, is constantly growing. The notch thus produced
has been already described, and is well seen in Fig. 16, d.,
Pei:

The extensions of the vesicle on each side of this notch are
the epicardial sacs of the adult animal. When the peribranchial
sacs are completely formed the epicardial sacs are merely their
posterior extremities.

And now a word concerning the relation of the epicardium
to the heart. In attempting to show that the epicardium is
the same structure in all Tunicates, and to make use of it asa
basis for classifying the group, Garstang has tried to escape
the difficulty of its being in some cases connected with the
heart, while in other cases it is not, by supposing that such
connection is secondary.

In my account of the origin of the heart in Goodszrza I have
shown that it arises from the right epicardial pouch, and I
can certainly see no reason for regarding this method of origin
as secondary. In fact, it was partly this consideration that
made me speak of the pouch as an epicardium in my preliminary
communication.

Concerning the epicardium in Pevrophora, after describing
the relation of the bud to the stolon, I have said in my prelim-
inary paper: “It is obvious that there cannot be in Perophora
an epicardium corresponding to the structure called by that

No.1.] BUDDING IN GOODSIRIA AND PEROPHORA. 219

name in C/avelina, since in this species the epicardium is
connected with the branchial sac.”

Lefevre (95) says: ‘No epicardium is present; in this
respect Perophora differs strikingly from Clavelina and some
other Ascidians.” So far as the Jdlastozooids are concerned
these unqualified statements are, I believe, fully justified. It
must, however, be remembered that we do not know how the
stolon originates from the embryozooid, and until we are
informed on this point I must place a certain reservation
on my assertion of the entire absence of the epicardium in
Perophora.

If we accept Garstang’s view that the relation of the heart
to the epicardium is secondary, then the fact that the heart
arises on one side of the body, while the stolonic septum is
attached to the peribranchial sac of the other side in Pevophora,
would be of little weight against supposing an epicardium to be
present in this species. But I have already shown that this
author’s conjecture is contradicted by the evidence of Goodsi7za,
if he would still maintain that the pouches described in this
species are homologous with the epicardium of Clavelina.

In the present state of our knowledge on this point, then, I
am a long way from conceding, as Garstang thinks we must,
“that these modifications of the epicardial tubes provide a
sound basis for a true and genetic classification of Tunicate
budding.” That, when considered in connection with various
other developmental and anatomical facts, it is of prime import-
ance in interpreting the budding, cannot be questioned. But
the attempt to make it, in itself, a basis for classifying the Com-
pound Ascidians can hardly be more satisfactory than one-
legged classifications ever are in zoology.

220 RITTER. [Vou. XII.

D. SUMMARY OF RESULTS.
GOODSIRIA DURA.

1. The budding is pallial, and the buds arise far forward on
the parent zooid. In no case has more than one bud been
seen on the same parent.

2. No “budding zone” is recognizable in zooids on which
buds are not developing.

3. The buds become wholly separated from the parent
zooids at a very early stage, z.e. before any differentiation of
organs begins; and at a considerably later time they become
secondarily connected with the vessels of the test.

4. The ampullae of the testicular vessels do not produce
buds.

5. The formation of the branchial and peribranchial sacs,
and of the digestive tract from the primitive inner, or endo-
dermic, vesicle of the bud does not differ in any essential
particular from that of all other Compound Ascidians.

6. The common Ax/age of the pericardium and heart is de-
rived from the wallof the endodermic vesicle by an imperfect
evagination that does not become fully separated until the
ventral partitioning folds which separate the peribranchial sacs
from the branchial sac have reached back to the region where
the heart is forming.

7. The ganglio-hypophyseal Ax/age arises as a gutter-shaped |
evagination from the dorsal side of the endodermic vesicle. As
this closes off to become the hypophyseal duct, the ganglion is
produced from the mass of cells that forms the last of the con-
nection of the duct to the endodermic vesicle. The ganglion,
therefore, in this species as in Botryllus, lies ventral to the
hypophyseal duct.

8. The youngest sexual cells observed were found free in
the body space of the buds, so that in all probability they pass
from parent to bud as is the case in Botryllus. But in none
of the material available for study were the elements present
in sufficient quantity and development to make it possible to
give a complete history of them. The ‘polycarps”’ appear to

No.1.] BUDDING IN GOODSIRIA AND PEROPHORA. 221

be confined to a single row on each side of the endostyle, and
not far from it.

PEROPHORA ANNECTENS.

1. The inner or endodermic vesicle of the bud is derived
from the stolonic septum; and from this are derived the
branchial and peribranchial sacs, and the digestive tract in a
manner in all essentials similar to that of all other Ascidian
buds.

2. The peribranchial sacs of the developing blastozooid are
formed in such a manner that the stolonic septum is connected
to the left one of these, and not to the branchial sac, as has
been hitherto supposed.

3. The pericardial Az/age arises almost certainly from the
wall of the inner vesicle, though there remains some doubt
whether or not it may be produced by an aggregation of
mesenchyme cells. But if this is so it is still probable that its
ultimate source is the endoderm, since the mesenchyme cells
that seem to enter into it are themselves probably produced
by the endoderm.

4. The ganglio-hypophyseal Az/age was originally produced
from the endodermic vesicle as an evagination, as it is in the
buds of numerous other Ascidians. This primitive method of
origin is, however, found in an occasional individual only at
the present time, the evagination having, in most cases, been
replaced by a migration of cells.

For the most part this migration takes place at the point at
which the ganglion will be ultimately situated; but it may
occur at other points more or less remote from this, and the
coming together of these migrating cells makes it appear as
though the Az/age is produced from mesenchyme cells. Or,
as in the case of the pericardial Az/age, we may consider that
mesenchyme cells do participate in the formation of the Ax/age,
but that these cells are derived from the endodermic vesicle.

222 RITTER. [Vor. XII.

In GENERAL.

1. It is now established beyond question that in some, at
least (and Goodsiria may be taken as one of the best instances
of this), of the Compound Ascidians the outer layer of the bud
contributes much less to the structure of the adult blastozooids
than it does to the adult embryozooids. This is most conspicu-
ously seen in the case of the nervous system, for this is cer-
tainly produced from the outer, or ectodermic layer, of the
embryo, while it is as certainly produced from the inner layer
of the bud. Whether we call this inner layer endoderm or
not, the fact of chief importance remains that the same layer
produces most of the organs of the zooid, among which are
included the digestive tract and the nerve ganglion.

2. The anomalous course of development of the bud is due
to the fact that the ectoderm is at no time in the life of the
bud an undifferentiated, or embryonic layer. It is from the
very outset and always a fully formed organ, its function
being to secrete the cellulose matrix of the test. The inner,
or endodermic, vesicle of the bud is, on the other hand, in the
completest sense, an undifferentiated, or embryonic layer. By
this purely physiological cause the inner layer has been made
to produce structures, most important of all the nervous
system, which in the embryo are produced by the ectoderm.

3. Illustrations of the potency of physiological influences to
profoundly change the usual course of development are found
in the budding of other animals, one of the most instructive of
which is the medusa Rathkea octopunctata, where the change
has been in the opposite direction from that in the Ascidians ;
for here the endoderm takes no part in the formation of the
bud, the whole blasto-medusa being produced from the maternal
ectoderm.

4. The evidence now in hand, drawn from adult structure
and from the blastogenesis strongly tends to the conclusion
that the budding of Goodsiria and Botryllus, represents a type
that is genetically independent of the type represented by
Perophora. In other words, that the type of budding rep-
resented by Goodsiria, has originated independently of the

No.1.] BUDDING IN GOODSIRIA AND PEROPHORA. 223

type represented by Pevophora. But more knowledge of the
embryology of each genus is requisite before this or any other
conclusion respecting their relationships will be fully warranted.

5. In neither Goodstvza nor Perophora is there an epicar-
dium comparable with the structure called by that name in
Clavelina and many other Compound Ascidians.

6. The budding of Goodstria greatly strengthens the con-
clusion justified by adult structure that the Polystyelidae and
Botryllidae are very closely related. The two families will
probably have to be ultimately united into one, but it is not
best to do this until we know the embryology of both groups
much more fully than we yet do.

UNIVERSITY OF CALIFORNIA, BERKELEY,
Nov. 11, 1895.

224 RITTER. [Vor. XII.

BIBLIOGRAPHY.

'16 Savicny, J. C. Mémoires sur les animaux sans vertébres. II Partie.
Paris, 1816.

43-50 CaRuS, V. On the Zodlogy of the Scilly Islands. Proc. Ashmol.
Sc. Vole le

'69 METCHNEKOFF, E. Entwickelungsgeschichtliche Beitrage. Ueber
die Larven und Knospen von Botryllus. S8ull. Acad. Imp. Se.
St. Petersbg., tome XIII, p. 291.

'72 GIARD, ALFRED. Recherches sur les synascidies. Archives de
Zoologie Expérimentale et Générale, tome 1, also Théses, pp.
1-204.

'74 GIARD, ALFRED. Note sur quelques points de l’embryogénie des
Ascidies. Assoc. Franc. pour Vavancem. dad. Sct., tome III, pp.
432-458.

'74a KOWALEVSKY, Pror. A. Ueber die Knospung der Ascidien. Archiv
fiir mikroskopische Anatomie, to. Band, pp. 341-470.

'74b KOWALEVSKY, Pror. A. Sur le bourgeonnement du Perophora
Listert Wiegm. (Traduit du Russe par le Prof. A. Giard.) Revue
des Sciences Naturelles, 1874.

'74 LACAZE-DUTHIERS, HENRI DE. Les Ascidies Simples des Cétes de
France. Archives de Zoologie Expérimentale, tome III, pp.
257-330.

’'81 JuLIN, CHARLES. Recherches sur l’organisation des Ascidies simples.
Sur Vhypophyse et quelques organes qui s’y rattachent. Archives
de Biologie, tome II, pp. 59-126, and 211-232.

'82 DELLA VALLE, A. Recherches sur l’anatomie des Ascidies com-
posées. Arch. Ital. de Biologie, tome II, 1882, pp. 9-49.

HERDMAN, W. A. Report on the Tunicata collected during the
Voyage of H.M.S. Challenger. Part I, Ascidiae Simplices.

’84 VAN BENEDEN ET JULIN. Recherches sur le développement postem-
bryonnaire d’une Phallusie (Phallusia scabroides nov. sp.). Archives
de Biologie, tome V, pp. 611-638.

MAURICE, CH. ET SCHULGIN. Embryogenésie de L’Amaroeceum
Proliferum. Axnales des Sci. Nat., Zool., tome XVII.

ULJANIN, DR. BAstLius. Die Arten der Gattung Doliolum im Golfe
von Neapel und den angrenzenden Meeresabschnitten. Fauna und
Flora des Golfes von Neapel. X Monographie.

’°8S RouLe. Recherches sur les Ascidies Simples des Cétes de Province
(Phallusidées). Annales du Musée d'Histoire Naturelles de Mar-
seille. Zoologie 2, 1884-85, pp. 1-197.

85a SEELIGER, O. Die Entwicklungsgeschichte der socialen Ascidien.
JSenaische Zeitschrift fiir Naturwissenschafe. 18. Band, pp.
45-120.

No.1.] BUDDING IN GOODSIRIA AND PEROPHORA. 225

'85b SEELIGER, OSWALD. Die Knospung der Salpen. Jena. Zettschr.,

86

87

’88

‘89

90

"91

92

93

94

1885, pp. 573-677.

HERDMAN, W. A. Report on the Tunicata collected during the
Voyage of H.M.S. Challenger. Part II, Ascidiae Compositae.

VAN BENEDEN ET JULIN. Recherches sur la morphologie des Tuni-
ciers. Archives de Biologie, tome V1, pp. 237-476.

Maurice, M. CHARLES. Etude monographique d’une espéce d’Ascidie
composée (Fragaroides aurantiacum n.sp.). Archives de Biologie,
tome VIII, pp. 205-4096.

SHELDON, LILIAN. Note on the Ciliated Pit of Ascidians and its
Relation to the Nerveganglion and so-called Hypophyseal Gland ;
and an Account of the Anatomy of Cynthia Rustica. Quarterly
Journal of Microscopical Science, vol. 28, New Series, pp. 131-148.

SEELIGER, OswALp. Zur Entwicklungsgeschichte der Pyrosomen.
Jenaische Zeitschrift fiir Naturwissenschaft. 23. Band, pp.
595-658.

LAHILLE, F. Recherches sur les Tuniciers des Cétes de France.
Theses, Toulouse, pp. 1-530.

HERDMAN, W. A. Revised Classification of the Tunicata. /ournal
of Linnean Society, Zoology, vol. XXIII.

HERDMAN, W. A. Note on Atrial, or Circumcloacal, Tentacles in
the Tunicata. Read at British Association, 1892.

Oka, A. Ueber die Knospung der Botrylliden. Zeztschrift fiir
wissenschaftliche Zoologie. 54. Band, pp. 521-547.

SALENSKY, W. Beitrage zur Embryonalentwicklung der Pyrosomen.
Zoologische Jahrbucher, Abthetlung fiir Anatomie und Ontogénie
der Thiere. 5. Band, pp. 1-98.

Brooks, W. K. The Genus Salpa. A Monograph with fifty-seven
Plates. With a supplementary paper by Maynard M. Metcalf,
Baltimore. Zhe Johns Hopkins Press.

Hjort, J. Ueber den Entwicklungscyclus der zusammengesetzten
Ascidien. J7itthetlungen aus der Zoologischen Station zu Neapel.
10. Band, 1891-1893, pp. 584-617.

Pizon, M. A. Histoire de la Blastogénése chez les Botryllides.
Annales des Sciences Naturelles. Tome XIV, pp. 1-386.

WILLEY, ARTHUR. Studies on the Protochordata, I. Quarterly Journal
of Microscopical Science. Vol. 34, New Series, pp. 317-360.

CAULLERY, Maur. Sur le bourgeonement des Diflosomidae et des
Didemnidae. Compt. rend. Ac. Sci. Paris. Tome 119, No. 8,
PP: 437-439.

Hype, Ina H. Entwick. einiger Scyphomedusen. Zezfsch. f. wiss.
Zool., 58. Band, pp. 531-565.

GOTTSCHALDT, Ros. Die Synascidien der Bremer Expedition nach
Spitzbergen im Jahre 1889. Jenatsche Zeitsch. fiir Naturw. 28.

Band, pp. 343-369.

226 RITTER. [VoL. XII.

MAYER, ALFRED GOLDSBOROUGH. An Account of some Medusae
obtained in the Bahamas. Bull. Mus. of Comp. Zoology, vol.
XXV, No. II, pp. 235-242.

Ritter, W. E. Tunicata of the Pacific Coast of North America, No.
1. Perophora annectens, n. sp. Proc. Cal. Acad. Sci., vol. IV,
Second Series, pp. 37-80.

SEELIGER, O. Tunicata, Bronn’s Klassen und Ordnungen des Thier-
reichs. Dritter Band, Supplement.

'95 BraEM, F. Was ist ein Keimblatt? Bzologisches Centralblatt. Band
KV, Nos. 01,02) and. 13;

CuuNn, C. Biologische Studien tiber pelagische Organismen. I. Die
Knospungsgesetze der proliferirenden Medusen. Szdliotheca Zoolo-
gica, Heft 19, pp. 3-51.

GARSTANG, W. Budding in Tunicata. Sczence Progress, vol. III,
March.

Hyort, JOHAN. Beitrag zur Keimblatterlehre und Entwicklungs-
mechanik der Ascidienknospung. Avxatomischer Anzeiger, 10.
Band, No. 7, pp. 215-22

Hjyort, JOHAN, U. FRL. BONNEVIE. Ueber die Knospung von
Distaplia Magnilarva. Anatomischer Anzeiger, 10. Band, Nr. 12,
pp. 389-394-

LEFEVRE, GEORGE. On Budding in Perophora. /ohus Hopkins
University Circulars, No. 119, June.

RitTER, W. E. On Budding in Goodsiria and Perophora. <Axa-
tomischer Anzeiger, 10. Band, Nr. 11.

'95a RITTER, W. E. Some Facts and Reflections drawn from a Study of
Budding in Compound Ascidians. British Association for the
Advancement of Science, Ipswich Meeting, 1895.

No.1.] BUDDING IN GOODSIRIA AND PEROPHORA.

227

DESCRIPTIVE LETTERS.

Anterior.

Ampullae of testicular blood
vessels.

Atrial siphon.

Bud.

Bud Ax/lage.

Blood corpuscles.

Branchial sac.

Branchial siphon.

Branchial stigmata An/age.

Branchial epithelium.

Branchial stigmata.

Body space.

Blood vessel.

Cloaca.

Cloison, or stolonic septum.

Cloacal epithelium.

Dorsal.

Dorsal partitioning fold.

Dorsal lamina.

Ectoderm.

Ectodermic, or testicular
blood vessels.

Ectodermic vessels project-
ing into body space.

Endostyle.

Endocarp.

Epithelium of peribranchial
sac.

Female polycarps.

Ganglion.

Ganglionic evagination.

Heart.

Hypophysis Ax/age.

Hypophysis duct.

Internal longitudinal bars.

tnt.
1n.UES.
2p.
l.coe.
ld.
l.pb.s.
Ess
m.en’c.

MeS.LaS.
MeES.VEC.

oe.

Intestine.

Primitive inner vesicle.

Internal papilla.

Lacteal coecum.

Lacteal duct.

Left peribranchial sac.

Lacteal system.

Male polycarp.

Gastric mesentery.

Rectal mesentery.

Oesophagus.

Ova.

Posterior.

An/lage of peribranchial sac.

Pericardium.

Pericardial AnJage.

Partitioning fold.

Polycarp.

Remnant of connection be-
tween bud and _ parent
zooid.

Rectum.

Remnant of evagination.

Right peribranchial sac.

Right ventral partitioning
fold.

Siphonal ectoderm.

Siphonal folds.

Stomach.

Stolon.

Testa.

Testa cells.

Transverse vessel.

Testis.

Ventral.

Ventral partitioning fold.

228 RITTER.

EXPLANATION OF PLATE XII.

GOODSIRIA DURA.

The figures were drawn with the aid of a camera lucida and a Leitz microscope,
excepting when otherwise specified.

Fic. 1. A colony, natural size, completely covering the surface of a piece of
seaweed. No buds are present in this colony. Not camera, but as faithful a
reproduction as possible.

Fic. 2. A colony, also on a piece of seaweed. The Ascidiozooids much less
closely crowded than in the preceding. Buds in various stages of development
are seen, and some of the very young ones are not as near the margin of the
colony as are some of the older ones. The ampullae of the testicular vessels are
seen, particularly near the margin of the colony. Not camera, but as exact as
possible. X 4.

Fic. 3. Small fragment of a thin colony removed from its substratum and
examined as a transparent object in clove oil. X 87.

Fic. 4. Digestive tract with a portion of the branchial sac attached. The
longitudinal folds of the stomach and the single spiral groove of the intestine are
seen. X 87.

Fic. 5. Camera sketch of a section of a colony showing sections of three
blastozooids in various stages of development. The top of the figure corresponds
to the top or dorsal surface of the colony ; it is consequently seen that the blood
vessels of the test, ec.ves., are at a deeper level than are the Ascidiozooids.
xX 87.

Fic. 6. Transverse section of an almost fully formed blastozooid, the section
being well toward the posterior end of the body. X 120.

Fic. 7. An optical section of two ampullae of the testicular vessels which
contain a closely packed mass, 4.c., of blood corpuscles. This, together with the
thick ectodermal wall of these vesicles, give them somewhat the appearance of
young buds. Drawn from a specimen cleared in clove oil. X 87.

Fic. 8. A single male “polycarp” attached to the mantle. The short vas
deferens is here seen. X 120.

Fic. 9. Optical section of an ascidiozooid with a bud, éd., still attached to it.
This is the zooid from which the section shown in Figs. 12 and 13 were cut. The
direction of the section is indicated by the line 4.4.“ The specimen was cleared
in clove oil. Not camera, but as faithful a representation of the object as
possible.

Fic. to. Optical section of a zooid with a young bud, éd. Cleared in clove
Oil mie toe

Fic. 11. Section of the earliest stage seen in the formation of a bud. As
seen, the ectoderm is not yet modified at the point where the bud is forming, the
bud Ax/age, bd.a., being confined to the parietal wall of the peribranchial sac ;
by.st.a. are points where stigmata of the present zooid will form. X 210.

f H _ f i
| f
s if _ \ /
j | |
; j
ed -
a =
eae Ei a
va, - '
2 4
‘hk
‘ i e
My
: \
j om
+e f t
J ‘ -
et i
4

ee

—<

Lith Anst x Werner Winter, Frankfurt at

230 RITTER.

EXPLANATION OF PLATE XIII.

Fic. 12. Section of the budding zooid shown in Fig. 9, the section corre-
sponding to the line 4.4.’ X 87.

Fic. 13. From the same series as 12, but far enough posteriorward so that the
inner layer of the bud is not continuous with the maternal endoderm. At ams.
are seen two ampullae of the testicular vessels which are in close contact with the
young bud, and the vessel, ec.ves., is still more closely pressed by the bud. X 2tIo.

Fic. 14. Transverse section of a bud not yet separated from its parent.
X 210.

Fic. 15. Section of a zooid with a bud, éd., almost severed, there being only
a remnant, ~c., of the connection between bud and parent. The bud is here
entirely undifferentiated as to the organs of the future zooid.

Fic. 16. Optical section of a young bud; the intestine, zz, is established,
and the folds that will ultimately separate the peribranchial sacs from the
branchial sac, are barely indicated at .£ The ganglio-hypophyseal evagination,
hy.cl., is represented as projected on the plane of the section. In reality it
would not be seen at the focus at which the rest of the bud is shown. The begin-
nings of the two so-called epicardial pouches will be observed at Y. Seen from
the dorsal side. X 87.

Fic. 17. A bud somewhat more advanced in development than the preceding.
The specimen was cleared in clove oil. It is seen from the dorsal side, but is
slightly rotated to the right. The right partitioning fold, 2.f, is well seen. The
distribution of the blood cells in the body space is shown. X 87.

Fic. 18. A bud considerably more advanced than the last, the peribranchial
sacs nearly complete. Seen fromthe dorsal side. Viewed as a transparent object,
but not shown in optical section. X 87.

Fic. 19. A still older bud, the differentiation of the organs almost complete.
Method of treatment and view similar to the last three. X 87.

Fics. 20 and 21. Sections of a bud in which the peribranchial sacs, 47.5.a., are
barely begun. The sections are not quite at right angles to the antero-posterior
axis of the bud, and as a consequence the sacs do not both appear on the same
section. X 210.

Fics. 22-24 (and 25, Pl. XIV). Four transverse sections of a bud about
corresponding in its stage of development to the bud shown in Fig. 17. Fig. 22,
the most anterior of the series, passes through the position at which the branchial
siphon, ér.szf., is soon to form. By Fig. 23 it is seen that the folds which are to
separate the branchial from the peribranchial sacs extend along the ventral as
well as along the dorsal side of the primitive inner vesicle, vf anda. In the
last of the four sections, Fig. 25, no trace of the folds appears. This section
passes through the beginning of the atrial siphon, at.st#. X 210.

4 iz >
é
J
: —_
:
: o
=
-~ es
=
oT —
0 Ll
7
< ’
oe }
yh a
tl *

‘

a

q
tc]
2

oto
oo &>

%

inves

°eoeo0e

rpb.s.

™, r y

bs.

tg
Ss
3

brs

er ba

a uf

Tih Anst.v Werner Winter, Frankfare eM.

232 RITTER,

EXPLANATION OF PLATE XIV.

Fics. 26-29. A series parallel to the last, but from a more advanced bud.
The stage of development is about midway between the buds shown in Figs. 17
and 18. X 210.

Fics. 30 and 31. From the posterior end of a bud nearly as far advanced as
the one shownin Fig.19. The sections show that the branchial sac, 67.s.,is wholly
closed off from the peribranchial sacs, 7.4.5. and /.0.s5., and Fig. 31 shows that the
peribranchial sacs unite behind the branchial sac to form the wide atrium, az.
X 87.

Fic. 32. Shows the relation of the digestive tract to the peribranchial sac
before it has become pushed into the latter. X 120.

Fic. 33. Shows the digestive tract pushed fully into the peribranchial sac.
X 210.

Fic. 34. Section of the digestive tract in the practically adult condition. The
lacteal coecum, Z.coe. (this lettering should point to the smaller of the two circular
structures ; the lithographer has moved the letters), the lacteal system, /.s.,
and spiral fold of the intestine, pointed to by the index line, zwt¢., are seen.
X 210.

Fic. 35 (and 36, Pl. XV). Two sections showing stages in the development
of a siphon. Fig. 35 shows that the siphon begins by an evagination of the
endodermic vesicle. X 508.

> h
have

= | “0 5)® CO

iy i)
| brs

comnt

mes. gas.

ee *
— _ _ —— —- ~—- —d
Tah Anst v Werner &Winter PrankrerteM

‘
' (
- ~=_1
: =
; ;

234 RITTER.

EXPLANATION OF PLATE XV.

Fic. 36. A much more advanced stage; shows the peculiar folding, s.é and
s.f’., that is produced before the siphon is complete. X 210.

Fic. 37. Shows two of the testicular vessels, ec.ves’., projecting into the body
space, 4.s., of the developing bud. X 21o.

Fic. 38. Shows two vessels, ec.ves., opening into the body space. X 210.

Fics. 39-42. Relate to the origin of the pericardial vesicle. Fig. 39 shows a
transverse section passing through the posterior portion of the animal, and at
r.pb.s’. is seen the pouch of the right peribranchial sac (the epicardial pouch)
from which the pericardial vesicle is produced. Fig. 39, X 210; the others,
X 508.

Fic. 43. Cross-section of a bud in about the same stage of development as
the one shown in Fig. 16. The section of the gutter-like evagination, Ay.d., of
the ganglio-hypophyseal duct, is here very distinct. X 210.

Fics. 44-47. Four sections of a series passing from before backward of the
ganglio-hypophyseal duct. They are almost, but not quite, at right angles to the
long axis of the duct. Fig. 44 shows the mouth of the duct, and Fig. 47 shows
that at its posterior end, or near the end, the duct still communicates with the
branchial sac, ”.¢.0.

Both the ectoderm and the cells in the body space of these sections are repro-
duced with special care. Figs. 44, 45, and 47, X 508; Fig. 46, X 720.

Fic. 50. Cross-section of the duct and ganglion nearly fully formed. X 508.

Fic. 51. Cross-section of the endostyle showing a young female polycarp.
X 210.

Fics. 53 and 54 (Pl. XVI). Two early stages in the formation of male poly-
carps. X 508.

«
\
-
mi,
‘
*~
—
os
°
a

a

fo
- 7

®. f
284), o¢ 9
A |
V4.8) 915 648 eb®a® aeee”
al v 46
Gee (OK = ER Oe:
i @ %, ee : 7‘
“ee ® 2 : (
a. El
e”
ed) ®%56 2 2. x 2 ao 39 %) ec - aS: |
Taree (ys Pile »~ |
pea! Vv
ad hyd |
gare 47, |
45. pee Be SSlSea5e
€ 0 Se ae B59) eo Faz, ee.
Cae se S Sle, se? ) 235%
pals ie, 3,3" : “Sy ‘ ( .
Sse ° |
I bs ‘4 ores Sn . =
y 928
ray %, a Qe: @ & ; e
LOM s
"= ; He, y * ~
9g 29,09 PRP pmersen®® re.0. *
Vv ~

— — —— = = a "Tih Aust. Werner aWinter, Frankfiert 2.

4

236 RITTER.

EXPLANATION OF PLATE XVI.

Fic. 48. A longitudinal section of the ganglio-hypophyseal duct at a slightly
more advanced stage than the one shown in Figs. 44-47. It will be noted that
here the communication of the duct with the branchial sac, 7.¢.0., the remnant of
the evagination, is almost closed. The differentiation of the ganglion, 9/., from
the duct and from the wall of the branchial sac is complete at its anterior end, but
not at its posterior end. X 720.

Fic. 49. Cross-section of a duct and ganglion after the ganglion is fully
separated, but is still very small. X 720.

Fic. 52. An ovum found alone and free in the body space. X 1200.

Fic. 55. A multinuecliated mass floating free in the body space ; probably(?)
an early stage in the formation of a male polycarp. X 508.

PEROPHORA.

Fic. 56. Small portion of a colony of P. axnectens almost entirely covering a
zooid of Clavelina. The specimen is the fully compounded variety, and forms a
thin encrusting layer on the surface of the C/avelina. The tips of the stolons, s/o.,
and young buds, éd., are shown at the margin of the colony. Not camera. X 2.

Fics. 57-64. From aseries of transverse sections extending from before back-
ward of P. aznectens, illustrating the method by which the bud is connected to the
stolonic partition by its left peribranchial sac. The point of connection is seen in
Fig. 60. X 138.

Fic. 65. Transverse section of a P. Listeri bud, slightly more advanced in
development than the preceding ; illustrating, likewise, the connection of the left
peribranchial sac to the stolonic partition. In this section, however, the series
passes from behind forward, so that the apparent reversal of right and left does
not occur. X 138.

Fic. 66. Shows a case in which the zooid is fully separated from the stolon.
This is the section of a complete series in which the stolon approaches most
closely to the bud, but here the separation is complete. P. azmectens. X 210.

Fics. 67 and 68. Show the development of the pericardial vesicle. Fig. 67
shows its position with reference to the point of attachment of the inner vesicle to
the stolonic septum. Both are from the same series of sections. PP. annectens.
Fig. 67, X 72; Fig. 68, X 330.

Fic. 69. A later stage in the formation of the pericardium, but still before its
separation from the endoderm. X 720.

Fic. 70. From the same series as the last to show the relation of the
forming pericardium to the point of attachment of the stolonic septum to the
primitive inner vesicle. X 120.

Fic. 71. A condition of the pericardium similar to that shown in Fig. 69, but
slightly more advanced in development. X 500.

4 = ~~ LEN.

ety -—

1. ne oat

)

~~ ith Anst.z Werner kWinter, Frankfort 3M.

10

i) a" iA
gM) aes)

May
ane

ia NL ae al we
a as S Oo hae’ : Die) Cry | ' : vn

4
mil bot

Wa

238 RITTER.

EXPLANATION OF PLATE XVII.

Fic. 72. An early stage in the formation of the ganglio-hypophyseal Ax/age
before its separation from the endoderm. Camera drawing as far as possible with
Leitz. oc. 5, obj. 7.

Fic. 73. Shows a stage in the formation of the ganglion somewhat more
advanced than the preceding, the whole section drawn in order to show the distri-
bution of the blood or mesenchyme cells in the body space. X 210.

Fic. 74. From the same series as Fig. 73, more highly magnified, showing the
supposed evagination to form the ganglio-hypophyseal Az/age. X 508.

Fic. 75. Longitudinal section of the ganglion and duct, the former almost
separated from the latter. X 250.

Fic. 76. Transverse section of the duct, with its dorsal wall considerably
thickened for the formation of the ganglion. X 508.

Fics. 77 and 78. Two sections in which cells appear to be breaking away from
the endoderm, and passing into the body space to become mesenchyme cells.
Both of these cases are far from the position where either heart or ganglion will
form. X 720.

Journal of Morphology Vol. XI. : PUL.AVIL

Juth. Anst.vWerner &Winter, Frankfurt vil

ON DHE SMALLESE) PARTS: OF SEENTOR. CAP:
ABLE OF REGENERATION; A CONTRIBUTION
ON VIE, LIMITS OF DIVISIBIEITY: OF ULIVING
MATTER.

FRANK R. LILLIE.

In experiments on the power of multiple development of the
ovum the result has been reached, that a portion of less volume
than one-fourth that of the normal ovum does not possess the
capacity of producing an embryo or larva, though it may a
gastrula (see postscript); while a portion equal to one-fourth or
more of the volume of the normal ovum may, under suitable
conditions, produce a gastrula and finally a larva of correspond-
ing relative bulk. This result has been attained by Loeb,!
Wilson,? Driesch,? Morgan,* and Zoja.®° Wilson found only a
single larva of Amphioxus of one-fourth the normal size and
that showed several defects ; and Driesch has not, so far as I
am aware, mentioned any pluteus of less than one-quarter size.
Morgan has recently published the results of his studies on
the power of multiple development of the echinoderm ovum.
In this he shows that “the volume of the smallest gastrula
which can be produced from fragments of the egg falls below
gz the volume of normal gastrulae. The volume of the frag-
ments of the egg which produced such gastrulae, varies
between ;{, and =, of the volume of the ovum.” (Summary

q
p. 124 loc. ct¢t.) But these smallest gastrulae were unable to

1 Jacques Loeb. ‘On the Limits of Divisibility of Living Matter.” In Biol.
Lectures of Marine Biol. Lab. for 1894. Boston, Ginn & Co. Also in Archiv fiir
Ges. Physiologie von Pfliiger. Vol. LIX. 1894.

2 E. B. Wilson. “ Amphioxus and the Mosaic Theory of Development.”
JOURNAL OF MorpPHoLocy. Vol. VIII, No. 3. 1893. ?

8 Driesch. “Entwicklungsmechanische Studien.” III-VI. Zeztschr. f. wiss.
Zool. Bd. LV,p.9. 1893.

«Studies of the Partial Larvae of Sphaerechinus,” by T. H. Morgan, in
Roux’s Archiv fiir Entwicklungsmechanik der Organismen. Bd. Il, H. 1. 1895-

5 Sullo sviluppo dei blastomeri isolati dalle uova di alcune meduse (e di altri
organismi) per il Dr. Raffaello Zoja; in Roux’s Archiv fiir Entwicklungsmechanik
der Organismen. Bd. II, H.1. 1895.

240 LITE. [Vou. XII.

develop further. Morgan himself says (p. 117): ‘The smallest
pluteus which I have found measured 7 X 8, and the normal
form in the same dish 12 X 15. Another larva at the begin-
ning of the pluteus stage measured 6 X 7. The larvae have
apparently one-eighth the volume of the normal, and correspond
in size very nearly to the pluteus figured by Loeb. If, how-
ever, we compare these small larvae with the larvae derived
from isolated blastomeres, the conclusion is forced upon one
that these plutei have in all probability come from fragments
of the egg having only about one-half to one-fourth of the volume
of the egg.’ Inasmuch as the test proposed for the limits of
divisibility rests upon the capacity for complete development
to an embryo, or larva properly so-called, it is only these last
figures of Morgan’s that demand consideration. It would
seem from these that Loeb (/oc. czt.) has made his figure, one-
eighth, too low, not having taken in account the fact, emphasized
by Morgan, that the growth of the small blastulae, gastrulae
and plutei is less rapid than that of the normal.

Zoja’s results on the separation of the blastomeres of the
ova of certain medusae must also be considered. In the sum-
mary, p. 32, we find the following remarks: “ Medusae; Die
Entwicklung der getrennten Blastomeren (§ und } Ei von
Liriope mucronata, Geryonia proboscidalis, und Mitrocoma
annae; 4, 1, 4, =4, Ei von Clytia flavidula und Laodice cruci-

2 OE) 7B) 62
ata) ist ganz genau in allen ihren Phasen wie diejenige des
ganzen Eies.” — “Es bildet sich endlich immer eine schwim-

mende Larva, aus zwei Gewebe bestehend, die von jener,
welche aus + Ei hervorgeht, nicht unterscheidbar ist, ausser in
den Dimensionen.”

From this it appears that a 54 blastomere of Clytia and
Laodice is able to develop into a swimming larva. But from
the next statement I judge that the development cannot go
farther; this is: “ Bei Clytia zeigten } und } Ei auch die voll-
standig entwickelte idroide Form, und bei Liriope gab } eine
kleine runde Medusa, in welcher die vier primaren Tentaculi
normalerweise im Kreuz angeordnet waren.” Thus a fourth
blastomere is the smallest portion capable of complete develop-
ment.

No. 1.] THE SMALLEST PARTS OF STENTOR. 241

There are three possible explanations of this failure of such
small parts to develop:

1. That the whole organization of the species cannot be
included in so small a space. Briefly, defictent organization.

2. That so small a volume of matter cannot fulfill the
mechanical conditions consequent on cell-division, formation of
a segmentation cavity, invagination, and so forth, owing per-
haps to increased surface tension (Driesch), not to mention
other conceivable alterations of the extrinsic factors of develop-
ment.

3. That such a small part “is not able to set free that
amount of energy which would be required for its transforma-
tion into a gastrula or a pluteus.’”’ (Loeb.) !

The third explanation seems to me inadequate; because such
masses may continue to live for a considerable period of time
and display an amount of energy in atypzcal form changes and
rapid ciliary movement, which would suffice to produce the
phenomena of normal development, did not other factors
(included in the first or second of the above alternatives)
prevent. Moreover, it is well known that exceedingly minute
protoplasmic bodies, very much smaller than one-eighth the
echinoderm ovum, may produce a relatively enormous amount
of energy: e.g. bacteria and spermatozoa. Finally we do not
know how much of developmental energy is of intrinsic and
how much of extrinsic origin. We are limited, then, to the
jirst two alternatives.

Now, in the regeneration of a unicellular organism those
mechanical conditions consequent on cell-division, formation
of a segmentation cavity, invagination, and so forth, are not
required to be fulfilled. Surface tension and other extrinsic
factors of development of Metazoa have not been shown to
exercise a controlling influence in the regeneration of such an
organism.” It occured to me, therefore, that the cz/zate Infusoria

1 Morgan’s explanation, that the failure to develop is due to inability to pro-
duce a sufficient number of cells for the next ontagenetic stage, will come under
the first or second of these alternatives, according to the general point of view.

2 Of course it is possible for any one to maintain that extrinsic forces do

control the regeneration. But the burden of proof rests upon the maker of such
an assumption.

242 LILLTE: [Vor. XII.

offer conditions for decision between these two alternatives.
If it should be shown that a nucleated portion of the body
below a certain minimal size is incapable of regeneration, the
first alternative would receive support. If, on the other hand,
the smallest nucleated fragments of the body are capable of
regeneration with restoration of the normal form, the first
hypothesis would fall, and the second tend to be established.

My material consisted of two species of Stentor, viz.: S.
polymorphus and S. coeruleus. The former species occurred in:
immense profusion on decaying leaves of the water-lily in a
small pond near Ann Arbor, the latter appeared in considerable
numbers in a small aquarium which had stood in the laboratory
for six or seven weeks, and contained gatherings from a swamp.
This species is more favorable for experimental work than the
former, because the protoplasm is transparent, enabling one to
see the nucleus readily in the living animal. In S. jpoly-
morphus the body is rendered almost perfectly opaque by the
presence of immense numbers of symbiotic unicellular Algae,
the so-called zoochlorellae, which either hide the nucleus com-
pletely from view, or permit mere momentary glimpses of it.
On account of the ease of procuring any desired supply of S.
polymorphus my work was done chiefly on this form; but the
results were checked on S. coeruleus, and were practically the
same for both species.

To reach the desired result it was necessary to find or devise
a method by which nucleated fragments of every possible size,
beginning with a portion not much larger than a single node
of the nucleus, could be produced in large numbers; for
reliable quantitative results can be reached only by observation
of a large number of cases of regeneration. For this purpose
I tried the method of shaking which has yielded such admir-
able results with the animal ovum in the hands of Wilson,
Driesch, Morgan, and others, and found it to succeed to per-
fection. Ifa number of Stentors are put in a small vial about
one-third filled with water and shaken quite violently from five
to twenty times (S. coesnuleus requires to be shaken only about
five times; SS. polymorphus ten to twenty times), and then
examined under a low power of the microscope, one sees that

No. I.] THE WSMALLE SL! PARLS (OFM SLEW LOR: 243

the animals have been broken into numerous fragments of
every possible size and shape. In the field of the microscope
there are present at the same time naked nodes of the nucleus
either single or united in groups of two or three, and parts of
the body, both nucleated and unnucleated, ranging in size
through every possible gradation from 25m in diameter to
about 200u. Most of the latter are being driven hither and
thither by the action of the cilia with which they are covered ;
and many of them are of the most bizarre and curious
shapes: T-shaped, Y-shaped, or provided with other arm-like
processes, or of forms impossible to describe; but most of
them are of more regular form, triangular, quadrilateral, oval,
and spherical.

The moniliform character of the nucleus in these species of
Stentor insures that a large proportion even of the smallest
pieces receive at least some part of the nucleus. In order to
satisfy myself that such is the case, I killed and stained the
whole of one lot of S. polymorphus, which had been shaken as
described, about fifteen minutes after the operation. The
stained material was then mounted in balsam and measure-
ments were made of the smallest nucleated pieces. Some of
the measurements were as follows :

1. Naked nodes of nucleus, spherical or oval; 20-25p.

2. A spherical piece 31m in diameter containing a single
node of the nucleus. Nucleus excentric. Protoplasm a thin
cortex.

3. A spherical piece 37m in diameter; contained a single
node of the nucleus.

4. A spherical piece 37m in diameter; contained two nodes
of the nucleus.

5. A spherical piece 40 in diameter; contained six nodes
of the nucleus.

6. A spherical piece 50m in diameter; contained one node
of the nucleus.

7. Aspherical piece 50u in diameter; contained two nodes
of the nucleus.

8. A spherical piece 50m in diameter; contained four nodes
of the nucleus.

244 IEG EISL TD [VoL. XII.

g. A spherical piece 66m in diameter ; contained a single
elongated node of the nucleus.

10. A spherical piece 69 in diameter; contained seven
nodes of the nucleus.

Other similar pieces containing one or more nodes of the
nucleus were seen; some were of course not spherical, but I
have given the spherical pieces as easier to compute the
volume. Of larger nucleated pieces there was no lack ; they
were very numerous, as one would expect. There were, of
course, numerous unnucleated masses of protoplasm of various
sizes, but few large pieces. In fact, the majority of the pieces
below 100m in diameter were unnucleated; but the above list
shows that a good many of such small pieces contained one or
more nodes of the nucleus. I did not attempt to ascertain
what was the proportion of nucleated to unnucleated pieces of
such small size; but it must have been quite large, perhaps
one to ten or even more.

One further remark as to my methods. When any doubt
existed in my mind as to the presence of parts of the nucleus
in examples noted, the specimen was killed and stained,
generally in Schneider’s aceto-carmine, and the actual condi-
tion of the nucleus thus determined with certainty. This of
course involved the sacrifice of a great deal of material.

In consequence of the often curious and asymmetrical shapes
of the pieces produced by shaking, I expected to obtain valu-
able results on the teratogeny of the Stentors for comparison
with the results of Balbiani and Johnson, but I have been
almost completely disappointed in this respect. When regener-
ation of a piece takes place at all, it almost always happens
that a single more or less perfect animal of typical form results.

RESULTS.

In this paper I shall speak only of results obtained on the
smallest parts capable of regeneration, leaving other questions
suggested by the experiments for future consideration. From
numerous experiments, involving many hundreds of S. poly-
morphus, it was found that the smallest parts capable of

No. I.] THE SUALLEST PARIS ‘OF STENTOR. 245

regeneration possess the volume of a sphere of about 804
diameter. A lesser number of experiments on a smaller
number of S. coeruleus yielded results almost identical. In
the following list I give measurements of some of the smallest
Stentors found. After measuring in a more or less expanded
condition the animals were made to contract, when they assumed
almost the form of a sphere; the diameter of the sphere was
then measured, and this measurement was used for comparison.

1. Stentor coeruleus. 45% hours after shaking. Regener-
ation was complete or nearly so. I could see the adoral spiral
sinking into the oesophagus, the mouth, and contractile
vacuole. When expanded the form was quite typical. There
were two separated nodes of the nucleus present.

Measurements. None were made of the expanded animal.
Diameter of contracted animal (spherical) 90p.

2. S. polymorphus. 67 hours after shaking. Regeneration
complete.

Measurements. a. Expanded, 257 in length; 80m across
frontal field. 6. Diameter of contracted animal (spherical)
8Op.

3. S. polymorphus. 70 hours after shaking. Regeneration
complete.

Measurements. a. Expanded, 257m in length; 84m across
frontal field. 0%. Diameter of contracted specimen (spherical)
87.

4. S.polymorphus. 96 hours after shaking. Regeneration
complete. The animal was sluggish and did not expand well.

Measurement. Diameter of contracted specimen 75y.!

I have measurements of a number of Stentors of slightly
larger size than the ones given; but the smallest Stentors
were very scarce. By far the greater number of nucleated
parts, which possessed a spherical diameter of less than 1oop,
were incapable of regeneration, or at any rate did not regener-
ate. However, but a single example is sufficient to show that
a portion of the volume of the example in question is capable
of regeneration.

1 My note-book expresses a little doubt about this specimen, but it was cer-
tainly under 80.

246 EILEIES [Vor. XII.

The volume of the smallest Stentor found was thus equal to
a sphere of somewhat less than 80m in diameter. Not one of
the hundreds of smaller nucleated parts regenerated, though I
found one part, 7Im in diameter in spherical form, which had
assumed a fairly typical form of semi-contraction and possessed
a single bead of the nucleus; anterior and posterior ends (or
foot) were thus recognizable, but there was neither oesophagus
nor adoral membranellae present. Even if we admit this as
regenerated, which I do not, it does not essentially alter the
final result.

My conclusion is, therefore, that nucleated parts of Stentor
polymorphus of less volume than a sphere of 80 (approxi-
mately) in diameter are incapable of regeneration; nucleated
parts of greater volume are capable under favorable conditions
of complete regeneration.

The main results hitherto reached on the merotomy of the
Protozoa can be summarized as follows :

1. Cytoplasm without nucleus is incapable of regeneration
(Nussbaum, Gruber, Verworn, Balbiani, and others). This I
can confirm. (Verworn has shown that the isolated central
capsule of Zhalassicola nucleata from which the nucleus has
been removed is capable of partial regeneration, but it soon
goes to pieces. Gruber has shown that if a Stentor in process
of fission be transversely divided so that the posterior part
receives no nucleus, this part is nevertheless able to regen-
erate.)

2. Nucleus without cytoplasm is incapable of regeneration.
(Verworn, Balbiani.) This also I can confirm.

3. Portions of the body consisting of nucleus and cytoplasm
are capable of regeneration. To this [ must add: provided that
the amount of cytoplasm exceed a certain minimal volume
(which in the case of Stentor at any rate is quite considerable).

This amounts to a demonstration of Verworn's view that
vegeneration tn the Protozoa ts due to the reciprocal interaction
of nucleus and cytoplasm. Organization resides in the cytoplasm
as wellas in the nucleus. How otherwise are we to explain the
fact that a difference in the amount of cytoplasm alone (equiv-
alent to the difference in volume of two spheres of SO and 70

—_

No: t.)] THE, SMAELEST' PARTS OF STENTOR. 247

micromtillimeters respectively) determines the occurrence of regen-
eration ?

As regards the bearing of the results on the limits of divisi-
bility of living matter: we are not concerned here with the
question of the ultimate constitution of protoplasm, its com-
position of any ultimate vital elements whatsoever, but merely
with the question propounded by Loeb, ‘‘ What is the order of
magnitude of the smallest particle that can show all the
phenomena of life?” In the case of the animal ovum as
already noted this is about one-fourth of its volume, if we
include development as one of the phenomena of life. Cer-
tainly development includes all the phenomena of life. In the
case of Stentor the volume is relatively considerably less as the
following calculation will show:

The volume of the smallest perfect Stentor polymorphus,
which I was able to produce, was equal to that of a sphere of
about 80 diameter; the average volume of the Stentors
used in the experiments was equal to that of a sphere of
about 230m, as I determined from a series of measurements of
animals killed in a weak killing-fluid, and thus completely con-
tracted. That is, the ratio of the diameters of the smallest
and the average Stentor is about 1:3; or the ratio of volume
to volume is about 1:27. That is to say that the smallest
Stentor which can be produced is about one twenty-seventh of
the volume of the average Stentor.! This number is of course
a mere approximation, but it certainly will not be made any
greater by subsequent investigation, though it may be lessened
somewhat.

8 In the case of S. coerudeus the figures are different: the smallest measurement
which I have of this species is g0u; the average is 280u; thus the ratio of the
smallest to the average is about 1:3 in terms of the diameter, or 1:27 in terms of
volume. I believe, however, that it would be possible to produce a smaller S.
coeruleus by working over a larger amount of material. I do not think that there
is much difference in the absolute size of the smallest Stentors which can be pro-
duced, whether one uses the largest or smallest normal specimens. If e.g. the
average size of a lot of large Stentors were 320u, the smallest specimen which
could be produced would still be 804. The ratio of volumes would then be 1:64.
Of course this does not necessarily mean that 64 Stentors could theoretically
be produced at one time from a single one, for I doubt that the nucleus could un-
dergo that amount of division.

248 TELE. [VoL. XII.

In any case this relation forms a striking contrast to that
found in the development of the animal ovum, where a portion
of less volume than one-fourth that of the ovum does not
develop into an embryo (see postscript). It has been very
generally found that a portion (of the two- or four-cell stage)
equal to one-eighth the volume of the ovum never develops
farther than the gastrula stage.

In the case of the animal ovum, again, parts slightly smaller
than the minimum necessary for the complete development may
undergo partial development. In Stentor we have a parallel
phenomenon: parts of less than 80 spherical diameter may
undergo partial regeneration, but are unable to complete it.

It seems to me that neither increased surface tension due to
diminution of surface area, nor yet any other external factor,
is responsible for this failure of small pieces of Stentor to
regenerate. The cause lies within; and I do not believe that
it is to be sought in an insufficient production of energy.
For such small pieces may live for days, constantly producing
and expending energy in the ordinary processes of metabolism.
I am forced, therefore, to the conclusion that the organization
of these parts is in some way deficient. There ts probably for
cach species of animals a minimal mass of definite size consisting
of nucleus and cytoplasm within which the organization of the
species can just find its latent expression. This is the mznzmal
organization mass.

In the case of the Protozoa the size of this minimal mass is
that of the smallest part capable of complete regeneration.
But I do not believe that in the Metazoa the minimal organiza-
tion mass is that of the smallest part of the ovum capable of
developing into a normal embryo; for undoubtedly the in-
fluence of external factors is of the greatest importance here.
I would conceive then that in the Metazoa this hypothetical
minimal organization mass is smaller than any part yet ob-
served to develop into a normal embryo. Still, from my
results on Stentor, I believe that it is of such a size as to be
easily visible under a low power of the microscope.

ZOOLOGICAL LABORATORY OF THE
UNIVERSITY OF MICHIGAN, Dec., 1895.

No.t.] THE SMALLEST PARTS OF STENTOR. 249

Postscript. — After sending the above article to the editor
I had access to Boveri’s recent paper entitled “Ueber die Be-
fruchtungs- und Entwickelungsfahigkeit kernloser Seeigeleier
und tiber die Moglichkeit ihrer Bastardirung,” published in
Bd. II, Heft 3, of Roux’s Archiv fiir Entwickelungsmechantk
der Organismen, Oct. 22, 1895. Boveri states that the smallest
dwarf larva which he obtained came from a fragment which
could not have measured more than ;!, the volume of the intact
ovum “bei ungiinstiger Rechnung.” MHis conclusion is:
“Das Fragment des Seeigeleies bis herab zu einer Grésse von
ag des urspriinglichen Kivolumens besitzt die formative Wertig-
keit des ganzen Eies.”’ This is in marked contrast to the
results of the other authors quoted, none of whom have found
a figure less than $. The difference may be due in part to the
fact that Boveri shook the ova before fertilization, while the
other experimenters performed this or an analogous operation
after fertilization ; although this does not seem very probable.
If the exact proportion of the minimal organization mass to the
whole ovum be a matter of any importance, very great care in
the estimation of the volumes of dwarf larvae would seem to
be necessary, taking into account the differences in thickness
of the layers in dwarf and normal larvae, and also the relatively
slow increase in volume of the former.

The figure which I have found for Stentor is but little
lower than that of Boveri for the animal ovum, and this approx-
imation suggests interesting comparisons.

od

ORGANIC VARIATION AS A CRITERION OF
DEVELOPMENT.

THOS. H. MONTGOMERY, Jr., Pu.D.

CONTENTS.

PAGE
EN PRODUCT ORY = hc.cetetccse ssi vse csestedcdes 208s dicispa0- le cca ne asnsttessecnoepnes napeeeceaeenes weno 251

I. DEFINITION OF VARIATION; CORRELATION OF ORGANS; PROGRES-
SIVEVAND! (REGRESSIVE, IDEVELOPMENT. ¢...2-..-ccsescesees secuestseeessseeeesee 253

II. CONTINUING ORGANIC DEVELOPMENT IS ALWAYS ACCOMPANIED BY
PUPAE ICT gs Red Ue ceo coco oA en sence scence en ea eee 259
A. Certain Criteria of Continuing Developmtent.........-...cccecce co 259
TES Wy 22) 2028 Sree eRe ar eee ae Nn cA ons Sd ek Ie ge 264
C™eDinech Injercucesaprom the: Tabulated: Data ete 290
TONG THE ORIGING OF IVAARTATION yee pole. a. eee eee 293
IV. ON VARIATION AS A CRITERION OF DEVELOPMENT........-:::cc:ceeeeeeeesees 298

INTRODUCTORY.

THE object of the present paper is, firstly, to give a tentative
explanation for the origin of variation, deducible from the law
of the concomitance of variation with continuing development.
The attempt will then be made to show that variation is caused
indirectly by change of environment, and directly by the dis-
turbance of correlation of the organs, resulting from the change
of environment. And secondly, the attempt will be made to
determine whether, in a given organism, the amount (or degree)
of variation, and the manner of its occurrence, can furnish us
with criteria for judging its lines of development. The only
postulates necessary for the treatment of the problem of varia-
tion from this point of view are: (1) the concomitance of
variation with continuing development, (2) the correlation of the
organs of an organism, as necessary for its existence, and (3)
the influence exerted upon the organism by its environment,
necessitating a degree of adaptation to its environment.

252 MONTGOMERY. [VoL. XII.

It is not my intention to review the many theories already
advanced to explain the nature and origin of variation, which
would be a task too extensive for the scope of my present
article. There are two well-known theories, each of which has
latterly been more or less modified in regard to the origin of
variation : the first teaches that variation is caused more or less
directly by the environment ; while, according to the second,
variation is the result of an inherent tendency on the part of
the organism to vary. Now, in regard to the last-named theory,
it may be said with truth that, though much may be explained
on the assumption of “inherent tendencies,” there is no empirical
proof of the existence of such tendencies ; and further, variation
is not explained by the assumption of an inherent tendency to
vary, until the origin and nature of the inherent tendency itself
be explained. Reference may also be made to the theory of
Weismann, that variation has its origin in conjugation. My
own theory, as will be seen in the following pages, inclines
somewhat to the doctrine of the origin of variation as caused
by the influence of the environment, but is a new departure
from the Lamarckian theory, inasmuch as I consider variation
to be possible only under a temporary state of independence
of the several organs, when their complete correlation has
been disturbed by a change in the environment.

The recent admirable work of Bateson! has shown clearly
the importance of a careful comparative study of the phenomena
of variation, for the understanding of the problems of morphol-
ogy. In his book he has treated principally of the phenomena
of variation in their relation to certain laws of bilateral and
radial symmetry ; but the problem of the origin of variation he
dismisses by stating that “Inquiry into the causes of variation
is as yet, in my judgment, premature” (p. 78). My reason for
attempting the solution of a problem so intricate and difficult
in its nature is the need of approaching the question from a
new point of view; and in this attempted solution I have
endeavored to keep within the line of facts as much as possible,
and to avoid making unnecessary theoretical assumptions.

1 Wm. Bateson: Materials for the Study of Variation, treated with especial
regard to Discontinuity in the Origin of Species. London, 1894.

No: 1-] ORGANIC VARIATION. 253

Although the doctrine of Natural Selection will be but little
mentioned in the following pages, it is nevertheless advisable
to give a clear definition of it. Darwin states (“Origin of
Species,’’new American edition from sixth English edition, 1886,
p. 63) : ‘This preservation of favorable individual differences
and variations, and the destruction of those which are injurious,
I have called Natural Selection, or the Survival of the Fittest.”
If we take this definition, and eliminate from it the sentence,
‘‘and the destruction of those which are injurious,’ we have
before us perhaps a true explanation of the origin of species.
But the elimination of this passage seems necessary, since it is
a debatable question to what extent the “destruction’’ of the
unfavorable variations can proceed. Darwin himself discussed
the preservation and destruction of variations, and left untouched
the problem of their origin ; but in the “ Origin of Species”’ he
makes two statements which are of interest here: “It seems
clear that organic beings must be exposed during several gen-
erations to new conditions to cause any great amount of varia-
tion ; and that, when the organization has once begun to vary,
it generally continues varying for many generations” (p. 5).
«Unintentionally he [man] exposes organic beings to new and
changing conditions of life, and variability ensues ; but similar
changes of conditions might and do occur under nature ”’ (p. 62).
From these two quotations we may conclude that Darwin con-
sidered variation to have its origin in change of environment, —
in which view he probably followed Lamarck.

I. DEFINITION OF VARIATION ; CORRELATION OF ORGANS ;
PROGRESSIVE AND REGRESSIVE DEVELOPMENT.

Bateson applies the following definition to variation (/.c., p. 3) :
“To this phenomenon, namely, the occurrence of differences
between the structure, the instincts, or other elements which
compose the mechanism of the offspring, and those which were
proper to the parent, the name varzazzon has been given.” But
since cases are known in which wholly normal offspring have
descended from parents which were not normal in all respects, we
cannot consider the offspring in such cases to present variations

254 MONTGOMERY. [VoL. XII:

from the specific type, even though they may differ from their
parents ; so that it is advisable to seek a more general definition.
Accordingly, organic variation may be defined as growth above
or below (¢.e. beyond) a given norm ; and organic variability,
the power of the individual organism to produce such variation.
Thus we can only then speak of variation when at a particular
point in the ontogeny the growth of a given organ in one indi-
vidual is greater or less than the normal growth at that stage.
It remains necessary, however, to apply distinctive definitions
to the terms “normal” and “abnormal’’; and, although it is
not possible to give strictly distinctive definitions to such relative
ideas, it is, however, generally understood that such characters
are zormal as are presented by the greater part of the indi-
viduals of a given species, and such adzormal as are pre-
sented by a much smaller percentage of individuals. Though
no really distinctive definitions can be given, the relative mean-
ings of normal and abnormal are sufficiently understood, which
is all that our definition of variation demands. And even in
cases of a more or less perfect intergradation of the degrees of
development of a given organ, in a large number of individuals
of a species, it seems to be always possible to show that the
limits of variation of the large majority of individuals lie within
a certain circumscribed compass ; so that here normal and
abnormal degrees of variation may be distinguished.!

By the expression “ Correlation of the Organs,” is understood
the state of mutual dependence of the organs, after their divi-
sion of labor has been brought about by the process of evolu-
tion ; each has its own particular function to perform, but the
fulfilment of this function is not sufficient for its existence,
since rather it would be unable to perform its own function
without the aid it derives from the other organs. But further,

1 The term variety is often used ambiguously, as synonymous with variation, or
as equivalent to the idea subspecies (geographical race). In the strict sense, how-
ever, the term variety is applicable only to the whole individual, and not toa single
organ of it, and therefore is not equivalent to variation, which is any growth above
or below the normal. A variety is then, sevsz stricto, synonymous with the term
subspecies ; but in order to avoid any possible ambiguity, which has risen from the
wrong use of the word variety, I shall in the following pages avoid adopting it,
and shall use instead subspecies, or geographical race.

No. 1.] ORGANIC VARIATION. 255

while this physio- and morphological correlation of the organs
aids each organ in the fulfilment of its proper function, it
simultaneously acts as a restraint upon the exertions of the
vital processes of the particular organ ; for the particular organ
does not functionate merely for the maintenance of its own
existence, but primarily for that of the whole organism, and
when it has fulfilled the demand of the whole, its correlation with
the other organs would cause a temporary cessation of its activity.
Thus the correlation of the organs exerts a restraint, —acts as
a regulator, upon the amount which each shall perform. And
when the correlation is perfect, we must assume that each organ
can normally perform a certain fixed amount, and no more nor no
less. Since the performance of a physiological function results
in morphological change, showing the direct correlation of the
function and structure of an organ, the result follows that, if
the amount of physiological action performed by an organ is
determined by the correlation of the organs, the amount of
morphological change must be determined also by the correlation
of the organs. This fact is important for the establishment of
the deduction, which will find its treatment further on, that
variation can appear only when the complete correlation of the
organs has been disturbed. And since the degree of perfection
of the division of labor between the several organs is propor-
tional to the amount of differentiation of the organism, it is
correct to conclude, that the completeness of the correlation of
the organs stands in a direct proportion to the degree of differ-
entiation of the organism, — the higher the organism the more
perfect is the correlation of its organs; and vice versa, the
lower the organism is structurally, the less intimate is this
correlation, z.c. the more independent the several organs are.
It will be well here to analyze and compare briefly the ideas,
progressive and regressive development. Either process may
modify a given organ in regard to its structure (chemical and
morphological), its size, position, and, in meristically arranged
organs, its number. Progressive development tends to com-
plicate or further differentiate its chemical and morphological
structure, to change its position and dimensions, and (subject
to certain limitations) to increase its number in a meristic

256 MONTGOMERY. [-VoL. XIT.

series; while regressive. development (degeneration, Réck-
bildung) tends to simplify the structure, to change the position
and dimensions, and (subject to certain limitations) to decrease
its number in a meristic series. Since either process may pro-
duce a change of position of the organ involved, such a change
furnishes no criterion whereby to judge the kind of develop-
ment, until it can be determined whether the direction or
amount of change of position differs according as the mode of
development is progressive or regressive. And whether change
of the dimensions gives us a criterion of the kind of develop-
ment is also doubtful; although, since it is the general rule
that increasing complexity of structure is usually accompanied
by actual increase in size, it might be concluded that an
increase in size of an organ denotes frequently the action of
progressive development. But since it is only a general rule,
and by no means a law without exceptions, that increase in
size goes hand in hand with increasing structural complexity,
it would be safer to eliminate change of dimension from our
criteria for distinguishing between the two modes of organic
development.

We find, however, a certain criterion for estimating the kind
of development, in change of structure; for an increasing com-
plexity of structure is distinctive of progressive development,
while, on the other hand, a decreasing complexity is the sign
of regressive development. Further, an increase in the num-
ber of meristically (segmentally or metamerically) arranged
organs is a criterion of progressive development, as a decrease
in the number of similar organs is of regressive development, —
provided that, in each case, no structural changes are simul-
taneously taking place. This standpoint will hardly be dis-
puted, for we consider an organism A to be morphologically
higher than an organism #, when A possesses a larger number
of organs in a given meristic series than does 4, even though
these organs are otherwise structurally equivalent in A and BZ.

But it is necessary to consider the case, when a progressive
meristic development is acting together with a regressive
structural development, or vice versa: as e.g. in the carpus of
the Ichthyosauria, where the phalanges are meristically pro-

No. 1.] ORGANIC VARIATION. 257

gressive, but regressive in regard to specialization, in compari-
son with their ancestral forms. The question before us is,
then: When in an organ there is at work a progressive
structural development, simultaneous with a regressive meristic
development,’ is the organ to be regarded on the whole as
progressive or regressive ? The answer to this is to be gained
by determining whether structural complexity is of greater or
of less morphological importance than is meristic change.
Now the consensus of opinion among biological investigators
would show that structural complexity (both chemical and
histological) is of much greater morphological importance than
is mere meristic development, — this assumption being, indeed,
a necessary preliminary in any attempt to homologize different
organisms. For to pick out an example at random, who would
venture the opinion that Branchipus occupies a higher morphol-
ogical position than Astacus, simply on the ground that it
possesses more numerous extremities ; and would not rather
conclude that Astacus is the higher organism, because its
extremities are structurally more differentiated ? Therefore, if
structural modification is of greater morphological importance
than numerical (meristic) modification, then when progressive
structural development is accompanied by regressive numerical
development, the organ as a whole is to be regarded as develop-
ing progressively ; and conversely, when regressive structural
development is simultaneous with progressive numerical de-
velopment, the course of development of the organ is to be
considered regressive.

It still remains to accentuate an apparent law, which
is generally recognized, with reference to this frequent con-
comitance of numerical and structural development. Appar-
ently a progressive structural development frequently causes
a regressive numerical development, as we find when, by
the coalescence of previously separate meristic organs (2.e.
through their regressive numerical development), a com-
pound organ is produced, which is higher morphologically
than was any one of the previously separate organs. How-
ever, a numerical reduction of the units of a meristic series
can proceed, and perhaps does so more usually, without

258 MONTGOMERY. [Vor. XII.

coalescence ; and such a regressive numerical development
is also usually associated with a progressive structural
development of those units of the meristic series which are
retained. Examples of such cases may be found in abundance,
and it is sufficient here to refer to the lateral line of sense-
organs in the Vertebrates, where a progressive development of
certain of the individual organs is concomitant with a reduction
in the number of the units comprising the series. And witha
view to many analogical cases, which it is unnecessary to give
in detail here, since any zoologist may recall many to mind,
the law will be found to be of general application, that pro-
gressive structural development is furthered by regressive
numerical development, since in this way greater centralization
of the forces of growth would ensue. And although regressive
structural is sometimes concomitant with regressive numerical
development, as exemplified in the case of certain parapodia of
sedentary Annelids, I recall no case of the concomitance of
progressive structural and progressive numerical development ;
but I would not imply by this that such a concomitance can-
not or does not occur, but rather that such a concomitance is
of so infrequent occurrence as not to render invalid the
general law just mentioned. My reason for accentuating this
law of the usual concomitance of progressive structural with
regressive numerical development is in order (1) to characterize
concisely a relation, which, although already recognized, has
not yet been awarded much attention from zodlogists ; and (2)
to emphasize a point which may serve as a criterion of pro-
gressive development.

By the term development is meant here any organic process
of change acting in the organism; when such a change tends to
further complicate the structure, it has been termed progres-
sive development (evolution, Hxtwzckelung); when it tends to
simplify the structure, regressive development (degeneration,
Rickbildung). Speaking generally, development leads towards
(1) the production of new species, or (2) towards the extinction
of already present species ; obviously, development cannot be
conceived as holding a species stable (z.e. unchanging), since
development always implies an organic change.

Nooti] ORGANIC VARIATION. 259

After these preliminary explanations, we may next consider
the phenomena of continuing development.

II. ContTINUING ORGANIC DEVELOPMENT IS ALWAYS ACCOM-
PANIED BY VARIABILITY.

Although this postulate may seem at first sight to be a mere
enunciation of a well-known biological axiom, and a sine gua
non of the theory of development, it is nevertheless of great
importance for arriving at a true conception of the nature of
variation; and although evolutionists in general will grant
with Darwin that for the action of Natural Selection the
occurrence of variations is necessary, yet to my knowledge no
one has particularly accentuated the fact of this actual con-
comitance of variations with continuing development. Indeed,
most biologists have accepted this fact, without a critical
inquiry into its fundamental importance. In my last paper!
I laid particular stress upon this point, by saying (p. 483):
“ Now I consider this variability in the number of the eyes
of the freshwater forms to be explained by the general
law, that all organs (and propter hoc all organisms) which are
undergoing progressive or regressive development tend to be
variable.’ In the present paper I hope to substantiate the
validity of this “general law’ by data from another source.

A. Certain Criteria of Continuing Development.

In order to prove the assumption that continuing progressive
and regressive development is always accompanied by varia-
bility, it is necessary to produce a series of facts, showing that
organs (or organisms) which are undoubtedly in a state of con-
tinuing development always evince variability. But, although
examples of variation may be found in abundance, it is obviously
difficult to prove conclusively that a given organ (or organism)
is at a given time influenced by a continuing process of de-
velopment. Accordingly, for each example to be cited, we

1 “The Derivation of the Freshwater and Land Nemerteans, and Allied Ques-
tions.” JOURNAL OF MorPHOLOGY, XI, 2, 1895.

260 MONTGOMERY. [VoL. XII.

must demonstrate satisfactorily that it is undergoing a process
of development.

The question to be solved is then, first of all: What are our
criteria of continuing development? Progressive and regres-
sive development having been sufficiently characterized, it
remains necessary to produce criteria, whereby we can deter-
mine whether a given organ (or organism) is at a particular
time dominated by a process of development, or whether the
organ (or organism) is not being influenced by a particular
developing agency, either progressive or regressive. We may
now consider briefly three reliable criteria of continuing develop-
ment, namely, (1) domestication, (2) the presence of geograph-
ical races (subspecies), and (3) migration; no doubt other
criteria may be found, but these three are sufficient for our
present purposes.

(1) Domestication may be taken as a criterion of continuing
development, since all organisms in a state of domestication
are being more or less continuously selected by man, with a
view to their adaptability for certain uses. The development
induced by human agency is also very energetic, since man’s
uses for domesticated animals and plants are manifold, and
since he frequently introduces changes in their environment.
And further, as we know in many cases that the length of
time necessary for the production of a new “ breed”’ has been
comparatively short, we must conclude that not only was the
action of the development continuous, but also that it must
needs have been very energetic.

(2) The presence of geographical races may also be considered
a criterion of continuing development. A species is said to pre-
sent geographical races or subspecies when in different portions
of its breeding area particular forms occur, differing mainly in
color and dimensions, but which are all connected together by a
more or less perfect series of intergradations, and all of which
may breed together fertilely. Any one at all conversant with
the geographical distribution of animals or plants knows how
frequently wide-ranging species are differentiated into a number
of geographical races, and that the number of such races stands
usually in a direct ratio to the extent, or diversification, of the

N@2 i] ORGANIC VARIATION. 261

range of the species. Now if the Darwinian doctrine of evolution
be true, all such races have descended from one ancestral form ;
but I think that we may go still further, and postulate that
wherever one geographical race grades insensibly into another,
there the agency of development must be still continuing.
For supposing A, 4, C to be three intergrading geographical
races inhabiting contiguous areas a, b,c. In area a, together
with the individuals of race A, will always occur some individuals
of races B and C, which have migrated from @ and ¢ into a.
Now these individuals, which have migrated from 4 and ¢ into
a, must adapt themselves to the new environment of the area a,
if they would compete successfully with the individuals of race
A ; and thus a continuous development of a considerable num-
ber of the individuals of the species must proceed, tending to
produce favorable adaptations in the struggle for existence, —
this struggle being probably keenest where the areas a, 6, and
c overlap. It is still a bone of contention between systematic
ornithologists whether individuals of a race B are ever found
in the area proper to a race A, or vice versa; but the cases
where this is so, as e.g. Dendroica palmarum and its variety
hypochrysea, are so numerous as to warrant the conclusion that,
wherever the geographical lines of demarcation between the
respective breeding areas are not strongly marked, there
must be a considerable interchange of wandering individuals.
And it seems to be always the rule, that the indefinitely broad
area of demarcation is peopled promiscuously by individuals of
the contiguous races. Only when the lines of demarcation
are formed by high mountain ranges, deserts, or great water
expanses would there be little or no interchange of individuals ;
but when the boundaries of the areas of the several forms are
so comparatively impassable, the various forms usually do not
perfectly intergrade, and hence are to be classed rather as sep-
arate species than as races of one species. Therefore it is cor-
rect to conclude that in a widely ranging species, split up into
a number of intergrading geographical races, a continuous
agency of development is at work, leading to the readaptation
of the migrating individuals to new environments.

Darwin (“Origin of Species”) has ably argued the point

262 MONTGOMERY. [Vor Xi.

that where the adjoining areas of geographical races overlap,
the struggle for existence would be keenest, so that the indi-
viduals occupying this intermediate area would in time become
extinct. But he has overlooked the fact that until such exter-
mination has been brought about, z.e. as long as the races con-
tinue to intergrade, the intermediate area would continue to be
the vortex of development, and the sharp struggle there would
itself instigate the wandering of individuals into the adjoining
areas, where again they must adapt themselves to new environ-
ments. Now, when in any species the individuals occupying
the areas transitional between those proper to the several races
have become exterminated, the races must cease to intergrade,
fewer individuals will continue to wander into other areas, and,
the struggle for existence becoming less sharp, the main factor
in the process of development would disappear. But, as we
have shown above, when the races cease to intergrade, and
become more distinctly pronounced, they can sensu stricto be
no longer termed races, but rather distinct species. Accord-
ingly, the presence of geographical races being correctly con-
sidered a criterion of continuing development, we should expect
our data to show, as indeed they do, that, other factors being
equal, species with geographical subspecies evince a greater
amount of variation than do species which present no geograph-
ical races.

It is noteworthy that extensive periodical migrations act as
a restraint upon the production of geographical races. And
I account for this fact! by assuming that the migratory
species, being influenced in winter by an environment to some
extent different from that which it experiences in summer, must
be equally adapted to both environments (equally, at least, if it
remains under the influence of both environments for the same
length of time), and hence, the winter environment exerting a
restraint upon the production of adaptations suited to the
summer environment alone, such a migratory species would
not be capable of producing geographical races to the same

1In a paper which has not yet appeared, but which will be published in the
American Naturalist for June, 1896, dealing with migration as a check upon
geographical variation in birds.

No. I.] ORGANIC VARIATION. 263

degree, as would a non-migratory species with an equally exten-
sive breeding area.

(3) Extensive migration may be taken as another criterion of
continuing development. (By the term extensive migration, I
mean, as will be explained later, a regular periodic migration
through a considerable distance, — 30° lat. or more.) For, as
was shown in the preceding paragraph, a migratory species in
wandering from its summer to its winter home, or vice versa,
is brought into contact with a different environment, necessi-
tating a certain amount of readaptation ; therefore, there must
be at work a more or less continuous process of development,
leading towards readaptations. Thus, to use a well-known
example, a man accustomed to spend his annual holiday abroad,
on arriving at his destination experiences the lassitude prepara-
tory to his becoming acclimated, and experiences the same
physical sensations on his return. Accordingly our data should
demonstrate that species which undertake periodic migrations
through long distances should evince a greater amount of
variation than do stationary species, other factors being equal
in amount.

Other criteria of continuing development might be mentioned,
but as the three already given are well founded and sufficient
for the furtherance of our deductions, we shall deal but briefly
with a fourth. If Wallace’s theory (‘ Darwinism”’) be true,
that secondary sexual characters are most accentuated in those
species where the sexual impulses are strongest, so that the sexual
impulse may be considered as an important if not sole agent in
their production, then we might consider the presence of strongly
marked secondary sexual characters as a criterion of continuing
development, by regarding the sexual impulse itself as a more
or less continuing impulse to development. In other words,
the degree of development of secondary sexual characters would
stand in direct proportion to the continuousness and energy of
the sexual impulse. And since secondary sexual characters are
often different in otherwise closely allied species, being as a
rule the least reliable (morphologically speaking) of specific dif-
ferences, they must be regarded as of comparatively recent
origin in each species, and thus be considered characters prob-

264 MONTGOMERY. [Vo. XII.

ably still under the agency of a process of development. But
though this reasoning may be plausible, it is based upon Wal-
lace’s assumption that the production of such characters is due
to the agency of the sexual impulse ; and rather than bind myself
to such a theory, I would leave the case still disputable, whether
or no the presence of noticeable secondary sexual characters
should be taken as a criterion of continuing development.

IB: Data.

It now remains to produce data in support of the thesis
that individual variation is always concomitant with continu.
ing development ; and to do this, it must be demonstrated
successively that (1) variation is always predominant in those
domestic animals which have been most carefully selected by
man; (2) in such species as are divisible into geographical
races; and (3) in those species which undertake extensive
periodical migrations.

Variation in domesticated animals is very marked, and espec-
ially so in those forms which have been most carefully selected
by man. It is of interest to compare the diversity of breeds
of the dog with the fewer breeds of the cat ; the former is of
greater practical use than the latter, and man has subjected it
consequently to a greater diversity of conditions of life. Whether
a greater amount of individual variation is evinced by the
domesticated animals than by their allies in a state of nature,
cannot as yet be answered with certainty, since, as Bateson
(/.c.) observes, the phenomena of variation in the wild forms are
not known to the same extent. Cope (‘ Origin of the Fittest ’’)
mentions the peacock and the Guinea fowl as forms which have
not been rendered variable by domestication; but these two
may be classed as the wildest, and least carefully bred by human
selection, of any of the domesticated birds. However, without
coing further into the much-discussed question of variation
under domestication, the fact is sufficient for us that animals
show considerable individual variation under domestication,
and that the amount of variation is greatest in those species
which have been most influenced by human selection; and we

No. I.] ORGANIC VARIATION. 265

have found domestication to cause a more or less continuous
development.

In order to test the correctness of the assumption that indi-
vidual variation is most marked (1) in those species which pos-
sess geographical races, and (2) in those species which undertake
extensive migrations, I have examined nearly all the species of
North American birds with reference to individual variation in
some or all of the following dimensions: culmen of the bill,
wing (from carpal joint to tip of longest primary), tarsus (so-
called, but really tarso-metatarsus), whole length (from tip of
bill to tip of tail), and tail (from the pygostyle to tip of the
longest rectrix). It was my original intention to personally
undertake all the measurements, and with that object in view
I commenced a series of detailed measurements upon the bird-
skins in the collection of the Academy of Natural Sciences of
Philadelphia. Unfortunately for me, however, this collection
did not offer large enough series of individuals of all the species
desired, and not having the opportunity nor time to study other
large collections, —namely those at Cambridge, New York,
and Washington, —I was obliged to desist from further personal
examinations. In lieu, then, of such direct examination, I have
taken Robert Ridgway’s excellent “ Manual of North American
Birds, 1887 ” (first ed.) as my authority for the extremes of indi-
vidual variation, in regard to the dimensions specified, of the North
American species of birds. And here I would express my
hearty gratitude to Professor Ridgway for his liberality and
generosity in allowing me to make use of his valuable data. In
speaking of the measurements given in his work, Ridgway
states (p. ix): “Whenever practicable, they have been taken
from large series of specimens, and the extremes given as well
as the average. . . . In the case of closely allied forms, or
where distinctive characters are largely a matter of dimensions
or the proportionate measurements of different parts, care has
been taken to measure, whenever possible, an equal number of
specimens of the several forms to be compared; and specimens
in abraded or otherwise imperfect plumage, as well as young
birds, have been excluded. When there is any marked sexual
difference in size, the number of males and females measured

266 MONTGOMERY. [Vou. XII.

of allied forms has also been made as nearly equal as possible.”
The degree of individual variation in regard to the dimensions,
according to Ridgway, is therefore based (for most of the
species) on large series of specimens of adult individuals, in
unworn plumage; and as this distinguished ornithologist’s
work is regarded as a standard by taxonomists, the accuracy of
his measurements cannot be questioned. And as such a large
series of data is the result of years of painstaking work, I may
be pardoned for not attempting such a labor in the limited time
at my disposal. I have taken these measurements as given by
Ridgway, with necessarily the exclusion of such extremes of
variation as were based upon a very small number of individuals,
and have computed the percentage of variation for each given
dimension, expressing the difference between the extremes of
variation as a percentage of the minor term of variation. In this
way I have deduced the percentage of variation in the dimensions
of the larger part of the species and subspecies of North Ameri-
can birds (together with those of a considerable number of exotic
species, of casual or possible occurrence within our boundaries) ;
or, altogether, the species and subspecies of fifty-six families,
the only omissions being the following small families : 77o-
gonide, Alcedinide, Momotida, Cotingide, Hirundinide, Am-
pelide, Laniide, Coerebide, Motacillide, Cinclide, Certhide,
Syluiide. These latter families have been omitted, because
their respective scarcity of species would hardly warrant com-
parisons. Accordingly, using the measurements given by Ridg-
way as my basis, I have computed the percentage of individual
variation for one or more of the five dimensions specified, for
the greater number (approximately 600 or more) of species
and subspecies of North American birds. It is unnecessary to
reproduce in this paper these measurements for all the families,
which would only result in the needless occupation of too much
valuable space in this JouRNAL ; accordingly, for purposes of
comparison I will present tables of variation for those families
only, in which the diagnostic characters of most of the species
are furnished by the measurements, and for which, therefore,
the percentages of variation, based upon such necessary accurate
measurements, may be considered as accurate as possible.

No. 1.] ORGANIC VARIATION. 267

Together with the degrees of variation, will be given for each
species also, as briefly as possible, the range of migration and
breeding area. These facts have been extracted principally from
Ridgway’s work (/.c.), and from the recent work of Witmer Stone.!
Birds with a migration range of 30° lat. north and south, or a
corresponding distance east and west across the continent, I have
classed as extensive migrants; birds with a smaller migration
range, as migrants; those which do not undertake regular peri-
odic migrations, but occasionally accomplish wanderings of
considerable extent, as zrregular migrants ; and finally, those
which migrate not at all, or, as the meadow lark and crow,
migrate through only short distances, as ves¢dents. Such a
classification according to the range of migration is necessarily
an arbitrary one ; as is shown e.g. by the migration of certain
species from high mountain ranges to the adjoining valleys in
the winter season, a case which could not be classed as an
extensive migration, although each such species meets with a
considerable change of environment. This classification has
been used merely as a convenience for computing the relation
of the amount of variation to the extent of migration of the
species. In other words, the extent of migration and the
breeding area have been given for each species in order to
learn the laws of the amount of individual variation in its
relation to the environment.?

1 Witmer Stone: The Birds of Eastern Pennsylvania and New Jersey, etc.
Philadelphia, 1894. I would here express my gratitude to my friend Mr. Stone for
his valuable aid in helping me to determine the migration and breeding ranges
of certain species; and also for the facilities offered me to study the bird collec-
tions of the Philadelphia Academy of Natural Sciences.

2 The nomenclature adopted here for the species of birds is that employed by
the American Ornithologists’ Union, with the emendations contained in its
supplementary lists. I was unable to consult the second edition of this work,
which appeared after this paper had been sent to press.

268

MONTGOMERY.

[Vou. XII.

The following abbreviations will be employed in the tables:

= variation under I %.

= variation between 1% and 1.5%.
= variation between 1.5% and 2%.
= variation of 2% or more.

resident.

= irregular migrant.

regular migrant.
extensive migrant.
district.

= tropical.
= temperate.

gulf.

St:
N.E.
A.
C.A.
Miss.
BIC.

centre (or central).
island.

interior.

river.

valley.

state.

New England.
America.

Central America.
Mississippi.
British Columbia.

n., €., S.. W. = north, east, south, west (or their adjectives).

All other abbreviations for states and countries are those commonly employed

in the U. S.
TABLE I. VARIATION IN
a ee Wey
i) : 5 = i
SPECIES. a ee eg |
Sale< <
5/2] e |BAle
Rallus crepitans Gmel. De ad Rail Wedel ean
R. c. saturatus Hensh. wre] | | 30
R. obsoletus Ridgw. we | * | & |
R. elegans (Aud.) call aad Meal ead
R. beldingi Ridgw. #e | eR | KK]
R. tenuirostris (Lawr.) seek] | eK
R. virginianus Linn. ee | & | & [seHRR
Porzana carolina (Linn.) werk] 2 | 3% bee
P. jamaicensis (Gmel.) SHRP HH] [FERRE
P. noveboracensis (Gmel.) seEHE BRE] pom
Ionornis martinica (Linn.) * | * |
Gallinula galeata (Licht.) * | & | me | we
Fulica atra Linn. wee | HR |
F. americana (Gmel.) seek] | HR poe

THE RALLIDA.

MIGRATION.
BREEDING AREA.

M. Atlantic marshes n. to
Long I.

R. coast of La.

R. coast marshes from L. Cal.
to Ore.

M. freshwater marshes of e.
WES:

R. e. coast.of L. Cal:

R. c. & w. Mex.

E. M. whole N. A. n. to Hud-
son B.

E. M. n. U. S. northward.

HE: M. s. U.S: ns to’ Mass. &
Ore.

E. M. n. U. S. to Hudson B.
w. to Utah.

M. nearly whole trop. & warm
temp. A.

M. whole trop. & temp. N. A.
toLba Ge.

M. n. Eurasia.

E. M.n. U.S. to Greenland &
Alaska.

No. 1.] ORGANIC VARIATION. 269
TABLE II. VarIATION IN THE FALCONIDA.
& Bae M
SPECIES. 2 ie Q S g i IGRATION.
S|] < fil < BREEDING AREA.
OO} |e |EAl a
Elanoides forficatus (Linn.) a | abe FREE] 9066 M. trop. & warm temp. N. A.
Ictinia plumbea (Gmel.) ad R. s. Mex. to Paraguay.
Rostrhamus sociabilis (Vieill.) | *« | ** #* R. Mex. to Argentine Rep.
Circus hudsonius (Linn.) g * * | x* | E. M. whole of N. A.
2 ¥* ek |
Accipiter velox (Wils.) $ #E | | 58K E. M. whole of N. A.
° ee | eK | ee
A. cooperi (Bonap.) a # | HE beer R. temp. N. A. & Mex.
$2) * | % |
A. atricapillus (Wils.) é #e | RK M.n:e. N. Aon! of US:
2 * |
Parabuteo unicinctus (Temm.) |++**) eee RG oo Ac
P. u. harrisi (Aud.) S| * | He] & beer RoC. A. &ss Us Se
QO) * | & | we |
Buteo borealis (Gmel.) S |e | ae [HERE] He R. e. N. A. w. to Gt. Plains.
© | ee | HE] OK |
B. b. harlani (Aud.) Oo] & | ee jee] & R. G. St. & lower Miss. vall.
Q| ** | & fee] x
B. buteo (Linn.) sk | | 28H | 286 R. n. E. Hemisphere.
B. abbreviatus Cab. B lene] & | xe | x R. ne i5s Ae toss Nex cas. Cal:
O| we] * | * | *
B. swainsoni Bonap. S| xe | He | HR x R. w. N. A. from Alaska to
Argentine Rep.
© | 6% | 906% | eK | &
B. brachyurus Vieill. S| ¥* | ee | HF R. trop. A. n. to e. Mex.
B. latissimus (Wils.) é % bee] He E.M e. N. A. to Saskatchewan.
QO | | peeeR] oe
B. lineatus (Gmel.) | HHeK HK HK] 3 M. e. N. A. w. to Gt. Plains.
O| we] * | & |e
B. 1. elegans (Cass.) a * R. Pacific coast of U. S.
2 *
B. albicandatus sennetti
Allen ¢ ee | * R.e. S. A. tos. Tex.
Q| ee] * |
Urubitinga urubitinga (Gmel.) * | x R. Costa Rica to Argentine
Rep.
U. ridgwayi Gurney see] & | eK R. Guatemala & s. Mex.
U. anthracina (Licht.) S| * | +e] x R. trop. A. n. to s. Ariz.
Q| * | ** | ee
Archibuteo lagopus (Briinn.) ¢ * ¥ M. n. E. Hemisphere.
** *
A. ferrugineus (Licht.) * M. w. U.S. n. to Saskatche-
wan.
2 +e
Aquila chrysetos (Linn.) ¢| * | * | * | x R. n. portions of N. Hemis-
phere.
ON casei ees ae |) ae
Halizetus albicilla (Linn.) | * | ¥«* | x! * M. n. portions of E. Hemis-
phere.
O| we | * | * |
H. leucocephalus (Linn.) Bee jy atiek eee) dk R. whole of N. A., Kam-
schatka.
© peek] dank | ek [see

270

MONTGOMERY.

[VoL. XII.

TABLE II. — Continued.

SPECIES.

Thalassoztus pelagicus (Pall.)

Falco islandus Brinn.

F. rusticolus Linn.

F. r. gyrfalco (Linn.)
F. r. obsoletus (Gmel.)
F. mexicanus Schleg.
F. peregrinus Tunst.

. p. pealei Ridgw.

. deiroleucus Temm.
. albigularis (Daud.)
. regulus Pall.

. columbarius Linn.

. c. suckleyi Ridgw.

soy ys yy Y

. richardsonii Ridgw.

F. fusco-ccerulescens Vieill.

F. sparverius Linn.
F. dominicensis Gmel.
Polyborus tharus (Mol.)

P. cheriway (Jacq.)
P. lutosus Ridgw.

g
=
g
2
3
2
3
2
$
2
g
2
$
~
3
3
3g
$
3g
=
g
2
g
?
3g
“
$
SY
$
?
$
?
$
?

z a
D a! <
O12 ]e
* *
* * *
ES * *
** a RRR
RE * **
* * *
ee * RK
* * E
* eK *
* * RE
* * PRK
* FRR | RK
RK | RK *K
eK * *
* * *
* eK
** *
FRR RRR
* *
ee * *
* * *
* * *
* * *
* * eK
* * *
HK) OK *
* * *
** | HR *
** * **
IBEEE F8RRK
eK | RK *
** * eK
* *

SRK | RE RK
RK) FW | FH
* * *

HK

| TAIL.

MIGRATION.
BREEDING AREA.

M. Kamschatka sea-coasts.
R. circumpolar regions.

M. extreme n. N. A. & Eura-
sla.

I. M. n. Europe & arctic A.
M. coast of Labrador.
RoweUi o.1S. ton Miex:

M. Europe and portions of
Asia.

R. Aleutian Is. & coast of Ore.
R. trop. A. n. to s. Mex.

R. trop. A. n. to n. Mex.

R. Eurasia.

E. M. N. A. chiefly n. of U.S.
M. n. Cal. to Sitka.

M. int. of N. A. from Col.
northward.

R. trop._A. n., tous. exe &
N. M.

M. whole temp. N. A.
R. Cuba & Hayti.
Rofoa Ae

R. s. U. S. to Ecuador.
R. Guadelupe I.

No. I.]

TABLE III.

ORGANIC VARIATION.

271

VARIATION IN THE PICIDA.

SPECIES.

Campephilus principalis (Linn.)| *

C. p. bairdi (Cass.) %
C. imperialis (Gould) ecad
C. guatemalensis (Hartl.) aK
Dryobates villosus (Linn.) **
D. v. leucomelas (Bodd.) sek
D. v. audubonii (Swains.) +
D. v. maynardi Ridgw. HERE
D. v. harrisii (Aud.) ebm
D. v. jardinii (Malh.) EK
D. pubescens (Linn.) HORE
D. p. gairdnerii (Aud.) *
D. borealis (Vieill.) +
D. scalaris (Wagl.) ¥
D. s. parvus (Cabot) eee
D. s. bairdi (Scl.) ¥%
D. s. lucasanus (Xantus) *
D. s. sinaloensis (Ridgw.) +
D. s. graysoni Baird ee
D. nuttallii (Gamb.) *
D. arizone (Hargitt) *%
Xenopicus albolarvatus (Cass.)
Picoides arcticus (Swains.) *e
P. americanus Brehm oad
P. a. alascensis (Nelson) +
P. a. dorsalis Baird HK
Sphyrapicus varius (Linn.) *
S. v. nuchalis Baird *
S. ruber (Gmel.) *
S. thyroideus (Cass.) HEE
Ceophlceus pileatus (Linn.) _ [s+
Melanerpes erythrocephalus
(Linn.)
M. formicivorus (Swains.) +
M. f. bairdi Ridgw. HK
M. f. angustifrons Baird *
M. torquatus (Wils.)
M. elegans (Swains.) 2H
M. carolinus (Linn.) HEE
M. rubriventris (Swains.)
M. pygmezeus Ridgw.
M. aurifrons (Wagl.) eK
M. uropygialis (Baird) HERE
Colaptes auratus (Linn.) **
C. chrysocaulosus Gundl.
C. chrysoides (Malh.) aad
C. cafer (Gmel.) +e
C. c. saturatior Ridgw. He
C. rufipileus Ridgw. ee

x | Winac.

eee eeee wee tig gee be ds

* * *

gx * * *

* * * * ¥ x « & EF x

ROE

| TARSUS.

WHOLE
LENGTH.

> |

KERR E* Ee x x

MIGRATION.
BREEDING AREA.

. Ss. Mex. to Costa Rica.

- U. (S./except’s. St:

. N. A. w. to Alaska.

. Atlantic & Gulf St.
ahamas. [s. to Mex.
2 w. U.S.e. to Rocky Mts.,
. Mex. s. to Muncie

Ss.
e. N.

+

See! ends of Mex. to U.S.
see Calk
. Mex.

.s. Ore. & n. w. Mex.
. mts. from Wash. Terr. to
saCalt [Wie S:
n. N. A. s. to border of
.n. N. A. e. of Rocky Mts.
.? Alaska to Gt. Slave Lake.
2 Rocky Mts. from Kodiak
to N. M.
Mi nes, NavAvens of UEss:
2 Rocky Mts. of U.S.
coast from Alaska to Cal.
w. U.S. to Rocky Mts.

OSes ool alle Sta

whole of N. A.
M.e. U.S.
R. s. e. Mex. to Costa Rica.
R. Mex. & contiguous U. S.
Reese waeals
R. w. U. S. e. to Rocky Mts.
R. s. & w. Mex.
M. e. U. S. w. to Rocky Mts.
R. Yucatan.
R. Cozumel I.
R. n. e. Mex. & s. Tex.
R.s. Ariz. dc\CalilvanGalsiwe
Mex.

‘(| M. e. N. A. w. to Gt. Plains.
R. Cuba. [ Sonora.
R: Ss: es Gali We GaleisseAniz:.
R.2 w. U.S. & n. Mex. e. to

Rocky Mts.

R.? coast from Cal. to Sitka.
R. Guadelupe I.

2)2 MONTGOMERY. [VoL. XII.
TaBLE IV. VARIATION IN THE TYRANNIDA.
é o) g . - Rs MIGRATION.
DERE et ch |p Ge cera BREEDING AREA.
O/}F |e (Bale
Milvulus tyrannus (Linn.) ¢ HK werk) | R. Mex. to S. A. & Lesser
Antilles.
Tyrannus tyrannus (Linn.) * #* | | E. M. N. A. e. of Rocky Mts.
T. magnirostris D’Orb. * * | R. Cuba & Bahamas.
T. dominicensis (Gmel.) * | * ¥* | xxx! R. W. Indies, coast of G. of
Mex.
T. crassirostris Swains. le ¥* | R. Mex.
T. melancholicus couchi
(Baird) | «| + x* | #*| R. Guatemala & Mex. tos. Tex.
T. verticalis Say ¥e | eK x*| * | M. w. N. A. e. to Gt. Plains.
T. vociferans Swains. * | * x | x | R. Guatemala, Mex., L. Cal.
Pitangus derbianus (Kaup) * | * x* | * | Ron. S. A. to Rio Grande.
P. bahamensis Bryant * | * | * ¥* | R. Bahamas.
Myiozetetes texensis (Giraud) * * | * | R. Colombia to n. Mex.
Myiodynastes luteiventris Scl. | #* | * * | * | R. Mex. to Panama.
M. audax (Gmel.) * | * * | R. Cayenne, Trinidad, Tobago.
M. a. nobilis (Scl.) wee | OX ¥* | R. Costa Rica s. to Ecuador.
M. a. insolens Ridgw. ae | xe%! R. s.e. Mex.
Myiarchus mexicanus (Kaup) |x*+*| ** | ** | * | ** | R. Guatemala n. to e. Mex.
M. m. magister Ridgw. eee] ee | * | & | x* | M. w. Mex. to s. Ariz.
M. crinitus (Linn.) wx | xe | * | % feox] M. ec. U.S. to Canada.
M. cinerascens Lawr. wee | x | % | % | 9%! M. w. U.S. e. to Rocky Mts.
M. nuttingi Ridgw. wee] * | * ¥* | R. s. Mex. to w. Costa Rica.
M. brachyurus Ridgw. Sehr (23 » | R. Nicaragua.
M. yucatanensis Lawr. * | * | * * | R. Yucatan.
M. sagrae Gundl. * | * | * * | R. Cuba.
M. lucaysiensis Bryant * | * | x * | R. Bahamas.
M. lawrencei (Giraud) wee] & |x ¥* | R. s. Tex. to Guatemala.
M. 1. olivascens Ridgw. xe | xe | & | * | * | M. w. Mex. tos. Ariz.
M. flammulatus Lawr. |e * | R.s. w. Mex.
Sayornis phoebe (Lath.) * x* | x | M.e. N. A. n. of Gulf St.
S. nigricans (Swains.) * xx | * | M. coast from Ore. to Mex.
S. saya (Bonap.) * * | x | M. w. N. A. n. to Saskatche-
wan.
Contopus borealis (Swains.) _ |#«+*) ##*| * | ¥* |xe*) E.M. n. U. S. northward.
C. pertinax Cab. he * | * | R. Guatemala to s. Ariz.
C. virens (Linn.) seek] Hee | He | x* | He | Me. U.S. n. to Canada.
C. richardsonii (Swains.) week | ae | 9 | & | ee] E. M. w. N. A. to int. of B.C.
C. brachytarsus Scl. x | ® | * | R. Yucatan & s. Mex.
C. bahamensis Bryant ae |Pseal ae x* | R. Bahamas.
C. caribzeus (D’Orb.) * * | R. Cuba.
Empidonax albigularis (Scl.) | ** | * | * x | R.s. e. Mex. & Guatemala.
E. difficilis Baird S| oe | eH] * ¥* | M. w. U.S. n. to Sitka.
2 + +*
E. flaviventris Baird & Hee] & | * * | E. M. n. U. S. northward.
* KE
E. bairdii Scl. * | * | * * | R. s. & e. Mex.
E. salvinii (Ridgw.) * |e] * xe] R. highlands of Guatemala.
E. acadicus (Gmel.) S| ee | ee | ae | I. M. e. U.S.
2 * *
E. pusillus (Swains.) fl] xe | * | | ® | we | E. M. w. N. A. to Sitka.
QD] we] | | | oe

No. I.] ORGANIC VARIATION. 273
TasLe IV. — Continued.
z alles
Guna es 5 g 5 : g i MIGRATION.
b|—| < a] < BREEDING AREA.
0/2 |e |BAl a
E. p. traillii (Aud.) S| * [xxl] * * | E. M.n. U.S. northward.
2 * *
E. minimus Baird B | ee | ee | eee x*« | E. M.n. U. S. northward.
° * *
E. hammondi (Xantus) o| | * | xe] * | * | E.M. w. N. A. n. to L. Slave
Lake.
2 * +k
E. wrightii Baird Gl xe] & | x * | M. w. U.S. to Rocky Mts.
2 * *
E. fulvipectus Lawr. * * | R. s. Mex.
E. fulvifrons rubicundus (Cab.
& Hein.) | * | * | * ¥* | R. s. Mex.
E. f. pygmzeus (Coues) * | x* [xe] * | * | R.s. Ariz. to w. Mex.
Pyrocephalus rubineus mexi-
canus (Scl.) * ** | * | R. Guatemala to s. U. S.
Ornithion imberbe (Scl.) * | * | * #* | R. C. A. tos. Tex.
O. i. ridgwayi Brewst. * | x | «| x* | xe | Ro? w. Mex. tos. Ariz.

“TABLE WV,

Pica pica (Linn.)

P. p. hudsonica (Sab.)

P. nuttalli Aud.
Psilorhinus morio (Wagl.)
P. cyanogenys Gray

P. mexicanus Riipp
Cyanocitta cristata (Linn.)
C. c. florincola Coues.

C. stelleri (Gmel.)

s. frontalis Ridgw.

s. annectens (Baird)

s. macrolopha (Baird)

s. diademata (Bonap.)

s. coronata (Swains.)
phelocoma floridana (Bartr.)
woodhousei (Baird.)

. insularis Hensh.

. californica (Vig.)

. c. hypoleuca (Ridgw.)
. sumichrasti Ridgw.

. couchi (Baird)

. sieberii (Wagl.)

A. s. arizone Ridgw.

>>> >>> braANA|NOA

Xanthoura luxuousa (Less.)
Perisoreus canadensis (Linn.)

VARIATION IN THE CORVID&.

eerered irs

sxnee 22 ¥
Kee ee ose, de

a

eee Eee Fx

xxx FEF

** RE
FE RRR RRR
FORK | HK | HK
HE | HK
RK *
* | RE
WRK | HK | RK
KR | WHE |
WK | OF | RR
RK | WK | RK
* WE | RK
* | WRK |
* *
* RK
FORK | RRR | AK
* aE | HK
* * *
RK! OK **e
* * Ed
*
*
* a *
* eK *
* HE
3K | HE | RE

.n. & c. Europe.

.w. N.A. from N. M. to Ariz.
weal:

.e. Mex.

e. Mex. & coast of Honduras.
. s. Mex. to Costa Rica.

.e. N. A. n. to fur countries.
Fla.

n. w. coast from Cal. to
Sitka:

Sierra Nevada.

n. Rocky Mts.

s. Rocky Mts. to n. Mex.

highlands of c. Mex.

s. Mex. to Guatemala.

Fla.

middle province of U.S. s.

to Mex.

Santa Cruz I.

coast from s. Cal. to Ore.

Sule@ale

s. Mex.

L. Rio Grande vall.

R. s. Mex. & southward.

R. n-w. Mex. & adjacent Ariz.
& N. M.

R. e. Mex.

R. N. E. to Minn., n. to arctic

regions.

booed des rieents sls

274

MONTGOMERY. [VoL. XII.

TABLE V. — Continued.

SPECIES.

P. c. nigricapillus Ridgw.
P. c. fumifrons Ridgw.
P. c. capitalis Baird

P. obscurus (Ridgw.)

Corvus corax Linn.
. c. Sinuatus (Wagl.)
. C. principalis Ridgw.

. c. behringianus Dybowski
. cryptoleucus Couch.

. americanus Aud.

a. floridanus Baird

. Caurinus Baird

. ossifragus Wils.

. mexicanus Gmel.
Nucifraga columbiana (Wils.)
Cyanocephalus cyanocephalus
(Wied)

OleVelevem cle leneie!

TABLE VI.

Ex Ex F

MIGRATION.
BREEDING AREA.

TAIL.

» | ** | R. coast region of Labrador.
were) eek] R. Alaska.

#*| * | R. Rocky Mts. from Ore. into
IBarAG
wee | He | Ro n. Cal. & n. Sierra Nevada
toeba iG:
* | R. Eurasia.
sex eee! R. w. U. S. to Guatemala.
R. n. N. A. from Greenland to
Alaska.
* | R. Commander Is.

R.s

xe | x | R.s. w. U.S. & tablelands of
Mex.
xxx] 4] R. ce. N. A. except arctic dists.
** | R.s. Fla.
* || R. Wash. Terr. to Kodiak.
#4* | xx] R. coast from Long I. to La.

weew] x | R. w. Mex.
* | * | R.w.N.A.from Ariz. toAlaska.

x*| * | R. w. N. A. between Rocky
Mts. & Sierra Nevada.

VARIATION IN THE ICTERID&.

Dolichonyx oryzivorus (Linn.

~—

Molothrus ater (Bodd.)
M. a. obscurus (Gmel.)

Callothrus zeneus (Wagl.)

Xanthocephalus xanthoceph
lus (Bonap.)

Agelaius pheeniceus (Linn.)
A. p. sonoriensis Ridgw.
A. p. bryanti Ridgw.

A. gubernator (Wagl.)

A. assimilis Gundl.

400340034003 = #040 094099 B HOON) 0840 OY

*

%

x * * % x

* | E. M.e. N. A. in U.S. w. to
Gt. Plains.

*

*

* |! M. U.S. & s. Can.

* | x | R. Mex., L. Cal., contiguous
Uns:

* | * | R. Rio Grande to Panama.

*

* | * | M. marshes of w. U. S.

** *
sexx joe] M. nearly whole temp. N. A.
RK RE

ye | * | R. n. w. Mex., s. Cal., lower

Col. val.

* | *« | R. Bahamas & s. Fla.

* *

¥* | #*! R. vall.of Ore. & Cal. into Mex.
eK eK

* | RyCuba.

No. I.]

ORGANIC VARIATION.

275

TABLE VI. — Continued.

gi
SPECIES. a
2
O
A. tricolor (Nutt.) S| x
Q| *
Sturnella magna (Linn.) S Hee
Q|
S. m. neglecta (Aud.) S| +
Q|
S. m. mexicana (Scl.) S| +e
: ee
Quiscalus quiscula (Linn.) ¢| *
2
Q. q. aglazus (Baird) o| *
Q. q. eneus (Ridgw.) S|
£
Q. macrourus Swains. G |
@ |
Q. graysoni Scl. o| x
Q| *
Q. major (Vieill.) |
Q |
Q. tenuirostris Swains. .°)

TABLE VII.

*

*

*eEx e Eee EE x

& MIGRATION.
z BREEDING AREA.
H
we | R. coast vall. from s. Cal. to
w. Ore.
RK
R. e. N. A. n. to Canada, w. to
Gt. Plains.
R. w. N. A. from Manitoba to
w. Mex.
R. s. Ariz. & e. Mex. to Costa
Rica.
** | M. Atlantic slope of U.S. from
N. E. southward.
x* | R. G. coast from Fla. to La.
#* | M.c. N. A. n. to N. E.
*
** | R. s. Tex. to Nicaragua.
RK
wee) R. w. Mex.
*
¥* | R.s. Atlantic & G. coast of U.S.
*
ax! R. c. Mex.

VARIATION IN THE FRINGILLID&.

Coccothraustes vespertina

(Coop.)} *
Pinicola enucleator (Linn.) eee
Pyrrhula cassini (Baird)

Carpodacus purpureus (Gmel.)

$

C. p. californicus (Baird) as

C. cassini Baird

C. mexicanus (Miill.) a

C. m. frontalis (Say) é

C. amplus Ridgw.:

Loxia curvirostra minor
(Brehm.)

L. c. stricklandi (Ridgw.)

L. leucoptera Gmel.

Leucosticte griseonucha
(Brandt)

RE

PRE

wr

WE

RR

*

xe! I. M. w. N. A. n. to B. C.
* | I. M.n. Eurasia.
* |M.? n. Alaska & portions of
Siberia.
x | M.e. U. S. northward.
* | M. coast from s. Cal. to B. C.
x%* | M. w. U.S. from B. C. to Mex.
* | Re. & s. Mex.
#««| M. w. U.S. from 40° lat. to
Mex.
** | R. Guadelupe I.
woe) M.n. N. A. e. of Gt. Plains.
*« | R.s.w. U.S. & Mex.
M. n. U. S. northward.
we! M.? Aleutian, Prybilof & Com-
mander Is.

276 MONTGOMERY. [VoL. XII.
TaBL_e VII. — Continued.
Z wo =
Sorel ase = alles MIGRATION.
Seed etal [eiaiiy Petal ete tS BREEDING AREA.
O|;2 |e |Bala
L. tephrocotis littoralis (Baird) |++**| ** | ** | #* |) M. coast mt. ranges of n. w.
N. A.
L. atrata Ridgw. ¥* | #* | * | * | x | M. (summer range unknown).
L. australis (Allen) were] ee | He | He | eee] M. high mts. of Col.
L. arctoa (Brandt) * M.? n. e. Asia.
Acanthis hornemannii (Holb.)
o | eR] | * | M.n. Greenland &e. Arctic A.
Ol «| * | * *
A. h. exilipes (Coues) & * | * | M. circumpolar continental re-
gions.
Q| x | * | #*
A. linaria (Linn.) S| wee] * | x x* | M. n. portions of N. Hemis-
phere.
© |e] & | *
A. 1. holbcelli (Brehm) S [eee] * | » | M. n. coasts of Eurasia, por-
tions of Alaska.
© |e] * | x #x
A. 1. rostrata (Coues) B He] | He ¥* | M. s. Greenland.
© | peek] 9 | doe *
Spinus psaltria (Say) * ¥* | x | M. w. U.S. from n. Cal. to Col.
S. lawrencei (Cass.) ial * | x* | M. Cal.
S. notatus (Du Bus.) + * |xe*| R. highlands of s. Mex. &
Guatemala.
S. atriceps (Salv.) * | R. Guatemala.
S. pinus (Wils.) * xee| x | E. Mon. U. S. northward.
Carduelis carduelis (Linn.) week] * #e*| x | R. Eurasia.
Plectrophenax nivalis (Linn.) ¢| * | * * | ** | M. circumpolar regions.
Q| * | * * | *
P. n. townsendi Ridgw. S| xe | me x* | R. Alaska, Prybilof, & Com-
mander Is.
QO; * | * *
P. hyperboreus Ridgw. $ * ¥en | Me ELallete
| * | * * | *
Calcarius lapponicus (Linn.) ¢ * x* | x« | E. M. circumpolar regions.
2 * *
C. pictus (Swains.) as * * E. M. int. of Arctic A.
C. ornatus (Towns.) * M. Gt. Plains n. to Saskatche-
wan.
Pooceetes gramineus (Gmel.) | #«| #*| * [eee] x | M. ec. N. A. from Va. to Ont.
P. g. confinis (Baird) seek] eee | | ee [vee M. w. N. A. from B. C. south-
ward.
Ammodramus princeps
(Mayn.)| **« | ** | «* | «* | ** | M. Sable I.
A. sandvichensis (Gmel.) seek] ee | xe | * | oe | M. n. w. coast from Una-
laschka south.
A. s. savanna (Wils.) 4% | ee | eee] ee | ee! Mon. U.S. to Labr.
A. s. alaudinus (Bonap.) * | * | x* bee ae] EE. M. w. N. A. n. to Alaska.
A. s. bryanti Ridgw. | OR ee ¥* | R. salt marshes of San Fran-
cisco Bay.
A. beldingi Ridgw. seek] xe | * | x |v! R. salt marshes of s. Cal. &
laCale
|

ORGANIC VARIATION.

277

TaBLe VII. — Continued.

WHOLE

No. 1.]
Beales
SPECIES. alo aline
Seid
O}F le
A. rostratus Cass. % | 86 | ee
A. r. guttatus (Lawr.)
A. bairdii (Aud.) we
A. savannarum passerinus
(Wils.)| «| #* | *
A. s. perpallidus Ridgw. wee | * | x
A. henslowii (Aud.) *
A. lecontei (Aud.) **
A. caudacutus (Gmel.) o| | * | x
Q| * | | x
A. c. nelsoni Allen ASW ees Ihe
QO| * | * | x
A. maritimus (Wils.) ¥e | ® | x
A. nigrescens Ridgw. tated Mikal
Chondestes grammacus (Say) HE
C. g. strigatus (Swains.) a
Zonotrichia querula (Nutt.) +
Z. leucophrys (Forst.) HK] | x
Z.1. intermedea Ridgw. xe | #* | x
Z. 1. gambeli (Nutt.) SPORE) 8K ee
Z. coronata (Pall.) **
Z. albicollis (Gmel.) **
Spizella monticola (Gmel.) **
S. m. ochracea (Brewst.) ¥e
S. socialis (Wils.) **
S. s. arizonz Coues *
S. pusilla (Wils.) +e
S. p. arenacea Chadb. *
S. pallida (Swains.) **
S. breweri Cass. *
S. atrigularis (Cab.) *
Junco aikeni (Ridgw.) S| * |x]
QO) * | * | *
J. hyemalis (Linn.) C5 eo Wea bs
Oiler ae
J. h. carolinensis Brewst. Gulley |) seal ee
Ola |)
J. h. oregonus (Towns.) S| ¥* | ee |
Q| * | we] *
J. caniceps (Woodh.) Baty ae | ae ||
O| * | He |
J. cinereus (Swains.) * | ee]

LENGTH.

EK | **

HE

*

MIGRATION.
BREEDING AREA.

M. coasts of s.Cal:, Il. Cali, &
Sonora.

R. Cape St. Lucas.

M. Gt. Plains from Da. to Sas-
katchewan.

M.e. U.S. & s. Canada.

M. w. U.S. e. to Gt. Plains.

M. e. U.S. to Gt. Plains, n. to
Ont.

| E. M. Gt. Plains from Da. to

Manitoba.
M. coast from Me. to N. Ca.

M. marshes of Miss. vall.

M. coast from Mass. to Tex.

R. s. e. Fla.

M. Miss. vall. n. to Mich.

M. w. U. S. e. to Gt. Plains.

M. e. Gt. Plains from Mon. to
Manitoba.

E. M. high mts. of w. U.S. to
Labr.

E. M. Alaska & Mackenzie r.
basin.

M. coast mts. of Cal. to B. C.

M. n. Cal. to Norton Sound.

M. n. U. S. northward.

M. Labr. & Hudson B. region.

E. M. Alaska.

M. e. N. A. to Gt. Slave Lake.

E. M. w. N. A. n. to 60° lat.

M. e. U.S. & s. Canada, w. to
Gt. Plains.

M. Gt. Plains from Tex. to
Wyoming.

E. M. Gt. Plains n. to Sas-
katchewan.

| M. w. U.S. e. to Rocky Mts.

R. Mex., L. Cal.
I. M. Rocky Mts. in Col. & Wy.

M. Me. to Alaska.

| R. s. Alleghany Mts.

M. coast from Cal. to Sitka.

| M. Rocky Mt. distr.

R. highlands of Mex.

278 MONTGOMERY. [VoL. XII.
TasLE VII. — Continued.
Z| leat
A =i] a faél 3 MIGRATION.
PECIES = = Paella: BREEDING AREA.
O};F | a IFH| &
J. c. dorsalis (Henry) S| * |) * * | R. s. Rocky Mts.
Q| * | * | * *
J. c. palliatus Ridgw. Gale | aes #*| R. s. Ariz. & adjacent Mex.
QO] * | * | ad
J. alticola Salv. * | * | * * | R. highlands of Mex.
J. annectens (Baird) S| * | * | x* | M. Ft. Bridger northward.
Q| * | xe | er
J. insularis Ridgw. a ee |x * | R. Guadelupe I.
2 * | * *
J. bairdi Belding ¥e | * * | R. mts. of s. L. Cal.
Amphispiza bilineata (Cass.) * xe | * | M.s. w. U. S. & Mex.
A. mystacalis (Hartl.) * | R.s. Mex.
A. humeralis (Cab.) * #* | R. s. Mex.
A. belli (Cass.) Ay * #* | R. Cal. to Cape St. Lucas.
2 * **
A. b. nevadensis (Ridgw.) ¢ * xe | M. w. U.S. from Mex. to Mon.
2 * +
Pencza zstivalis (Licht.) %» |x | * | we | xe | R. Fla. & lower Ga.
P. a. bachmani (Aud.) #4 | 90 | 9 | x] * | R. s. Atlantic & G. St.
P. mexicana (Lawr.) * | x] * | % | a] R. Mex.
P. botteri Scl. * * | R.s. e. Mex.
P. cassini (Woodh.) * | * | * | «| * | M.s. w. border of U. S.
P. ruficeps (Cass.) Mi] ae |e || anil ren || Call:
P. r. boucardi (Scl.) ee | ee | 90] & | & | R. Mex.,s. Ariz., N.M., L. Cal.
P. carpalis Coues * * | * | R.s. Ariz.
P. notosticta Scl. & Salv. Meal ae (lla * | Rois. Mex.
Melospiza fasciata (Gmel.) % | x% | | ae [ee | R. ec. U.S. & Brit. Prov. n. of
40° lat.
M. f. montana (Hensh.) * [eee] * | xe [ee) R. Rocky Mts. w. to Ore. &
Nev.
M. f. heermanni (Baird) see | ee | | & fee! R. int. of Cal.
M. f. samuelis (Baird) #* | eee] & | & lex%] R. coast of Cal.
M. f. mexicana Ridgw. x | xe | * |OR. Ss. Mex:
M. f. fallax (Baird) * | xe | ee | x | ee | R. Ariz.
M. f. guttata (Nutt.) % | xe] & | Hee le) R. coast from Ore. to Van-
couver.
M. f. rufina (Bonap.) * ee | &* [eH HH) R. coast of s. Alaska.
M. cinerea (Gmel.) xe | % | * | x) | R. Aleutian Is.
M. georgiana (Lath.) * #* | # | M. e. U. S. to Labr.
M. lincolni (Aud.) eee #« |e] E. M.n. U. S. northward.
Passerella iliaca (Merr.) * | x | * [xx] x | E. M. G. of St. Lawrence to
Labr. & Alaska.
P. i. unalaschcensis (Gmel.) x4 | soe] x | & | 4 | EE. M. coast of Alaska.
P. i. negarhyncha (Baird) peek, ee | ee | He | Hee | R. mts. of Cal.
P. i. schistacea (Baird) week) & | & | | ® | M. Rocky Mts.
Embernagra rufivirgata Lawr. * | * | | * | R. RioGrande vall. southward.
E. r. crassirostris Baird Pe #* | R. s. Mex.
E. r. verticalis Ridgw. gee | | * | R. Yucatan.
Pipilo erythrophthalmus
(Linn.) + ¥x% | x*| Ms. St. to B. A.
P. e. alleni Coues * xe] *# | R. Fla.
P. maculatus Swains. I all | *« | R. c. Mex. to Guatemala.

Nos 13]

TABLE

ORGANIC VARIATION.

299

VII. — Continued.

SPECIES.

. m. arcticus (Swains.)
. m. megalonyx (Baird)

. m. oregonus (Bell)
consobrinus Ridgw.
carmani Lawr.
. macronyx Swains.
chlorosoma Baird
. chlorurus (Towns.)
rutilus Licht.
fuscus Swains.
. f. mesoleucus (Baird)
. f. albigula (Baird)
. f. crissalis ( Vig.)
aberti Baird
ardinalis cardinalis (Linn.)
c. superbus Ridgw.
c. igneus (Baird)
c. coccineus Ridgw.
c. yucatanicus Ridgw.
c. saturatus Ridgw.
. carneus (Less.)
. pheeniceus Gould
Pyrrhuloxia sinuata Bonap.
Habia ludoviciana (Linn.)
H. melanocephala (Swains.)
Guiraca czerulea (Linn.) é
G. c. eurhyncha Coues é
G. cyanoides concreta

(Du Bus)
Passerina parellina (Bonap.)
P. amoena (Say)
P. cyanea (Linn.)

QOONOONON Ne oO

. versicolor (Bonap.)
. v. pulchra Ridgw.

. ciris (Linn.)

. leclancheri Lafr.

. rosite (Lawr.)
Sporophila morelleti (Bonap.)
S. torqueola Bonap.

S. corvina Scl.

Euetheia bicolor (Linn.)
Spiza americana (Gmel.)
Calamospiza melanocorys

laciiaeine)ino}iqo)
O30

Stejn.

% %
x * * ¥

* * | WING.

ee seme

ee KKK HR KH KK HH EE He «EE wT

$x aex Bu fe & x

i

x % * x * * ¥ x x | Tarsus.

* * * FE x ¥

MIGRATION.
BREEDING AREA.

M. Gt.Plains to Saskatchewan.

M. Rocky Mts. from L. Cal. to
Wash.

M. coast from Cal. to B. C.

. Guadelupe I.

. Socorro I.

vall. of Mex.

s. Mex.

. Rocky Mts. n. to Ore.

s. Mex.

Mex.

N.M. & s. Ariz.

jena:

Cal.

N. M. & Ariz. to Col.

e. U. S. n. to 40° lat.

w. Mex. to s. Ariz.

I. Gals

e. & c. Mex.

Yucatan.

Cozumel I.

s. w. Mex.

: N., coast of Ss As

Mex. & contiguous U. S.

M.n. U. S. & Canada.

.w. U.S. e. to Gt. Plains.

wees WICtSy

. w. U.S. n. to Col. & Cal.

an

SSEM PPA P APPAR RPE

R. C. A. to e. Mex.

R. s. & e. Mex.

M. w. U. S. e. to Gt. Plains.

M. e. U.S. & s. Canada to Gt.
Plains.

R. e. Mex. to Tex.

R. L. Cal. & w. Mex.

M. s. Atlantic & Gulf St.

R. s. w. Mex.

R. s. Mex.

R. Rio Grande to Costa Rica.

R. w. Mex.

R. e. Mex. to Costa Rica.

R. Bahamas.

E. M. e. U.S. to Rocky Mts.

M. Gt. Plains from Kan. to

beyond U.S.

280 MONTGOMERY. [Vov. XII.

TABLE VIII. VARIATION IN THE VIREONIDA.

z ae rye
Bolom s Flee MIGRATION.
VEGEES 5 5 < ee E BREEDING AREA.
Vireo altiloquus barbatulus
(Cab.)| ** | * * | ** | R. Cuba, Bahamas, s. Fla.

V. olivaceus (Linn.) | * a | * | E. M.e. N. A. n. to Hudson B.
w. to Rocky Mts.

V. flavoviridis (Cass.) * | * |! R. Rio Grande to Upper Ama-
zon.

V. cinereus Ridgw. * | * R. Cozumel I.

V. philadelphicus (Cass.) * | xex| E. M. e. N. A., chiefly n. of
Wits:

V. gilvus (Vieill.) x | | * | a | x | E. Moe. N. A. n. to Hudson B.
w. to Gt. Plains.

V. g. swainsoni (Baird) * | x | * | x Jaxx] E. M. w. U.S. ec. to Rocky
Mts.

V. flavifrons Vieill. * we | xxx] EDM. ec. U.S. n. of middle
Sts. w. to Gt. Plains.

V. solitarius (Wils.) % | * | * [ee] *« | E. M. ec. N. A., chiefly n. of
U.S;

V. s. cassinii (Xantus) ae | % | | x | * | M. w. U.S., chiefly on Pacific
coast.

V. s. alticola Brewst. we] & | & * | M. higher s. Alleghany Mts.

V.s. plumbeus (Coues) xe] & | * #* | M. Rocky Mt. dist. of U.S.

V. atricapillus Woodh. * * | x | M.s. Gt. Plains n. to Kan.

V. noveboracensis (Gmel.) * | * | * | a | x | E.M.e. U.S. w. to Rocky Mts.

V.n. maynardi Brewst. NU |e * | R. Key West.

V. bellii (Aud.) * | x | * [xeex! « | M. Gt. Plains n. to Da.

V. b. pusillus (Coues) thee [759 | ee || ae) | ee Ress Cals Wa Cal. Ariz,

V. ochraceus Salv. we | * | R. Yucatan to Guatemala.

V. huttoni Cass. we | * | * | x |] * | R. Cal.

V. h. stephensi Brewst. wre] & | *® | * | * | R. Mex., Tex., L. Cal., Ariz.

V. pallens Salv. * * | R. w. coast of Costa Rica &
Nicaragua.

V. vicinior Coues 3798 |! 3 oa =| Rots. Cale, “Ariz: Ne Ms new
Mex.

V. gundlachi Lemb. * | * | R. Cuba.

V. hypochryseus Scl. * * | R.s. w. Mex.

Hylophilus decurtatus (Bonap.) ¥* * | R. e. Mex. & Guatemala.

H. ochraceiceps Scl. * * | R. s. Mex. s. to Costa Rica.

TABLE IX. VARIATION IN THE MNIOTILTIDA.

Mniotilta varia (Linn.) +e sex] eee] E. M. e. N. A. from Potomac
r. to Hudson Bay.
Protonotaria citrea (Bodd.) * #* | * | M. Gulf St. & lower Miss. vall.
Helinaia swainsonii Aud. a | 906K | eK bee] ee] ML. Gulf St. & lower Miss. vall.
Helmitherus vermivorus
(Gmel.) ¢| * | * wee |e! M. e. U.S. n. to about 40°.
Helminthophila celata (Say) * wee] * | E. M. n. N. A. from Rocky

Mts. to Alaska.

No. 1.] ORGANIC VARIATION. 281

TABLE IX. — Continued.

Z| | 3 |az
SrECES: rs Z E : Z| 5 eee ak
S|2F/a Ale '
H. c. lutescens Ridgw. * * | * | M. coast from s. Cal. to Ko-
diak.
H. ruficapilla (Wils.) $ * #*| x | E. M.e. N. A. from n. U.S. to
Hudson Bay.
H. r. gutturalis Ridgw. As * * | M. w. U.S. from Rocky Mts.
to coast.
H. virginze (Baird) * * | * . mts. of w. U. S.
H. lucize (Coop.) * ied Maal Ariz., s. e. Cal., Sonora.
Compsothlypsis americana
(Linn.) * Sed Ral M. e. U. S. to Canada.
C. nigrilora (Coues) ¥* xe | lower Rio Grande vall.
C. insularis (Lawr.) * * Tres Marias Is.
C. graysoni Ridgw. * * Socorro I.
Dendroica tigrina (Gmel.) HER M.n. N. A. to Hudson Bay.
. olivacea (Giraud) el Tex. to Guatemala.
zestiva (Gmel.) ** see | M- ie: Sen: IN-VA®
. petechia (Linn.) * + W. Indies, Cozumel I.
cerulescens (Gmel.) * ee | M.e. N. A. from n. N. E.

northward.

. coronata (Linn.)
audubonii (Towns.)
maculosa (Gmel.)

. ceerulea (Wils.)

pennsylvanica (Linn.)

castanea (Wils.)
striata (Forst.)

dominica (Linn.)
d. albilora Baird
blackburniz (Gmel.)

graciz Coues

decora (Ridgw.)
nigrescens (Towns.)
chrysoparia Scl. & Salv.

virens (Gmel.)
townsendi (Nutt.)
occidentalis (Towns.)
vigorsii (Aud.)

. kirtlandi Baird

. discolor (Vieill.)
palmarum (Gmel.)

bP SYDUDUY SoOy YoU Uy » Dyed bEUD

p. hypochrysea Ridgw.

Seiurus aurocapillus (Linn.)
S. noveboracensis (Gmel.)
S. n. notabilis (Grinn.)

.n. U.S. northward.

.M. w. N. A. n. to B. C.

. M.n. N. E. to Hudson Bay.
M. Miss. vall. to Allegha-

5
=
oO
wn

.e. U. S. & Canada n. of
“40° lat.

. M.n. N. E. to Hudson Bay.
M. n. N. E. & Labr. to

>

e-
es
»
7)
a
»

. s. Atlantic St.

. Miss. vall. n. to Gt. Lakes.
M. e: N. A from ins U7)S.;

northward.

.s. Ariz. & N. M., Mex.

.s. Mex. & Graremale

.w. U. S.n. to Ore. & Col.

c. Tex. & highlands of Gua-

]

temala.

M.n. e. U. S. northward.
M. w. N. A. n. to Sitka.
w. U. S. near coast.

e: U: S. ne tolOnt:

oe eas ee rene n aoe
2

M.e.U.S.n. to Mich. &s. N.E.

E. M. int. of N. A., n. to Gt.
Slave Lake.

M. coast from Nova Scotia to
Hudson Bay.

E. M. e. N. A. to Alaska.

E. M. n. e. U. S. northward.

E. M. w. N. A. to Miss. r. &
Alaska.

282

MONTGOMERY.

[Vou. XII.

TABLE IX. — Continued.

SPECIES.

S. motacilla (Vieill.)
Geothlypsis formosa (Wils.)
G. agilis (Wils.)

. philadelphia (Wils.)

Q

. macgillivrayi (Aud.)

0340 O3+0 05

trichas (Linn.)

t. occidentalis Brewst.
melanops Baird
beldingi Ridgw.
rostrata Bryant
tanneri Ridgw.

coryi Ridgw.

- speciosa Scl.
poliocephala Baird

. palpebralis Ridgw.

. caninucha Ridgw.
Icteria virens (Linn.)

I. v. longicauda (Lawr.)
Sylvania mitrata (Gmel.)
S. pusilla (Wils.)

S. p. pileolata (Pall.)

S. canadensis (Linn.)

QAAAAAAHHAD a

O03 O05

Setophaga ruticilla (Linn.)
Cardellina rubifrons (Giraud)

Ergaticus ruber (Swains.)

Basileuterus culcicivorus
(Licht.)

B. belli (Giraud)

B. delatrii Bonap.

B. rubrifrons (Swains.)

z | CuLMEN.

* * *

eee be Pee eee be Fe EE we EEE wx | Wane.

* ¥ * | TARSUS.

Be)
oO} 4
ER| 3
Bale
xe | x
see | HK
see | JOR
*E
#K
HE
*
*
FREE)
*
oe
*
¥*
*
HK
*
x | x
*
+
wee | x
HORE] -&
**
¥*F | *
xe |
*
| RK
* | x

MIGRATION,
BREEDING AREA.

. S.n. to Gt. Lakes.
SUE Sato SayNeeie
anitoba.
. northward.

i

New Providence I.
Obaco I.
Ae If

e. Mex.
w. Mex.
e. Mex. & Yucatan.
Guatemala to Costa Rica.
wens. S. n. to Ont.

to INE.
. S. northward.
. to Kodiak.
.n. to L. Winni-

rs ee ee Ho pbb

R. highlands of Guatemala to

S-- ANZ.
R. highlands of e. Mex.

R. Veragua to e. Mex.

R. Guatemala & e. Mex.
R. Guatemala to Panama.
R. s. Mex.

TABLE X. VARIATION IN THE TROGLODYTIDA.

Oroscoptes montanus (Towns.)
Mimus polyglottos (Linn.)

M. lawrencei Ridgw.

M. gracilis Cab.

M. gundlachii Cab.
Galeoscoptes carolinensis
(Linn.)
Mimodes graysoni (Baird)
Harporhynchus rufus (Linn.)

PREK

* *

ee

RK

HK

M. Artemisia Plains of w. U.S.

R. 38° lat. to Mex. & Bahamas.

R. s. Mex.

R. Atlantic coast from Yuca-
tan to Honduras.

R. Bahamas, Cuba, Jamaica.

E..M: e.. N. A.'n. to'54° lat.

R. Socorro I.

M. e. U. S. to Rocky Mts., n
to Ont.

ORGAMWIC VARIATION.

283

TABLE X. — Continued.

No. 1.]
cae
SPECIES. 5 z
5 |e
H. longirostris (Lafr.) Sed ia
H. guttatus Ridgw. me lle
H. cinereus Xantus * | *
H. bendirei Coues * | *
H. curvirostris (Swains.) pe | *
H. c. palmeri Ridgw. eee |
H. c. occidentalis Ridgw. ee | *
H. redivivus (Gamb.) ead ea
H. lecontei (Lawr.) ead ied
H. crissalis (Henry) Saal iia
Heleodytes brunneicapillus
(Lafr.)] + | *
H. affinis (Xantus) ¥e | x
Salpinctes obsoletus (Say) wee)
S. guadeloupensis Ridgw. * | *
Catherpes mexicanus (Swains.)|#+* +++
C. m. conspersus Ridgw. per]
Thryothorus ludovicianus
(Lath.)| «++ | +
T. 1. miamensis Ridgw. se | eK
T. albinucha (Cabot) * | *
T. bewickii (Aud.) #e | *
T. b. spilurus (Vig.) #ee | *
T. b. bairdi Salv. & Godm. _|*##*) +
T. brevicaudus Ridgw. * | *
T felix Scl- * |
T. lawrenci (Ridgw.) * | *
T. maculipectus Lafr. ee | *
T. m. umbrinus Ridgw. * | *
T. m. cannobrunneus Ridgw. *
Troglodytes insularis Baird * | *
T. beani Ridgw. we | *
T. aédon (Vieill.) ee | eK
T. a. parkmanii (Aud.) eed Mee
T. intermedius Cab. ¥* | *
T. brunneicollis Scl. * | *
T. hiemalis Vieill. ae | ob
T. h. pacificus Baird ¥e | *
T. alascensis Baird xe | *
Cistothorus stellaris (Licht.) | «** | »*
C. polyglottus Vieill. | allie
C. palustris (Wils.) Le) 40m

C. p. paludicola Baird

WHOLE
LENGTH.

as MIGRATION.

Z BREEDING AREA.

cal

¥* | R. e. Mex. & s. Tex.

* | R. Cozumel I.

eRe uGailk

wre | R. s. Ariz.

* | R. Mex., s. Tex. & s. N. M.

xe RACSSeAniz:

* | R. coast of w. Mex.

***| R. Pacific coast of Cal. & L.
@alké

¥* | R. vall. of Colorado and Gila

rivers.

#* | R. N. M., Ariz., s.Utah, s. Cal.,

neleaCale

* | R. s. w. border of U. S.

I Reis. LauEale

x* | R. arid distr. of w. U. S. s. to

Guatemala.
x*| R. Guadelupe I.
* | R. Mex., s. Tex.
seek) Ros. w. U.S. n. to Ore.
wee] Mn. e. Mex. & e. U.S. to 40°
lat.

wee | R.s. e. Fla.

» | R. Yucatan & Guatemala.

#% | M. e. U.S. n. to 40° lat., w. to
Gt. Plains.

#*% | R. coast from w. Mex. to B.C.
xe R. table-lands of Mex. to Kan.

Rk. Guadelupe I.

x | R. w. Mex.

x | R. Tres Marias Is.

* | R.s. Mex.

* | R. Guatemala.

x | R. Yucatan.

x | R. Socorro I.

x | R. Cozumel I.
wore) M. e. U. S. & Canada, w. to

Miss. vall.

xx | M. w. U.S. s. to Vera Cruz.

#* | R.s. Mex. to Costa Rica.

x* | R.s. e. Mex.
xe) M. n. e. U. S. northward.

x* | M. coast from s. Cal. to Sitka.

x R. Aleutian & Prybilof Is.

» | M.e. U. S. w. to Gt. Plains.
wee] R. e. trop. A. from Mex. to
| Brazil.
|#ex| M. e. U. S. & Brit. Prov.
| x | M. w. U.S. to Rocky Mts.

284 MONTGOMERY. [VoL. XII.
TaBLE XI. VARIATION IN THE PARIDA.
ge] .| $186 M
SPECIES. s|/zZi2iegi a B pe
Syl leet eee m| < REEDING AREA.
O/B] a |RAle
Sitta carolinensis Lath. xe | x we] ae | Roe. U.S. & Brit. Prov.
S. c. aculeata (Cass.) ee | oe ee] eee] Ro w. U. 5S. into Mex.
S. canadensis Linn. * sae M. chiefly n. of U.S.
S. pusilla Lath. ee ** R. s. Atlantic & Gulf St.
S. pygmza Vig. * ee R. w. U.S. to Mex.,e. to Rocky
Mts.
Parus atricristatus Cass. al wx! R. e. Mex. to s. Tex.
P. a. castaneifrons Sennett * * | R.e. Tex.
P. inornatus Gamb. * | * | x | 8 |ae%! R. coast from s. Cal. to Ore.
P. i. cinerascens Ridgw. Socal ao) a Rs: le Gal.
P. i. griseus Ridgw. wee] & [ae | % | ee | R. Rocky Mt. distr. of U.S.
P. wollweberi (Bonap.) * | #* | #* | R. highlands of Mex.to s. Ariz.
P. gambeli Ridgw. a we] * | R. mts. of w. U.S.
P. meridionalis Scl. x | * | a | * | R. highlands of Mex. n. to s.
Ariz.
P. carolinensis Aud. x |e] * fae] Roe. U.S. s. of 40° lat.
P. atricapillus (Linn.) x | x [ee] * | M.e. N. A. n. of 40° lat.
P. a. occidentalis (Baird) wee] * [ose] ee | Ron. w. coast distr. of U.S.
P. a. septentrionalis (Harris) ¥ | #e* Seer doer | R. Rocky Mt. distr. from N. M.
to Alaska.
P. cinctus obtectus (Cab.) * | * | * | * | R.e. Siberia & n. Alaska.
P. hudsonicus Forst. ¥% | d% [seek foe) Rv? n. N. A. e. of Rocky Mts.
P. rufescens Towns. sex | eet] ae | se] R. coast from Ore. tos. Alaska.
P. r. neglecta Ridgw. ae | HEE ¥* | R. coast of Cal.
Psaltiparus minimus californi-
cus Ridgw. we | ee | ae | see] R. Cal.
P. m. grindz (Beld.) * * | R. s. L. Cal.
P. plumbeus Baird * | # | | * | R. Rocky Mts. from Col. to s.
Ariz.
P. melanotis (Hartl.) x | * * | R. highlands of Mex. & Guate-
mala.
Auriparus flaviceps (Sund.) * ax* [sex | R. arid distr. of n. Mex. & con-
tiguous U. S.
Chamea fasciata Gamb. we la #*| R. coast of Cal.
C. f. henshawi Ridgw. ae | 38 | oe se | R. int. of Cal.

TABLE XII.

VARIATION IN THE TURDIDA.

Myadestes townsendi (Aud.)
M. obscurus Lafr.

M. o. occidentalis Stejn
M. o. insularis Stejn
M. unicolor Scl.

Turdus mustelinus Gmel.
T. fuscescens Steph.

FRR

* }

*

*

FHK

* SE
*
*
we | ORE

M. mts. of w. U. S. to B. C.

R. highlands of e. Mex. & Gua-
temala.

R. c. & w. Mex.

R. Tres Marias Is.

R. highlands of s. Mex. & Gua-
temala.

M. e. U. S. to Mass.

E. M. e. N. A. between 40° &
50° lat.

No. 1.]

ORGANIC VARIATION.

285

TABLE XII. — Continued.

SPECIES.

. f. salicicolus (Ridgw.)
. alicize Baird
. a. bicknelli (Ridgw.)

. ustulatus (Nutt.)

u. swainsonii (Cab.)

. aonalaschkez Gmel.:

. a. auduboni (Baird)

. a. pallasii (Cab.)

. iliacus Linn.

Merula migratoria (Linn.)

H84888 48H

M. m. propinqua Ridgw.

M. confinis (Baird)

M. flavirostris Swains.

M. graysoni Ridgw.
Hesperocichla nzvia (Gmel.)
Cyanecula suecica (Linn.)

C. wolfii Brehm

Saxicola cenanthe (Linn.)

Sialia sialis (Linn.)

S. s. azurea (Baird)

S. s. guatemale Ridgw.
S. mexicana Swains.

S. arctoa (Swains.)

+O OF OHO 0303400

eK

MIGRATION.
BREEDING AREA.

M. Rocky Mts.

E. M. Labr. to Arctic coast.

E. M. Catskill Mts. to Nova
Scotia.

E. M. Pacific coast n. to Sitka.

Ei. Me. NA na of Unis.

M. coast from Cal. to Kodiak.

M. Rocky Mts.

M. n. e. U. S. northward.

M. n. Eurasia.

M.e. N. A. n. to Hudson Bay
& Alaska.

M. w. Us S.n. to) B. C.7e..to
Rocky Mts.

Ress Uacalt

R. w. & s. Mex.

R. Tres Marias Is.

M.w. N.A.n.to Behring Strait.

M. n. Eurasia.

M. c. Europe.

M. n. portions of N. Hemis-
phere.

M. e. U. S. w. to Rocky Mts.

R. highlands of Mex. & s. Ariz.
R. highlands of Honduras &
Guatemala.

Ri ws Ue Sin. toy be Cees ato
Rocky Mts.

M. Rocky Mts. n. to Gt. Slave
Lake.

The following table (XIII) represents the percentage of
species and subspecies examined, of fourteen families, evincing

variation of at least 1.5% in two dimensions.

Only such species

and subspecies enter into the computation for which Ridgway’s
measurements express variation in at least three of the five
dimensions ; and the percentages have been deduced separately
for non-migratory, migratory, and extensively migratory species

and subspecies.

286 MONTGOMERY. [Vou. XII.

TABLE XIII.

NON-MIGRATORY EXTENSIVELY MIGRA-
MIGRATORY SPECIES.
SPECIES. TORY SPECIES.
FamILy.
ace kes Per cent. Number Per cent. Nomber Per cent.
examined. examined. examined.

Anatidze 3 333 31 35-4 17 41.1
Rallidze 3 66.6 8 2 5 80
Scolopacidz I Too 21 42.8 22 50
Falconidz 25 32 10 40 4 25
Picidze 36 19.4 8 2 I fo)
Tyrannidz 29 10.3 13 23 8 25
Corvidee 42 30.9 ° fe)
Icteridz 17 17.6 5 40 I fo)
Fringillidze 59 22,7 64 26.5 1 40
Vireonidee 9 fo) 5 fe) 7 14.2
Mniotiltidz 14 fo) 19 26.3 24 B38
Troglodytidz 37 21.6 II 63.6 I 100
Paridz 22 36.3 I fo) fo)
Turdidze 9 PUI 15 333 5 60

The following table (XIV) represents the percentage of
species- and subspecies examined, of nine families, which
exhibit an amount of variation of at least 1.5% in two out of at
least three dimensions examined. These percentages are com-
puted separately: for (1) distinct species (z.e. without geograph-
ical races), and (2) geographical races. In this computation
necessarily Melospiza fasciata, e.g., is considered a race, just as
is MZ. f. montana.

ee ee eS ene,

TABLE XIV. i
Distinct SPECIES. GEOGRAPHICAL RACES. A
FamILy.
Number Number
; Per cent. : Per cent.
examined. examined.
Anatidze 56 25 12 B33
Rallidz 12 50 2 100
Falconidz 26 34.6 13 30.7
Picidze 19 Tb 7, 26 19.2
Corvidze 17 2355, 24 Be°8
Fringillidz 60 21-0 74 32.4
Troglodytidz 31 22:5 19 47-3
Paridze II 272 17 29.4
Turdidz 14 14.2 16 43-7

|

No. 1.] ORGANIC VARIATION. 287

The following four tables (KX V—XVIII) give respectively the
percentage of those species and subspecies examined, in each
family enumerated, showing an amount of individual variation
of at least 2% in the particular dimension (culmen, wing, tarsus,
or whole length). Only such families are computed, for which
Ridgway’s data express individual variation in the particular
dimension for at least ten species.

TABLE XV.

Comparative tabulation of percentage of species examined, with variation in the
length of the cu/men of at least 2%.

+ ocd
og
Famuy. | 222 | 0% | 5% | 1% | 10% | 20% | 30% 50%
A %
Vireonidz 18 side
Tyrannidz 43 13.9
Mniotiltidz 14 14.2
Icteridz 31 16.1
Corvidee 35 16.6
Turdide 24 16.6
Falconidz 64 17.1
Fringillidz 109 17.4
Troglodytide| 47
Picidze 42
Rallidz 16
Paridz 10
TasBLeE XVI.

Comparative tabulation of percentage of species examined, with variation in the
length of the wig of at least 2%.

5.29
oO
FamILy. E 22 0% 20% | 30% | 40% 50%

als
Vireonide 26 °
Mniotiltida 71 fo)
Paridz 26 fo)
Turdide 32 fe)
Icteridz 43 fo)
Corvide 42 fe)
Tyrannidee 63 fo)
Fringillidze 181
Troglodytide | 49
Picide 48
Falconidz 89

Rallidz 16

288 MONTGOMERY. [Vou. XII.

(PABLED OV LI.

Comparative tabulation of percentage of species examined, with variation in the
length of the ¢arsus of at least 2%.

+ 3 z
FamIiy. E oe 0% 10% | 20% | 30% | 40% | 50%

Zs §

Vireonidz 14 fe)

Paridze 20 fe)

Tyrannidee 34 °

Icteridz 2 fe)

Fringillidz 99

Corvidze 34

Turdidz 25

Mniotiltidz 18

Rallidze 16

Troglodytide | 46

Falconidz 78 |

TABLE XVIII.

Comparative tabulation of percentage of species examined, with variation in the
whole length of at least 2%.1

he ies
Famity. E oo 0% 5% 1%
ag
aa
Picidze 31 3:2
Tyrannidz 28 7-1
Icteridz 39
Fringillidze 118
Vireonidz 16
Falconidee 53
Paridz 22
Turdidz 20
Corvidee 34

Troglodytide | 34
Mniotiltidze 53
Rallidze 12

1 Whenever a species has been considered in the calculations of Tables XV-XVIII, for which
Ridgway gives the extremes of variation in a particular dimension separately for the sexes, I have
computed and added the percentage of each sex as an equivalent to that of a species. This method
would hardly impede the general accuracy of the percentages.

No. 1.] ORGANIC VARIATION. 289

TABLE XIX.

Exhibiting variation in relation to sex, the figures in the vertical columns refer-
ring to the number of species and subspecies examined.

$ 2
FamIiLy. Larger, with Smaller, with Larger, with Smaller, with
greatest greatest greatest greatest
variation. variation. variation. variation.
Culmen.

Falconidz fo) 10 13 fe)

Trochilidz I 6 fe) I

Icteridze 6 fe) ° 4

Fringillidze 7 3 ° I

Wing.

Falconidze ° 19 16 °

Trochilidz 2 5 I fo)

Icteridz 8 fo) ° 5

Fringillidze 13 Oo I 9

Tarsus.

Falconidz 2 18 8 I
Trochilidz

Icteridz 3 fe) fo) 2

Fringillidz 6 fo) I 6

Whole Length.

Falconidz ° II 9 °

Trochilidz fo) 5 2 °

Icteridz 5 fo) ° 8

Fringillidz I oO fo) 4

54 Hi sr 41

Totals, ae +

ata ——10)2)
TABLE XX.

Differing from the preceding table in so far, that in addition to the four families
therein enumerated, all species and subspecies of other families are computed, for
which Ridgway has given individual extremes of measurements for each sex.

$ &
Dimensions. Larger, with | Smaller, with Larger, with | Smaller, with

greatest greatest greatest greatest

variation. variation. variation. variation.
Culmen 15 19 15 7
Wing 41 26 20 21
Tarsus II 20 9 Io
Whole length 16 17 12 14
83 82 56 52
Totals, + —_-—_-

290 MONTGOMERY. [VoL. XII.

C. Direct Inferences from the Tabulated Data.

(a) It is the rule that, in genera comprising more than one
species, those species which inhabit small or insular breed-
ing areas do not evince as much individual variation in the
dimensions as do species with more extensive and diversified
breeding areas. This fact becomes at once apparent after a
study of the Tables I-XII, when a comparison is made
between species inhabiting small islands, or other restricted
districts, and those which have a much wider distribution. But
few exceptions are to be found to this rule.

(6) It is the rule that species with geographical races, when
the latter differ from one another in one or more dimensions,
evince a greater amount of individual variation than do species
which are not divided into such races, provided that the breeding
area is approximately equal in extent and diversification in both
cases. Thus of the nine families tabulated in Table XIV, in
all, with the single exception of the Falconide@, a greater per-
centage of geographical races evince variation to the amount
of 1.5% in two dimensions, than of species which are not split
into races. If, however, the geographical races of a species
differ from one another mainly in color (as e.g. those of Cardina-
lis cardinalis), and less or not at all in dimensions, then as a
rule they do not evince a greater amount of variation in the
dimensions than do species without geographical races.1_ In
Table XIV, further, the percentages in favor of geographical
races may still be increased, when we consider that many
species which are as yet regarded as distinct, may in the
future be classed by ornithologists as subspecies, — resulting
in a subtraction from the left hand column of percentages,
and an addition to that on the right.

(c) It is the rule, subject to the preceding two “laws,” that
migratory species evince a greater amount of individual varia-

1 Though I have not investigated in birds the facts of individual co/or varia-
tion, —a kind of variation of which it is obviously difficult to determine the
amount, I would be inclined to conclude, @ f77or7, that the races of a species differ-
ing from one another mainly in color would present a greater amount of color

variation than would species which are not divisible into color races, other
factors being equal in both cases.

No. 1.] ORGANIC VARIATION. 291

tion than do non-migratory species ; and species which under-
take extensive migrations, a greater amount than species
which make migrations of less magnitude (Table XIII). This
fact is conformable with our “law” a, since migratory species
have as arule more extensive breeding areas than have non-
migratory species. Thus of the 110 species which undertake
extensive migrations, entering into the computation of table
XIII, 104 (94.5% inhabit breeding areas of comparatively great
extent, while but 6 (5.4%) inhabit small areas.}

(Z2) It is the rule, that males exhibit a greater amount of
individual variation in the dimensions than do females of the
same species or subspecies. For of the 223 computed meas-
urements of both sexes in Table XIX, in 131 cases (58.7%) the
males show the greater amount of variation, and in 92 cases
(41.2%) the females, —a difference of about 17.5% in favor of
the males; and of the 273 measurements of both sexes com-
puted in Table XX, in 165 cases (60.4%) the males show the
greater amount of variation, and in 108 cases (39.5%) the
females show the greater amount,—a difference of about
20.8% to the advantage of the males.

(ce) It is the rule that there is less variation in the length
of the wing than in the length of the culmen, tarsus, or whole
bird. This fact becomes at once apparent, by comparing the
“curve” of variation expressed in Table XVI with the
“‘curves’’ of variations of the other three dimensions (Tables
DV OWT, OSV EDT):

We may now consider the support given by these five
“laws” to the thesis, that continuing development is always
accompanied by variability.

To recapitulate briefly : it follows from the data given, that
the greatest amount of individual variation occurs, as a rule, in
those species occupying the most extensive breeding areas ;
that of two species occupying breeding areas approximately
equivalent in extent, the one divided into geographical sub-

1 These six species are Ammodramus lecontei, Spizella monticola ochracea,
Passerella iliaca unalaschcensis, Chen caerulescens, C. hyperborea, Turdus aliciae
bicknelli ; of these, only the first and third evince variation to the amount of 1.5%
in two dimensions.

292 MONTGOMERY. PVOL. XII

species evinces a greater amount of variation than the “stable”
species ; and that species which undertake extensive migra-
tions exhibit more individual variation, other factors being
equal, than do species which do not migrate. Now in the first
part of this second section, we have found that the presence
of geographical races (subspecies) and migration are two
criteria of continuing development. Therefore, the fact that
the amount of variation is greater in migratory species, and in
species which exhibit geographical races, than in non-migratory
species, or than in species which present no geographical sub-
species, is a sufficient proof for the assertion that continuing
development is always associated with variability. In other
words, individual variation is greater in amount in those
species which we must consider under the influence of a con-
tinuing process of development than in those species which
we must consider as being influenced by no process of develop-
ment at all, or by a much less energetic development. And as
we have found variation is as a rule more marked in migratory
than in non-migratory species, and in extensively migratory
than in less migratory species ; and further, that as a rule,
greater individual variation is present in the several races of a
species which exhibits a large number of races than in the
races of a species exhibiting a smaller number, — therefore we
must conclude that the amount of individual variation stands
in a direct ratio to the activity and energy of the operating
process of development. In short, the logical sequence from
the facts given is plainly that the amount of individual
variation stands in a direct ratio to the degree of complex-
ity of the environmental forces which influence the organism.
A species with an extensive breeding area, or one which
meets with environmental changes in the course of its
migrations, is more variable than a species with a restricted
and little diversified breeding area, or than a non-migratory
species, which does not come into contact with new environ-
ments.

The fact that the dimensions of birds are more variable in
the males than in the females is interesting, as offering a
parallel to the case in man, where, too, the males are more

No. 1.] ORGANIC VARIATION. 293

variable.! Also in domesticated animals, the males are more
variable (Darwin). Why the wing should be less variable
than the other dimensions, is difficult of solution on any other
ground, than that the wing has but one main function, while
the tarsus and the bill are put to a diversity of uses, which
would result in the two latter being more variable.

III. ON THE ORIGIN OF VARIATION.

The doctrines of Lamarck and Darwin teach that the
organism is more or less adapted to its environment ; even
Bateson (/.c., p. 10) grants a certain amount of such adaptation,
though in the main he militates against the assumption of any
complete degree of adaptation. We may then start out from
the assumption, for the correctness of which there is strong
evidence, that it is a biological law for the organism to be more
or less completely in correlation with its environment, in order
to insure its existence. It is even permissible to go further,
and assert that its chance of existence will stand in direct pro-
portion to the degree of its adaption to the environment. (By
the term environment, which is used here in its full sense, is
meant the sum total of all the external forces acting upon the
organism.) Therefore it is obvious that the chance of the
organism’s existence must depend upon its degree of correla-
tion with the environment, or, in other words, with its readiness
of response to the external conditions. This law being so
reasonable, and so thoroughly in accord with our modern
biological ideas, it is not necessary to enter into any further
discussion of its plausibility.

For the purpose of advancing further deductions as to the
primal cause of variation, we may proceed from the considera-
tion of a species which, we have reasons to conclude, is, com-
paratively speaking, perfectly adapted to its environment. We

1 Tt is not impossible, that in birds, as in man, the female may be more con-
servative and less progressive than the male, and passing a more (physiologically)
monotonous existence than the latter, is less influenced by the struggle for
existence, and accordingly is less variable structurally. This suggestion has, how-
ever, no more value than that of a mere comparison.

294 MONTGOMERY. [Vor. XII.

may take, for example, a species inhabiting a small island,
which is comparatively uniform in character throughout its
extent (in vegetation, etc.), and at all seasons of the year
(climate, etc.). Such insular species are very numerous among
land-inhabiting animals of tropical distribution. The environ-
ment influencing the insular species being then so uniform and
unchanging in its action upon it, the species could more easily
and quickly adapt itself to it, than if the environment were
changeable. Taking then an insular species, which, we have
reasons to suppose, has inhabited a particular island for a com-
paratively long period, —and there are usually certain criteria
whereby we may judge whether its residence there has been of
long duration, — we must suppose that it had time to become
adapted to its environment; and if we have equally good
reasons for supposing that the character of the environment
itself has not changed, it must seem probable that its adaptation
to the environment is comparatively perfect. Granting sucha
species, accordingly, to be closely adapted to its environment,
let us consider what changes, if any, would occur in the
organism, if the environment should change. And it may be
remarked just here, that marked changes have been recorded
by geologists and others in different districts, as is well known,
such as a surface rising or sinkage, changes in vegetation due
to a prolonged drought, the invasions of organisms strange to
the region, destruction of life caused by epidemics, etc.
And in fact, in many districts where the environment appears
to the human sense to be practically unchanging, a slow and
gradual change may nevertheless be taking place.

Now such an organism is adapted to its environment, at the
same time that its various organs must necessarily be cor-
related to one another. By correlation of the organs is meant,
as was explained in the introductory part of this paper, their
mutual dependence upon, and concerted physiological action
with, each other. Thus, in the case under question, we must
treat together the two facts: (1) the adaptation of the organism
to its environment, and (2) the physiological and morphological
correlation of its organs. When a change of environment
occurs, z.e. when consequently a new and different environment

No. I.] ORGANIC VARIATION. 295

commences to influence the organism, the latter cannot at first be
adapted to this new environment, but must become, so to speak,
out of touch with it. This change of environment must then
influence the organism, primarily, by disturbing or interrupting
the correlation of its organs. For even should the change of
environment interrupt directly the physiological action of only
one of the organs, this particular organ would no longer be
capable of acting in concerted harmony with the other organs,
and thus the correlation of the whole would be disturbed. For
example, if the customary food supply become exhausted, so
that the organism is compelled to seek nourishment of a dif-
ferent kind, not only in the immediate intestinal cells must a
physiological (and consequent morphological) change ensue,
but also the structure and function of each organ, indirectly
dependent upon the intestine’s action, must become modified.
And, pari passu, if the external forces acting upon a sense-
organ assume a different direction, not only must this particular
organ become physiologically modified, but indirectly also the
nervous system, and all organs in functional communication
with the latter; or if a certain organ become diseased, its
normal action must become to some extent impeded, which
would exert an influence upon the functions of the other
organs. In fact, whatever organ be directly influenced by a
change in the environment, a modification of the functions and
structure must result in those organs which are in correlation
with it.

Accordingly, after a change in the environment, a temporary
disturbance of the correlation of the organs must result. When
we speak in this way of an “interruption ”’ or of a “disturbance”
of the correlation, we mean that the mutual dependence of the
several organs upon one another is reduced, so that they become
to an inverse extent independent of one another. Now I con-
sider that the origin of variation is to be found in this condition
of temporary independence of the organs, which independence
is caused by change of environment, resulting in the partial
interruption of the organs’ correlation. For each organ, after
the disturbance of such mutual correlation, becomes more az‘o-
dynamic, —more indep¢éudent of the restraining influences of

296 MONTGOMERY. [VoL. XII.

the others, and consequently its own forces of growth and
action may operate more freely and to greater extent than was
possible in the previous condition, when it was held in check
by the state of correlation with the other organs. In other
words, after any change of the environment, that is, after any
consequent interruption of the correlation of the organs, each
organ becomes temporarily more autodynamic than it was before,
and the comparatively independent action of its vital forces
may result in the production of abnormalities, which are known
as variations. To what extremes these variations may go in
amount or extent will be considered in the next section.

The question arises at this point in the argument : why, when
through the interruption of the correlation of the organs a
certain degree of temporary independence of each of the latter
ensues, should any of the particular organs exert this indepen-
dence in the production of structural variations? Now we
have seen that the chances of existence of an organism stand
in a direct proportion to the degree of its adaptation to the
environment ; and further, that a complete correlation of its
organs is necessary hefore a perfect adaptation to its environ-
ment is possible, thereby its chance of existence depending
upon the degree of correlation of its organs. Accordingly,
when the correlation of the organs has been disturbed by a
change in the environment, it has but a small chance of exis-
tence until this correlation is restored. This deduction can
hardly be questioned, because it is difficult to conceive of an
organism existing when the correlation of its organs is greatly
impaired. It follows, therefore, that the organism must attempt,
in its fight for life, to restore this correlation, which is obviously
a step necessary for its becoming adapted to the new environ-
ment. Plainly, then, after the correlation of the organs has
become more or less interrupted, the several organs would exert
their degree of temporary independence of one another in such
a manner as to restore the correlation. Bearing this point in
mind, and remembering in this connection the previous assump-
tions of our argument, we conclude: that organic structural
variations are the morphological results of physiological exertions
on the part of the organism, to restore that complete correlation

NO: 12] ORGANIC VARIATION. 297

of its organs which had been disturbed by a change of environ-
ment. Even in the case of a disease attacking an organism,
may we not regard any abnormal growth produced by the
afflicted organism itself, to be the structural result of vital
processes in the organism striving to restore the correlation of
its organs? Indeed if we consider, which we must in view of
the facts at hand, that correlation of the organs is a physio-
logical necessity for the existence of an organism, then, if this
correlation be disturbed by a change of environment, we are
logically forced to conclude that the organism must make the
attempt to restore this correlation, and that structural variations
would be the result of such a physiological exertion.

A similar explanation of the origin of structural variations
can be reached also from another standpoint : when a change
of environment disturbs the correlation of the organs, the cor-
relation when restored must differ to some extent from the
previous state of correlation, since the new environment exerts
a different influence upon the functional activity of the organs.
But, as a different state of correlation cannot be conceived as
existing in the same structural unity, certain changes of struc-
ture are necessarily involved, and these we term “variations.”

The statistics given in this paper on variation in birds show
that species with restricted breeding ranges and which are non-
migratory in habit are, as a rule, much less variable with regard
to dimensions — other factors being approximately equal — than
are species occupying more extensive and more diversified areas,
and which undertake periodical migrations of considerable mag-
nitude. We are justified in concluding from these facts, that
as a rule the amount of variation stands in direct ratio to the
degree of environmental diversity of the inhabited area, z.e. to
the amount of change of environment which influences the
organism. And, as the degree of interruption of the correlation
of the organs must be in direct proportion to the amount of
change in the environment, the greater variation in those species
of extended and diversified habitats is easily and solely explain-
able on our deduction, that the greater change in the environ-
ment should cause greater physiological independence between
the several organs ; and therefore the structural changes ensu-

298 MONTGOMERY. [Vou. XII.

ing from the physiological exertions to restore the correlation,
should be greater than in species of comparatively restricted
and uniform habitats which experience neither so many nor so
great changes of environment.

Recapitulation. — Organic variation owes its origin indirectly
to change of environment. For it is necessary for the existence
of an organism that its organs be correlated physiologically and
(consequently) morphologically, and no adaptation to its environ-
ment, a factor which is also necessary for its existence, can be
brought about until the organs become correlated. Now when
a change occurs in the environment, this change checks the
normal action of one or more of the organs, and influences indi-
rectly the others, so that the correlation of the organs becomes
temporarily disturbed or interrupted. The degree of disturbance
in the correlation of the organs, probably stands in a direct
ratio to the amount of change in the environment ; and accord-
ing to the statistics given above, the amount of variation cer-
tainly stands, as a rule, in direct proportion to the complexity
of the environment. Naturally, when an interruption of the
correlation occurs, the several organs become to such a degree
independent of one another as is the extent of its disturbance :
the greater the disturbance of the correlation of the organs,
the more autodynamic the individual organs become. In order
to adapt itself to the new environment, the organism must first
restore the correlation of its organs. Now the several organs,
being no longer held in strict restraint by the agency of a com-
plete correlation, make use of their temporary degree of physi-
ological independence in order to restore this correlation. Any
structural changes resulting from the exertions of the compara-
tively unrestrained (independent or autodynamic) physiological
forces of the organs, to restore their correlation, are organic
variations.

IV. On VARIATION AS A CRITERION OF DEVELOPMENT.

In the preceding pages I have tried to analyze briefly the
processes of progressive and regressive development, and from
a study of the facts of variation in birds, which show that the

INO: 23] ORGANIC VARIATION. 299

amount of variation stands in a direct proportion to the amount
of change in the environment, to advance a theory of the origin
of variation, —as due directly to the physiological exertions on
the part of the organism, to restore the correlation of the organs
which had been disturbed by a change of environment. This
theory is sustained by the facts given here, and if it be corrob-
orated by future studies, we may use it as a standpoint from
which to review the phenomena of organic variation, as offering
criteria for the study of the processes of development.

The question often recurs to the biologist engaged in com-
parative anatomical investigations, why in a certain species a
particular organ should be structurally variable, which is emi-
nently stable in allied species. It has, thus far, been the
method of the biologist to attempt an explanation of this
variability by reasons derived from his assumptions as to the
phylogenetic origin of the group, and of the different forms
comprising it. But may we not, conversely, acquire some
understanding of the hitherto unknown, or but hypothetically
conjectured, development and homologies of an organ, by start-
ing out from the facts of the phenomena of variation themselves?
This is a line of research inaugurated by Bateson, and which
may in time afford important results.

First of all, it is well to recall to mind the two factors neces-
sary for the existence of the organism: (1) its adaptation to
the environment, and (2) the correlation of its organs. When
the environmental forces become more complex in their action,
the intimacy of the correlation of the organs being more or less
in a direct proportion to the degree of complexity of these
external forces, as will be shown later, —then the structural
development occasioned by such a change of environment must
be a progressive one, if the organism would maintain its adapta-
tion tothe environment. But if, on the other hand, the change of
environment is tending toward a simplification of the previously
complex action of the environmental forces, then the organism
must undergo a regressive structural development in order to
remain in adaptation to the environment. In other words, if
the environment is becoming more complex, the structure of the
organism must also become more complex in order to insure its

300 MONTGOMERY. [Vou. XII.

adaptation ; while if the environment is becoming less complex,
then the structure of the organism must become less intricate.
For if the environmental action become more complex while
the structure of the organism remains as it was, or becomes
more simple, then the latter can no longer continue to exist ;! or
if the environment become less complex in its action while the
organism’s structure either remains as it was or becomes more
complicated, its extinction must be brought about. These
relations of changes in the environment to changes in the
structure of the organism, with reference to their effect on
the existence of the latter, may be graphically represented as
follows :

( + Progress
development of -= continued existence of organism.
ahs ip roy
complexity of + Regress.
environment development of ¢= extinction of organism.
organism
+ Progress.
development of -= extinction of organism.
mea Sat
complexity of + Regress.
environment development of -= continued existence of organism.

organism

Accordingly, when the phenomena of variation are more
clearly understood, and the direction and quantitative amount
of change in the environment can be determined, then it will
be possible to predict the future of a given organism.

Now variation, as I have tried to explain it, expresses a want
of correlation between the several organs; and such an inter-
ruption of the correlation can be caused only by the agency of
a change of environment. Accordingly, it is permissible to
state of a variable organ that it is not in complete correlation
with its fellow organs ; and consequently, further, that some

change is occurring, or has taken place, in the environment.

SD)

1 A possible example of such a case is that afforded by the now extinct group
of Ammonites.

INO Ie ORGANIC VARIATION. 301

Therefore, in order to explain the presence of variation in a
certain species of a group not present in the same organ of
closely allied species, we must compare the conditions of the
environment of the one with those of the others. Thus,
regarded from this standpoint alone, when variation is per-
ceptible in organ x of species A, but not so in organ + of a
closely related species 4, we may conclude that organ x of
species A is being influenced by some change of environment
which is not affecting the corresponding organ x of species B.
May we not consider that the particular organ in A is com-
mencing to develop in a new direction, while the organ in B is
remaining unchanged? By this would be merely shown, how-
ever, that whenever variation is noticeable, the organ evincing
it is tending towards an ultimate structural modification, due to
the fact of a change of environment already taking, or having
taken, place.

What light can the phenomena of variation throw upon the
phylogeny of organisms? I consider that it may be possible
to decide, with a certain degree of certainty, whether a given
species is developing progressively or regressively at the present
time, and whether in the near past it has progressed or degener-
ated; by basing our conclusions as to its course of development,
present and past, upon the direction and degree in which the
variation appears. This may seem to bea bold assumption, but
if the views expressed in this paper upon the nature and origin
of variation be probable, we may yet learn that the study of
variation may furnish valuable criteria for estimating the facts
of phylogeny. As Bateson observes (/.c., p. 6), two criteria of
phylogeny upon which much confidence is misplaced, namely
the ontogeny and the direct study of adaptation, are by no
means infallible ; so that to-day we have only the criterion of
the facts of paleontology, —a criterion which Bateson fails to
mention, with which we may feel ourselves secure. And this
being the case, we should gladly avail ourselves of a further
criterion, namely, the phenomena of variation.

In comparing the antagonistic states, progressive and regres-
sive development (cf. Section I), it was found that progressive
development leads towards a more complicated structural modi-

302 MONTGOMERY. [VoL. XIT.

fication, and regressive development to a structural simplification
of the organism. Now, as is generally conceded, species are
maintained in the struggle for existence by the preservation of
favorable individual variations, z.e. (to my mind), variations
which are favorable, in as much as they tend to produce a
complete correlation of the organs; and all such structural
variations are necessarily either more complex or more simple
than the normal. The preservation of favorable individual
variations which are more complex would result in the produc-
tion of a more highly differentiated species ; and, on the other
hand, the preservation of those which are less complex would
result in the formation of a morphologically less differentiated
species. Accordingly, we must first determine whether the
variations are above or below the normal, — more complex or
more simple. For if the variations are more complex, then
if they should be preserved a more highly organized species
would be evolved ; and if they are structurally simpler, and
should be preserved, a less highly organized. Similarly, judging
from the paleontological remains of a series of individuals of
a now extinct species, it might be possible, after a careful investi-
gation into the nature and amount of individual variations
exhibited by them, to conclude whether a more highly or a less
highly organized form, if any, had been produced. Thus we
should expect, that a given species A occurring in the Liassic
beds, presenting individual variations (osteological, e.g.) more
complex than the normal, would be represented in the Triassic
by a more highly differentiated species, if by any.

But the study of variation, thus far considered, gives us
criteria for only the future development of the species, so that
it remains necessary to seek criteria from the phenomena of
variation for the past phylogenetic stages.

And, firstly, it is desirable to determine as far as possible
what limits there are to individual variation. It seems to be
well ascertained that there is a limit to such variation, but
where that limit may be placed for a given organism, or what
organic law fixes it, is very difficult of experimental proof.
Since we find the amount of variation to be in direct proportion
to the amount of change in the environment, the question is,

No. I.] ORGANIC VARIATION. 303

in other words, how great a change of environment the organism
can withstand without serious injury. Now from a number of
carefully made experiments, as noticeably the recent observa-
tions of Davenport and Castle,! we find that an organism which
would be killed by a sudden change of temperature of 10° C.,
may become acclimated to that amount of increase in tempera-
ture if the change is made gradually. This fact proves that an
organism can withstand only a certain maximum amount of
sudden change of environment, while if the change be greater
than this maximum amount, death ensues. Still another fact is
important in this connection : lowly organized forms are, as a
rule, more wéderstandsfahig than highly organized forms; it
suffices, as an example, to call to mind the great changes of tem-
perature which are not injurious to certain disease germs and
swarm-spores, but which would cause the sudden death of a worm
or mammal. From these facts we may conclude: (1) that as a
given organism can withstand only a certain maximum amount
of change in the environment, and since the amount of variation
stands in a direct ratio to the amount of change in the environ-
ment, therefore the organism can produce only a certain
maximum amount of variation; and (2), since lowly organized
forms can withstand greater changes of environment, as a rule,
than can more highly organized forms, that the former can, as
a rule, produce a greater amount of variation than the latter
can. And these facts coincide perfectly with the general physi-
ological law that the more differentiated the organs become
structurally, the more intimate and complex becomes their
correlation ; for more highly differentiated organisms, with a
more complex correlation of their organs, are unable to produce
variations to the same amount as can more lowly organized forms
which have a less intimate correlation of their organs. Further,
the correlation of the organs in lowly organized forms, being
less complex, can be restored sooner after a change of environ-
ment than the correlation can be restored in more highly dif-
ferentiated forms after the same amount of change. Thus we
find that the amount of variation depends upon the degree

1“On the Acclimatization of Organisms to High Temperatures,” Arch. f.
Entwicklungsmech. d. Organismen, Bd. II, 1895.

304 MONTGOMERY. [Vou. XII.

of change in the environment, and upon the degree of dif-
ferentiation of the organism ; but that a certain maximum
amount of variation cannot be exceeded by the organism, and
this amount seems to differ for different organisms. Variations
must continue to be produced until the correlation of the organs
is fully restored, when the restraint exerted by this acquired cor-
relation upon the physiological processes of the several organs,
would prohibit the production of further variations. Therefore,
if the variations continue to be produced through a long period
of time, we must conclude that the correlation of the organs
has been greatly disturbed, which is equivalent to saying that
a comparatively great change has occurred in the environment.
If, however, but one considerable change has occurred in the
environment and the latter remains thereafter unchanged, then
the longer the period of time is which has elapsed since this
change, the nearer at hand will be the restoration of the cor-
relation of the organs, and consequently the less will be the
amount of variation. This will serve further to elucidate the
deduction made in a former paper! of mine (p. 483): that
“the amount of variability above or below a given mean will
stand in inverse ratio to the length of time in which the devel-
opment (progressive or regressive) has acted upon the given
organ.” If the change of environment be comparatively slight,
the restoration of the correlation of the organs might be fulfilled
in a single generation ; but if the change had been more marked,
this correlation might not be restored until after the lapse of
a large number of generations, during which time the produc-
tion of variations would continue, though their amount would
decrease as the time became longer.

Here, then, the phenomena of variation may furnish us with
a criterion for deducing, to some extent at least, the previous
conditions of existence, if not also the phylogenetic stages of
some organisms. For we may briefly consider, e.g. the fresh-
water Nemerteans, which are undoubtedly of marine origin.
In studying the comparative anatomy of this group of worms,
I was struck by the fact that while the nearest marine allies of

1 “The Derivation of the Freshwater and Land Nemerteans, etc.,”” JOURNAL
OF MorpPHoLoecy, XI, 2, 1895.

No. I.] ORGANIC VARIATION. 305

these freshwater species possess almost invariably four large
eyes, the freshwater forms, on the contrary, have a larger
number of eyes, varying from four to eight, which are also
smaller than those of the marine species. How is this vari-
ability in the number of eyes of the freshwater forms to be
explained ? Now, variability is engendered, in our view, indirectly
by change of environment ; and we know that the species in
question have changed their environment by migrating from
bodies of salt water into freshwater rivers and lakes. (Or,
in certain cases, they are inhabitants of lakes which were origi-
nally of marine character, but have become fresh.) The num-
ber of the eyes of the marine species being very stable, we
conclude : (1) that the correlation of the organs in the marine
forms is comparatively complete, and that, therefore, they are
well adapted to their environment ; and (2) that no variability
being perceptible, they are neither at present giving rise to new
species, nor are they themselves of recent origin. On the other
hand, the eyes of the freshwater forms being very variable in
number, we conclude for these : (1) that the correlation of their
organs is not perfect, and hence that they are not fully adapted
to their environment ; (2) that this variability must have been
caused by a change of environment within a comparatively
recent period of time, since the variability is still continuing ;
and (3), that as the numerical variation of the eyes is above the
normal four (the number 6-8 being in fact more frequent than
4 or 5), the species is tending to evolve a form with a greater
number of eyes than its ancestors possessed (progressive
numerical development in relation to the eyes).

A similar case occurs to me in regard to the individual varia-
tion in number in the rectrices (stiff tail-feathers), of certain
North American species of grouse. . Table XXI (placed at
the end of the paper) expresses this variation for Ceztrocercus
urophasianus, and for Dendrapagus obscurus with its two
geographical races.!_ According to Mr. Clark’s observations,
these are the only North American species of grouse evincing
such variation ; the number of rectrices being stable in Den-

11 am indebted to the kindness of my friend Mr. Hubert Lyman Clark, of
Baltimore, for the communication of the facts embodied in Table X XI.

306 MONTGOMERY. [ VoL. X11.

drapagus canadensis, D. franklini, and in Bonasa, Lagopus,
Tympanuchus, and Pediecetes. Now it is interesting that the
subgenus Dendrapagus of the genus Dendrapagus is limited
to North America, and its species are variable in regard to the
number of the rectrices ; while such variation apparently does
not occur in subgenus Cazace of the same genus, which inhabits
both Eurasia and North America. Of the other genera, Bonasa
and Lagopus have the same geographical distribution as Canace,
while Z7ympanuchus, Centrocercus, and Pediecetes are limited to
North America. Those individuals evincing this numerical
variation of the rectrices must have been influenced by a con-
siderable change of environment, since the variability has
continued through a large number of generations. Now we
can hardly suppose that this change of environment has occurred
within the areas occupied by these variable species in North
America, since in that case we must suppose that the other
species occupying the same areas must have become similarly
modified. Accordingly the species of Dexdrapagus which pre-
sents this variation must be of comparatively recent occurrence
within its present habitat ; and its variability would be caused by
having changed its former habitat within a comparatively recent
period, by migrating from its former area (Eurasia ?) to its pres-
ent geographical position in America. On the other hand, the
non-variable North American species of this genus (namely
canadensis and frankliniz) must be regarded either as not
having experienced such a change of environment, or as
having experienced it in a much remoter period, having in
the meanwhile restored the correlation of their organs, 2.e.
adapted themselves to the new environment. Regarded from
this standpoint, the variation in the number of the rectrices
may serve to elucidate the origin and present distribution of
the Zetraonide.

These considerations are offered as mere suggestions for
explaining how variation can serve as a criterion of development.
We have found that the data of variation offer a certain mean
for determining whether the development is progressive or
regressive, depending upon the fact whether the variations are
above or below the normal. When the phenomena of varia-

Not] ORGANIC VARIATION. 307

tion are better understood, for this study is at present nearly
virgin ground, we will probably be enabled to deduce from it
criteria, which will offer certain and valuable aid in the study
of phylogeny. And another important line of investigation
must become the study of the environment, and of the changes
of the latter in their influence upon the organism. A con-
siderable number of experiments have recently been made in
this line upon the ontogenetic stages of organisms, and a
smaller number upon the adult organism; and a careful
analysis of the results of such experiments may give valu-
able aid in the discrimination of methods for the study of
variation.

A final word in regard to regressive development. It is as
yet an undecided question whether regressive development (or
the action of Natural Selection during this mode of develop-
ment) can result in the total disappearance of an organ, or
whether a structural rudiment must remain. According to our
conclusions upon the nature of variation, the latter would be the
more probable view, and forthe following reasons. Forthe action
of a regressive development, the occurrence of variations below
- the normal are necessary, so that if these be preserved, a less
differentiated (z.e. retrogressive) type of organism must be pro-
duced. The occurrence of variations in regressive development
is due, just as in progressive, to the physiological exertions
of the organism to restore that correlation of the organs which
had been disturbed by a change of environment. Now if the
change of environment had been a great and sudden one, and
resulted in a less complex environment, the organism, in order
to become adapted to the new environment, must produce
structural variations which are less complex than the normal in
order to bring about a less intimate and complex correlation of
the organs than had previously existed. Such variations can
obviously be produced only as long as the organs continue to
to be physiologically active ; and necessarily their action must
cease before the organs disappear. In other words, it is impos-
sible for variations to lead to the total disappearance of an
organ, since the very production of variations is dependent
upon the physiological exertions of the organ. Thus we must

308

postulate for regressive, as has been shown for progressive
development, a certain maximum amount of variation, the

MONTGOMERY.

passing of which would cause death.

TA

BLE XXI. INDIVIDUAL VARIATION IN THE NUMBER OF THE

RECTRICES OF CERTAIN N. A. TETRAONID.

NuMBER OF INDIVIDUALS

She Sat EXAMINED.
Centrocercus urophasianus 2a¢ : fo sae
(Bonap.) 3 ae 19 ves
[ iS
I
Dendrapagus obscurus (Say) 623,499, 2 (sex?)
L 266,19
1g
D. obscurus fuliginosus I juv. 3
Ridgw. 3 i on I (sex ?)
466, gee 4 (sex ?)
[ ie
D. obscurus richardsonii 266, 4 (sex?)
(Sab.) I*(sex ?)
1¢

NuMBER OF
RECTRICES.

16
18
20
16

17
18

WIST’ . *-STITUTE OF ANATOMY AND BIOLOGY,

PHILADELPHIA, Jan. 4, 1896.

4
c
;

;
aS
il
3
a

ere Y ee

PRY TRS
OF oy 7
EDITED BY
Cc. O. WHITMAN,
po Head Professor of Biology, Chicago University.
: WITH THE CO-OPERATION OF
oe EDWARD PHELPS ALLIS.
Reee es Milwaukee.
Vor. XII. FEBRUARY, 1897. No. 2.
BOSTON, U-S:A-:
GINN & COMPANY.
AGENT FOR GREAT BRITAIN: AGENTS FOR GERMANY: AGENT FOR FRANCE:
EDWARD ARNOLD, R. FRIEDLANDER & SOHN, JULES PEELMAN,
37, Bedford Street, Strand, Berlin, N.W., 189, Boulevard Saint-
London, W.C. Caristrasse 11. Germain, Paris.
“BY e

CONTENTS OF No. 2, FEBRUARY, 1897.

PAGES

I. The Egg of Amia and its Cleavage. . . . 309-354

C. O. WHITMAN AND A. C. EYCLESHYMER.

II. On the Modes of Development of the Mesoderm
and Mesenchym, with Reference to the Sup-
posed Homologies of the Body Cavities. . 355-366

Tuos. H. MONTGOMERY, JR.

III. Some Spinning Activities of Protoplasm in
Starfish and Echinus Eggs . . . . . 367-389

GWENDOLEN FOULKE ANDREWS.

IV. The Origin of the Egg Centrosomes. . . . 391-394
A. D. MEap.

V. Development of the Human Coelom . . . . 395-453

FRANKLIN P. MALL.

VI. A Contribution to the Study of Variation . . 455-484

HERMON C. BUMPUS.

Che Atheneum Press,
GINN & COMPANY, BOSTON, U.S.A.

Volume XII. February, 1597. Number 2.

JOURNAL

OF

MOR HOLL OGe

SHE EGG OF AMIA AND ITS CLEAVAGE:

C. O. WHITMAN anp A. C. EYCLESHYMER.

PREFACE.

THE search for the eggs of Amia, begun by Louis Agassiz,
and renewed from time to time by others, was first successful
May 1, 1887, when the senior author of this paper found nest
and eggs in Pewaukee Lake, Wisconsin. There has been no
lack of interest in the search on the part of American natural-
ists, but a number of reasons conspired to delay the discovery.
The nature of the breeding-grounds and the habits of Amia
were little understood, and those most interested in securing
the material lived at long distances from the more promising
localities.

The difficulties in the way of the first discovery seem to be
little appreciated by those who have recently collected these
eggs and made free use of information well known to have
originated with one of us, but with no acknowledgment, except
to parties who were either wholly innocent of anything but bor-
rowed knowledge of the subject or had acquired their first
information under our guidance.

310 WHITMAN AND EYCLESHYMER. [VoL. XII.

So far as we have been able to learn, no eggs of Amia had
ever been collected or seen before the date above given. No
one knew where to look for them, whether in deep or shallow
water, under stumps and logs, or in open places, on sandy or
marshy bottom. Whether they were to be found free or adher-
ing to roots or leaves, scattered among grass and reeds or col-
lected in nests or beds, no one could foretell. Fishermen had
seen the young in swarms along the shores of many of the Wis-
consin lakes, but they knew nothing of the nesting habits and
could only guess at the time of spawning.

Amia is very shy and nocturnal in habit. Its nests are not
designed to catch the eye. One who had never seen them
might pass over dozens and fail to notice them. They are often
concealed beneath supernatant grass or reeds, and appear, at
first sight, like natural depressions. The eggs agree so closely
in color with the ground that the inexperienced observer must
get close over the nest to see them, and even then he may
fail if the water is the least clouded with mud or rippled by the
wind.

The male fish, which alone guards the nest, lies motionless
at the approach of a boat, and often allows it to pass over him
without stirring. If frightened by a jar in the boat or by the
movement of the oars, he leaves the nest, but in doing so gen-
erally raises a cloud of mud, and darts off under its cover,
stealthily, but with wonderful rapidity. The nest and eggs are
often thus effectually concealed from view. If this behavior
were something special and peculiar to the care of the nest, as
one might suspect on first acquaintance, it would serve to indi-
cate approximately its location ; but one soon learns that not
every streak of mud in the wake of an Amia leads to a nest.
Indeed, one may search for hours on such trails and find nothing.

Amia is fond of shallow and quiet bays, where the water first
gets warm in spring, and hiding-places are easy to find among
the reeds and logs. In such places they spend the day in quiet
concealment. When approached in a boat pushed forward with
great care to avoid noise and disturbance of the water, they
may keep their places, if concealed from view, letting the boat
pass by or even over them without moving. Sometimes they

pl ee

NO W235) (THE TEGG OF AMIA AND TIS CLEAVAGE. 311

steal away, under cover of plants, so quietly as to escape obser-
vation. If suddenly surprised they dash off with such speed
that one may get no distinct image of the fish, and see only a
streak of muddy water marking its path. A few fish escaping
in this way often leave the water of a small bay so cloudy that
the search for eggs has to be abandoned until the mud settles.

The difficulties of the search for these eggs were, then, not a
few, even in Pewaukee Lake, which is undoubtedly one of the
most favorable localities, and thus far the only one where they
have been found in abundance. In lakes where the water is
constantly cloudy, as in the small lakes south of Chicago, the
difficulties are so great that, with all our experience in collect-
ing these eggs, we have not yet succeeded in finding a single
nest, after looking for them for two seasons. There must be
breeding-grounds somewhere in these lakes, for Amia is abun-
dant, both young and old.

After collecting for several seasons at Pewaukee, and after
much searching in the rivers and other lakes of Wisconsin and
Illinois, we are not greatly surprised that the eggs of Amia
escaped early discovery. But how did it happen that the dis-
covery once made was not forthwith announced, and followed
up with the usual haste to secure priority? Although the fact
was not announced in print, no secret was made of it, and it
soon became known to many American and European natural-
ists. As to “priority,” no apology need be offered for indif-
ference. ‘Priority’? weighs less and less as investigation
weighs more, and, so far as the egg of Amia is concerned,
there appears to be little reason to envy either the methods or
the results of priority-seekers.

To those who have known where these much coveted eggs
could be found in abundance, and who have courteously re-
frained from helping themselves, out of respect for first claims,
a word of explanation is due for the failure to bring forth
results at an earlier date.

When Mr. Allis started his Lake Laboratory at Milwaukee,
under the direction of Mr. Whitman, it was suggested by the
latter that the eggs of Amia and Necturus would be two as
interesting subjects for embryological research as could be

312 WHITMAN AND EVCLESHYMER. [VoL. XII.

found in the lakes of Wisconsin. It was decided, in case the
eggs could be found, that Mr. Allis should devote himself to
Amia, and Mr. Whitman to Necturus. Mr. Allis knew the
lakes and the fishermen, and was thus most helpful in direct-
ing to favorable localities. It was not long before the eggs of
both Amia and Necturus were found. During the first season
only a few Amia eggs were obtained, but the young were to be
had in great numbers. It was decided, therefore, that Mr.
Allis should begin his study with the development of the “lat-
eral line”? system, for which there was abundant material, and
wait until the next season before attempting to trace the devel-
opment of the egg. Mr. Allis became deeply absorbed in this
study, and when the time came for collecting eggs again, the
work had progressed to a point where it could not to advantage
be laid aside for the tempting study of the eggs. A third sea-
son came, and still there was no room for eggs, although a
fairly complete series had been collected by Whitman and his
assistant, Dr. Patten. At length the work on the lateral line
was brought to a close; but just then Mr. Allis’s health broke
down, and he was advised by his physician to go to Europe for
rest and for expert treatment of his eyes. From that time to
the present he has found it necessary to prolong his stay, but
not without maintaining a private laboratory, and, by the aid of
assistants, bringing another anatomical work on Amia, of great
extent and value, to completion. Beginning with the second
year of his absence, Dr. Ayers had charge of the laboratory in
Milwaukee, with Eycleshymer, Strong, and Nomura as assist-
ants, and continued the collection of Amia eggs. A large
number of elegant drawings of the egg were made by Mr.
Nomura, and the various stages were systematically sectioned,
and several cases of slides were prepared by Eycleshymer and
Strong, ready for study the moment Mr. Allis should be able
to return. The second plate of the present paper represents
some of Mr. Nomura’s work, directed by Dr. Ayers and Mr.
Allis, and generously placed at our disposal by Mr. Allis.

The material for the study of the embryology of Amia, now
in the possession of Mr. Allis, covers all stages, and the work
already done upon it is considerable. In order to keep this

NO:2.) (27 EGG OF AMIA) AND) TTS CLEAVAGE: 203

and related work moving, Mr. Allis kept the Lake Laboratory
open for the first four years of his absence, and then closed it,
with the expectation of reopening it as soon as he could resume
personal direction. All the while he has been pushing forward
the anatomical work on Amia (now completed and in press),
and has thus prepared himself in quite an exceptional way for
a thorough study of the development.

Those who know how faithfully Mr. Allis has pursued his
work under unexpected difficulties, and how he has continued
for years to divide his income in support of the investigations
of others, and in maintenance of a national medium of publica-
tion, — those who are aware of all this may wonder that a
report could get into circulation to the effect that the Amia
material was being unfairly monopolized. The implication
extends to all who have participated in the discovery, collec-
tion, and elaboration of this material (Whitman, Patten, Ayers,
Eycleshymer, Strong, and Nomura), for all have abetted the
crime, in so far as they have refrained from snatching the mate-
rial themselves or assisting others in such business. The prize
of priority, coupled with dishonor and theft, was not a distinc-
tion coveted at the Lake Laboratory. If to find what others
have failed to find, and to devote life and fortune to its investi-
gation, is unjust monopoly, then how might biology prosper in
the exchange of footpads for monopolists ?

Those who, knowing all the circumstances, have nevertheless
given currency to this censure, or allowed it to pass undisputed,
have something less than generous instincts to be proud of.
Mr. Allis’s great crime reduces itself to the misfortune of hav-
ing had to submit to delay in his work, to the neglect of all
tintinnabulous advertising, and to contempt for the notoriety
accorded to prolific scribbling and priority hustle. To all cour-
teous applications for Amia material Mr. Allis has freely re-
sponded to the extent that his resources and work would permit.

It remains to deal briefly with Dr. Fiilleborn’s mission for
the collection of the eggs of Amia and Necturus. From what
has already been stated, it will be clear that there were certain
claims and obligations in relation to this material which were
not to be entirely overlooked. Without a word of previous

314 WHITMAN AND EYCLESHYMER. PVOL cur:

notice or consultation with the parties primarily interested,
Dr. Fiilleborn came to us with the request for direction and
aid in getting the eggs. Under the circumstances, we could
only decline, explaining the reasons which stood in the way,
and pointing out some of the proprieties in the case which
should have received attention. It is quite possible that Dr.
Fiilleborn had not been previously informed of these, as was
to be expected that he had been by Dr. Virchow, under whose
instruction he was acting.

When, in 1893, Dr. Virchow applied to several parties for
assistance in finding the eggs of Necturus and Amia, he was
referred to Mr. Whitman. Learning that work on these eggs
was already in progress, Dr. Virchow courteously explained that
he only wished to study the “ Dotterorgan,” and that he would
not make any use of the material which would conflict with the
work on general development already begun by Mr. Whitman.
With that understanding, the information desired was freely
and fully given, and a pretty complete series of drawings of the
egg of Necturus was shown to him.

Next came Dr. Fiilleborn, under advice from Dr. Virchow,
to take advantage of the information confidentially obtained by
the latter, in collecting the same material for purposes well
known to conflict with studies in progress here. The mystery
of the “ Dotterorgan”’ was now disclosed, and no further com-
ments seem to be required.

OECOLOGICAL OBSERVATIONS.

1. Previous Papers.

There are only four works to which we can refer the reader
for information on the habits, eggs, nests, and larval history of
Amia. Mr. Allis! has given a series of drawings of the larva,
to illustrate the successive changes in form and the develop-
ment of the lateral-line organs. Dr. Filleborn? has recently

1 Edward Phelps Allis, Jr.: The Anatomy and Development of the Lateral
Line System in Amia Calva. Journ. of Morph., Il, p. 463. 1888.

2F. Fiilleborn: Bericht iiber eine zur Untersuchung der Entwickelung von

Amia, Lepidosteus und Necturus unternommene Reise nach Nord-America.
Sitzungsberichte der Akad. ad. Wiss. zu Berlin, XL, pp. 1057-1070. Oct. 25, 1894.

Nov2 LHEVEGG, OF AMIA AND TIS S€CLEAVAGE. 315

reported his experience in obtaining the eggs, together with
observations on the habits, times, and places of breeding. This
author describes the floating islands in Pewaukee Lake and
Fowler Lake and their labyrinth of canals, and speaks as if
Amia was so select in its choice of breeding-ground that its
nests were to be found only in these canals. That is not the
case, however, in either of the lakes just mentioned, and it
is quite certain that Amia breeds in many places where
there are no floating islands with canals, —for example, La
Belle Lake, which Dr. Fiilleborn cites as not containing a
single nest.

The spawning season is said to have extended from the
beginning of May to the first days of June in the summer of
1894. According to our observations, the best time for collect-
ing generally falls between the middle of April and the end of
the first week in May. We have collected eggs the last week
in March and the first week in June, but these extremes mark
unusually early or late seasons.

The time for hatching is estimated to vary between six and
fourteen days. This is a variation wide enough to include the
truth without hitting it in either direction. Surely one season’s
experience ought to have reduced both margins of this conjec-
ture, and furnished an answer to the question how long a time
is ordinarily required to hatch these eggs.

The nests described as ‘not yet filled with eggs,” in which
male Amiae were found, were probably nests in which the young
had hatched. The observation furnishes no ground for the
belief that the male alone makes the nest. The statement
that the nest is always open to the sunlight does not accord
with our experience, and the opinion that young and old seek
deep water after the first of June is entirely erroneous. The
assertion that nests containing “nur verschimmelte Eier”’ were
still guarded by males is another error due to superficial exami-
nation. Such eggs are frequently found in nests containing
newly hatched larvae, and deserted nests are occasionally occu-
pied as convenient resting-places during the day. It does not
follow because an Amia is found in a nest that it must be there
for the purpose of guarding it.

’

316 WHITMAN AND EYCLESHYMER. [VoL. XII.

The statement that the male takes its brood to the shore in
order ¢o sun itself and them would adorn a nursery tale better
than the Proceedings of an Academy of Sciences. Dr. Fiille-
born does not appear to have discovered that Amia prefers the
shadows of evening and the darkness of night to the glare of
the sun. Amia places its nest near the shore for warmth, not
for light; it leads its young along the shore for food,* not for
basking in sunlight. It is towards evening and in the early
morning that the swarms of young may be seen to best advan-
tage. As they grow older and more wary they may shift their
feeding-ground, not to deep water, but to new places along the
shore. Their seeming to disappear is accounted for by the
fact that the individuals of a brood wander more widely apart
as they grow older and require more food, and at the same time
they become increasingly shy, and hide themselves beneath the
banks, or whatever lies nearest, long before a boat in motion
can be brought within seeing distance. The boat must come
to rest and the observer must sit motionless by the half-hour if
he desires to see young Amiae feeding in June. A slight jar
of the water is enough to send hundreds of these young fish
out of sight in a flash.

The mistake in supposing that young Amiae retreat to deep
water soon after June 1 led Dr. Fiilleborn quite astray on
another point of considerable importance. The conclusion
that the young go straight on to become more and more like
the adult in both form and color is about as wide of the mark
as it could be. The changes in color exhibited at different
ages during the summer and autumn are as characteristic and
remarkable as the changes in plumage among birds.

Dr. Fiilleborn is careful to note that the young larvae with
large yolk-sacs, before leaving the nest, have a distant resem-
blance to tadpoles, but he neglects to mention that this resem-
blance becomes even more striking after the yolk-sac has
disappeared and after the nest has been abandoned. The
larvae, shortly after leaving the nest, while moving about in a
dense swarm, resemble tadpoles so closely in form, color, and

* The food of the young consists mainly of daphnia and cypris, as Dr. Ayers
informs us.

NON) LAEVM EGG OF AMIA AND US CLEAVAGE. Bey

motion as to be easily mistaken for them when seen at a dis-
tance of some yards, as from a boat. A stranger to the fish
and its habits might wonder at the closeness of the swarm, but,
if he did not stop to examine, he would take them for tadpoles.

Dr. Fiilleborn was in such haste for priority that he could
not stop to settle the important question as to whether the first
cleavage grooves actually reach the vegetative pole of the egg.
«The first grooves,” he says, ‘‘ appear indeed to reach the vege-
tative pole; but the investigation has not advanced far enough
to admit of definite statements.” This question might have
easily been settled on the living egg, and if the material
collected was so poorly preserved as not to admit of positive
statements on this point, it may be more “complete” than
instructive.

Only a little over four pages of this preliminary paper were
devoted to Amia,— certainly enough to show the need of more
careful observation.

The third paper to be mentioned in this connection is that
of Bashford Dean,? who affirms several times that he has fully
confirmed the statements of Fiilleborn on the “general” and
“spawning habits of Amia.’’ Perhaps the above remarks are
not superfluous after such confirmation. As Dr. Dean seems
to have depended a good deal upon what ‘‘the fishermen stated,”
the confirmations offered are not always very confirming.

It should be said, however, that the fishermen of Pewaukee
and Oconomowoc have had many opportunities for instruction
since 1887, and some of them are now better informed on the
habits of Amia than they were found to be at that time. Some
of the statements credited to them are entirely accurate, some
are crude guesses, and others pure “‘fish stories.” The following
will serve as examples:

“The fish were observed depositing their eggs as early as
April 25, and before the rst of May the spawning appeared to
have been generally completed.” Although no authority is
quoted, this is evidently a statement based upon a fisherman’s
story. It so happened that we collected eggs and larvae from

8 Bashford Dean: The Early Development of Amia. Quart. Journ. Micr. Sct.,
XXXVIII. February, 1896.

218 WHITMAN AND EYCLESHYMER. [VoL. XII.

the same locality during the week preceding Dr. Dean’s visit
in the spring of 1895. On the 12th day of May, between 4.00
and 5.30 A.M., a number of broods of larvae were caught, the
largest of which measured 25 mm. These must have been at
least 35 days old, as a glance at our table, p. 327, will show.
These facts show beyond question that eggs were laid during
the first days of April.

«At the height of a ‘run’ as many as a half-dozen nests, as
fishermen stated, were found to occur within the space of a few
square yards.” This needs confirmation. We have never found
the nests in such close proximity, — never more than four or
five in a single bay, and usually rods apart. The small bays
along the shore of Pewaukee Lake generally furnish ground for
but one, or at most two or three nests.

«Immediately before spawning, z¢ zs sazd that the fish divide
themselves into parties, each comprising a female and several
males, and that they then circle about nearer and nearer the
shallows.” True to the extent that the female may be con-
tended for by two or more males. They are on the ‘“shallows”’
to begin with, and hence do not need to “circle about”’ to get
nearer to them.

To say that “the fish divide themselves into parties’’ com-
ports well with a fisherman’s yarn, and shows how useful such
yarn may be in darning up gaps in observation.

“The mode of building a nest is in some ways doubtful ;
fishermen state that the spawning party prepares it by constant
circlings before the time of spawning, and this view seems
entirely corroborated by a careful examination of the newly
made nest; the soft weeds and rootlets appear bent and
brushed aside in a way that gives it somewhat the appearance
of a crudely finished bird’s nest.”

That the nest is built, and built with considerable care in
many cases, is evident enough. ‘The mode of building,”
however, is, indeed, doubtful; no less doubtful at the end
than at the beginning of the fisherman’s tale.

«“ And it seems evident that nests are prepared sometimes
well in advance of spawning, for several were noted by the
writer which were occupied by the fish for a number of days

Noi29 (LHe EGG OF AMIA AND TES CLEAVAGE. 319

before the eggs were deposited.”’ Dr. Dean does not give us
any information as to the number or sex of the fish which he
observed in these nests. From his statement, page 414, viz.:
«“ By the time of my visit, however, the spawning season had
practically ended,” we are led to suggest that these nests might
have contained newly hatched larvae. We have never observed
either the male or female occupying a nest for a number of
days, or even one day, preceding deposition.

«The mode of depositing the eggs appears to be entirely
similar to that described by the present writer in the case of
the gar-pike. The spawning fish leaves the nest from time to
time, returning in company. The eggs and milt are emitted
simultaneously. The fishes apparently rub close together,
since scales are found scattered in the nest bottom, as noted
by Fiilleborn, and as now confirmed by the present writer.”
How much of this is to be understood as personal observation
the author seems to avoid making clear. The words ‘“‘appears”’
and “apparently” look more like conjecture than anything
else. Leaving the nest from time to time could hardly apply
to the gar-pike, which has no nest. ‘From time to time”
suggests that oviposition occupies considerable time; and the
author conjectures that in some cases as many as twelve hours
are thus consumed. In this he is probably much mistaken, as
will appear later on. The occurrence of scales in the nest has
an entirely different explanation from the one above suggested,
as an observation made by Dr. Eycleshymer, to be related
presently, will make evident.

The nest represented on page 419 is so great an improve-
ment on nature’s art that one might readily conclude that it
was formed by “constant circlings.” The essential thing in
an Amia nest is fine rootlets or grass for holding the eggs free
from mud. If the fish finds a surface adapted to its needs it
may use it without making any excavation whatever. Such
nests are often found along the submerged edges of the float-
ing islands, or “bogs,” as they are locally called. Along the
shore of the lake the nests are sometimes placed beneath a log
or stump, but most frequently by the side of a clump of reeds.
Sometimes the fish finds rootlets only around the edges or on

320 WHITMAN AND EYCLESHYMER. [VoL. XII.

one side of its nest, in which case no eggs will be found on the
bottom. When excavations are made the depth may vary from
a few inches to a foot, depending upon how much mud and
weeds have to be removed in order to get the surface required
for the eggs. The nest is generally from 50 to 60cm. in
diameter, and more or less circular, varying considerably in
adaptation to the conditions of the ground.

The author’s remarks on the breeding habits end with a bit
of romance funny enough to claim a place beside the story of
Dr. Estes, which will be given presently.

Dr. Dean says: “A fine nest of eggs was found to be
entirely deserted at a time when the young could not have
been older than twenty-four hours. The closest search in and
about the nest revealed no trace of their whereabouts, although
from their larval habits it was thought that they should surely
be found attached to the neighboring weeds, or deep in the
mass of root fibres and detritus of the nest bottom. They had
evidently left the nest in a body, and were nowhere in the
immediate neighborhood. It was plausibly suggested by Mr. G.
W. Kosmak, who then accompanied the writer, that they had
been taken away by the male fish, attached to him by their
sucking-discs. It is certain that when the male reappears it is
with a swarm of nestlings; but they are now well grown
(5 to 1% inches).”

The earliest paper dealing with the habits of Amia to which
we have to refer is one cited by Dr. Goode from the Sports-
man’s Gazetteer, pp. 324-326, 1887.4 It is here that the story
of Dr. Estes, above alluded to, is found.

«The best description of the habits of this fish,” says Goode,
‘“‘is here quoted from the pen of Charles Hallock:

««They take frogs, minnows, and sometimes the spoon.
Their habitat is deep water, where they drive everything
before them. They are very voracious and savage. Their
teeth are so sharp and their jaws so strong that they have
been known to bite a two-pound fish clean in two the very
first snap. They are as tenacious of, life asthe celeb

* Geo. Brown Goode: Natural History of Useful Aquatic Animals, pp. 659,
660. 1884.

See

No..2:] ZHE EGG OF AMIA AND ITS CLEAVAGE. 321

young, when about six inches long, make a famous bait for
pickerel and pike. To use it, run the hook into the mouth
right up through the centre of the head, through the brain,
cast a hundred times, catch several fish, and at the end of
three to six hours he will kick like a mule. Put one hundred
in a rain-barrel and you can keep them all summer without
change of water. For the aquarium the young have no equal,
and on account of the spot in the tail are quite attractive; but
nothing else but snails can live in the tank. He will killa
lizard or any other living thing the instant it touches the
water.’

«Dr. Estes says: ‘I have sent these young dogfish hun-
dreds of miles for the aquarium. It is only necessary to keep
them in water, a change scarcely being required. The adults
are the great “jumpers” of the lake. On certain days they
are to be seen in all directions jumping clean out of the water,
and turning complete somersaults before again striking. They
Spawn in May and June among the grass and weeds of the
sloughs, i£ they can reach them in time. As soon as the
spring rise comes, usually in May and June, and connects the
inland sloughs with the lake (Pepin), they run up and over
into the sloughs, deposit their eggs, and remain near the beds
and young just as long as they can and not be shut in by the
receding water. The eggs hatch in eight and ten days, the par-
ents remaining with the brood two or three weeks, if possible,
but will leave them much sooner, if necessary, to save them-
selves. The young will not make any effort to escape to the
lake until the next season, when, if an opening occurs, they
come pouring out in countless numbers. At this time we take
them by stretching the minnow seine across the opening and
raising it when full. They are now from three to six inches
long, fat and chubby. J come now to mention a peculiar habit
of this fish, no account of which I have ever seen. It is this:
While the parent still remains with the young, uf the family
become suddenly alarmed, the capacious mouth of the old fish will
open, and in rushes the entire host of little ones ; the ugly maw
zs at once closed, and off she rushes to a place of security, when
again the little captives ave set at liberty. If others are conver-

322 WHITMAN AND EYCLESHYMER. [ViOr.w cn:

sant with the above facts I shall be very glad; if not, shall feel
chagrined for not making them known long ago.’ ”

The author of this amusing story is evidently quite blind to
the part played by hisfancy. The fish has no such strange habit,
and the wonder is that any sane mind could have invented such
a prodigious fabrication without realizing that the whole thing
was a myth. The facts from which the story was concocted
are: (1) The fish sometimes opens its mouth on approach (as
if threatening an attack); (2) the young suddenly disappear (in
the mud and weeds of the bottom); (3) the old fish darts off
and is quickly lost sight of ; (4) the old and young are found a
little later at a short distance from the place where they were
first seen.

The young fish disappear so quickly that the observer fails
to see where they go; the reappearance of the brood in com-
pany with the parent is taken as proof that he took them with
him, and the open mouth as proof that they were swallowed.

2. Behavior during Spawning.

Although the early cleavage stages of Amia have been
repeatedly collected, it was not until May 12, 1895, that the
behavior of the fishes immediately preceding and during spawn-
ing was witnessed. The observation leaves certain points in
doubt, but fills some of the gaps in the records hitherto made.

Starting at daybreak on the morning of May 12th, we soon
arrived at the “bogs.” After taking a number of broods of
larvae, we were slowly passing through a narrow channel which
in its deepest portion did not exceed 60cm. This channel was
filled with dead grass from the preceding year, while the growth
of the present season was barely visible above the surface of
the water. On one bank the green grass grew in profusion,
on the other the dead rushes formed a thicket. While resting
for afew moments a disturbance of the water not more than
3m. distant attracted our attention; stepping to the bow of
the boat we saw four large Amiae lying ina nest. Seeing us
they left the nest and disappeared in the surrounding grass.
Soon three of them returned, but after a few brisk movements

No:2i) (2H BGG OF AMIA AND TTS) CLEAVAGE. 322

again retreated, this time making short circuits in the adjoining
grass and frequently thrusting their snouts above the water.
We now easily ascertained that the party consisted of two
males and one female.* After a short interval the three again
returned, when a fierce battle for supremacy ensued between
the males. They approached from opposite sides of the nest
and locked jaws in a most ferocious manner. Their struggles
were so violent that a cloud of muddy water soon arose and
obscured them from view. When the female, which during the
battle had remained concealed at the side of the nest, again came
into view, the victorious suitor rushed at her and began to bite
her sides with so much vigor that a number of scales were
detached. The two then swam slowly about the nest, keeping
their bodies in close proximity. At short intervals this move-
ment was interrupted by momentary periods of vigorous activ-
ity. This lasted some five or six minutes, when they departed.
Believing this to be due to unavoidable movements of the boat,
we thought it best to retire and leave them undisturbed.
Returning after twenty minutes, and moving with greatest
care, we approached and found the male lying at full length in
the nest. Occasionally he would move slowly around the area,
keeping up a very rapid movement of the pectoral and pelvic
fins. We expected that our approach would frighten him ; but
not until the boat was directly over him did he show any signs
of vacating. It was now plainly to be seen that the nest was
filled with freshly deposited eggs. It is probable, though not
certain, that no eggs had been deposited at the time when we
first saw the fishes in the nest. The oviposition most probably
took place after the battle of the males, and during the time
(5-IO min.) the female was in the nest with the victorious
male. That the average period of deposition is brief can
hardly be doubted, since in most cases the eggs of a nest are
found in the same or nearly the same stage of development.
In one nest, however, we found some of the eggs in a late stage
of cleavage, while in others the embryo was already visible.

* The male is easily distinguished by a dark spot in the upper portion of the
caudal fin. This spot is fringed by a ring of dull orange, which during the
breeding-season assumes a brilliant hue.

324 WHITMAN AND EYCLESHYMER. [ VoL. XTi:

We inferred that two females sometimes deposit in the same
nest. This view is corroborated by the fact that in this case
the number of eggs was about double the average number
deposited by a single fish.

The eggs were scattered over an area about 50 cm. in diam-
eter, the greater portion being deposited upon the grass lining
the sides of the nest.*

Dr. Ayers has kindly given us the following observations on
the breeding habits of Amia, which supplement the foregoing
statements :

«The nest of Amia is built after pairing, and during the pre-
liminary steps to egg-laying. The nest is not a premeditated
structure, but merely the result of the movements of the fish
in and about the place selected for spawning during the period
of sexual excitement, which ends in the emission of eggs and
sperm. How long the movements continue I am unable to say
positively, but from many observations I conclude that they
usually last from two to three hours. The nest-forming process
generally begins at early dawn, and the eggs are cast about sun-
rise. I have observed some cases, however, in which the eggs
were deposited in the afternoon.

«« Amia usually selects grassy beds on which to deposit its
eggs. At the time of spawning, the dead grass of the previous
season is readily swept loose by the powerful swimming move-
ments of the excited pair, thus leaving the peaty stem and root
surface for the eggs to fall upon. All loose bits of stems and
leaves are swept out of the nest, and with them most of the
spores of fungi and algae which were in the area.

“In May, 1892, a nest was found at 3.30 P.M. one day, in

* In collecting the eggs we found it most convenient to clip off the blades of
grass or rootlets bearing them. They can thus be easily carried alive to the labo-
ratory, or placed at once in preserving fluids. In cutting the grass and rootlets,
one may avoid the discomfort of immersing his arm by using long-handled shears.
These may be improvised by making shears of two pieces of board about 1 m.
long. The short ends of these are then inserted into the handles of ordinary
shears and fastened by winding with wire; the blades will thus be nearly at right
angles to the wooden handles. It may not be superfluous to add that we have
found it convenient to mark nests by means of stakes, to the upper ends of which
were attached pieces of cloth of different colors. The value of such a guide will
be apparent to those who have attempted to revisit nests.

NOmZz ih Lie wE GG, Of AMITA AND TES. {LEA VAGE. 325
which the pair were still engaged in spawning. Eggs were
taken from the nest and at once placed under observation, and
the following record made: Three sets of eggs were put in
glass dishes, and Mr. Nomura, Mr. Eycleshymer, and myself
observed independently the segmentation of these fresh-laid
eggs. The first segmentation groove began at 5.30 p.m. The
second meridional began at 6.30 in those eggs which were
placed in the sun after the beginning of the first furrow, and
at 7 o'clock in those eggs kept in the shade out of the
sun’s rays.

“The third and fourth meridional furrows began simulta-
neously at 7.30. The first equatorial to be seen made its
appearance at 8.25, and completed the circuit at 8.40. The
next set of eight peripheral meridional furrows began at g
o'clock, and the second apparent equatorial furrow began at
9.30.

‘«« Amiae are to be found in Oconomowoc Lake during the win-
ter months, living in schools closely huddled together in the
bottom of pockets or shallow depressions of the gravelly bed of
the lake, among the water weeds. These fish may be speared
from a boat, or, if the lake is frozen over, through holes in the
ice. They lie so close together that occasionally two individ-
uals are impaled on the fish spear by one throw. When thus
disturbed they scatter from their resting-place, moving out a
short distance, to return quickly after the first few disturb-
ances. When alarmed by repeated thrusts of the spear, they
remain away for hours at a time. Why they should be so tena-
cious of these resting-places will probably be explained when
we know why they choose them in the first place. It is pos-
sible that the temperature of the water is higher in these places
than elsewhere, owing to the entrance of spring streams into
the lake bottom.”

3. Some Further Data.

In some of the eggs taken from this nest, and kept under
close observation, the first indication of the first cleavage ap-
peared at the upper pole at 7.54, z.c., 2 hrs. 24 min. from the
supposed time of deposition. At this time the first cleavage

326 WHITMAN AND EYCLESHYMER. (VoL. 2xie

had already begun in many of the eggs, while in others it had
not yet appeared.

At the end of 10-15 hrs., 100-150 cells have been formed
in the upper hemisphere, while in the lower, 15—20 grooves are
present, of which 8-12 have reached the pole.

To distinguish a stage which marks the beginning of gastru-
lation is somewhat difficult; if the stage shown in Cut 12
might be considered as such, the time would be 35-40 hrs.
after deposition. The process covers a period of 30-40 hrs.

In 70-80 hrs. after deposition the embryo appears. On the
7th or 8th day its first movements are visible; these occur
at long intervals, and last only afew moments. The periods
of activity become more frequent and vigorous until the 8th
or oth day, when the envelope is ruptured and the larva escapes.
The larva at this time measures 5-6 mm. Its body is pale
flesh color (without pigment), while the large yolk sac is deep
sepia. The larvae, as soon as they are hatched, attach them-
selves, by means of their adhesive discs, to the first object met,
and adhere so firmly that it requires some effort to shake them
loose. It is impossible at this time for the larva to move its
load of food yolk. Once losing its hold, it falls heavily to the
next object, where it again attaches itself.

At the time of hatching one frequently sees the male fish care-
fully guarding his nest, in which there apparently is not a single
living egg or larva, the only evidence of their former presence
being the fungus-covered débris scattered here and there. A
careful search, however, among the rootlets will enable one to
discover the newly hatched larvae.

During the roth or 11th day the larva reaches a length of
6-8 mm.; pigment begins to appear in the retina and other
parts of the head. The pectoral fins have shifted from the
horizontal to an oblique position; the movements are less vig-
orous but more frequent.

In 12-15 days a length of 9-11 mm. is reached. They are
now deeply pigmented with dark brown. The pectoral fins
have taken a vertical position. The yolk sac has changed from
the spherical to an ovoid form, with its smaller end posterior,

*&

NoOr2Z|\) CHE EGG OF AMTIA AND TTS “GLEAVAGE. B27)

and is soon absorbed. About this time the larvae are ready to
leave the nest with the male fish. Frequently we have seen
the entire brood, generally numbering several hundred, huddled
so closely together that they formed a black mass moving along
in close company with the parent. If disturbed, the parent
rushes off, setting the water in commotion by its movements,
and the brood scatter to conceal themselves at the bottom. In
a short time, if one remains perfectly quiet, he may see the
larvae come forth in small groups from their hiding-places, and
perhaps the return of the parent.

The following table shows the rate of development. The
cleavage times are based upon observations made on a single
egg, while the times recorded beyond late cleavage are based
upon data obtained by following the development of a second

ere

DD:
AGE IN Hours. LENGTH IN MM. STAGE OF DEVELOPMENT.
_— -- Oviposition, beginning of
2.24 — First cleavage, “ ‘
3.28 — Second “ a
4.20 — Third et cs as
5:22 — Fourth “ e sc
6.20 _- Fifth He % a
7-25 = Sixth it a
10. — Late .$
TGs — Blastula
40. —_ Gastrula, early
70. a Embryo visible
160. — Embryo begins to move
200. 5-6 Embryo hatches
250. 6-8 Appearance of pigment
360. 9-10 Yolk-sac absorbed
480. 12-14 Pelvic fins appear
720. 17-20 —-
CLEAVAGE.

Our observations on the cleavage comprise : first, a study of
surface phenomena, followed continuously on the living egg ;
secondly, a comparison of corresponding stages as traced on
prepared material ; and thirdly, a study of serial sections.

In following the cleavage of the living egg we found it con-
venient to proceed as follows : The eggs still attached to blades

es

328 WHITMAN AND EYCLESHVMER. [Vora

EXPLANATION OF CUTS 1 TO 6.*

% ‘ : ; 4
Curt 1.— Vertical section of the mature ovarian egg, showing the calotte, the

germinal vesicle, and a portion of the yolk.

Cur 2.—Vertical section of an egg at the beginning of the second cleavage.
The position of the section is shown by the dotted line in Diagr. A. It shows
the thickness of the calotte, the depth of the first cleavage groove (I) and the
vacuolar spaces.

Cur 3.— Vertical section in the stage of the third cleavage, in the plane of the
dotted line in Diagr. B. It shows the depth to which the first (I) and third
(III) cleavage grooves have extended, having now cut through the calotte and
become continuous with the vacuolar spaces, which have enlarged and increased
in number.

Cur 4.— Vertical section of an egg in the same stage, but with the third verti-
cals (III) arranged as in Diagr. C. The section passes near the margin of the
calotte (dotted line, Diagr. C). The cleavage grooves have barely cut through
the calotte, and the vacuolar spaces are less numerous than at the centre of the
egg.

Curt 5.— Horizontal section of an egg in nearly the same stage. In this case
one of the third set of grooves (a), instead of taking a radial direction like the
others, occupies the position shown in Diagr. D. The section passes just below
the calotte, and shows that the cell cut off by this grooveis still in continuity with
the yolk.

Curt 6.— Vertical section of an egg in the stage of the fourth cleavage, in the
plane of the dotted line in Diagr. E. It will be observed that the circular
groove passes somewhat obliquely, and apparently cuts off two cells at the right.
An examination of another section of the same egg (Cut 7) shows, however, that
centrally these cells are in direct continuity with the underlying yolk.

* Roman numerals indicate the order of the cleavage grooves.

NO Zi ie GG OF AMA AND TTS GEEAVAGE. 329

Journ.Morph. Nos.1-6.

330 WHITMAN AND EYCLESHYMER. [Mor exci:

of grass or rootlets are placed in shallow watch-glasses and
held in a fixed position by weighting with small pebbles. The
watch-glasses are then placed upon a mirror fastened to the
stage of a dissecting microscope. We could thus observe
the changes occurring on opposite sides of the egg without
disturbing its position.

The fixing fluids which have given the more satisfactory
results are: Flemming’s fluid, Perenyi’s fluid, chrom-osmic,
picro-acetic, and picro-sulphuric. For surface views chrom-
osmic fixation gives most perfect pictures, the osmic blacken-
ing the cleavage grooves so that they stand out in bold relief.
Another excellent method is surface staining with Delafield’s
haematoxylin, which may be employed with any of the above-
named fixatives. For material which is to be sectioned, picro-
acetic and Perenyi have given best results. Owing to the
crumbling of the yolk we have been obliged to use celloidin as
an imbedding mass. Serial sections by the method described
by the junior author® have been used for the most part.
Staining in section with Ehrlich’s haematoxylin and Mayer’s
alcoholic carmine has proved most satisfactory.

The freshly deposited egs (Fig..2) is firmlyfixed tovthe
object with which it first comes in contact by means of the
threads of the villous layer, which are elongated over the lower
hemisphere of the egg membrane. The egg is oval in profile
view, measuring in its longest diameter, including the mem-
brane,’ 2/5 to 3mm. ; in its.shortest, 2 to 255mm: The upper
pole of the egg, ‘calotte”’ of Fiilleborn, “ germ-disc””’ of Dean,
is light yellowish brown. This calotte shades off at its margin
into the dark grayish brown of the yolk. Near the centre of
the calotte there is a single micropylar orifice, as shown in
Figs. 21 and 22, which remains visible for some time (Figs. 12
and 16). We have frequently noticed that the margin of the
calotte extends farther toward the equatorial region on one side
than on the other, as shown in Fig. 2.

The calotte is already present in the full-grown ovarian egg
shown in Cut 1. It is not of quite uniform thickness, but is a

5 A.C. Eycleshymer: Notes on Celloidin Technique. American Naturalist,
vol. XXVI, pp. 354-358.

Noi2.) ZHE EGG OF AMIA AND ITS CLEAVAGE. 331

little thicker on one side than elsewhere. The germinal ves-
icle lies eccentrically beneath the deeper side, mainly in the
yolk, only its upper surface projecting into the calotte. There
are thus indications of a definite orientation in the egg before,
as well as immediately after, fertilization.

First Cleavage. — In the egg shown in Fig. 3 the first cleav-
age appeared at 7.54 A.M. (2 hrs. 24 min. after deposition). A
slight flattening of the animal pole precedes the appearance of
this groove, which generally divides the calotte into nearly
equal portions. The groove travels more and more slowly as it
passes beyond the margin of the calotte. Its decreasing rate of
progress in the egg figured is indicated by the following per-
centages, based upon the length of the are traversed at succes-
sive intervals of about 1 hr. each, as compared with the entire
circumference. At 8.53 (Fig. 3) the arc described equals 36%
of the entire circumference ; at 9.49 (Fig. 4) 60%; at 10.50
(Fig. 5) 77% ; at 11.55 (Fig. 6) 89% ; between 11.55 and 12.58
(Fig. 7) 100%. In most cases this cleavage is not complete
until after the appearance of the fourth and sometimes even
the fifth set of grooves.

The first groove is most frequently meridional, as shown in
Figs. 3, 21, 22, and 23. In some cases, however, it may depart
so far from a meridional that the segments formed are quite
unequal (cf. Dean, /c., p. 425), as shown in Fig. 9. The ends
of the groove may meet at the lower pole so as to form a con-
tinuous straight line, or in such a manner that an obtuse angle
is formed (Figs. 14, 17). Sometimes they do not meet at all,
and are only united by a portion of the second, or even third,
cleavage grooves.

At the time the second groove appears the first has a depth
of about one-half the thickness of the calotte, as shown in Cut
2 (made in plane of dotted line of Diagr. A). The path of
this groove is definitely premarked to a point about twice this
depth, terminating in an irregular vacuolar space at the level of
the nuclei. The depth of the cleavage at this time strongly
reminds one of the conditions seen in teleostean eggs with a
perfectly defined blastodisc. The calotte certainly represents
a concentration of germinal material, analogous to the blasto-

332 WHITMAN AND EVCLE SEH VIlia [VoL ik

disc seen in the meroblastic egg of the teleost. One might
say that the calotte represents a nascent blastodisc, —a blasto-
disc 72 statu nascendi, so to speak. It differs from the blas-
todisc, as seen in pelagic teleostean eggs, chiefly in not
containing the whole, or almost the whole, of the material to
be used in the formation of the embryo. The Amia egg is
holoblastic, but represents an advance upon the condition seen
in the egg of Acipenser, in the direction of the meroblastic
egg of the typical teleost.

At this stage one or more irregular cavities are found in the
upper hemisphere, between the centre of the egg and the
calotte. One of these is shown in Cut 2. Just when and how
these cavities arise we cannot say, but they are present in the
first, as well as the later stages of cleavage. Sooner or later
they become continuous with cleavage grooves, and in many
cases unite in a common cavity. As these cavities appear in
eggs prepared in different ways, collected in different seasons
from various nests, it seems certain that they are not to be
considered as artificial.

Second Cleavage. — The two grooves of the second cleavage
usually begin at the same point of the upper pole and extend
meridionally at right angles to the first groove (Figs. 4, 22, 23).
Their progress is essentially the same as that of the first. In
the egg shown'in Fig. 4 these grooves appeared at 8.58; at
9.40 they had extended slightly beyond the margin of the
calotte, and three to four hours later they had completely
encircled the egg.

Variations in the formation of these grooves are not uncom-
mon. The point of origin may be at a greater or less distance
from the pole (Fig. 10). Instead of a right angle, acute and
obtuse angles may arise, the grooves crossing in the form of an
X (Fig. 11). Again, the furrows may not arise at the same
point, but at points more or less widely separated, as in Fig. 12.
They do not always begin at the same time; one may precede
the other by a considerable interval, reaching near the lower
pole before the other has passed the equator (Fig. 14). Ovcca-
sionally the two unite at the lower pole to form a continuous
line running at right angles to the first; at other times they

NGO. 2 VEEG OF AMIA AND IS *CLEAVAGE,. aees

form acute or obtuse angles, as at the upper pole, or they may
even terminate at points widely separated, as in Fig. 17. In
some cases the grooves may not even reach the lower pole, but,
diverging from a meridional plane, unite with the first groove
at almost any point between the margin of the calotte and the
lower pole.

Third Cleavage. — By the time the second grooves have
passed just beyond the margin of the calotte the third set of
grooves appear. Ina majority of cases these are vertical, and
occupy the positions shown in Figs. 5, 13, 15, 23, 24, and 25.
They generally all depart from one or the other of the first two
meridionals, thus giving rise to a distinctly bilateral appearance.
In the egg followed (Fig. 5) the first of this set originated in
the first meridional at 9.50. The second was formed in the
adjacent quadrant, on the opposite side of the first meridional,
at 9.51. The third appeared one minute later in the adjacent
quadrant on the same side of the first meridional. The divi-
sion of the remaining quadrant began at 9.54. The two furrows
shown in Figs. 5, 6, and 7 later reached the lower pole, but the
one first formed turned aside, running into the first meridional
at some distance from the lower pole, while the remaining one
united with the second meridional at about the level of the
equator.

It often occurs that one or more of the set depart from the
first meridional, while the rest depart from the second, or w2ce
versa (Figs. 13 and 24). Frequently one observes the condi-
tion shown in Fig. 15, where all the grooves of this set pass in
planes nearly parallel to the first or second meridional, giving
rise to a bilateral form comparable with the 8-cell stage of the
teleostean egg.* All these grooves may occasionally depart from
a common point at the upper pole, and cut each quadrant of the
calotte into two approximately equal portions, in which case
the cleavage assumes a radial form such as has been described by
Agassiz and Whitman ® in the pelagic fish egg. Rarely, if ever,
do these grooves divide the quadrants equally at the lower pole.

* Dean (No. 3, p. 425) speaks of this condition as if it were the general rule.

6 Agassiz, A., and Whitman, C. O.: On the Development of Some Pelagic Fish
Eggs. Proc. Amer. Acad. Arts and Sci., XX, pp- 23-75. 1884.

334 WHITMAN AND EYCLESHYMER. [VoL. XII.

Cut 3 represents a vertical section of an egg at a stage just
preceding the fourth cleavage. Its position is shown by the
dotted line in Diagr. B. The section runs parallel with the
second meridional, and cuts through two of the cells delimited
by the third verticals (III). The first meridional has at this
time cut entirely through the calotte and deeply into the under-
lying yolk. The third set of grooves have likewise extended
far into the yolk. These grooves usually pass in vertical
planes ; at times, however, they are more or less inclined ; we
have sections in which one or more of them pass so obliquely
that the cells are deeply cut, but in no case have we found
any of them entirely severed from the underlying yolk. Occa-
sionally one finds eggs in which one of these grooves (III®)
occupies the position shown in Diagr. D, Cut 5. From sur-
face study one might be in doubt whether this groove is verti-
cal or horizontal. Serial sections of three such eggs show that
the groove is vertical, and that the segment in each case is
continuous with the underlying yolk.

In Cut 5 a horizontal section of such an egg is shown which
passes just beneath the calotte. All the third cleavage grooves
(III) except one (III) are arranged radially, alternating with
the two primary grooves. It is noteworthy that all these
grooves widen into fissures which have a common vacuolar
centre, giving rise to a star-shaped figure.

Cut 4 exhibits a vertical section of the same stage, but at
right angles to the plane of Cut 3. The grooves on the surface
of the egg are shown in Diagr. C, and the position of the |
section is indicated by the dotted line. The section is taken
near the margin of the calotte. Sections nearer the centre of
the egg show that the first and second grooves have reached a
much greater depth, and have become continuous with cavities
similar to those shown in Cut 3.

Dean (No. 3, p. 426) states that “the segmentation cavity
takes its definite origin at this stage; in the region of
the animal pole the blastomeres are separated from the
underlying yolk—the germ disc by a narrow fissure, which
has been found to arise in the cleavage planes of the
animal pole.”

INOm 2a AGG (OTE ANIMA. AUNT AG Sae Cis IZA V AGE. 335

Passing over the singular construction of the second state-
ment, which we confess to being utterly unable to comprehend,
we are led to make a remark on the “segmentation cavity.”
It seems evident that Dr. Dean is predisposed to interpret his
sections of the Amia egg in correspondence with what he finds
laid down by Dr. H. V. Wilson in regard to the egg of Ser-
ranus. The discovery of a ‘segmentation cavity” in his
Fig. 23 seems to have been thus inspired (compare this figure
with Wilson’s Fig. 15).

If the section shown in Dean’s Fig. 23 passes in the plane
‘indicated in his Fig. 4, the cell- zis cut very near ats central
margin, and a little obliquity, either in the plane of the section
or in the plane of the bounding groove, might be sufficient to
account for the appearance. In other words, the section may .
represent only a chip from the central margin of cell a.

We must differ with Dr. Dean in respect to what seems to
be his notion of “the segmentation cavity.’’ The fissure, or
fissures, thus designated by Dean, as a glance at his Figs. 27—
30 will show, represent nothing more than ordinary inter-
cellular spaces. Such spaces, as all the world knows, are not
peculiar to any egg, and are not to be confounded with ‘“ ¢he
segmentation cavity.” Dr. Dean says in one place (p. 424)
“the segmentation cavity is practically wanting.” Neverthe-
less he speaks later of its “definite origin,’ and calls special
attention to it in his figures. In one case (p. 441) it is called a
“flattened segmentation cavity.”

Dr. Dean represents the segmentation cavity as beginning
with the second cleavage but as taking “ definite origin”’ at the
time of the third cleavage. In Dean’s Fig. 24, representing a
section of an egg in the stage of the fourth cleavage, the “‘seg-
mentation cavity’ is shown as intercellular spaces, which he
admits “might perhaps be regarded as artifacts.’’ We would
suggest that these. spaces have a different meaning. The
elongated nucleus shown in one of these cells is very clear
evidence that what we have called the horizontal cleavage is in
progress. At the time of the cleavage a constriction takes
place at about the level of these spaces, and it is to these con-
strictions that they probably owe their origin. There is another

336 WHITMAN AND EYCLESHYMER. [Vor xr

important point to be noticed in connection with this figure.
The central blastomeres are represented as cut off from the
yolk below. We feel confident that if the entire series be
examined the cells will be found to be still continuous with the
underlying yolk.

The cleavage cavity in Amia has rather an interesting mode
of origin. For the first appearance of this cavity is in the form
of central vacuoles, which are present as early as the first
cleavage groove. The cleavage grooves, in many cases at
least, expand into broad fissure-like cavities as they approach
the centre of the egg (Cuts 3, 5, 15), where they become con-
tinuous with the earlier cavity or cavities. As the cleavage
runs on these spaces enlarge, flow together, and thus often give
rise to quite a spacious cleavage cavity, such as shown in Cuts
10, 17, 19. In some cases the cavity is so much reduced that
it becomes inconspicuous.

Fourth Cleavage. — The fourth cleavage marks an interesting
epoch, as it is the first to take a direct part in the formation of
small cells at the animal pole. For the complete delimitation
of the cells another cleavage is necessary, namely, a horizontal
cleavage which cannot be seen from the surface and which
affects only the eight cells bounded superficially by the fourth
cleavage grooves, or by what we may call the circular groove.
The latter usually appears midway between the pole and the
margin of the calotte. It represents, strictly speaking, eight
distinct grooves, one for each of the eight primary segments,
but these usually run together in such a way as to form a con-
tinuous groove, encircling the pole and bounding a polar field
which may have a circular form (Figs. 6, 25, 26) or a more
elongated oval form (Figs. 16, 18). In one egg (Fig. 6) we
have traced the order and direction of these grooves, noting
also the time and place of origin as indicated in the figure.

There are many variations in this cleavage. It often hap-
pens that the eight constituent grooves run at various angles
and are discontinuous, so that the polar field is bounded by a
more or less irregular or zigzag line (Figs. 16, 18) ; sometimes
one or more of these grooves take a meridional direction, as
that shown in the upper part of Fig. 18.

Nove LAE VWEGG OR SAMIA, AND TRS VOGEAVAGE. aay

This circular groove looks externally as if it might actually
cut off polar segments, but as an examination of sections (Cuts
6, 7, 13, and 15) will show, the groove descends more or less
obliquely, or centripetally, so that when completed its inner
deeper boundary would be lost in confluent vacuolar spaces.
As the plane of the groove probably never passes below the
vacuoles, one might say that it describes an axial mass in the
upper hemisphere approximating in shape an inverted truncated
cone. Cut 61s taken in the plane of the dotted line in Diagr. E.
The circular groove appears to cut off two cells; this appear-
ance is due to the section passing near the margin of the polar
field. In Cut 7, which passes parallel with the second cleavage
(plane of dotted line, Diagr. F.), these cells are still continuous
with the yolk below. Cuts 13 and 15, taken along the dotted
lines of Cut 7, further illustrate this point. It will be seen
that at their bases the segments lying within the fourth cleav-
age grooves (IV) are likewise continuous with the yolk. Cut
8, taken at the level of the dotted line in Cut 7, shows the
central grouping of the vacuolar spaces and the depth of the
cleavage grooves at this time.

Dean (No. 3, p. 427, Fig. 24) figures and describes the
“central blastomeres”’ of this stage as ‘separate from the
underlying germ-yolk.” In describing the next cleavage we
shall explain how the central cells are cut off.

Fifth Cleavage. — The fifth cleavage comprises two distinct
sets of grooves. One set appear on the surfaces of the eight
primary segments in the form of meridionals. In Fig. 26 they
are just beginning, and are more advanced in Fig. 27. The
other set extend horizontally in the eight cells of the polar
field, and of course are not visible from the surface. This
cleavage thus brings about the 32-cell stage, and makes the
central portion of the calotte two cells deep. The origin of
yolk cells is thus not limited to the margin of the calotte, as in
pelagic teleost eggs, but extends to all the cells beneath those
of the polar field. These cells go on budding off cells to the
germ-disc. It is rather interesting to see this horizontal cleav-
age appearing with the passage from the 16 to the 32-cell stage,
because in the corresponding stage in the pelagic fish egg it

338 WHITMAN AND EYCLESHYMER. Vion: Aaike

EXPLANATION OF CUTS 7 TO 12.

Cut 7.— Vertical section of the same egg as that from which the section
shown in Cut 6 was taken. Its position (dotted line in Diagr. F) is much
nearer the centre of the egg. The section shows that the cells apparently cut off
in Cut 6 are continuous with the yolk, the oblique character of the fourth (IV)
cleavage grooves, and the elongated nuclei of the central cells.

Cur 8.— Horizontal section of an egg in the same stage, in the plane indicated
by the line 8—8 shown in Cut 6. It shows the central grouping of the vacuoles,
and the depth to which the grooves have at this time cleft the yolk.

Cut g.— Vertical section of an egg in the stage shown in Fig. 8, Pl. X VIII.
The central cells have been cut by both horizontal and vertical grooves. Nuclei
are present in the yolk.

Cur 10.— Vertical section of an egg in the stage of early blastula, showing
the cleavage cavity and the extent to which the yolk is now segmented.

Cur 11.— Horizontal section of an egg in the same stage, in the plane of the
dotted line 11-11 in Cut Io.

Cur 12.— Vertical section of an egg in a late blastula stage. The calotte has
begun to extend marginally over the yolk, and the epidermal layer is beginning to

| differentiate.

No.2.] THE EGG OF AMIA AND ITS CLEAVAGE. 339

Journ. Morph. Nos. 7-22.

340 WHITMAN AND EVCLESHYMER. [VoL. XII.

takes effect only in the four central cells of the 16-cell stage.
The marginal cells are divided by verticals which correspond to
what happens in the eight yolk segments of the Amia egg.

Variations in this cleavage are numerous. It often happens
that one or more of the central cells are cut by grooves passing
parallel with the circular. Again, some of the grooves, instead
of taking a horizontal or a circular form, may pass vertically,
as indicated both by surface views and the elongation of the
dividing nuclei. Whether the grooves cutting the central cells
shown in Fig. 7 are to be interpreted as variations, or whether
they represent a part of the sixth cleavage, we are unable to
say. In this egg the central grooves appeared shortly after
the appearance of the grooves dividing the marginal cells, with
which they soon became continuous.

Sections of this stage are shown in Cuts 9, 14, 16, and 18.
In Cut 14 the plane of elongation of the nuclei in the large
marginal cells confirms what has been described from surface
views. In Cuts 16 and 18 the elongation of the nuclei (V)
lying within the circular groove (IV) shows that the next
division of these cells will take place in a horizontal plane.
While this is undoubtedly the usual plane of division, there are
exceptions. In each of the sections shown in Cuts 14 and 15
one cell of the eight delimited by the fourth cleavage shows its
nucleus elongated in such a direction that the resulting cleav-
age will be circular. In other sections the elongation of the,
nuclei indicates that vertical cleavages may also replace the
horizontal. Cut 9 represents a vertical section of an egg in
the stage shown in Fig. 8. The horizontal cleavage has evi-
dently taken place, and in addition to this another set of verti-
cals have appeared. In this section it was impossible to trace
the cleavage grooves through the yolk.

Later Cleavage. — The next cleavage, which might be called
the sixth, is shown in Figs. 8 and 28, and consists in a general
way of two sets of circular furrows, one of which appears
between the first circular and the upper pole, the other between
the first circular and the margin of the calotte. In addition to
these, new verticals have arisen in some of the marginal
segments.

NG IZ) re waGG “OF AMITA AND TESSCLEAVAGE, 341

Figs. 29 and 30 represent eggs of about the same stage
as that shown in Fig. 28. In Fig. 29 a radial symmetry is
noticeable in the arrangement of the smaller cells, while in
Fig. 30 a bilateral grouping is evident. Fig. 31 represents a
third egg in about the same stage, showing the position and
extent of the furrows, which at this time have reached the
vicinity of the lower pole. A somewhat later stage is shown
Ineipien es 2

Cut 10 represents a vertical section of an egg about ro hrs.
after deposition. The calotte is now from four to six cells
thick. Beneath the central portion is the segmentation cavity,
which at this time has become greatly enlarged through the
confluence of the vacuolar spaces. In some eggs (e.g., Figs.
18, 19) at this stage the cavity is so much enlarged that it
appears like that seen in Amphibian eggs. The floor of the
cavity is made up of large yolk segments. A horizontal sec-
tion (Cut 11) of the same stage taken in the plane of the
dotted line of Cut 10 shows the number and depth of the
grooves at this level; above this level, near the equator,
the grooves are more numerous.

Cut 12 represents a vertical section of an egg 35 to 4o hrs.
after deposition (late blastula). The calotte, which has now.
begun to extend over the yolk, consists of thickly crowded
spherical cells which marginally pass abruptly into the large
yolk segments, while in the central portion they gradually
increase in size and lie loosely scattered. The outer layer of
the calotte is distinctly differentiated in that the cells are elon-
gated and more densely granular. The entire yolk is irregu-
larly cleft, the cells forming the lower portion are roughly
polygonal and grade off into the large yolk spheres which lie
at the centre.

THE RELATION OF THE EMBRYO TO THE FIRST Two
CLEAVAGE PLANES.

The elongated form of the egg of Amia, in a closely applied
envelope, prevents rotation about its minor axes. It is there-
fore a favorable egg for ascertaining what effects, if any, gravity

342 WHITMAN AND EYCLESHYMER. (Mor. ATh

EXPLANATION OF CUTS 13 TO Ig.

Cur 13.— Oblique section of an egg in the stage of the fourth cleavage, just
before the fifth cleavage. The plane of the section is represented by the dotted
line 13-13 in Cut 7. The section shows on one side that the fourth cleavage has
not yet cut off the central cells.

Curt 14.— Oblique section of the same stage, passing along the line 14-14 of
Cut 6. The section shows the plane of elongation of the nuclei in the marginal
segments which are soon to be divided by a set of verticals, forming a part of the
fifth cleavage.

Cur 15.— Oblique section along the line 15-15 of Cut 7. The section
shows that the circular groove (IV) becomes continuous below with the vacuolar
spaces.

Cuts 16 and 18.— Vertical sections of the calotte from different eggs in the
stage of fourth cleavage. The sections show the vertical elongation of the nuclei
of the central cells preparatory to the horizontal cleavage which is to divide
them.

Cut 17.— Oblique section of an egg in a stage a little earlier than that
shown in Cut 19. The section passes in the plane indicated by the line 17-17
in Cut 19. The section shows an exceptionally large cleavage cavity. It also
shows numerous yolk nuclei lying at the inner ends of the large yolk segments.

CuT 19.— Vertical section of a typical blastula. The cleavage cavity in this
egg is also exceptionally large. Some of the large yolk segments may be seen
dividing at their inner ends, the cells thus derived being continually added to the
calotte.

343
Nos. 1319.

NOs EGG, ORV AMIAS AUVD, TEES MCIMEAVAGE.

Journ. Morph.

O85
ents
taal

950 hor

er aooods D
6,8

a! fo8 eo,
Veeoes

344 WHITMAN AND EVCLESHVMER. [Vor tit:

may have on the direction of cleavage, and for determining the
relation of the early cleavage planes to the median plane of

20

Cw

the embryo. Dean (no. 3, p. 422) found “in the early stages of
segmentation that the cleavage planes occurred in the normal
way when the position of the egg was reversed.” This is true

Nomzelae PEGGlOFr AMIA AND: TS GEEAVAGE. 345

not only for inverted eggs, but for eggs placed in any position
whatsoever. It seems to follow that gravity can have no
directing influence on the cleavage.

In order to ascertain whether there is any constant relation
of the embryonic axis to either of the first two cleavage planes,
eggs were fixed in given positions by weighting, as before men-
tioned, and a sketch of the early grooves was carefully made in
each case. These grooves are easily identified for a long time
in the lower hemisphere of the egg, even as late, in some cases,
as the early stages of gastrulation. As the sketches made at
successive intervals showed no movement of the egg during all
this time, it seems probable that the position of the egg
remained practically unchanged up to the time when the
median plane of the embryo was ascertainable. In some cases
accidental markings on the surface of the egg remained in a
fixed position until the embryo was well defined.

Our observations were made on three sets of eggs. The
results are tabulated below and illustrated by the diagrams
shown in Cut 20.

First Series. — At 8 a.m. May 12, 1895, ten eggs in which
the first grooves had just appeared were fixed in position and
sketches made at successive intervals. Two of the eggs died
during gastrulation, leaving eight in which the embryos were ~
apparently normal. The results were as follows :

Eco. EMBRYO VISIBLE. ANGLE WITH First Groove.
I 6 A.M. May 15 67°
2 6 A.M. ‘ 0°
3 5 A.M. ‘s 7iSe
4 Died during gastrulation
5 5 A.M. May 15 75a
6 5 A.M. es 0°
7 PAC IM ise ine ce 45°
8 Died during gastrulation
9 5 A.M. May 15 B09

10 6 A.M. ‘e om

Second Series. — At 9 A.M. on the same day, a second lot of
ten eggs in which the second grooves were well under way were
fixed in position. Nine of these formed normal embryos.

346 WHITMAN AND EYCLESHYMER. (Von. XII:

Ecc. Empryo VISIBLE. ANGLE WITH First Groove.
II 5-30 AM. May 15 702
12 5 AM. st 35°
13 5 A.M. ss 45°
14 5 A.M. = 88°
I5 Died in late cleavage
16 5aA.M. May 15 Ac.
17 5 AM. s ae
18 5 A.M. « 7O4
19 5 A.M. She
20 5 A.M. ss 85°

Third Series. — At 10.30 A.M., on the same day, a third lot
of five eggs, in which the third set of grooves had begun, were
fixed in position. Only three of these formed embryos.

Ecc. Embryo VISIBLE. ANGLE WITH First GRoove.
21 6 a.M. May 15 go°

22 Died

23 Died

24 GyicM teks 30°

25 Sol 5eA-Ma : {O°

The results may be summarized as follows:
1. Angles formed by median sagittal plane of embryo and
first cleavage plane :

In 10%, coincidence.

In 10%, less than 5°.

In 20%, greater than 5° and less than 45°.
In 10%, 45°.

In 35%, greater than 45° and less than 85°.
In 10%, greater than 85° and less than go°.

In 5%, 90°.

2. Angles formed by the median sagittal plane of the embryo
and the second cleavage plane:

In 5%, coincidence.

In 10%, less than 5°.

In 35%, greater than 5° and less than 45°.
In 10%, 45°.

In 20%, greater than 45° and less than 85°.
In 10%, greater than 85° and less than go°.
In 10%, 90°.

NoN2t) PHEVEGG OF ‘AMIA AND TTS CLEAVAGE. BAN,

CONCLUDING REMARKS.

Dr. Dean (No. 3, p. 425) states that he “as taken espectal
care to verify his observations on the meroblastic character of the
cleavages of Amia. During the first cleavages several hundred
living eggs were examined, with a view of determining holoblas-
tic variations. These, however, did not occur, nor were there
found, even by the most favorable means of tllumination, traces
of what might be construed as surface furrows traversing the
yolk region of the egg. Inno case did a marginal cleavage pass
below the rim of the germinal disc.”

If an author can say all this of an egg that is plainly holo-
blastic, how can we accept his testimony against Balfour, to the
effect that the egg of Lepidosteus is meroblastic? If this egg
be holoblastic, its cleavage might be said to agree very closely
with that of Amia. It would be of interest to know when the
first horizontal cleavage occurs. Up to the 8-cell stage the
cleavage grooves may have practically the same order and
arrangement in Amia, Lepidosteus, Acipenser, and the tel-
eost. In passing to the 16-cell stage we usually get four cen-
tral cells, inclosed by twelve marginal ones in the teleost.
This arrangement, according to Dean’s figures, holds both for
Lepidosteus and Acipenser. In Amia we get ordinarily the
circular groove, resulting in eight small cells in the polar (cen-
tral) field, surrounded by eight large cells. Thus far the three
eges agree in having no horizontal cleavage. The passage to
the 32-cell stage introduces, in both Amia and the teleost, the
first horizontal cleavage; but in the former we have eight cells
dividing in this way, in the latter, only four. The resemblances
in other respects would lead us to expect this form of cleavage
at the corresponding stage in Lepidosteus and Acipenser. In
that case these ganoids would agree with the teleost in having
four central (polar) cells thus divided.

In the more regular forms of holoblastic cleavage, as is well
known, the first horizontal, or the equatorial, as it is often
called, comes in as the third cleavage. Is this cleavage, when
it occurs as the third, homologous with the horizontal, which
appears as the fifth in the teleost, Amia, and probably other

348 WHITMAN AND EVYCLESHVMER. [Vor XIE

ganoids? How are we to determine homology in such a case ?
To suppose, as some authors have done, that the first horizontal
cleavage of the frog’s egg is skipped in the case of the fish egg,
is nonsense. There is no skipping in cleavage so long as the
division of the cytoplasm follows regularly the order of nuclear
division. If nuclear division runs on unaccompanied by seg-
mentation of the egg, as in the insect egg, then we may speak
of skipping cleavage; if cleavage does not appear until after
three or more nuclear divisions, and then begins and takes the
regular course, splitting into two, four, eight, etc., as in the
crab’s egg, according to Paul Meyer, we may speak of deferred
cleavage. In the fish egg the cleavage is neither deferred nor
skipped. Where, then, is the homologue of the third cleavage
of the frog’s egg? If position and relation to the axes of the
future embryo are to decide homology, then there is no homo-
logue before the fifth cleavage, and no certainty of any even
then. To assume homology between a cleavage of the 4-cell
stage and one of the 16-cell stage is to venture into a hopeless
tangle of anachronisms. If order of succession is to be our
criterion, then first is first, second is second, third is third, and
so on, without regard to the position of the planes of cleavage
or the fate of the cells. In either case, the attempt to homolo-
gize cleavages leads to contradictions and confusion. Homol-
ogy that means nothing beyond the second cleavage, or that
may fail even at the first cleavage, is its own negation. The
wide range of variation in cleavage in many eggs, its total sup-
pression in many others, and the possibility of changing its
course by artificial means, without affecting the final result, all
go to show, as it seems to us, that homologies do not depend
upon cleavage planes. Even the first and second cleavages,
which appear, at first sight, to agree so closely in widely differ-
ent eggs, may cross the future lines of homology at various
angles. If there is any homology in cleavages worth talking
about, it must be because they take definite and constant parts
in defining homologous organs or areas. Something of this
sort has been claimed for the first cleavage. On further exam-
ination the claim proves to be ill founded. Often, as early as
the 8 and 16-cell stages, the first groove is transformed into a

NOm2s |New GG, OF AMPA WAND STS CLEAVAGE. 349

zigzag line, running irregularly, now to one side, now to the
other, of the median plane.

If homologies do not depend upon the form of cleavage, and
may even unfold without any cleavage at all, how are we to
explain the fact that in many cases (e.g., annelids, molluscs)
cleavage is so constant, and so early becomes coincident with
lines of homology? What determines this coincidence? This
question has been discussed by one of us elsewhere, and we
need not repeat here the conclusions reached.

350 WHITMAN AND EYCLESHYMER.

REFERENCES.

6 AGassiz, A., AND WHITMAN, C. O. On the Development of some
Pelagic Fish Eggs. Proc. Amer. Acad. Arts and Sci., XX, pp.
23-75. 1884.

1 ALLIS, EDWARD PHELPS, Jr. The Anatomy and Development of the
Lateral Line System in Amia Calva. Journ. Morph., Il, p. 463.
1888.

3 DEAN, BASHFORD. The Early Development of Amia. Quart. Journ.
Miucr. Sci., XXXVIII, pp. 413-444. February, 1896.

5 EYCLESHYMER, A. C. Notes on Celloidin Technique. Am. WVaz.,
XXVI, pp. 354-358. 1892.

2 FULLEBORN, F. Bericht uber eine zur Untersuchung der Entwickelung
von Amia, Lepidosteus und Necturus unternommene Reise nach
Nord-America. Sztzungsberichte der Akad. ad. Wiss. zu Berlin,
XL, pp. 1057-1070. Oct. 25, 1894.

4 GoovbE, GEO. Brown. Natural History of Useful Aquatic Animals,
pp. 659, 660. 1884.

en jase | : A oe ee
7 7 ; i vet a
at » f ib ee (Mp
a oe car gos 4 ee,
” Sepa eae id = te roel 4
ny ay) by. ee ad Or i i : ; ; gen

an a ee

Y Mike .

rahe: or ik
eee bh, wi

aS

4 sh #1 eee dt 19

ae
Tat
¥

an?
Vie atu. *° lao)

a" Bsiit: lo teatln ft he

er Ms

vee *
a, ; ae
f

ir ., es ce hy ites i
Ste !

ae vA iy vy oy wi

es ih ; i SALT. Hae gy -
Se ; , ee?
t vy rere PU) ear i oan, i wa, -
a aalirhgll atin OT es A

Itte

Sathorn

WHITMAN AND EVCLESHYVMER:

iS)
on
No

EXPLANATION OF PLATE XVIII.

All figures, excepting 8, 14, 17, and 20, drawn from living eggs. The numerals
affixed to the cleavage grooves give the time in hours and minutes at which the
grooves reach the points indicated by the dotted lines. The egg shown in
Figs. 3-5 remained in a constant position.

Figures 9-20 are so arranged that the first cleavage groove has always the same
position.

Fic. 1. Unsegmented eggs attached to grass. Natural size.

Fic. 2. Profile view of the unsegmented egg, showing natural color of the
egg and the villi by which it is attached. Made by J. Nomura. X Is.

Fic. 3. Profile view of the egg 3 hrs.°23 min. after deposition, showing
position and extent of the first vertical groove. X 14.

Fic. 4. Profile view of same egg at 4 hrs. 19 min. The second verticals
have extended slightly beyond the margin of the calotte. X 14.

Fic. 5. Profile view of the same egg at 5 hrs. ro min., showing the position
and extent of the first three sets of verticals. X 14.

Fic. 6. Profile view of the same egg ‘at 6 hrs. 20 min., after the appearance
of the circular groove. X 14.

Fic. 7. Profile view of same egg at 7 hrs. 23 min. The fourth vertical
cleavage is in progress. X 14.

Fic. 8. Profile view of same egg at 8 hrs. 25 min., showing additional
circular grooves, one set within, the other without, the first circular groove.
X 14.

Fic. 9. Showing a case in which the calotte is unequally divided by the first
groove. X 13.

FIGS. 10, 11, 12. Views of the upper pole, showing variations occasionally
observed in the position of the second verticals. X 13.

Fic. 13. View of the upper pole, showing variation in position of one of the
third verticals. X 14.

Fic. 14. View of the lower pole of same egg. X 13.

Fic. 15. View of the upper pole, showing another variation in the position of
the third verticals. X 13.

Fic. 16. View of the upper pole, showing a variation in the position of the
grooves of the first circular cleavage. X 13.

Fic. 17. View of the lower pole of the same egg, showing points at which the
second verticals terminate. X 13.

Fic. 18. View of the upper pole, showing a broken circular groove. X 13.

Fic. 19. View of the upper pole of an egg in a stage somewhat later than the
one shown in Fig. 8. X 13.

Fic. 20. View of the lower pole of same egg. X 13.

PL XVII

]
|
|
|

=f
+ = 7
‘
e
—__
© ~
~

354 WHITMAN AND EYCLESHVMER.

EXPLANATION OF PLATE XIX.

All the figures were drawn from material fixed in chrom-osmic and preserved
in 80% alcohol. Magnification about 16 diameters.

Fic. 21a. View of the upper pole of the egg, showing the position of the first
groove and micropylar orifice.

Fic. 21b. Profile view of the same egg.

Fic. 22a. View of the upper pole of the egg at the beginning of the second
vertical cleavage.

Fic. 22b. Profile view of the same.

Fic. 23a. View of the upper pole, showing symmetrical third verticals.

Fic. 23b. Profile view of the same.

Fic. 24a. View of the upper pole in same stage as Fig. 23, showing an
asymmetrical position of one of the third verticals.

Fic. 24b. Profile view of the same egg.

Fic. 25a. The formation of the first set of circular grooves.

Fic. 25b. Profile view of the same.

Fic. 26a. The first set of circular grooves completed, and the fourth set of
verticals beginning.

Fic. 26b. Profile view of the same.

Fic. 27a. Fourth set of verticals well advanced.

Fic. 27b. Profile view of the same.

Fic. 28a. Shows the addition of two new sets of circular grooves, one within,
the other without, the first circular groove.

Fic. 28b. Profile view of the same.

Fic. 29a. A little later stage, showing another set of circular grooves outside
of those seen in Fig. 28.

Fic. 29b. Profile view of the same.

Fic. 30a. About the same stage. The cells show a bilateral arrangement.

Fic. 30b. Profile view of the same.

Fic. 31a. The lower pole at about the same stage as Figs. 29 and 3o.

Fic. 31b. Profile view of the same.

Fic. 32a. A later stage.

Fic. 32b. Profile view of the same.

DP. ey

———

ON StHE MODES, ‘OF DEVELOPMENT OF THE
MESODERM AND MESENCHYM, WITH REF-
ERENCE, TOs Tik, SUPPOSED HOMOLOGS
On thr BODY CAVITIES:

THOS Ee MONTGOMERY, JiR, PHD:

MESODERM is the term broadly applied to that embryonic cell
mass situated between the primitive ecto- and entoderm. The
mesoderm, z.¢., mesodermal cells, and not the extracellular gel-
atinous substance secreted in the archicoel by the ecto- and
entodermic cells, is derived either from one of these two layers
or from both of them. It is derived (1) from the ectoderm in
the Anthozoa, Porifera; (2) from the entoderm in Ctenophora,
Polycladidea, Nemertini (in part), Nematodea, Annelida, Hir-
udinea (?), Sipunculacea, Chaetognatha, Balanoglossus, Echin-
odermata, Copepoda, Insecta (?), Mollusca, Phoronis, Ectoprocta,
Brachiopoda, Entoprocta (?), Tunicata, Amphioxus; (3) from
both ecto- and entoderm: in the Nemertini (Desor’s type), also
perhaps in the Rotatoria and many of the Arthropoda (except
Copepoda, Insecta). The origin of the mesoderm from the
entoderm is thus the most usual.

The mesoderm may arise from the entoderm in three ways :
(1) by multipolar delamination ; (2) by unipolar delamination ;
(3) in the form of epithelial diverticula.

(1) The multipolar delamination, the so-called mesenchym
production (Mesenchymbildung), consists in the separation of
cells from the entodermal layer at various points on the inner
surface of the latter, which become situated in the archicoel, in
the gelatinous substance secreted by both ecto- and entoderm.
Such mesenchym production is typical for the Nemertini, Ech-
inodermata, and Ectoprocta.

(2) The unipolar delamination of the mesoderm from the
entoderm takes place through the so-called pole-cells of the
mesoderm, z.¢., one or two cells which divide off from the ento-
derm at one pole of the latter, and by the cleavages of which

356 MONTGOMERY. [Wor Sane

the mesodermic cell layer is produced. These pole-cells always
lie in the archicoel. Such a production of the mesoderm is
typical for the Annelida and Mollusca; further, in more or less
modified form, for the Nematodea, Copepoda, and Entoprocta.

(3) Finally, the mesoderm may be produced from the ento-
derm, in the form of epithelial diverticula, or the so-called
coelom sacks. Such coelom sacks may be (a) evaginations of
the gastrocoel, or (b) invaginations into the gastrocoel.

(a) The mesoderm is produced in the form of a coelom sack,
which is an evagination of the entoderm, and the coelomic cav-
ity of which is consequently a derivative of the gastrocoel: in
the Echinodermata (the vasoperitoneal vesicle), Balanoglossus,
Brachiopoda, Ascideae, and Amphioxus.

(b) The mesoderm is produced in the form of a coelom sack,
which is an invagination of the entoderm, and the coelomic
cavity of which is therefore derived from the archicoel, in
Sagitta and Phoronis.

The three types of development of the mesoderm delineated
above are not sharply distinct, since they differ from one
another rather in degree than kind, and are connected by inter-
mediate stages. Thus we read on p. 12 of Korschelt and
Heider’s Lehrbuch: “So verschieden diese beiden Arten der
Mesodermbildung (namlich Urmesodermzellen, Urdarmdiverti-
keln) auch scheinen mogen, so lassen sie sich doch auf ein ein-
heitliches Schema zuriickfiihren, wenn wir annehmen, dass im
ersteren Falle die Mesodermdivertikeln frihzeitig (als Urmeso-
dermzellen) den epithelialen Verband des Entoderms verlassen,
wahrend im zweiten Falle die Mesodermzellmasse vorlaufig in
dem epithelialen Verbande bleibt und erst spater durch die
Divertikelbildung zur Lostrennung gebracht wird” (cf also
p. 267). In fact, I would refer both the production of the
pole-cells and of coelomic diverticula to the more primitive
mode of formation of multipolar mesenchymatic migration.
By the mesenchym cells remaining in contact as an epithelium,
coelomic diverticula are produced, which are bipolar but multi-
cellular in origin and character. In the case of the two
““Urmesodermzellen,” I would explain, in agreement with
KKorschelt and Heider, the two pole-cells as two mesenchym

No. 2.] THE MESODERM AND MESENCHYM. B57.

cells, the mesenchym tissue being here unipolar and bicellular,
instead of multipolar and multicellular in origin.

Further, in meroblastic ova, from which the gastrulae are
not of the invagination type, but sterroglastulae, etc. (as in the
Cephalopoda, Sauropsida, Hirudinea, Turbellaria, most Arthro-
poda, etc.), the whole process of production is so much modified
by the amount of yolk in certain of the blastomeres, that the
mode of development of the mesoderm can with difficulty be
referred to the three more primitive modes of development
given above.

The examples given show how multifarious the development
of the mesoderm is in different animal groups.

Again, there is no sharp distinction between mesenchym and
mesoderm, but only a difference of degree; as is shown in a
large number of the newer embryological researches. Thus,
while the term “mesenchym”’ is generally applied to cells of
the third primitive embryonic layer which are not united
together in a continuous mass, and while “mesoderm”’ is
applied to cells of that layer when from the beginning they
are in contact with one another, forming a mass of cells (‘‘mes-
oderm-stripe’’), we find, in the case of the production of the
mesoderm through typical pole-cells, that the two cells may be
or may not be in contact, but that the mesoderm cells prolifer-
ated by each construct a solid cell mass. In fact, whether the
cells of the third embryonic layer are not united (‘“‘mesenchym’’)
or are united (‘‘mesoderm,” sensu strictiort) would seem to
depend upon two factors: (1) the comparative size of the arch-
icoel and (2) upon the modes of production of the third embry-
onic layer, which modes we have found to intergrade. And in
such groups as the Mollusca and Arthropoda, it is difficult to
decide whether the third embryonic layer is to be classed as a
“mesoderm” or as a “mesenchym.” Accordingly, these two
terms express but extremes of a series, and morphologically
cannot be sharply distinguished.

Lastly, as to the morphological value of the cavity enclosed
by the mesoderm cells. This cavity is a derivative of the arch-
icoel in all those forms where the mesoderm is formed (1) by
multipolar, mesenchymatic delamination, or (2) by pole-cells,

358 MONTGOMERY. [Vou. XII.

or (3) by invaginated diverticula of the entoderm (in Sagitta
and Phoronis). It is a derivative of the gastrocoel when the
coelomic sacks are evaginations of the entoderm (as in the
Echinodermata, Balanoglossus, Brachiopoda, Ascideae, and
Amphioxus). In the Ctenophora it is both gastrocoelic and
archicoelic in origin. This cavity is called a pseudocoel when
it is not lined by a continuous epithelium of cells, and a coe-
lom when it has such a lining. The pseudocoel of a turbella-
rian and the coelom of an annelid are both derivatives of the
archicoel ; but the coelom of Amphioxus is gastrocoelic in origin.
In the same animal, coelom and pseudocoel, when both cavities
occur together, may communicate. From these facts we may
conclude, that there is also no valid morphological distinction
between a pseudocoel and a coelom, when both are derivatives
of the archicoel. Whether, however, a morphological distinc-
tion can be drawn between the archicoelic coelom of an annelid
and the gastrocoelic coelom of an Amphioxus, is a question
which I believe has not yet been considered.

In regard to the ontogenetic derivation of the blood-vessels,
or blood lacunae, their endothelia, and the free blood corpus-
cles, the following points are of interest. These observations
are based mainly upon data given in the zodlogical text-books
of Korschelt and Heider, Hatschek, and Arnold Lang.

In the Nemertini the cavities of the blood-vessels and of the
rhynchocoel are archicoelic, and these are lined by endothelia,
from which the floating cells are derived. And, as I have
shown in a paper, ‘‘On the Connective Tissues and Body Cavi-
ties of the Nemerteans,” etc., to appear in Spengel’s Zool.
Jahrb., the gonads and their products are also archicoelic — at
least in all those forms in which the gonadal sacks are then first
formed, when the sexual cells become differentiated from the
somatic cells.

In the Chaetopoda and the Archiannelida the cavity of the
blood-vessels is archicoelic, their walls, from which the free
corpuscles are derived, splanchnopleuric in origin. In the Hir-
udinea, to cite Korschelt and Heider (p. 222): ‘Von den dor-
salen und ventralen Blutgefassstammen ist angegeben worden,
dass sie vom splanchnischen Blatt aus, durch Spaltung dessel-

No.2.] THE MESODERM AND MESENCHYM. 359

ben, ihren Ursprung nehmen”’; they stand in open communi-
cation with the pseudocoelic body sinuses, the latter having an
endothelial lining in the Rhynchobdellini, but not in the Gnath-
obdellini. The blood corpuscles of the leeches are probably
derived from the blood-vessels as well as from the so-called
“‘parenchym” (Lankester). In the Sipunculacea, there is
probably a communication between the blood-vessels and the
coelom, the splanchnopleuric peritoneum of the latter being
ciliated, as in the frog (Lang, p. 213).

Echinodermata: ‘‘Spaltbildungen im Mesenchym sollen die
Blutgefasse der Holothurie entstehen lassen. . . . Die Blut-
zellen hingegen sollen sich von den Wandungen des Hydro-
enterocoels losgelost und bei der Bildung jener Gefasse
betheiligt haben. Diese freien Zellen, welche sich sowohl
in der Leibeshohle, wie in den Ambulacral- und Blutgefassen
finden, wiirden also nach dieser Auffassung (Semon) nicht von
dem urspriinglichen Mesenchym abstammen” (Korschelt and
Heider, p. 286). In the Asteroidea: “In der zwischen Hydro-
coel-, Enterocoel-- und Darmwand gelegenen Mesenchym-
schicht bildet sich dort ein Spalt, welcher eine Auskleidung von
sehr flachen Zellen aufweist”’ (zbzd., p. 290). The blood-vessels
in the Crinoidea communicate with the coelom, and arise as
cavities in proliferations of the enterocoel (zdzd., p. 302).

In all the Arthropoda, as is well known, the blood-vessels
stand in communication with the body cavities. The heart of
the Crustacea is mesodermal in origin, its cavity archicoelic
(Korschelt and Heider, p. 376). In Limulus (zdzd., p. 528),
the heart develops from the mesodermal plates, and cells wan-
der into its cavity, which had their origin in its walls. In the
Scorpionidea (zbzd., p. 556), a segmented coelomic cavity arises
in the primitive mesodermal proliferation, the walls and free
corpuscles of the blood-vessels being mesodermal (is their cav-
ity then coelomic ?). As to the Araneina (zézd., p. 615), it isa
disputed point whether the blood corpuscles are merocytes or
whether they originate from the protovertebrae ; later they form
a solid chord on the dorsal side of the embryo; they separate
from one another again, when the heart becomes formed by a
coalescence of neighboring splanchnopleuric layers; thus the

360 MONTGOMERY. [VoL. XII.

cavity of the heart is archicoelic. In Peripatus (2dzd., p. 711),
the heart is formed by mesodermal cells, its cavity being arch-
icoelic. The blastocoel of the Myriapoda (zézd., p. 752) becomes
filled with yolk, while later mesodermal cells penetrate and sur-
round the yolk, producing a pseudocoel ; these cells accomplish
the formation of the blood-vessels, and perhaps also of the
heart. The body cavity of the Insecta is a product of the
primitive archicoel and the coelom ; the heart is formed of mes-
odermal cells of protovertebral origin, and the blood corpuscles
are also of similar derivation (zézd., pp. 818, 833, 834).

In the Mollusca the archicoelic (pseudocoelic) body cavity :
“stellt im Allgemeinen das Lacunen- und Sinussystem des
Korpers dar, in welches sich die Arterien 6ffnen, und aus
welchem die Venen, wo solche vorhanden sind, ihr Blut bezie-
hen. Sie ist ohne eigene Epithelwand”’ (Lang, p. 792). The
true coelom is much reduced, being represented only by -the
pericardial and gonadal cavities, and is bounded by its own epi-
thelium. The blood-vessels (a closed system only in the Ceph-
alopoda and certain Prosobranchs) have no endothelium (Lang,
p. 780). In the Lamellibranchiata, the heart, aortae, and gill-
veins are produced by the agency of mesodermal cells in the
primitive body cavity ; there is no communication between the
blood-vessels and the pericardial cavity (Korschelt and Heider,
pp. 971, 973). The heart in the Gasteropoda is derived from
the pericardium, while “die Gefasse entstehen als Liickenraume
im mesodermalen Zellmaterial der primaren Leibeshohle (Blas-
tocoel), also zunachst ganz unabhangig vom Herzen” (zdzd.,
pp. 1081, 1082). In Phoronis a coelom is well developed and
is bounded by an endothelium (Lang, p. 213). The blood-vessels
probably arise as lacunae in the splanchnopleuric mesoderm :
‘‘wahrend nach Cori das Gefasssystem des ausgebildeten Thieres
ein vollstandig geschlossenes ist, scheint in der Larve eine Com-
munication zwischen demselben und dem Kopftheil der Leibes-
hohle zu existiren. Im letzteren sollen die Blutkérperchen in
Massen angehauft ihre Entstehung nehmen” (Korschelt and
Heider, p. 1184).

In the Ascideae, the early coelomic cavities give way later
to a mesenchym, which fills the archicoel: “die spater in die-

ae

No. 2.] THE MESODERM AND MESENCH YM. 361

sem Mesenchym auftretenden Lacunen miissen ebenso wie die
Blutgefasse (welche . . . einer endothelialen Wand vollkom-
men entbehren) als Pseudocoel betrachtet werden. . . . Indem
einzelne Zellen des Mesenchyms frei werden und in das Pseu-
docoel gelangen, bilden sie sich zu Blutkorperchen um” (Kor-
schelt and Heider, pp. 1289, 1290). In Salpa the heart is of
mesodermal origin, and the vessels are lined with an epithelial
intima: “die Blutgefasse entstehen anscheinend als Liicken-
rdume innerhalb jenes gallertigen Bindegewebes, welches in
spateren Stadien die primadre Leibeshohle erfiillt” (cdzd.,
p. 1346).

In the Vertebrata the cavity of the blood-vessels is probably
archicoelic and their walls mesenchymatic in origin; but the
present views upon the derivation of the corpuscles and of the
blood-vessels themselves are very conflicting.

To recapitulate in regard to the origin of the blood-vessels
and blood corpuscles: the cavity of the blood-vessels (with the
exception of the heart in the Gasteropoda) is apparently always
archicoelic (blastocoelic), and never gastrocoelic nor coelomic.
The blood corpuscles in most of the animal groups are derived
from the endothelial lining of the vessels where such membrane
is present, but where it is absent, from the surrounding connec-
tive tissue elements. Accordingly, since the walls of the blood-
vessels may be mesenchymic (Nemertini?, Hirudinea Asteroidea,
Ascideae, Vertebrata?), or may be mesodermal (Annelida, Holo-
thuroidea, Limulus, Scorpionidea, Peripatus, Insecta, Lamelli-
branchiata), so the blood-vessels themselves may be either
mesodermal or mesenchymic, or both (e.g., Hirudinea).

The foregoing brief summary of our present knowledge on
the development of the various body cavities of the Metazoa
would lead to the conclusion, that particular differentiations of
this cavity in one group cannot be safely homologized with
similarly situated cavities in other groups. And the reason for
this is not far to seek. For, in the first place, the very differ-
ent modes by which the process of gastrulation takes place —
a process which seems to become modified by the mechanical
factors determining the previous cleavage of the egg — induces
very heterogeneous formation of both archi- and gastrocoel.

362 MONTGOMERY. [Vor. XII.

Thus, though under the term “archicoel”’ is understood the
cleavage cavity between the embryonic layers ectoderm and
entoderm, it is in reality also equivalent to the space between
any two neighboring blastomeres, as, ¢.g., the cavity between
two ectodermal or between two entodermal cells, or even
between an ectodermal and an entodermal cell. Further, since
this cleavage cavity is, in pre-blastula stages at least, in a direct
communication with the outside (as is well seen in the cleavage
of the Ctenophora and Tricladidea), its relation to the gastro-
coel is found to be close. And in the very frequent, and
perhaps most primitive, method of gastrulation, according to
which the entodermal cells delaminate from the ectoderm and
wander into the blastocoel, we find that the cavity enclosed by
the later entoderm, and thus comparable to a gastrocoel, is, in
fact, a portion of the earlier blastocoel. So, nearly all intergra-
dations may be found between the cleavage illustrated by the
triclad Turbellaria, where the blastomeres are at first isolated
(with consequently a large and open cleavage cavity), and the
sterroblastic cleavage exhibited, e¢.g., by the polyclad Turbellaria
or the Rotatoria, where the blastocoel is represented merely by
clefts between adjacent blastomeres.

Gastrocoel and blastocoel are nothing more than spaces
between or enclosed by the blastomeres, which, in early stages
at least, communicate with the outside. The gastrocoel may
be in certain cases a space lined by entodermal cells, divided
off from the primitive blastocoel, or it may be an extraneous
space bounded by such cells. In the same group of animals
both modifications may occur (e.g., in the Mollusca, the Crus-
tacea, and Turbellaria).

The embryonic body cavities known as blastocoel and gas-
trocoel are, therefore, not morphologically distinct spaces, and
only in certain cases (e.g., typical invagination gastrulation) are
they to be separated. For the mode of formation of both is
apparently dependent upon such factors as the mechanical pres-
sure of the yolk, etc. (other factors might well be at work,
which we have failed to recognize), that is, dependent upon the
process of cleavage; and the latter process, as is well known,
may present great differences in closely allied forms (as in the

No. 2.] THE MESODERM AND MESENCHYM. 363

Crustacea or Turbellaria), and hence no high morphological
value can be attributed to it.

Obviously then the mesoderm, as well as the spaces enclosed
or penetrated by mesodermal elements, cannot be granted impor-
tance in morphological classification ; which deduction is in
accord with a number of recent investigations, that stand in
opposition to the acceptance of the germ-layer theory. For
the development of the mesoderm and its cavities is, in its turn,
dependent upon the cleavage and gastrulation processes. Thus
a mesodermal pseudocoel as well as a coelom are usually blas-
tocoelic; but the coelom may be also gastrocoelic in origin.
Similarly there is no essential difference between formation of
the mesoderm by detached and isolated cells, and by coelom
sacks or epithelially united mesoderm stripes.

Accordingly, I am led to conclude that the body cavities in
different animal groups cannot be homologized merely on the
ground of apparent similarity of development; for the earlier
development and differentiation of these cavities must be
referred, directly or indirectly, to the modes of cleavage and
gastrulation, and the latter, as is well known, often differ widely
in closely allied forms.

The coelom of a vertebrate is frequently spoken of as being
homologous with that of an annelid, since it passes through an
apparently similar development. Now without stating or in
any way wishing to imply that the homology of the coelom in
this case is not correct, I would emphasize the point that the
similarity of development is itself not an adequate reason for
the homology. This standpoint should seem justifiable to any
one acquainted with the facts reviewed in this paper, which
tend to show how various the formation of mesoderm and its
cavities are in closely related forms.

Are, then, the body cavities possible of homologization ?
Comparison of the modes of early development shows that the
ontogeny is of little value in this connection; but it might be
thought that comparative anatomy could be of avail in the
search for homologies.

But, though often spoken of as such, a “ body cavity” can-
not be considered an organ, equivalent, e.g., to a brain or a

364 MONTGOMERY. [Vor. XII.

nephridium, but rather it must be compared to a system of
organs, since a number of different functions are performed by
it, and since it stands in a peculiar relationship to most, if not
all, of the bodily organs. The whole origin, manner of differ-
entiation, and extent of development of the body cavity, is to a
great extent dependent upon the mutual positions of the organs
as well as upon their degree of correlation. Therefore, any
change which causes a difference in the relative positions of
the organs must also effect some degree of change in the
diversification of the body cavity.

The body cavity being essentially a space or system of spaces
separating and penetrating the organs of the body, a common
possession of these organs, cannot be correctly homologized, for
the very fact that it is not comparable to any individual organ.
It would be as difficult to satisfactorily homologize these
cavities in different forms as to establish homologies between
states of correlation, or between mutual arrangements of
parts. And in case examples are required we may consider
the group of the Nemerteans, for it was a study of the body
cavities in this interesting group of worms which led me to
make the present examination into the question of possible
homologies.

In the Nemerteans, the body cavity occurs as (1) the rhyn-
chocoel (the large space surrounding the proboscis); (2) the
cavity of the blood-vessels ; (3) the perivisceral space situated
between the intestine and blood-vessels on the one hand and
the body muscular wall on the other. The portions of the cav-
ity mentioned under (2) and (3) are remnants of the cleavage
cavity (z.e., archicoelic in origin); but whether the rhynchocoel
also is archicoelic has not yet been definitely settled. Now in
a previous contribution, to which I have already referred, I com-
pared the pseudocoel in this group to the coelom of the Annelids,
and for the following reasons: the Nemertean pseudocoel —
that space the existence of which has, until recently, been ques-
tioned —encloses in the adult worm true mesenchym tissue con-
sisting of multipolar cells ; it is from certain of these cells that
the sexual cells are derived. Further, in all the species where
the gonadal sacks are not preformed, —and this is the case in

No. 2.] THE MESODERM AND MESENCHYM. 365

the majority of forms, —pseudoepithelia of these primitive
sexual cells arrange themselves in the form of paired and meta-
meric sacks, which are then the gonads. Thus the comparison
holds good, that in both Annelids and Nemerteans the sexual
products are derived from the lining of the perivisceral body
cavity. Inthe Metanemerteans the blood-vessels are completely
closed, while in the lower Nemerteans they are (in the head
region) in communication with the pseudocoelic spaces: in the
former as in the Chaetopoda, in the latter as in the Hirudinea.
Thus parts of the body cavity in the Nemerteans may be com-
pared to a true coelom, other portions to a pseudocoel (archi-
coel). The great difficulty of determining the homology of the
Nemertean body cavity is simply due to the fact that it unites
characteristics of a coelom and a_pseudocoel — formations
which, according to the mesenchym theory of Hertwig, had
formerly been supposed sharply distinguishable. So we find
the difficulty is in reality owing to the fact of the impossibility
of correctly homologizing such structures as body cavities ; for
not only may two (supposedly) different types of cavities be
present in the same species, but the two are frequently so
intermingled in it as to render their recognition almost im-
possible in our present state of knowledge. Thus we are not
in position to state whether the Nemertean body cavity is of
Turbellarian or of Annelidan character.

The general conclusion which I would maintain, then, is that
body cavities in different animal groups cannot be safely homol-
ogized, either from the ontogenetic or from the comparative
anatomical standpoint; though the latter method would seem
to be, in this matter at least, more reliable than the preceding.
The body cavity, whether as coelom or as pseudocoel, is not
comparable to any single organ or set of organs, but must be
considered as a structure of approximately equal economy to
all the organs. And in its early formation and later differ-
entiation probably most if not all of the organs take part,
beginning with the blastomeres as the earliest. Accordingly,
before seeking to homologize the body cavity, the morpho-
logical value of all the organs themselves must be learned, as
well as the morphological value of their mutual topography.

366 MONTGOMERY.

In my search for the homologies of the coelom, or pseudo-
coel, of the Nemerteans, a search which has led me to negative
results, I have been obliged to exceed the limits which I had at
first intended; but it is to be hoped that this brief critical con-
sideration of the morphological value of body cavities in general
may throw a new light on these structures.

Any one wishing to compare earlier views upon their nature
may consult, outside of the text-books of Balfour, and of
Korschelt and Heider, the following papers:

BALFour. On the Structure and Homologies of the Germinal Layers of
the Embryo. Quart. Journ. Micr. Sci., 20. 1880.

BUTSCHLI. Bemerkungen zur Gastraeatheorie. Morph. Jahrb., 9.

HAECKEL. Die Gastraeatheorie, die phylogenetische Classification des
Thierreichs und die Homologie der Keimblatter. /exa. Zeztsch., 8.

IpEmM. Nachtrage zur Gastraeatheorie. /dzd., 11.

HATSCHEK. Studien iber Entwickelung des Amphioxus. Avr. zool. Inst.
Wien, 4. 1881.

HERTWIG, O.AND R. Die Coelomtheorie. Versuch einer Erklarung des
mittleren Keimblattes. Jena, 1881.

LANKESTER. On the Primitive Cell-layers of the Embryo as the Basis
of Genealogical Classification of Animals, and on the Origin of
Vascular and Lymph Systems. Azn. and Mag. Nat. Hist., 11.
1873.

IpEM. Notes on the Embryology and Classification of the Animal King-
dom, etc. Quart. Journ. Micr. Sci..17. 1877.

RABL. Theorie des Mesoderms. Morph. Jahrb., 15. 18809.

WALDEYER. Die neueren Forschungen im Gebiet der Keimblattlehre.
Berlin. klinisch. Wochenschr. 1885.

WISTAR INSTITUTE OF ANATOMY AND BIOLOGY,
PHILADELPHIA, Pa., July 15, 1896.

SOME SPINNING ACTIVITIES OF PROTOPLASM
IN STARBISH AND SEA-URCHIN, EGGS.

GWENDOLEN FOULKE ANDREWS.

THE observations recorded here were begun at the Marine
Biological Laboratory of Wood’s Holl, in the summer of 1893.
For use of an “investigator's room” and other privileges
enjoyed there for the third time, I am indebted to the kind-
ness of the Director, Dr. Whitman.

While watching the protoplasm of developing starfish and
sea-urchin eggs under very high powers, certain curious filose
phenomena restimulated an interest that had for years con-
cerned itself much with the “thread-forming,” “filose,” or, as
I prefer to call them, the spzznzng activities of the living sub-
stance, especially as found among the Protozoa, where they are
of widespread occurrence outside the extremely large group of
protoplasts in which they are characteristic phenomena.

The eggs in which the supposedly typical phenomena were
watched were normal, so far as could be discovered by a series
of comparative observations and experiments on many other
specimens, some of which developed normally into quite ad-
vanced stages. Great care was used to make little or no com-
pression on these eggs, to keep the water fresh and plentiful
about them, to maintain an even temperature, and to hold them
but a few moments at a time under observation.

Other cases which were carelessly dealt with, or which were
immature, or over-fertilized, showed plainly an abnormal state
of their substance, which was visible in spinnings of a decidedly
different character from those of the normal specimens. These
cases were followed closely for long periods, and normal speci-
mens were made abnormal by heat, or pressure, or confinement,
so as to learn the peculiarities of such states, and, if possible,
distinguish with regard to the spinning phenomena between
these and the normal condition.

368 ANDREWS. [VoL. XII.

The observations at Wood’s Holl were made with a Beck
jz immersion, which had been carefully chosen from a number
for its flatness of field, definition, and lack of color, — with apo-
chromatic eyepieces.

At the Zodlogical Laboratory of the Sorbonne, Roscoff, Brit-
tany, where I was enabled by courtesy of M. Lacaze Duthiers
to make a series of similar observations, the instruments used
were all of Zeiss, — largest size stand, with Abbé condenser and
iris diaphragm ; 4.0 mm., and 2.0 mm. immersion, lenses ; with
4-12 apochromatic oculars.

While there proved to be nothing to correct in the first
observations, they were much extended and amplified by the
second series whose range of magnification brought to sight a
greater length of filament, also secondary spinnings, or ramifica-
tions of these, and still other primary filaments far beyond the
reach of the powers before used.

The sea-urchin eggs were from two genera ; those obtained
at Wood’s Holl, from Arbacia ; those used at Roscoff, from the
common, large, and finely colored species of Echinus found
along that coast.

The Arbacia eggs were perfectly normal in their initial
conditions, and in large proportion developed to a late larval
stage; they, as well as the sperm, were obtained artificially
from the animals. The Echinus eggs and sperm were also
obtained artificially, but, being more hardy, gave, it was
thought, for that reason more reliable data than Arbacia even,
although the results in the one case but confirmed those in the
other.

The starfish eggs used in both series of observations were of
the same genus, but of different species. In all the test cases
the eggs were laid and fertilized naturally in a tank shortly
after capture. Those collected at Roscoff were sometimes
from two or more females and two or more males depositing at
one time. The eggs were, however, protected from too great
admixture with sperm by instant separation, and afterwards
carefully washed and then kept in small quantities in
separate, large dishes in which the water was frequently
changed.

No. 2.] SPINNING ACTIVITIES OF PROTOPLASM. 369

In many of these dishes it was difficult later to find any eggs
which had not reached a normal, free-swimming, larval state
due at about that given time. The test specimens were kept
isolated of course, but compared with results given by random
specimens from the larger number.

In the eggs of Arbacia, I had seen the characteristic, striated
membrane produced, immediately after entrance of the sperm,
by formation of innumerable, delicate, thread-like processes
from a clear pellicular layer of the egg. Although the
resemblance between these threads and those formed in the
tuft of protoplasm which receives the sperm was very great,
I could not convince myself that they were not perhaps
of non-living matter and caused by mere exudation of a.glairy
substance, which, if issuing from small, pore-like openings,
could assume a thread-like shape. An appearance of partial
return in single instances of the substance towards the egg, did
not seem proof to the contrary, since it was thought that this
would be quite possible were the substance of the nature
suggested.

In the Echinus eggs, the question seemed to be solved by
several cases of polyspermy, in which the egg, seen to be
immature from the nuclear conditions, never formed a perfect
membrane, but merely made abortive attempts to do so, the
various portions of its surface having more or less success,
or making, perhaps, more or less exertion, in spznning. Of
the most typical instance of this sort, a camera lucida memo-
randum was made, showing the various protoplastic activities
of the pellicular substance, and also the several tufts which
mark the entrance, complete or partial, of several sperm.

The whole surface of this egg, soon after entrance of the
first sperm, was covered with a rather thick and albuminous-
looking material, like ectosarcal protoplasm, in which no vesic-
ular structure of Biitschli was visible, although there was a
deceptive appearance of such, caused, as it seemed, by optical
sections of papillose, or short, thread-like, processes.

This clear layer seemed at first to be only partially effective
in preventing the entrance of more sperm, and it responded to
the attempts of these with more or less pronounced tufts, and

370 ANDREWS. (Vou. XII.

showed wave-like modulations of its surface outline, possibly
due to accession of more material from the egg.

About the point of entrance of the first sperm, there arose
for some distance a host of delicate threads, closely resembling,
except in their instability and their unevenness of length,
those of the normal eggs. These threads had an optical effect
of being outspinnings of the outer clear layer of the egg, and
of the mere surface substance of this. At other regions there
was sometimes an effect of the processes piercing this and
arising from the inner or true surface of the egg itself. That
they did arise in some places from the outer layer was shown
by local formations of short, papillose roughenings of this,
which were seen first as mere superficial unevennesses but
which afterwards extended themselves so as to form true
threads.

That the outer layer and its products were, for the time
being at least, part of the living material of the egg, —a true,
ectosarc-like formation, — seemed proven by ramifications of the
threads in the more abnormal regions of spinning ; by unsta-
ble anastomosis with each other of such secondary, and also of
the primary, processes ; and by the general flux of their sub-
stance to and from the layer, as well as amongst each other,
in a manner most characteristic of protoplastic phenomena.
Such activities were noted in the tufts, and indeed there
were no characters optically discoverable which enabled one to
separate the two groups of phenomena in these abnormal eggs.
In normal cases the single tuft is formed from a delicate
covering of the egg, which has the same general character as
the thick, abnormal envelope, being merely more stable and
quiet.

At time of forming these areas and processes in the abnormal
type-instance the nuclear membrane had not yet been dissipated,
but lay somewhat to one side of the egg. The nuclear area,
although perfect, showed itself to be immature by the mode of
distribution of its elements.! The progress of this special egg

1 Certain progressive differences which mark the development or maturation of
this cell body are described in a forthcoming thesis on the living substance Zev se
in some of its more minute structural aspects.

MNor2.) WPIVNING ACTIVITIES OF PROTOPLASM. 371

was not followed directly, but was seen later to have had a
markedly degenerating course.

Wherever such abnormal processes of membrane formation
by spinning were watched, the processes differed from the
normal filaments in being more irregular and unrestrained in
protoplastic character ; there was less of direct thread forma-
tion, more of tufts or brush-like processes, which were unstable
in progress and even returned altogether to the egg covering.

In the normal eggs the spinnings were swift, straight, smooth,
direct, and continuous in their progress, neither ramifying nor
returning. They seemed more homogeneous optically, and had
not granules scattered along their course, as the abnormal
spinnings often had for a considerable distance. Normally
the threads came to be of about even length at about the same
moment, while in abnormal eggs they were at a given moment
unlike at almost all points of the periphery, the most perfect
being formed always about the region of entrance of the first
sperm.

In normal eggs, the threads, having reached a certain, but
rather variable, distance from the egg, appeared to fuse at their
tips and then to spread out their substance there, so as to form
a ceiling film. Increase of outflowing substance soon thick-
ened the pellicle thus made, and almost simultaneously the
threads themselves proceeded to fuse along their length, either
by access of material from the egg over them, their linear
extension having been to a great degree stopped, or by a
spreading out of their substance as they continued to be formed
from the rear; or perhaps even by exudations or secretions
from them.

The distance from the egg at which the threads began to
fuse at their tips was variable ; when it took place at but a
short distance, the film so formed was raised and extended by
continued elongation of the threads, the lengthwise filling in
beginning then or later, as the case might be.

And in this manner was a striated membrane seen to be
formed about sea-urchin eggs.

Starfish eggs. —It has been commonly observed that, just at
the region where the so-called polar globules are expelled, the

372 ANDREWS. [Vou. XII.

cytoplasm shows amoeboid movements before, and for a short
time after, these bodies are thrown off.

Under magnification of three thousand diameters, the lobe-like
protrusions of the substance here were seen to be still further
extended in filose processes, most like the outer portion of the
tuft which receives the sperm, but finer and more thread-like,
less protoplastically diffuse. ;

After the globules have been thrown off, these processes are
for a short time withdrawn, while the general surface of the egg
smooths itself out, restoring the general contour. For a short
time only, however ; for very soon the spinnings arise again
from the pellicular surface, now in the form of still more deli-
cate threads, or rays, which extend themselves towards the egg
membrane, and even at moments attach themselves to this.
They also surround the polar bodies with ramifications of
extreme minuteness.

At other points of the periphery similar spinnings spring up,
the egg being still in the single-celled stage, and thus is inaugu-
rated a phenomenon which from this time persists with varying
freedom until shortly before the closing of the opening into the
cleavage cavity, across which the polar bodies lie. Then all
peripheral spinnings cease, to burst out once more for a few
moments all over the pellicle when the free-swimming life of
the larva is begun, after which no more filose activities were
seen from the exterior of the embryo.

The threads, or rays, formed from the pellicle in the one-
celled stage are of extreme delicacy in the normal egg, and in
many cases are but just perceptible with the powers named,
except near their point of origin in the pellicle, where they are
thicker and less viscid in appearance.

They branch sometimes near their base and sometimes further
out from the egg, when in their more fluid states, for their
viscosity and their refractive quality vary from moment to
moment. Such branches take a somewhat wandering course
often, and anastomose, or interlace, with the secondary, or
even at times with the primary, processes near them, so as to
form unstable networks. The branches are often as thick as
the threads from which they arise.

Now2Z:). SPINVING ACTIVITIES OF PROTOPLASM. 373

The rays do not have always a radial direction as to the egg,
not even at the moment of their formation, but may make with
the surface of the cell almost any angle, even so acute an one
as to be actually tangential.

They are moved slowly at times, as if from a hinge, or ball-
and-socket joint, at their base and point of termination in the
egg pellicle. They are seen also to bend sharply and suddenly
at some point of their length, forming thus angles more or
less approaching right angles. And at all points of their
length they freely interchange varying states of viscosity ;
being here or there in the different filaments, or here or there
at different points in the same filament, either in a state of
fluid flux, or stiffly viscid to a point which may reach elastic
rigidity.

The spinnings from the pellicle continue without intermis-
sion during all the internal preparations for caryokinesis, and
for cell division, showing no constraint during aster formation,
nor even in actual cleavage, except for a moment just before
actual splitting of the surface begins, and just along the
region, or coming path, of that splitting, for here they are
more or less completely withdrawn.

As external cleavage begins, the rays nearest the actual split-
ting line show usually some agitated-looking bendings about,
both from their base and along their course. Sometimes a quiv-
ering movement runs along them, with roughenings of the ray
outlines exactly as is seen often in Heliozoa. At other times,
in the earlier stages of the egg’s history, such phenomena
were seen from moment to moment in different rays at other
regions.

The split of the first cleavage is now visible as a triangular
cleft in the optical contour of the egg. In the moment this
appears, there is a visible haste in the protoplasm of the
rounded, divergent sides of the two cells in process of forma-
tion, to renew the spinning activities. The rounded optical
edges may show slight amoeboid modulation ; and then, rap-
idly starting up, a number of processes extend themselves
from each opposing surface, towards the other. This is
quickly reached, for there is both haste and directness of

3/4 ANDREWS. [Vou. XII.

formation, and so the two halves become in fact re-united by
their substance before they are well separated.

Following hard upon the physical splitting of the mass apart,
similar spinnings from the newly separated surfaces spring up,
so close indeed at the heels of the cleavage, that there is often
scarcely the width of a half dozen such threads between those
most newly formed and the still fused edges of the cleft. By
the time the first cleavage is ended, and this does not take
up many moments, the path of liquid between the sister cells
is already crossed many times by minute and most delicate ray-
like extensions, and strands, and even skeins of more tenuous
threads.

During succeeding changes of contour, when the blastomeres
round themselves up and then again flatten, and still again
become slightly concave in the centre, yet oppose a more or
less plane surface to each other, the spinnings never cease,
although the processes are in a state of general instability.

Under a lower power, even one of 1500 diameters, the filose
processes are wholly invisible, the surface of the cell presenting
an optically unbroken surface, which to still lower powers looks
smooth even.

While the cells lie opposite one another thus, there is some
ramifying and divergent spinning among the threads, and anas-
tomosis of some of the finer secondary filaments, but this is
mostly near the central, slight hollowing out of the surfaces ;
those threads having their origin in the more viscous-looking,
outer portion of this plane tending rather to form bands and
strands of more refractive and direct character.

Later, when the cells begin to approach each other, the con-
necting filaments show a decided tendency to still more marked
directness, and smoothness, and viscosity of texture. The
lateral, branching spinning becomes rarer; and as the cells
approach each other the band and rod-like connecting fila-
ments become shorter and gain a look of tenseness; and flow
between the two cells of protoplasm by way of these threads
was not noticed.

The threads shorten, and as they shorten become thicker
and more highly refractive, — that is, certain ones, for some of

>

No. 2.] SPINNING ACTIVITIES OF PROTOPLASM. 375

the processes, notably those which spun latest, are now gener-
ally retracted.

The whole aspect of things is as if the blastomeres were
being drawn together by contraction of some of the threads.

Moderate pressure upon the cover glass, slowly increased,
tending to a mechanical forcing apart of the cells with more or
less flattening, had a most curious effect, for it seemed rather
to intensify and hasten the drawing together of the cells, and
to increase the refractiveness of the threads, intensifying, rather
than lessening, their optical characters.

The threads were not seen to be drawn out into finer and
less refractive filaments by such pressure,— as would naturally
be the case were their nature purely physical, like that of the
threads left when Bitschli forced apart masses of his viscid oil
foams, — but, as stated, if the pressure were slow and moderate,
they held and even emphasized their peculiar character. In
finally giving way, as sometimes happened when the pressure
threatened to rupture the cell walls, they broke short off and
then rounded their ends in a viscid, mucilaginous-looking
manner before retraction, which soon took place.

Another curious fact was, that such mechanical pressure
caused the egg to withdraw many peripheral processes, but
seemed at the same time to increase, or re-stimulate, new forma-
tions from the opposing surfaces of cells. Later, the peripheral
spinnings were renewed with even greater activity if the pressure
were taken away; or, in cases where the cells were forced apart
and the membrane remained intact, it being very plastic under
slow pressure, the peripheral spinnings were not only renewed
at points over the entire surface, but extended themselves
through unaccustomed distances until they reached the other
cells. These results were best gotten in the 8-16 cell stage,
— which seemed peculiarly to favor them.

In the natural course of events in the typical, normal eggs,
after the blastomeres were closely apposed to, or fused with,
each other for a time, the same response to pressure was given
by the cytoplasm of the new mass. /Pvessure seemed, in short,
to increase the physical resistance of the mass to crushing stress.
Since, if the cells so treated happened to be in that rhythmi-

376 ANDREWS. PVOE. XDI:

cally somewhat relaxed state between cleavages, they rounded
themselves up under intermittent pressure instead of flatten-
ing ; it is thought that the peripheral substance was instru-
mental in these changes.

After two sister cells are in close apposition to, and what may
now, perhaps, be called safely, physiological union with, each
other, there is offered to mechanical pressure an emphatic, and,
if the pressure be slowly augmented, an increasing resistance.
I found that shaking also increased the resistance to pressure.
In making these pressure experiments I used a large, screw-
adjustment compressorium, making marks upon the head of
the screw and upon the supporting plate to estimate in a rough
way the relative amounts of pressure. No doubt special devices
for this purpose would bring to light an important series of
differences.

After re-union of the first two blastomeres to form what
may be termed a dual mass, matters progressed in the usual
rhythmic way towards the second cleavage. During this time
the spinning phenomena were persistent over the general
periphery, the processes being longer and extending themselves
to greater distances, which they may easily do because of in-
crease in size of the egg membrane.

It may be difficult to convey a true idea of the extreme deli-
cacy of these spinnings from the egg pellicle. In attempting
camera drawings, I found that even a fine needle point came
much in one’s way in efforts to follow their outlines. There
were local thickenings and differences of aspect which could
not be portrayed so at all, and the chief meaning of these lay
in their very evanescence. The only way seemed to be, to
receive as truthful and passive a brain photograph as possible,
and. to translate this into terms of permanent line and shape,
getting with the camera a skeleton of relative lengths and gen-
eral local character of the lines of living substance.

The rays showed all the characteristic differences known in
Heliozoan rays; and when one says this, much is said of many
variations. They were freely produced and as freely returned
to the substance of the pellicle from which they arose.

All the while, the alveolar structure, which was clearly

Nou2iy “SPINMING' ACTIVITIES OP PROTOPLA SIA. 2V7

traceable in the protoplasm underlying the pellicle, and much
finer than that found in the more central portion of the mass,
remained optically undisturbed by these displacements of a
peripheral substance which formed indeed the so-called pel-
licular film of the structures of Biitschli.

In attaining a four-celled stage, the phenomena were prac-
tically a repetition of the first set described. After fusion, the
blastomeres rounded themselves up and formed a hollow mass.
The central cavity of this, which was the “cleavage cavity,”’ of
course, was filled with a network of interlaced and even anasto-
mosed filaments. And still the periphery of the now quadruple
mass spun as before.

The most active portion of the common periphery seemed to
be always at the rounded sides of the partial cleft which marked
the junction of each two, adhering blastomeres. The most
freely ramifying and anastomosing processes seemed to come
always from the inner surfaces which lined the cleavage cavity,
and this portion of the general substance showed greater fluidity,
which extended inwards some little distance in each cell.

During the rhythmic cell-flattenings, the very active portion
of the periphery just mentioned was always most extended and
flattened, so that under highest powers it showed a decided
prominence. This, under lower powers, might possibly be seen
as a slight modulation of the rounded surface-outline.

Here the filaments attaching themselves to the opposing sur-
face would seem to pull it somewhat out of place, and here,
as in the actual cleavage, there was a tendency to formation of
fusing filaments, or strands and bands.

So matters continued throughout the cleavage up to the time
of closure of the opening into the cleavage cavity, over, and
sometimes within, which the polar bodies lay.

At all times the cleavage cavity was crossed by a variable
number of threads, connecting distant cells as well as those
nearer together; each cell seeming desirous of connection,
though not persistently, with all other cells. The filaments
were long, more or less direct, connections which in many
cases ramified during their course, sending branches to several
cells, or spinning by the way a network whose filaments

378 ANDREWS. (Vor. XII.

interlaced with, or joined themselves to, other filaments or net-
works met with.

The greatest number of spinnings seemed to be from the cells
near the polar globules, and this region of the egg, it will be
remembered, was the first to spin as well as the most active at
all times. In this region was shown also most tendency among
the peripheral filaments to ramification and anastomosis, espe-
cially when they crossed the path of certain spinnings from
the polar bodies, shortly to be described ; most capriciousness
in their optical qualities, and interchange of fluid and viscid
states, with greatest inclination to stiff viscidity.

At time of closure of the cleavage cavity pore, the cells
about it flatten markedly, and at the same time become irregu-
lar in optical contour ; the spinnings are then very strong and
viscid-looking, and the protruding regions are seen to be attached
by them to like regions of other cells. There follows much the
same series of optical appearances among the filaments as when
sister cells are drawn together, and for these reasons it seemed
not unlikely that the filaments were actually instrumental in
closure of the space between the cells.

After a closure of the space is effected, the general “ ecto-
derm”’ cells, as they multiply, become gradually flattened exte-
riorly beneath their common pellicle, so that all seeming of
intercellular clefts is obliterated. At the same time their
inner ends protrude somewhat within the cavity, so that here
there are marked intercellular clefts.

These clefts are crossed from cell to cell by delicate, hyaline,
filose extensions of the cell pellicle; and from the ends of the
cells, which are plainly more fluid than the outer portion, are
produced other and much longer threads, which extend even
across the whole cavity, binding the most distant cells into
physiological and direct continuity.

Up to this time the external spinnings of the mass pellicle
have been progressively somewhat less profuse. When the
blastula has a sufficient number of cells for its free-swimming
state these outer spinnings rather suddenly cease, but burst
forth again over the surface of the pellicle as hair-like pro-
cesses which quickly show irregular, contractile, or waving

No. 2.] SPINNING ACTIVITIES OF PROTOPLASM. 379

motions. These become stronger, more rhythmic and or-
ganized in action, and soon form a sufficiently harmoni-
ous impulse to carry about the blastula as a free-swimming
organism.

After invagination, those cells which as entoderm project,
first as a plug and later as a tube, into the cleavage cavity,
spun strongly and in a free protoplastic manner from their
portions facing the cavity, which, like the same region of the
ectoderm cells, were more fluid in appearance. These ento-
derm spinnings crossed the cavity to the ectoderm cells and
in their course freely anastomosed in many places with the
spinnings of the latter.

After the mesenchyme cells were added to the group of
internal cells, these, by spinning in even more free and pro-
fusely protoplastic a manner, increased very greatly the number
of intercellular connections.

It now became difficult to distinguish the various groups of
spinnings from each other; in the majority of cases where
anastomosis had taken place, it was impossible, because new
centres of spinning were formed of accumulated substance
which might have come from anywhere. The network was also
most unstable from moment to moment. Yet there were
many tapering processes which could be connected with their
point of origin.

Larval stages were followed up to quite late periods, even to
time of formation of the proctodzum; and the internal spinnings
seemed at no time to pause or cease.

It should be pointed out that in the early stages of the seg-
menting egg the protoplasm which spins is not the least, but
the most, highly organized portion of the cytoplasm. It is also
the most viscous portion, as can readily be demonstrated by
physical experiment.

Polar Globules. — Perhaps even more strange than these
spinnings from the egg were similar activities shown by the
cytoplasmic envelope of the chromatin granules of the polar
globules.

During that short period of quiescence for the egg pellicle,
which was spoken of above as following the extrusion of the

380 ANDREWS. [Vou. XII.

polar bodies, thread-like processes with some wave-like change
of contour were noted on these little lumps, the egg being at the
time under magnification of three thousand diameters. These
initial protoplastic activities were extended, and multiplied from
other portions of the hyaline protoplasm enveloping the nuclear
substance.

Delicate threads and ray-like processes grew out on many
sides, and the shape of the whole mass suffered change from
moment to moment, both from the actual displacement thus
caused and from an apparent pulling here or there of the little
cells by their attached filaments and strands which had formed
actual connection with the egg pellicle or membrane.

It seemed to me to be from this time, z.e., when the threads
from the polar bodies effected reunion with the parent cell,
that the egg spinnings were again renewed. However this
may be, it was certain that from this time forward, throughout
the entire cleavage of the egg up to a late larval stage, the
egg and these bodies were united by their individual spinnings,
which formed an unstable, and at times evanescent, series of
interlaced and anastomosed networks. These more or less
surrounded the polar bodies and stretched across the cleavage
pore.

The processes from the polar bodies were finer in general
than the average egg filament, but this was variable, and along
the finest filaments would often pass little masses, relatively
large in quantity, of flowing protoplasm which might collect
here, or there, and form nodes as it were, and these became
centres of renewed and somewhat independent spinnings.

From moment to moment, the substance composing the net-
work, or its islands of protoplasm, would return wholly to the
polar bodies, to be again sent out in a new direction and to
assume a new form.

The globules seemed to delight, especially later on in the egg
history, in forming brush-like tufts, or skeins, of finest fila-
ments, which often assumed a curious, superficial resemblance
to a spindle, spreading at a little distance from their point of
origin and then again drawing together at their point of fusion
with some ray or strand, or with the nearest egg cell.

No. 2.] SPJVNMING ACTIVITIES OF PROTOPLASM. 381

By shortening of the filaments, the position of the globules
was changed, these being drawn here or there ; and later their
shape underwent strange changes which forced one to correlate
them with existing thread formations. Now they would be
seen as irregularly spindle-shaped, with bundles of radiating
filaments arising from each end of each body; then as spherical
masses, with Heliozoan-like rays extending from all sides nearly
alike ; then again as colonial groupings of separated minor
masses of protoplasm connected by filaments and having the
chromatin granules distributed to some extent among the
larger lumps; then again, with filaments almost entirely
withdrawn, as amoeboid shapes with delicate wave-like expan-
sions of substance flowing in protean manner about the
granules which were then perhaps collected together at some
one point.

Delicate vacuolations of variable sizes appeared and disap-
peared in the main masses and in their smaller colonizing
masses.

The protoplasmic granules were transported here and there
along the processes, and the filaments underwent all those
changes of optical quality which have been described in the
egg spinnings.

The change of position and grouping of the chromatin
granules were deeply interesting, for they were at times
scattered, then drawn in line, then variously grouped. But
I was unable to determine any coherent significance in these
differences at the time, and dare not so much as guess that
their arrangement was other than fortuitously and mechanically
altered by flux of the surrounding substance.

The granules’ presence, even singly, in any mass of the
cytoplasmic spinnings certainly seemed to be correlated with
more continuous, persistent, and in a way organized, displays
of this sort.

From all the surrounding cells, passing filaments gave of
their substance from moment to moment to the compound net-
work which compassed the bodies about, and it was not always
possible to know from what source a given portion of the
network had come nor whither it was bound. It was always

382 ANDREWS. [Vou. XII.

possible to know some portions of the network as distinctively
the polar bodies’ own.

The position of the polar bodies, with relation to their dis-
tance from the egg membrane and the egg itself, varied also
throughout the time of development up to this point.

Their substance, like that of the egg, frequently travelled
to the membrane and there adhered, spinning backwards or
anastomosing with such egg filaments as happened to be there.

At time of closure of the cleavage pore, the polar globules
were shut inside the blastula, chiefly, it seemed, by action of
the egg processes overpassing them and drawing together the
cells above them; yet possibly they assisted also by some
migratory movements of their own.

Inside the blastula, they could still be followed clearly for
some time, until at last, after being involved in the web of spin-
nings from the ectoderm cells and, in the gastrula, from the
protoplastic ends of entoderm cells, they were finally lost to
sight in the still further additions given off from the mesen-
chyme cells. After this point their fate could no longer be
followed, though it enlisted the strongest interest.

Towards the time of closure of the cleavage pore an increased
tendency to viscidity and to organized action of the spinnings
was shown by the strange geometrical positions maintained by
the lines of living substance.

In one instance where the conditions remained stable for some
moments, a camera drawing was obtained of a certain arrange-
ment which from a mechanical point of view was remarkable.
One line of very viscid-seeming protoplasm formed an arc
which would make part of a very large circle. The ends of this
terminated in the angles formed by sharp bendings of two long
rays having their origin in the two polar bodies then lying
quite apart ; and their termination in two cells of the egg which
were not adjacent but separated byanother cell. Thus it came
about that one saw a viscous fluid in linear extension, maintain-
ing the curve of a distinct arc, yet attached to the angles formed
by what were, mechanically, four contending directions of ten-
sion. It might have been open to one to suppose that the
angled lines were very viscid, and that the arc line was a less

No. 2.] SPINNING ACTIVITIES OF PROTOPLASM. 383

viscid, hanging line of protoplasm, had not the existing condi-
tions negatived this view. The arc line was in fact far more
dense and refractive, also thicker than the angled lines between
which it held its place ; and it did not hang.

Appearances were far more in favor of its being this line
which bent the longer lines into their angles. Yet in such a
state of tension, how explain the curve? I have dwelt at
length upon this single phenomenon because it is typical of
the difficulties which everywhere beset physical and mechanical
interpretations when one brings them into contact with the facts
of the living substance.

The hyaline substance of the polar bodies frequently flowed
along the filaments of the egg, and gathering together at some
point spun characteristic brush and skein formations, or made
nodes for diverse sorts of protoplastic phenomena.

It was a noteworthy feature of the polar spinnings that they
so frequently grouped themselves in spindle-shaped bundles,
and showed a marked vesicular structure in their substance
at most times except during certain changes of viscosity.

Concentric chains and lines of vesicles gave often by their ar-
rangement the aspect of a spindle to the whole polar body, which
the outline of the mass at the same time emphasized. In the
brush and spindle-like spinning products there was often a very
distinct chain-like arrangement of vesicles forming the threads.
These were of course the larger and more stable processes.

Description of the filose phenomena of these eggs is an
almost inexhaustible subject, and I give only the more patent
phenomena.

The activities of the polar bodies did not decrease as time
passed after their extrusion from the egg, but rather increased,
both as to amount and as to the controlled and ordered nature
of the phenomena.

In other words, instead of their losing at all their individu-
ality and becoming more like non-living and excreted substances,
they rather gained than lost in independence of action and the
vigor and vitality of their organized activities.

Whether this were due to some influence exerted upon them
by the intimate connection of the egg spinnings with them,

384 ANDREWS. [Vor. XII.

through which physiological relation they might possibly
share in the growing individuality of the major body, it is
impossible to say. I am inclined to think that although they
undoubtedly show strange freedom and initiative energy,
although they seem to possess all that is requisite for what
would be termed from a Protozoan standpoint, a complete
organism ; yet, since they are also an integral part of the egg
substance and of the complex system of egg phenomena, they
must share largely in the sympathetic interactions and co6per-
ative physical and physiological phrasings of the powers of the
whole material, as well as in a progressive subjugation of the
whole life-machine to an ever-increasing despotism of parts.

In the Echinus and Arbacia eggs also the spinnings were
seen between cells, although with far less freedom than in
Asterias. The sea-urchin threads were fewer, in comparison,
more direct, and less given to forming networks and side-
spinnings. They arose from a more distinct, and _ thicker,
hyaline covering of the egg, which seemed structureless under
the highest powers, yet the spinnings themselves showed
often a vesicular tendency and structure, albeit of the finest.
Over the sea-urchin filaments granules as well as little lumps
of protoplasm streamed at times. In the four-celled stage, the
small cleavage cavity was crossed by a number of threads
between the cells, and thereafter whenever, and for as long
as, interspaces between the cells could be seen, spinnings
crossed these and bound the associated cells into an intimate
union.

Here, as in starfish eggs, the most active and sensitive por-
tion in each cell was where it began to curve away from an-
other cell.

In both starfish and Echinus eggs, after any artificial separa-
tion of the cells by such pressure as did not rupture the egg
membrane, and even in some cases where this actually occurred,
the spinnings were increased in number and the length of the
threads was extended until connection was again established,
as if indeed they sought for their lost comrades.

To the very natural question whether these filose activities
of the eggs were not indeed abnormal phenomena, it is to

No. 2.] SPLNVNMING ACTIVITIES OF PROTOPLASM. 385

be answered that the same question having great weight in
the observer’s mind, every effort was made to learn the
truth.

Thus much was made sure. The eggs can be stimulated by
additional heat, by polyspermy, by adverse states of the water
such as are caused by too long confinement, by mechanical
pressure, or by being fertilized:when in a still immature con-
dition, to spin far more freely.

But the protoplastic phenomena in these cases are distinctly
different from those of what I was impelled to believe were
normal eggs.

The character of the formations, the quantity of the egg sub-
stance, and the use made of it during the normal and abnormal
states, are so different, so characteristic, that, after following
a dozen or two specimens of abnormal spinnings in all grades
of pathology, one feels an almost unshakable assurance in the
nature of the true and normal process, and would almost be
willing to diagnose the state of an egg by the very character of
its spinnings in any given case.

Immature eggs being fertilized, or over-fertilized, give them-
seives up to uncontrolled spinning activities which in many
cases end by the whole mass being drawn out of form and
position and distributed through the surrounding space in a
granular and unevenly vesiculate, or a partially structureless, net-
work of substance. The transported substance may mass itself
here or there, adhering perhaps to the membrane and thence
setting up from broken, or partly isolated, lumps and nodes of
protoplasm new centres of activity.

Traversing relatively large spaces, often by way of invisible
extensions of substance, the deported protoplasm accumulates
at this or that point; and then may disappear, to come again
within optical reach as a new current from nowhere flowing
into or onto some filament in plain sight ; or as small lumps
of substance, growing from some invisible source, suspended in
an otherwise empty space, their true source being some distant
filament, or the egg mass from which outflowing currents take
their way, and pass at some point from one’s power to trace
them optically.

386 ANDREWS. [Vou. XII.

Mature and properly fertilized eggs when heated too much
or kept too closely confined, cease their normal, delicate spinning
to burst into profuse transpositions of their substance, which
may cause even considerable distortion of their form, or, accord-
ing to the degree of abnormality produced, may be visible only
as excess of the usual processes. If not too much affected, such
slight abnormalities may be cured and the perfect larva form as
usual.

If the adverse conditions are continued, such eggs often cease
from all further cell division and exhaust their powers in perfect
orgies of spinnings, as just described for abnormal eggs.

Protoplasm artificially pressed from normal starfish eggs in
early stages spins characteristically like abnormal unruptured
eggs and cells, not like those in normal condition. Such
activities varied also their quantity and character in correlation
with certain rhythms of difference of physical quality which
marked the living substance during development of both star-
fish and Echinus eggs. These rhythms are treated of in the
forthcoming essay mentioned above.

There was every optical evidence of actual interchange of
substance by the cells of developing starfish eggs and also
Echinus eggs by way of the filose processes; for a procession
of granules along a filament for some moments in one direc-
tion, sometimes ended by abrupt withdrawal of the filament.
Whether these granules returned later along some other fila-
ment could not of course be determined ; but even so, they
would have been for a time inhabitants of the cell in which
they were left, and so also would the substance of whose pass-
ing they were clearest indications.

A most noteworthy thing was made plain by this series of
observations. The vesicular structure of Biitschli, and that
modification of it termed by him the “alveolar layer,” clearly
traceable under favorable optical conditions in these eggs, exists
and remains optically undisturbed for moments at a time during
such peripheral activities as can drain the egg of considerable
amounts of its substance, yet pass undetected under the
powers which have been commonly used for observing develop-
mental phenomena.

No. 2:]) SPINWVING ACTIVITIES OF PROTOPLASM, 387

Further, there is present and visible under these lower powers,
and capable of “ preservation,” a perfect, pellicle-like covering
of ectosarc-like substance which under quite high powers may
show only such surface roughening as would pass as finely
granular, or seem smooth even, and yet be extending itself all
the while in hundreds of processes which are most powerful
determining factors in the scheme of development.

In the case of the polar bodies as in the cell division, it
becomes plain that the separation or isolation of portions of its
substance by the developing egg is but an optical illusion; and
that the space separation between sister cells, which has been
taken to be actual and complete and unbridged by living sub-
stance, and that cavity or space between all the cells of the
blastula known as the cleavage cavity; are deceptive in the
highest degree, being in fact bridged by all the cells concerned
by means of extensions of their living substance.

At all times the cleavage of the mass of these eggs into por-
tions which simulate units, is seen to be but a mask for actual
continuity of the substance of the whole throughout all its
subdivisions in cell form.

More than this, the cell substance is seen to have some sort
of deliberateness, or purpose, in its spinning activities, for
where the space between blastomeres is artificially increased far
beyond the accustomed limit, it hastens to cross by unaccus-
tomed degrees of extension, not resting till it has reached the
missing masses and reéstablished union with them.

A reasonable summing up of the phenomena described, —
weighted by a host of subtle, modifying and restraining evi-
dences which in such delicate phenomena must always exist
over and above the description, — would seem to be as follows.

The facts point to a physiological drawing together of the
cells, rather than to any physical and chemical “cyto-tropismus”’ :

To a physiological, rather than a physical reaction to me-
chanical stimulus of pressure or shaking :

To a physiological, rather than a physico-chemical, cause of
the spinning activities :

To actual and physiological communion as well as physical
connection between cells after and during their formation :

388 ANDREWS. [Vor. XII

To some interchange of the protoplasm between cells, which
may be but temporary, or may be part of the organising phe-
nomena:

To an actual desire on the part of the cells for such physical
connection and physiological communion among themselves;
and a coherent attempt to regain and continue these conditions
after they have been artificially destroyed :

To a continued, and to great degree independent, existence
of the polar bodies after their extrusion: and to their acting as
part of the coalition of cells up to a late period, yet without self-
multiplication, and with marked distinctive features of their
own, which are peculiarly protoplastic.

The facts assert also:

That the peripheral substance of egg and cells is freely proto-
plastic, despite its appearance under less magnification of being
a smooth and stable pellicle :

That the ectosarcal formations of these Metazoan eggs are
possessed of tactile power such as characterizes similar modifi-
cations of the substance in protoplasts :

That considerable deportation of living substance can take
place from egg or cells in a manner invisible to even quite high
powers (as 1500 diameters), and yet to an extent producing a
measurable diminution in size of the mass from which it is
carried :

That, in short, the measure of the living facts cannot be taken
truly with anything less than our best optical resources, and
that even these fail of perfect adequacy :

That apart from the grosser organization of the egg, which
we have had spread out to some extent before us by reagents,
there exists a minuter mechanism which these fail to record, or
even hint to us; and that beneath the interactions of those
grosser masses which seem so complete as seen under lower
powers, there are supplementary and strongly causative inter-
actions of minuter portions of the substance composing them,
which are in part visible with our highest powers and which are
from some points of view curiously contradictory in seeming :

That while the grosser organization acts through more
definite and stable structures, the finer, and heretofore unseen,

No. 2.] SPINNING ACTIVITIES OF PROTOPLASM. 389

organization acts by unstable and freely protoplastic phenomena
and structures :

That while the phenomena pertaining to the organism as
composed of these grosser masses and structures we call cells, are
of vast importance; the phenomena pertaining to the organism
as composed of those minuter and unstable portions which go
to make up the cells, and which taken together throughout the
whole mass, as well as separately, go to make up the protoplasm
per se,— the living substance, as such ; — are of transcending
importance, since they prove to be control phenomena for the
former. In saying this I have in mind the familiar phenomena
of caryokinesis, and certain unfamiliar phenomena of re-arrange-
ments of the cytoplasmic substance which are treated of in the
forthcoming paper, and which undoubtedly strengthen the ex-
pression here of these conclusions.

Here I will limit myself to saying that whatever may be the
significance of the cell wall in the development of these eggs, it
surely cannot be thought a separator, in either a physical or
physiological sense, of the cell contents from other portions of
the common mass.

If we look upon the cell wall as a part of the machinery of
the embryo, the larva, the coming or immediately present
organism, that is, as an organ of the mass of living substance,
just as is the nuclear membrane, or any other local modifi-
cation for physiological purposes, we shall probably be nearer
the truth than if we keep to an earlier conception, and hold
that, for the living substance, cell walls a prison make.

BALTIMORE, October, 1896.

THE ORIGIN OF THE EGG CENTROSOMES.

A. D. MEAD.

Brown University, Provipence, R. I.

THE observations recorded in this paper were made upon
the eggs of the marine annelid Chetopterus pergamentaceus,
procured at the Marine Biological Laboratory at Woods Holl,
Mass., where I was enabled to work during the past summer
through the courtesy of its Director.

My best preparations were obtained by fixing the eggs with
picro-acetic acid, and staining with Heidenhain’s iron-alum
haematoxylin, followed by orange G. The slides were left in
the 4% iron-alum for half an hour, rinsed, and left in %%
hematoxylin for twelve hours. After drawing the color with
iron-alum, the slides were dipped in the aqueous solution of
orange G. Hermann’s fluid, Flemming’s fluid (weaker), and
a mixture of Hermann and formalin also gave satisfactory
results, though the staining was not so brilliant. Sublimate
acetic usually works havoc in the region of the astrosphere.
In a previous paper! I stated that “until the entrance of the
spermatozoon, the egg remains with the first maturation spindle
in the equatorial-plate stage.” The earlier stages were not
seen at’ that time.

During the past summer, however, on taking the precaution
to preserve the ovaries, together with the loose eggs, imme-
diately after dissecting them out into sea-water, I was able to
obtain a complete series of stages previous to the formation of
the first maturation spindle.

The cytoplasm of the smallest ovarian eggs is compact, and
this gives to the eggs when stained an almost uniform dark.
purple color. The increase in size is due in great measure to
the accumulation of yolk, the distribution of which is accom-
panied by noteworthy changes in the appearance of the cyto-
plasm. The yolk is laid down in the form of yellow-staining

1 Some observations on the maturation and fecundation of Chetopterus per-
gamentaceus, Cuvier. Journ. of Morph., vol. X, No. 1.

392 MEAD. [VoL. XII.

granules, which are more numerous at first near the periphery
of the egg. The granules are held in the meshes of a reticu-
lum formed of beaded strands of cytoplasm, whose purple color
is in striking contrast to the yellow yolk.

In most ovarian stages only a part of the cytoplasm presents
the loose reticular appearance ; the rest remains as dark purple*
masses. Occasionally sections show but one of these masses,
crescentic in outline, located near the nucleus. I presume
that this is equivalent to the “ paranucleus”’ or “ yolk-nucleus”’
of certain authors. These masses are not homogeneous, but
resolve themselves into a radially compressed cytoplasmic net-
work, the strands of which are
often clearly visible and are
continuous with the more open
network which contains the
yolk. At first they fill the
larger part of the egg outside
the nucleus ; but as the egg
accumulates yolk the meshes
at the periphery of the masses"
expand into an open reticulum
so as to occupy the increasing
dimensions of the egg (Fig. I).

Fie. I. — Section of ovary. Camera. Disso- Through this process of con-
lution of the paranucleus.

tinuous fraying out, the masses

become thinner and thinner and are finally completely resolved
into the cytoreticulum.

The reticulum can be traced with ease to the extreme peri-
phery of the egg, where in section one can follow a continuous
beaded line entirely around the egg. The nuclear membrane
is evidently a part of the same network. The general appear-
ance of the reticulum varies with the age, the older ovarian
eggs having the nodes the most pronounced. In eggs freed or
about to be freed from the surface of the ovary, the nodes
become less frequent and still more prominent, until at length
a large portion of the reticulum is transformed into a multitude
of small asters (Fig. II). Many of them, however, are by no
means diminutive, particularly those in the vicinity of the

No. 2.] ORIGIN OF THE EGG CENTROSOMES. 393

nucleus. All gradations in size occur, but at a certain stage
many of the larger ones are approximately equal both in size and
distinctness. Frequently one can count from fifteen to twenty
very distinct asters in a single section. These structures cor-
respond closely to “secondary mechanical centers” of Reinke,}
and I will call them the secondary asters. While yet distinct
from one another, they are often so near together that their
rays intercross.

It is not long before two of the asters become predominant,
— primary asters. Their rays increase in number and length,

Fic. II. — Section of egg free from ovary, Fic. III. — Section of egg in a stage later than
showing secondary asters. Camera. Fig. II. Only the two primary centers are
present. Camera.

apparently at the cost of the secondary asters, for the latter
gradually evanesce with the further development of the former,
and at length the cytoplasm possesses only two well-marked
centers of radiation (Fig. III).2 The primary asters lie near
the germinal vesicle, usually about go° apart, but their relative
position is subject to considerable variation in different eggs.
They are destined to be the asters of the first maturation spindle.
Each of them has for its center a perfectly definite deeply
stained centrosome (centriole), surrounded now by a lighter
area from which the rays diverge. The nuclear membrane

1 Reinke: Zellstudien II. Arch. f mikr. Anat. Bd. XL, p. 276. Peritoneal
cells of the larval salamander.

? The initial predominance and further growth of the primary asters certainly
appears to be due in part to the actual coalescence of smaller asters.

394 MEAD.

invaginates in the vicinity of each aster in the manner already
described by several authors.! It is afterwards resolved into
the cytoreticulum, though the huge nucleolus and other features
of the nucleus do not immediately disappear.

Definite chromosomes are seen to be attached to the achro-
matic fibrils of the two asters, and to take their position in the
equatorial plate in the ordinary way. The centrosomes become
doubled and are demonstrable in each aster as two exquisitely
clear dark dots in the midst of the light yellow centrosphere.
The spindle thus formed swings around to occupy a radial
position at the periphery of the egg, and remains in the
metaphase until the sperm enters. On the entrance of the
sperm the karyokinetic activity is resumed and the maturation
processes are completed.2 Usually, before fertilization, the
nucleolus of the germinal vesicle breaks up and disappears ;
the last traces of it are seen in the neighborhood of the
maturation spindle.

The foregoing observations convince me that the asters and
centrosomes in the Chetopterus ovum arise by a modification
of the cytoplasmic reticulum. The phenomena of their origin
and their relation to the secondary asters are similar to those
described by Reinke? in the tissue cells of the larval salamander.

Watasé says he has “seen in the egg of Macrobdella a series of
thirteen asters, ranging from a diminutive aster with a microsome
for its center to the normal aster with a veritable centrosome.” #

The fact that in Chetopterus the secondary and primary
asters were formed when the eggs were transferred from the
body-cavity of the worm into the sea-water, suggests a compar-
ison of these phenomena with the production of ‘artificial
centrospheres’”’ in sea-urchin eggs by adding more salt.®

1 Compare Wheeler: Myzostoma. Journ. of Morph. vol. X, No.1. Griffin:
Thalassema. TZyvans. V. Y. Acad. Sct., June 2, 1896.

2 Mead: Joc. cit.

3 Reinke: Joc. cit.
4 Watasé: Origin of the Centrosome. iol. Lectures, Marine Biological

Laboratory, 1894, p. 285.
5 Morgan: The Production of Artificial Centrospheres. Archiv f. Entwickel-
ungsmechanik, Bd. III, Heft 3.

DEVELOPMENT OF THE HUMAN COELOM.

FRANKLIN P. MALL.

Four years ago I wrote a general article on the coelom for
Wood's Handbook of Medical Science, in which was emphasized
the separation of the body cavity from the extraembryonic
coelom. Since then I have had opportunity to extend my
observation to the human embryo, and therefore make this
communication.

Unfortunately, there are no data regarding the beginning of
the coelom in the human embryo, and in all probability none
will ever be found. The smallest human ovum ever seen is
that described by Reichert.1 It was obtained from a woman
who had committed suicide, on account of pregnancy, forty-one
days after the beginning of the last menstrual period. It was
therefore presumably about thirteen days old. This ovum,
which is pictured in every text-book, was 5.5 X 3.3 mm. in
diameter, was surrounded by a zone of villi leaving two poles
bare, and contained in its interior a mass of cells measuring
1.5 X 1.75 mm. All the space between this inner mass and
the chorion is the coelom, and regarding its origin, we can no
more than speculate.

During the last few years three other human ova, slightly
larger than Reichert’s, have been cut into sections, thus per-
mitting a more careful study of their contents.2- The dimensions
and approximate ages of these embryos are given in the
table on the following page.

It is noticeable that in the three embryos just mentioned, as
well as in the remaining four of the table, the size of the whole
egg does not correspond with the size of the embryo, nor with
its age. I do not think that this great variation in the size of
the chorionic vesicle is altogether due to the method of harden-

1 Reichert: Abhandl. d. kgl. Akad. d. Wiss., Berlin, 1873.

2 Von Spee: His’s Archiv, 1889; Mall: Anatom. Anzeiger, 1893; Johns Hop-
kins Hospital Bulletin, 1893; and von Spee: His’s Archiv, 1896.

396 MALL. Vion.) excl:

TABLE OF YOUNG HUMAN OVA.

TIME BETWEEN

DIAMETER OF First LAPSED

OBSERVER. Empryonic Vesicuz, | DIAMETER OF OVUM. | “periop AND
ABORTION.
Malls(No. XI), 23%: Tes OG mm. TOL 2X 72min 13 days
ReICHEKt - mies seine oa I.75X1.5 se SoS ISSA se TAs
Von Spee (v.H.) . . 1.84X 1.083 “ Gs X, 1455), 5° tap
Von Spee (Gle.) . Za es2s ce TOs eTRO.G <6 T2 mus
Mall (No. XVI). 251) G21 ce iin eSetsy 1 13) uss
His (Lg.) .- 2.15X2 ‘s HS. oh. Gece T2hianes
Von Spee . 2.69 X 3 sé WSs 14 ‘ WAM as
Janosik . 3. x4 - 8. ee ost

* These are all of the authentic young human ova I can collect from the litera-
ture giving all of their measurements as well as the menstrual history of the mother.
In both of von Spee’s cases the time between the abortion and the end of the last
period is given; in embryo v. H. the time is given as “exactly five weeks,” while
in embryo Gle. “five weeks ” is given. If we estimate the duration of menstruation
as five days and its frequency twenty-eight days, then the time between the first
lapsed period and the abortion is twelve days, as I have given it in the table.

ing the specimen. Just at this time the growth of the chorion
is precocious, as is also the case in the dog,! rabbit,? and
monkey.

The papers by Bischoff and by Selenka are worthy of the
most careful study by every embryologist, and I take the liberty
of rearranging some of Bischoff’s data on the development of
the dog. His observations are very extensive, and give us the
basis for our present ideas of the passage of the ovum into the
uterine tube after fertilization. Unfortunately, they were
made before the time of sectioning specimens, yet they are
more complete than most researches relating to this subject
published since his time.

The portion I tabulate relates to the size of the embryonic
mass or vesicle, the size of the ovum, and its approximate age.
As far as I have been able to determine, these data taken from
the dog are still the most important ones with which we can

1 Bischoff: Entwicklungsgeschichte des Hundes Eies, Braunschweig, 1845.

2 Bischoff: Entwicklungsgeschichte des Kaninchen Eies, Braunschweig, 1842.

3 Selenka: Studien iiber Entwicklungsgeschichte des Thiere, Heft 5, Wies-
baden, 1892.

No. 2.| DEVELOPMENT OF THE HUMAN COELOM. 397

compare the human ovum. Embryologists are accustomed to
state that the age of a human ovum is to be reckoned from the
beginning of the first lapsed period, and I think that Bischoff’s
observation upon the size and growth of the dog’s ovum cor-
roborates this view. He found that the ova left the ovary
during the rutting period, but the exact date could never be
determined. Neither did the time of copulation determine the
ovulation. Asa rule, it took twenty-four hours or less after
copulation for the spermatozoa to reach the ovary, and about
the same time is required for the ovum to reach the beginning
of the uterine tube after ovulation. So if ovulation and copu-
lation took place at the same time, fertilization of the ovum
could not take place until twenty-four hours later.

In Bischoff’s tables he often rates the age of an ovum from
the first or from the last copulation, or from the beginning or
from the end of the rutting period. I have attempted to tabulate
his specimens from all four of these dates, but in none of the
attempts did the size of the ova correspond with their respective
dates. Often eggs of a given date were smaller and developed
to a less degree than ova presumably younger. After much
difficulty I finally constructed a table in which the size of the
ovum and its age correspond. A number of the ova published
by Bischoff were obtained from the same animal by removing
half of the uterus at one time and the remaining half the next
day. In each portion a number of ova were found, and they
were usually of about the same stage of development. By this
method of procedure it is possible to determine very accurately
the growth of the ovum from one stage to one twenty-four
hours later. So, by gradually plodding through the specimens
published by Bischoff, it was possible for me to correct his
data completely. It is remarkable, as the table shows, how
slowly the development takes place in the early stages, and
about ten days are required before the ovum is one millimeter
in diameter. On the fifteenth or sixteenth day the ovum is
about as large as the human ovum described by Reichert (see
table).

Similar results can also be obtained from the various papers
published on the rabbit’s embryo, Its development, however,

398 MALL. (Vox. XII.

is considerably more rapid than the dog’s, as the period of
gestation is but thirty days.

Recently Selenka has given some of his results relating to
the development of the monkey. The most valuable specimen
relating to the early development of higher animals was unfor-

TABLE OF AGE AND SIZE OF THE DOG’S OVUM.
(COMPILED FROM BISCHOFF.)

DIAMETER OF
AGE. Deere Os EMBRYONIC STAGE.
f Mass.

I day. 15 mm. I cell.
2 days. sta as 2 cells.
3 “ce 14 “cc 4 “ce

4 “cc 16 “ I2 “c

5 “ .16 “ec 64 “c
(Xo shes Mulberry.
7 “ .20 “ ity

8 “cc .21 “ “c

9 “ 28 “ “
On 3 oeto) et: .07 mm.

meee G. ie 6 juley
2 “ 2. “ 2 “cc
13 “ 3: “ +43 “
14 “ 4. “ 5 “cc
I 5 “ 5. “ee he “
16 “ 5: “ 2. ““
164 6c 6. ““ 3. “

It has been somewhat difficult to compile this table, as Bischoff’s measurements
are all given in Paris lines. My measurements are taken in great part from his
figures, and I think that these are very accurate.

tunately lost, but its age and dimensions are preserved for us,
and are of value in the determination of the age of human ova.
The ovum came from a monkey kept in confinement which was
killed eight days after copulation. If we estimate one or two
days required before fertilization, this ovum cannot be over
seven days old. This suggests that the early stage of this
variety of monkey is developed more rapidly than that of the
dog.

o

No. 2.] DEVELOPMENT OF THE HUMAN COELOM. 399

DEVELOPMENT OF THE MONKEY. (FROM SELENKA.)

DIAMETER OF

DIAMETER OF RTERTONIC

Ovens VESICLE.
Semnoplthecussmaurus, ~~) 312 = 1. i. s I.5 mm. -3 mm.1
Semnopithecus pruinosus ....... copy 35
@ercocebusjcynomolgus) 5+. 5. 2 igo ay et
@ereocebusicynomolgus, 2 2 8 os 3. ss Osman Ziq kes

The pictures Selenka gives indicate that the development of a monkey’s
ovum is identical with that of the human ovum. At any rate, the few specimens
Selenka publishes give results which are equal to the great number of specimens
of human ova which have been published. This only indicates that many of the
interesting problems relating to early human development will probably be solved
by the study of the monkey’s ovum. There is but little doubt now that young
monkeys’ ova will soon be obtained for study.

MATERIAL EMPLOYED.

During the last few years I have obtained a number of young
human embryos from physicians in different portions of the
United States, and to them I am under all obligation for the
present study as well as for some others which are to follow.
Nearly all of the specimens which I give in a table are well
preserved, and a number of them are preserved excellently.
All of the specimens were stained in alum carmine, and with
the exception of Nos. XVII, XLIII, and LVII were cut trans-
versely. These three were cut in sagittal sections.

All of the specimens were hardened in alcohol, the value of
which method I have repeatedly emphasized to my friends, and
do continue to emphasize to those who may preserve specimens
for my use in the future.?

1 Not an embryonic vesicle, but only a disc.

2 Neurenteric canal present.

8 Embryologists usually recommend that human embryos should be hardened
by placing them in dilute alcohol and then gradually increasing the strength of the
alcohol. It has been my experience that by this treatment the specimen is injured
by maceration due to the weak alcohol. A few years ago I emphasized the fact
that the whole ovum should be placed in a large quantity of strong alcohol as soon
as possible. It should be handled as little as possible before hardening it, thus

preventing mechanical injury. By leaving the ovum closed the alcohol must first
penetrate the chorionic and amniotic fluids before it reaches the embryo, and thus,

400 MALL. [Vou. XII.

Nearly all of the embryos were drawn or photographed to
scale and then carefully cut into sections from ten p to fifty
thick. I find that for purposes of reconstruction it is a mistake
to cut the sections very thin. Yet in small specimens, as Nos.
XI and XII, the specimens were cut thin to permit of careful

LIST OF EMBRYOS STUDIED.

LENGTH IN MILLIMETERS.
No. LSS FROM WHOM OBTAINED.
Vv. Bt N. B.

XI — — Dr. Kittredge, Nashua, N. H.
XII PA _ Dr. Ellis, Elkton, Md.
Ill 252 — Prof. His, Leipzig, Germany.
XIX 5:5 4.5 Dr. Williams, Baltimore, Md.
SVT Cie Fe Dr. Douglas, Nashville, Tenn.
II 3 Te Dr. C. O. Miller, Baltimore, Md.
IV — oh Dr. Williams, Baltimore, Md.
XLII 15. 13. Dr. Booker, Baltimore, Md.
VIII 7s 14. Dr. Ritter, Brooklyn, N. Y.
IX 17. 14. Dr. Eycleshymer, Chicago, II.
Vv 18.5 17 Dr. Kittredge, Nashua, N. H.
XLII 18. 15. Dr. Wills, Los Angeles, Cal.
XVII 18. 16. Dr. Cottrell, Louisville, Ky.
XXVIII 19. 18. Dr. Sewall, Denver, Col.
VII 19.5 18. Dr. Booker, Baltimore, Md.
XXII 20. 18. Dr. Snively, Waynesboro, Penn.
VE 22% 20. Dr. Howard, Cleveland, Ohio.
VI 24. — Dr. C. O. Miller, Baltimore, Md.
x 24. 20. Dr. W. S. Miller, Madison, Wis.
X<OEV 28. 15) Dr. Douglas, Nashville, Tenn.
XXXIV 80. 60. Dr. Ellis, Elkton, Md.
XLVIII 130. 110. Dr. Wilson, Worcester, Mass.

without placing the embryo first in weak alcohol, it naturally passes through the
successive dilutions of alcohol before it is completely hardened.

It is very injurious to these delicate specimens to be wrapped in cotton before
they are sent by mail or express. A perfect method is to place the preserved
specimen in a bottle filled completely with alcohol, thus imitating the condition
of a foetus in utero. If there is no air or cotton in the bottle containing the
embryo it is almost impossible to injure the embryo by shaking it.

Since I have emphasized this method of preservation (Johns Hopkins Hospital
Bulletin, 1893), I have obtained a number of specimens excellent in every respect.
These specimens are not distorted, nor macerated, nor shrunken.

1 V. B. and N. B. indicate the length of the embryo measured from the vertex
to the breech and from the nape of the neck to the breech, respectively.

No. 2.] DEVELOPMENT OF THE HUMAN COELOM. 401

cytogenic studies also. In most of the specimens photographs
or an additional series of sections were made of the chorion and
amnion in order to study the variation of these structures.

Embryos XI,! XII, and II? were completely reconstructed
in wax by the method of Born. Nos. IX, VI, and X were
reconstructed in part by Born’s method and finished by His’s
method of reconstruction. The abdominal viscera of Nos. VI,
IX, X, XXXIV, XLV, and XLVIII were modeled by Born’s
method.

The mechanical portion of reconstruction has been simplified
to a great extent by a special apparatus used in the Anatomical
Laboratory,? which enables us to employ a modeler. The sec-
tions are projected upon a screen, to which the wax plate is
attached. By working in a dark room with this apparatus it is
easy to direct a modeler to draw the outlines accurately. He
can then cut them out, and all that remains to be done is to
pile the pieces and then blend them.

THE CoELOM IN YOUNG OVA.

All of the young human ova which have been described con-
tain within them a cavity, lined with mesoderm ; this is the
coelom, bounded by the somatopleure on the outside and by the
splanchnopleure on the inside. This arrangement, as shown by
a number of diagrams by recent authors, is very unlike the
appearance of the blastodermic membranes of many of the
lower mammals, and it is necessary therefore that we should
revise our conception of the formation of the amnion in the
human embryo.

The ova recently published by Graf Spee indicate that the
amnion must be formed very early, and, since it is completed
before the medullary grooves begin, we must admit now that
it is formed much the same as it is in many rodents, z.e., by
apparent inversion of the membrane. When Bischoff* first
described inversion of the membrane in guinea pigs it seemed

1 Mall: Anatom. Anz.
2 Mall: Journ. of Morph., vol. V.

8 Hoen: Johns Hopkins Hospital Bulletin, 1896.
# Bischoff: Entwickl. d. Meerschweinschens, Giessen, 7852.

402 MALL. [Vou. XII.

like a paradox, but, since the comparative methods of study
have been introduced, inversion only means that the amnion
is completed before the medullary groove begins to form.
This alteration of the development of the amnion and the
medullary groove makes the body of the embryo develop on a
concave surface instead of on a convex one, thus apparently
making the embryo inverted, as is the case in the guinea pig.

Closely associated with inversion of the blastodermic mem-
brane is the formation of an additional layer of cells, discovered
by Rauber,! the importance of which has been emphasized by
Selenka and others. Rauber’s layer is so marked in the rabbit
that it was at first believed to be the true ectoderm. The fate of
Rauber’s layer has not been sufficiently studied to interpret it
completely, and our ideas regarding it will not improbably
require some revision. In many rodents Rauber’s layer becomes
markedly thickened on one side of the ovum, forming a support,
or 7rager, for the ovum. The relation of Rauber’s layer to
the Trager is shown beautifully by Selenka? on Plate XVI of
his monograph.

The question which interests us here is whether the inver-
sion of the blastodermic membrane as well as the discovery of
Rauber’s layer aids us in advancing a theory of the develop-
ment of the germ layers of the human embryo, and thus in
turn to explain the large coelom as found in all of the earliest
human ova. I realize fully that any such effort will not be
final, yet I believe that it will aid us to understand better the
relation of the membranes as found in the human ovum.

In looking over the illustrations of the development of ani-
mals closely related to man, one is struck with the similarity of
the arrangement of the membranes to those described for the
human ovum by Graf Spee. One must compare only plates
XXXV-XXXVIII of Selenka’s® paper with the two plates
accompanying Graf Spee’s* to be convinced that the early
development of monkeys is almost identical with that of man.
Yet Selenka’s researches on monkeys do not help usa great

1 Rauber: Sitzungsber. d. Naturforcher Gesellsch., Leipzig, 1875.

2 Selenka: Studien iiber Entwickl. d. Thiere, Heft 3, 1884.

3 Selenka: Studien, efc., Heft 5, Erste u. Zweite Halfte, 1891, 1892.
4 His’s Archiv, 1889 and 1896.

No. 2.] DEVELOPMENT OF THE HUMAN COELOM. 403

deal ; they only show us that the monkey’s development is like
that of man. In monkeys we have also the precocious chorion
and the early amnion and the large coelom between the
umbilical vesicle and the chorion. The marked difference is
that the amnion is attached to the chorion along its dorsal side,
while in the human embryo this is only the case along the pos-
terior end of the amnion. The attachment of the amnion
along the chorion suggests that the embryonic plate separated
from the exterior of the ovum along this point, as Selenka
thinks he observed in a very young ovum only 1.5 mm. in
diameter. Unfortunately, the most valuable specimen was
injured in its preparation,! and Selenka did not trust himself
to give any illustrations of it.

With the amnion attached at its dorsal end to the chorion,
we understand why the entodermal end of the allantois must
grow around an angle to reach the chorion (Selenka, Plate
XXXVII, Fig. 5). Somewhat the same arrangement has been
described by Graf Spee? in his embryo Gle., but the curve is
by no means as marked, indicating that the attachment of the
embryo to the chorion is along its posterior end, as shown
by His? in his well-known diagram of the formation of the
amnion.

Regarding the very early stages of monkeys and man it is
better that we make comparisons with animals most nearly
related to them, and now we have careful studies of the very
early stages of Chiroptera at our disposal. I believe that
Selenka’s* study of the development of Pteropus edulis gives
us the key for the comparison of the formation of the blasto-
dermic membranes in mammals. Recent investigations by
Duval® on different families of Chiroptera appear to confirm
the work of Selenka on Pteropus.

In order to illustrate these points more clearly I have made
diagrams of three of Selenka’s figures of Pteropus. Fig. 1 is
from an ovum covered completely with two layers of cells,

1 Selenka: Studien, 1891, p. 201.
2 Graf Spee : His’s Archiv, 1896, Taf. I, Fig. 1.
8 His: Anat. mensch. Embryonen, Theil I, p. 171.

£ Selenka: Studien, 1892, p. 2009.
5 Duval: Jour. de l’Anatomie et de la Physiologie, 1895.

404 MALL. [Vo.. XII.

between which at one pole of the egg there is a mass of scat-
tered cells destined to become the permanent ectoderm. The
outer layer of cells has a tendency to grow into the form of
villi over the embryonic disc, while on the opposite side of the

P

Fic. 1. Fic. 2.

yor

efoe e
e ° °
SOL we eeees

Fic. 3.

Fias. 1-3. — Diagrams of the Development of Pteropus Edulis, after Selenka. Fig. 1 is Selenka’s
Fig. 2; Fig. 2, Selenka’s Fig. 5; Fig. 3, Selenka’s Fig.9. &, Rauber’s layer; P, placenta;
ec, ectoderm; ez, entoderm ; c#, chorion; az, amnion; # v, umbilical vesicle; #zes, meso-
derm ; coe, coelom; ad, allantois, with the arrow indicating the direction of its future
development.

egg it is composed of but a single layer of cells. Since this
outer layer remains well separated from the body of the
embryo throughout its development, and since it holds the same
position to the egg that Rauber’s layer does in the rodents, I

No. 2.] DEVELOPMENT OF THE HUMAN COELOM. 405

believe it to be identical with Rauber’s layer, and shall speak
of it as such. According to Duval this Rauber’s layer disap-
pears over the embryonic disc in the Chiroptera much as in
the development of the rabbit and the field mouse. This does
not necessarily contradict Selenka’s observations on Pteropus,
for the house mouse begins to develop like the field mouse,
but continues during the early stages in the same manner as
Pteropus does.

In the next stage the ectoderm has been converted into a
hollow mass of cells, Fig. 2, rather by a process of absorption
than by an invagination, as I have expressed it in the diagram.
The entoderm lines the whole interior of the egg, and surrounds
the ectoderm of the amniotic cavity. The ectoderm of the
exterior of the egg, Rauber’s layer, is again thickened over the
embryonic mass to form the placenta, as Selenka calls it, or
the 7réger, if we were discussing rodent embryology.

In the next stage, as expressed in Fig. 3, the mesoderm is
beginning to form, and has extended completely over the
amnion and partly over the umbilical vesicle. The entoderm
has retracted itself and touches the ectoderm ; only the chorda
dorsalis is yet to form. Between the amnion and the placenta,
or the Zager portion of Rauber’s layer, there is a marked
space, and the mesoderm does not come in contact with it. The
allantois grows as a bag into this space and attaches itself to
the thickened part of the ectoderm, as shown by Géhre! in his
figures. In the figure 3 accompanying Gohre’s paper he shows
the vesicular allantois attached to the support of the chorion
(black portion of my Fig. 3) leaving on either side of the embryo
acoelom. The allantois carries the mesoderm and vessels to
the villi of the chorion, and these in turn are imbedded in the
decidua of the uterus. In so doing the ectoderm of the chorion
receives a second layer of epithelium, and I believe that this
must account for the two layers of epithelium we have on the
chorionic villi of the human ovum. There has been much
written on the subject of the double layer of epithelial cells
of the human chorion, and I think that a glance at Godhre’s
figures 3 and 4, on Pteropus, as well as at Selenka’s figures 11

1 Gohre: Selenka’s Studien, etc., 1892, p. 218.

406 MALL. [Vou. XII.

and 12 (Plate XX XV) and figure 6 (Plate XX XVII) on monkeys,
will decide this question more definitely than all the many
discussions on the human chorion put together have done.

Having now selected from Selenka diagrams and descriptions
of the development of the germ layers of Pteropus, it is easier
for me to give a plausible explanation of the beginning of the
coelom in the human embryo. If the diagram I have given
in Fig. 3 is compared with Selenka’s figures 5 and 11 (Plate
XX XV) and figure 5 (Plate XX XVII) of the monkey, as well
as with the sections of young human ova published by Graf
Spee! and by myself,? one is struck with the great similarity
of the two groups of figures.

Fig. 14, given further on, is a diagrammatic outline of a
longitudinal section of a young human embryo published re-
cently by Graf Spee. It is the one marked v. H. in the table
of young human ova given in the beginning of this paper.
When, now, this section is compared with the transverse sec-
tion of Pteropus, in Fig. 3, the only marked difference is that
the umbilical vesicle in Pteropus has retracted, in order to make
the arrangement of the membranes as given for the human
embryo in Fig. 14.

In order to make the connection complete, I give hypotheti-
cal stages in Figs. 4, 5, and 6. Fig. 4 represents the human
ovum in the two-layer stage. The outer layer, or Rauber’s
layer, is complete as in the rodents and in Pteropus. The
inner layer, or entoderm, is also complete. Between the two is
the embryonic shield, or ectoderm of the future embryo. The
next figure, 5, shows the beginning of the mesoderm develop-
ing towards the tail end of the embryo, as this is the position
of the primitive streak, and as the head fold of the amnion
in many embryos is often only invested with ectoderm and
entoderm. A stage later, Fig. 6, finds the mesoderm envelop-
ing the umbilical vesicle completely, and also partly lining the
outer layer, R, of the ovum. The cavity between the two is the
coelom. At the tail end of the embryonic disc the mesoderm

1 Graf Spee: His’s Archiv, 1889 and 1896.

2 Mall: A Human Embryo of the Second Week, Anatom. Anz., Bd. 8, and

Early Human Embryos and the Mode of their Preservation, Johns Hopkins Hos-
pital Bulletin, 1893.

No. 2.]| DEVELOPMENT OF THE HUMAN COELOM. 407

of the somatopleure and splanchnopleure are still united, and
mark the place of the formation of the rudimentary allantois.

Having carried the development of the human ovum to this
stage by means of hypothetical stages, based upon the develop-
ment of Pteropus, I can now continue the description of the
development based upon observation.

Abnormal Ova. — Teratologists are accustomed to view a
group of abnormal states as arrested development, and in
recent years a number of abnormal human ova have been
studied by His,1 by Giacomini,? and others. Frequently in the

ee

Fie. 5. Fic. 6.

Fics. 4-6. — Hypothetical Stages of the Early Development of the Human Ovum. A&, Rauber’s
layer ; ec, ectoderm ; ex, entoderm ; »zes, mesoderm; # v, umbilical vesicle; coe, coelom;
all, position of allantois.

development of an ovum the embryo is destroyed completely,
or, according to Giacomini, may wander out of the ovum.
In these cases the ova are aborted. Frequently, however, a
portion of the embryo is not developed, or it dies and the
remaining portion develops for a time, and then the ovum is
aborted. I have now in my collection a beautiful example of
an ovum of apparently normal structure, the interior of which
is lined completely with an amnion, and in place of an embryo

1 His: Anatomie mensch. Embryonen, Heft 2, 1882, and Internationale Beitrige
zur wissenschaftlichen Medecin, Bd. 1, 1891.

2 Giacomini: Ergebnisse der Anatomie und Entwicklungsgeschichte, Bd. 4,
1895. The original papers of Giacomini are in the Archives Italiennes de Biologie,
vols. X VITI-XXII.

408 MALL. [Vor. XII.

there is only an umbilical cord. The ovum was aborted fifty-
four days after the first lapsed period, and was 30 mm. in
diameter. The cord was 2 mm. in diameter and 9 mm. long.
Its embryonic end seemed to be cut off abruptly, and was
covered with a small mass of round cells. I give this example
only to show that the embryo may be entirely wanting with a
perfect cord and membranes.

A large per cent of young ova which come into the embry-
ologist’s hands are abnormal. According to Professor His’s

TABLE OF ABNORMAL OVA.

DIAMETER OF

No. DIANE TEe OF AC Ue FROM WHOM OBTAINED.
XIII 8 mm. 1.4 mm. Prof. His’s No. XLIV, Leipzig.
XIV 30, << Babes Dr. Friedenwald, Baltimore, Md.
KX LG ae 2, «6 Dr. Williams, Baltimore, Md.
XXI 0} eG Dr. Cullen, London, Canada.
XXXVII 2 bine 2st Dr. Gould, Philadelphia, Pa.
LVIII oy GG Dr. Howard, Cleveland, Ohio.
XXXII sfoy Z5cO) is Dr. Booker, Baltimore, Md.
XXIV 20m — Dr. Miller, Baltimore, Md.
XXIX Bom — Dr. Booker, Baltimore, Md.
ION gon < — Dr. Watson, Baltimore, Md.

It does not necessarily follow that these embryos are all less than six weeks
old, for the menstrual history of the mother indicates that some of them must be
considerably older. This is one source of error in obtaining the high per cent of
abnormal ova among young embryos. The statistics will not be accurate until the
menstrual history accompanies the measurements.

experience over half of the ova less than three weeks old are
abnormal, while of those of four and five weeks one quarter are
abnormal. In my collection there are ten abnormal ova among
twenty-six ova which are less than six weeks old. Of these
ten specimens three contained no embryos at all, one (No.
XXXVII) contained the cord only, and six were of the
nodular form. Of this group of six, three contained a double
vesicle with a kind of fibrous capsule, to a great extent similar
to the mesoderm of the chorion. One of these is His’s Embryo
XLIV, which is frequently described in the books as a normal

NO: 2.| DEVELOPMENT OF THE HUMAN COELOM. 409

specimen, but which unfortunately is an abnormal one. My
interpretation of these three specimens (Nos. XIII, XIV, and
XX) is that the fibrous degeneration overtook the embryonic
vesicle after it had reached the stage of Graf Spee’s embryo
v. H., my Fig. 14. The remaining three embryos (Nos. XXI,
XX XVII, and LVIII) are of the vesicular form, and I believe
them to be especially valuable for the proper interpretation of
the early stages of development of the human coelom.

Nos. XXI and LVIII came to me as perfect specimens,
both having been hardened unopened, the first in strong
formalin and the second in strong alcohol. No. XXI was still
enclosed in its decidua,
and appeared to be a
normal specimen until it
had been cut into serial
sections. The embry-
onic vesicle proved to
be very large, and was
composed throughout of
two layers, an inner one

giving all the appearance
of the entoderm and an fig. 7. — Diagram of a Pathological Ovum which rep-
outer giving all the ap-
pearance of the mesoderm of the umbilical vesicle of young
embryos. The mesodermal layer contained within it islands
of blood cells, as are also present in normal specimens. The
whole vesicle was connected to the chorion with a mass of
mesodermal cells somewhat as shown in the diagrammatic
Fig. 7. The chorion and decidua appeared to be normal.

No. LVIII showed considerable change in the mesoderm of
the vesicle and chorion, giving somewhat the appearance of
fibroid degeneration rich in cells. The chorion was attached
to the vesicle by a strong pedicle, as shown in Fig. 7. The
vesicle itself was composed of two layers, an inner and contin-
uous one composed of one layer of cells, and an outer and
thickened layer appearing like the mesoderm of the chorion.
There were no indications of blood islands. In addition to
these two layers there was a third layer fairly well marked near

resents an Early Hypothetical Stage.

410 MALL. [VoL. XII.

the pedicle and between the vesicle and the chorion. With the
exception of the allantois canal, Fig. 7 is a diagram of this
specimen. No. XXXVII is much like No. XXI, but it did
not stain well as the specimen was a number of years old
when cut.

Giacomini! has described a number of similar vesicles, and
he expressly states that the vesicles had the structure of the
umbilical vesicle, but that
there was no trace of the
amnion present in any of
them. A number of other
vesicular forms have been
described, and in general
they all appear much like
the two specimens I have
given.

I do not think that it
is rash to assert that
these vesicles represent
an arrested development
of an earlier stage, which,
due to impaired nutrition,
or whatever it might have
been, simply allowed the
Pi, 6 Dlannade Senie ot Ballots Host embryonic vesicleto keep

drawn only on the upper side. ec, ectoderm; ex, On expanding. That this

entoderm ; wes, mesoderm ; 2 v, umbilical vesicle ;
coe, coelom; al, allantois ; a, amnion.

expansion can keep on is
already shown in the sim-
ple enlargement of the chorion after the embryo is distorted
or wanting altogether. We have in these specimens a thin
chorion with atrophic villi, and why can we not have an
expanded and atrophic embryonic vesicle if its development is
impaired? In this way I view specimen No. LVIII. It
represents a much earlier stage, which has simply expanded and
was ultimately aborted. In No. LVIII the embryonic vesicle
must have ceased its further development a week or so before
the abortion, about the time the coelom was beginning to

1 Giacomini: Ergebnisse d. Anat. u. Entwicklungsgesch., Bd. 4, S. 636.

No. 2.] DEVELOPMENT OF THE HUMAN COELOM. ATI

develop. At that time the fibrous degeneration enclosed the
embryonic vesicle as well as extended around the whole chorion
into all of its villi. This, then, arrested the further development
of the embryo, and the embryonic vesicle simply continued to
expand.

This idea is further strengthened by another ovum whose
history I published on several occasions three years ago.
The specimen is a good one, having been preserved fairly well,
and it has every indication of being normal. Since the speci-
men has been in my hands I have studied it over and over
again, have photographed many of the sections, and have
reconstructed it. At first it was very difficult for me to
interpret it, but finally it appears to me that something defi-
nite can be said regarding the arrangement of the membranes
and their relation to older as well as to the pathological and
presumably younger specimens.

Embryo No. XI, —‘*The woman, from whom the specimen
was obtained, is twenty-five years old, menstruates regularly
every four weeks, the periods lasting from four to five days.
She gave birth to a child Sept. 19, 1892, and had the first
recurrence of menstruation December 19. The second period
followed on January 25, and was very profuse; it lasted until
February 1. The next period should have begun about Feb-
ruary 22, but on account of its lapsing the patient concluded
that she was pregnant, and called at my office a few days later.
I did not examine her, but asked her to remain quiet and await
developments, as I thought possible that she might be preg-
nant. On the evening of March 1 she fell and sprained herself,
and during the same night had a scanty flow. The flow
recurred each day, and on the 7th of March she passed the
ovum. It was kept in a cool, moist cloth for twenty hours,
and when it came into my hands was at once placed in a large
quantity of 60% alcohol.” 2

The ovum is very large for its age, having a long diameter of
IO mm. and a short diameter of 7 mm. It is covered with villi
only around its greatest circumference, having two spots with-

1 Mall: Anatom. Anz., 1893, and Johns Hopkins Hospital Bulletin, 1893.
2 Letter from Dr. Kittredge, April 27, 1893.

412 WAIL: [VoL. XII.

out villi, as was the case with Reichert’s ovum. The villi of
the chorion are from 0.5 to 0.7 mm. long and are branched.

Upon opening the chorion it was found that the germinal
vesicle was situated just opposite the edge of the zone of villi.
About it was much coagulated albumen, which I did not remove,
and therefore could not obtain good camera drawings. The
portion of the chorion to which the vesicle was attached was
cut out and stained with alum cochineal and cleared in oil, but
even after this treatment it was impossible to obtain any clear
picture. The specimen was next imbedded in paraffin and cut
into “sections 10 thick. The {series proved to be pertect.
From the sections a reconstruction was made in wax, and the
accompanying Fig. 8 is a sagittal section of it.

The dimensions of the different portions of the vesicle are
as follows:

Diameter,of Stem .5) 00) ee ees eee eo 4am.
IEF OO Guiel. ofa 4 gid Joe 6 Sg! on! thd (Oye (&
BGesoVHd sy, CVEAVEHOLES 958 of og) a M6 oo) Bg oo ot Go) a
AVitelhwOyeaele 5b ol Holo wm a o op ov o Wo) w
isengethyofinvapinationy We) mee rnin i) tes wee rel Oc Ome
Widthyofnva cin atlony ames. mn ins i-tal ie ceatO-5ianse
Diameter of opening of invagination. . . . . 0.03 “

The sections and reconstruction show that the embryonic
vesicle is attached to the chorion by means of a stem (Lauch-
stiel). The vesicle itself is composed of two layers, between
which, at a distance from the stem, there are indications of
blood-vessels in the middle embryonic layer. Just beside the
attachment of the vesicle to its stem there is a deep, narrow
invagination of all layers of the vesicle. The walls of the
invagination are somewhat thicker than those of the surrounding
vesicle. The accompanying figures give the arrangement of
the embryonic layers in different portions of the vesicle. The
invagination is in no respect artificial, as suggested by Graf
Spee, as the curves are all sharp, and the layers of mesoderm
and ectoderm are very definitely outlined. The ectoderm has
the sharp contour of the ectoderm of other young embryos pub-
lished, and gives the pictures which are familiar to all embry-
ologists. The entoderm does not extend all around the sections
as I have pictured it, but has fallen off at some points, and

No. 2.] DEVELOPMENT OF THE HUMAN COELOM. 413

this explains why the figures here given do not correspond
exactly with those in previous publications. There cannot be
any doubt about my interpretation of the arrangement of the
embryonic layers in this specimen, nor do I think that it is
abnormal. Yet this last point will be decided in the near
future, I believe, and therefore it may be dropped until other
young embryos are described.

Within the stem of the vesicle there is a sharply defined
allantois, which communicates with the cavity of the vesicle

Fic, 12. Fic. 13.

Fias. 9-13. — Sections No. 43, 53, 68, 80, and 89 through the Embryonic Vesicle of Embryo No. XI.
Enlarged 33 times. The entoderm is a heavy line, the ectoderm is striated, and the mesoderm
dotted. a, amnion; X, cavity of the umbilical vesicle extending into the stem of the vesicle ;
&, Rauber’s layer as the ectoderm of the chorion.

just below the invagination of the ectoderm. The cavity of
the vesicle extends into the stem at a lower point, and it is
this invagination which Graf Spee believes to be the amnion.
Gladly would I agree with him were it not that there is no
evidence whatever of the presence of ectoderm at this point.
Throughout this invagination into the stem there are only two
thin layers of cells, one of which runs over into the mesoderm
of the chorion and the other into the entoderm of the vesicle.
Yet in this invagination the entoderm is not detected at any
point.

A414 MALL. iVor. Xi.

The ectodermic plate in the large invagination of the amnion
is very broad, but not of equal thickness throughout its extent,
and it ends very abruptly beyond the opening upon the surface.
As the opening approaches the stem, the cells of the ectoderm
are continued somewhat along its surface, as indicated by the
black line in Fig. 8.

All of the space between the embryonic vesicle and the
chorion is the coelom, and in this specimen it communicates
with the amnion. Whether this is transient or unusual can-
not, of course, be stated. Should further experience show that
the amnion is closed at an earlier stage than indicated in this
specimen, it would not materially affect my diagrams or obser-
vations. Graf Spee’s recent observation, Fig. 14, makes him
think that this is the case, but it is just as easy to interpret
the formation of the amnion in Fig. 14 from that in Fig. 8 as
by his theory.

The next stages in the development of the embryonic vesicle
are taken from Graf Spee, and they are of importance to eluci-
date the changes which take place preparatory to the forma-
tion of the body cavity. In Fig. 14, which represents the
younger embryo, the amnion is still surrounded completely
with mesoderm, as in embryo No. XI, represented in Fig. 8.
The mesoderm crosses the median line, as the sections given
by Graf Spee! show. The dorsal side of the amnion is cov-
ered with a very thick layer of mesoderm, as the closure of the
amnion in embryo No. XI would suggest.

From the stage represented in Fig. 14 it is easy to pass to
the older embryo represented in Fig. 15. Now the body of
the embryo is well marked, the neural folds are just beginning,
and the neurenteric canal has just been formed. The chorda
dorsalis is not yet separated from the entoderm, and the blood
islands encircle completely the umbilical vesicle and have nearly
reached the head end of the body of the embryo preparatory to
the formation of the heart.

It is not very difficult to imagine the embryonic vesicle of
Fig. 8 to be converted into the vesicle of Fig. 14. To be sure,
the invagination in Fig. 8 seems to be much larger than neces-

1 Von Spee: His’s Archiv, 1896, Plate I, Figs. 4, 5, 9, and Io,

No. 2.] DEVELOPMENT OF THE HUMAN COELOM. 4I5

sary, but variations of this kind are frequently encountered in
the study of embryology. In the diagrammatic outline of von
Spee’s embryo v. H. (Fig. 14), I have emphasized the variation
of the thickness of the ectoderm lining the amnion to corre-
spond with von Spee’s Figs. 7, 8, 9 in his recent publication.

Fic. 14. Fic. 15.

Fics. 14 and 15.— Longitudinal Sections of Two Young Human Ova, after Graf Spee. Enlarged 10
times. Fig. 14, Embryo vy. H.; Fig. 15, Embryo Gle. Just half of the chorion is drawn,
and the villi are outlined only over a portion of the ovum. #&, Rauber’s layer; a, amniotic
cavity ; «v, umbilical vesicle ; ez, entoderm; es, mesoderm; a//, allantois; c, chorda; # c,
neurentic canal; 4, position of heart.

His longitudinal section, from which my figure is taken, does
not emphasize this point, which I consider of importance in
this discussion.

In these two ova described by von Spee, the coelom is much
of the same form it was in embryo No. XI, Fig. 8, and there-
fore needs no special comment. Yet around the head end of
embryo Gle. there is a marked accumulation of mesoderm into
which the heart is to grow. In the illustrations of the section
of this embryo Graf Spee! pictures spaces in the mesoderm
which he believes to be portions of the body cavity of the

1 Graf Spee: His’s Archiv, 1889.

416 MALL. [Vot. XII.

embryo, that is, the cavity of the muscle plates, pericardial
cavity or peritoneal cavity. It is impossible to determine defi-
nitely which portion of the body cavity these spaces represent,
but I do not feel inclined to believe that what he marks peri-
cardial cavity in Fig. 23 can possibly represent it, for we are
to look for the pericardial cavity between the junction of the
pharynx and umbilical vesicle and the head end of the embryo.
This portion of the embryo is marked H in my Fig. 15, and
falls ‘anterior to.von’ Spee 's Hig. 16. Ven Spees Figs 16us
the 24th section of the embryo, beginning at the head, while
his Fig. 23 is the 81st section.

The various small spaces in different portions of the meso-
derm cannot be viewed as the real origin of the body cavities
without further discussion. In the von Spee embryo v. H. there
are indications already of small spaces in the mesoderm at the
border of the ectoderm of the embryo. Similar spaces are
described by Bonnet? for the sheep and by Selenka? for the
monkey. While von Spee and Bonnet believe that these
spaces belong to the coelom, Selenka simply designates them
heart, or vascular.

The blood-vessels are intimately associated with the coelom
in their early development, and it is easy to be led into error
without an abundance of material. Drasch® recently has
again emphasized this relation. He has shown in the chick
that the blood islands are separated from one another by a
number of closed spaces filled only with a fluid. These spaces
soon flow together to form the large slit-like coelom of birds.
The same condition of things has been shown to be true, but
from a very different method, by Budge. He injected the
blastoderm of the chick, and showed that the coelom was com-
posed of a network of spaces, which gradually flowed together
into the large coelom surrounding the embryo.

Of course in the young human embryos we have at our dis-
posal this stage of the process has long passed, but there is no
reason why a remnant of it should not exist at the point of

1 Bonnet: His’s Archiv, 1889.
2 Selenka: Studien, etc., Taf. XX XVIII, Fig. 35.

8 Drasch: Anatom. Anz., Bd. 9.
* Budge: His’s Archiv, 1887.

No. 2.] DEVELOPMENT OF THE HUMAN COELOM. 417

union of the umbilical vesicle with the body. The reason I
question von Spee’s interpretation of these small spaces in the
mesoderm in embryo Gle. is that I believe that all, or certainly
nearly all, of the body cavity is formed by an incorporation of
the extra-embryonic coelom within the embryo. What I have
observed in human embryos as well as in the injected speci-
mens of Budge shows that this must be true. These small
spaces in the mesoderm of the body may belong to the muscle
plates and the early blood-vessels, and certainly cannot play any
great part in the development of the body cavity. There is no
doubt whatever that the whole peritoneal cavity is simply
pinched off from the coelom of the outside of the body and
it is highly probable that the pericardial cavity and pleural
cavities are formed in the same way. The anterior mesen-
tery of the intestine has never existed in the human em-
bryo, and it is therefore needless to explain its mode of
disappearance.

My statements are based in great part on embryos Nos. III
and XII, and since No. XII is such a perfect specimen it is
well for me to describe it in greater detail. The embryo is
about the same age as Kollmann’s! embryo Bulle, which unfor-
tunately was never fully published. No. III is an embryo
given me by Professor His. This embryo had been torn from
the umbilical vesicle, and was injured in different portions of
the body. Yet the head end of it is fairly well preserved, and
it is of value in determining the growth of the body walls
covering the heart.

Embryo 2.1 mm. long. — The history of embryo No. XII is
as follows. ‘The woman from whom the ovum was obtained is
twenty-three years of age and has been married for three years.
She is avery intelligent woman, and her statements are reliable.
Her menstrual periods recur every thirty days. She had been
married some time before she became pregnant, and after pass-
ing two periods aborted July 6, 1893. She was unwell the 5th of
October and again on the 7th of November, this last period

1 Kollman: His’s Archiv, Supplement Bd., 1889, Plate V, Figs. 1 and 2; von
Lenhossék: His’s Archiv, 1891, Plate I; Kollman: His’s Archiv, 1891, Plate III,
Fig. 3.

418 MALL. [Vou. XII.

lasting five days. She passed her next period and on December
18th aborted the ovum.” }

The ovum was hardened in strong alcohol without opening
it first, and when it came into my hands its dimensions were 18

a
Om

win og

Fic. 16. — Outline Drawing of a Sagittal Section of the Model of Embryo No. XII. Enlarged 50
times. The heavy line is the aorta. The muscle plates are numbered for occipital, cervical, and
dorsal regions, respectively. ‘The mesoderm is striated. am, amnion; a, border between
fore-brain and mid-brain ; x and x’, extent of closure of spinal canal; S, Seessel’s pocket;
ch, chorda; 4' and 4”, first and second branchial pockets; 0 v, otic vesicle ; #z, mouth; 7,
thyroid; /7, pericardial space; £2, pharynx; ez, entoderm ; S 7,septum transversum; J, liver;
mc, neurenteric canal; ad/, allantois.

X 18 X 8 mm., that is, it was slightly flattened. It was com-
pletely covered with long villi. It was carefully opened, care
having been taken not to injure the embryo in any way. The

1 Letter from Dr. Ellis, Jan. 7, 1894.

No. 2.] DEVELOPMENT OF THE HUMAN COELOM. 419

coelom was filled with a clear fluid, and many firm shreds of a
fibrine-like body which obscured the embryonic vesicle greatly.
With much difficulty the embryo could be outlined, and these
drawings proved to be of great service in making the recon-
struction. The portion of the chorion to which the embryo
was attached and the embryo were stained in carmine and
imbedded in paraffin. The whole was cut into sections, at
right angles to the body, 10 yp thick.

Every other section was enlarged 100 times and drawn on
wax plates 2 mm. thick, and from them the model of the
embryo was made. The model gives the whole central
nervous system, the entoderm throughout its extent, the blood-
vessels, and the muscle plates.

The shape of the neural tube is given in the diagrammatic
outline. It was closed only along the middle of the body, being
open in front down to the beginning of the fourth muscle plate.
From the beginning of the fourth plate to the beginning of
the fourteenth it was closed, and from there on again it was
open. In the figure the portions between x and x’ indicate
to what extent the tube is closed. In Figs. 17 and 18 the
tube is nearly closed, while in Fig. 20 the tail end of the tube
is just beginning to separate from the ectoderm. The cephalic
end of the tube already clearly outlines the fore-brain, the
mid-brain, and the hind-brain; the constriction, Fig. 16, a,
indicates the junction between the first two. On the ventral
side of the fore-brain there are two marked pockets, one on
either side, just behind the neuropore, which are no doubt the
primary optic vesicles. It shows that in the human embryo these
are fully outlined before the brain has separated itself from the
ectoderm. Farther behind, very near the dorsal median line
and about in the middle of the head, there is a short pocket of
thickened ectoderm, the otic vesicle. Towards the hinder
end of the embryo the spinal cord communicates by means of
a solid band of cells with the entoderm, Fig. 20. At no point
in this communication is there a canal, so it must be viewed as
the last remnant of the neurenteric canal. The location is
opposite the twelfth muscle plate, or in the neighborhood of
what will later on be the position of the first rib. The chorda

420 MALE, [Vou. XII.

dorsalis extends to the neurenteric canal, but not beyond it.
There is no chorda in the tail end of the embryo.

—

Sess

yy

2
AY
gS j hie)

/

Fics. 17-20. — Sections through Embryo No. XII, as indicated by the Lines in Fig. 16. Enlarged
50 times. The black is the coelom within the body. QO! and 03, first and third occipital muscle
plates; Cl and C8, first and eighth cervical muscle plates; D!, first dorsal muscle plate ; a,
aorta; v, omphalomesenteric vein; ¢, thyroid; Z, liver: 4, pharynx ; 2, intestine ; 2 c, neuren-
teric canal ; #z x, membrana reuniens.

Throughout the central nervous system, immediately about
the central canal, there are many karyokinetic figures, showing

No. 2.]| DEVELOPMENT OF THE HUMAN COELOM. 421

that the specimen was excellently preserved. In the greater
portion of the neural tube the tissue is already marked by two
zones, a central one rich in nuclei, and a peripheral containing
none. This corresponds with the description already made
familiar to us by His.

The general shape of the whole central nervous system is
very unlike that of any other young human embryo ever pub-
lished. It circumscribes the greater portion of a circle, while
in the other human embryos of this size it makes more of a
straight line. I think that it is probable that this specimen
represents the normal, as it was not injured nor handled in
any way before it was cut into sections.

The entoderm, as the figures show, is already divided into
fore-gut, mid-gut, and hind-gut. The fore-gut makes the
pharynx, from which there are four diverticula on the dorsal
side, one on the ventral side, and two near the mouth. The
four on the dorsal side mark the first two branchial pockets on
either side of the embryo ; the two in front are Seessel’s pocket
and the entodermal portion of the mouth ; while the one on
the ventral side of the pharynx is the beginning of the median
portion of the thyroid gland (Fig. 17, t.).

At the junction of the pharynx with the umbilical vesicle
there is a large diverticulum into the septum transversum, Fig.
18 1, the beginning of the liver.

Within the tail end of the embryo, behind the neurenteric
canal, the hind-gut is enlarged considerably, and from it the
entodermal canal of the allantois arises.

The whole umbilical vesicle is covered with blood-vessels
which communicate with the veins and arteries of the embryo.
Near the origin of the liver there are two veins which collect
the blood from the umbilical vesicle and then enter the heart.
These are the omphalomesenteric veins. They with a num-
ber of their branches are shown in sections in Fig. 18, v.
The heart itself is broken, but there is enough of it left to
show that it is bent upon itself and contains a large cavity at the
point where the veins entered it. From the heart two arteries
arise and pass in front of the first branchial pocket, and each
follows the course as shown in black in the reconstruction.

422 MALL. [VoL. XII.

The aortae do not unite, but each sends a number of segmen-
tal branches to the umbilical vesicle along the tail end of the
embryo. These are, of course, temporary; they maybe
called collectively the omphalomesenteric arteries. As the
permanent omphalomesenteric artery arises more aboral than
any of these, it is easy to understand that most of them must
degenerate.

The sections show that there are fourteen muscle plates, all
of which are hollow and do not in any way communicate with
the body cavity in general. Kollman, who described an embryo
of this same age, numbers them from before backward, but I
think that they can be designated more definitely. Froriep?
showed that in all amniotic vertebrates there were a number of
muscle plates and dorsal ganglia formed in the occipital
region, and studied their fate in the chick and in the cow’s
embryo. Platt? has also followed the order of the origin of the
muscle plate in the chick, and found that the first division of
the mesoderm was between the third and fourth occipital plates.
The first three or four of these segments communicate in the
chick, according to Dexter,? with the coelom, and Bonnet ‘ has
found also that the same is true in the sheep. Bonnet’s fig-
ures (compare his Plate IV) show that a sheep’s embryo of the
same stage as embryo XII has muscle plates much more sharply
outlined than the human. In order to locate the muscle plates
more definitely I have made every effort to count the spinal
ganglia in embryo XII, but with no definite result. It is
impossible for me to define the spinal ganglia, as often they
are represented by a few cells only, then again as a band of
cells they extend over several segments. The same is true in
the occipital region. Had I been able to number them defi-
nitely it would still have been impossible to number the muscle
plates from them, for His® has shown that there is an occipital
ganglion in the human embryo as well as in the lower animals.

The fact that the muscle plates reach to the otic vesicle in

1 Froriep: His’s Archiv, 1883 and 1886.
? Platt: Bulletin of the Museum of Comparative Zodlogy, vol. XVII.
3 Dexter: Anatom. Anz., 1890.

4 Bonnet: His’s Archiv, 1880.
5 His: Abhandl. d. sich. Gesellsch. d. Wiss., Bd. XXIV.

No. 2.] DEVELOPMENT OF THE HUMAN COELOM. 423

embryo XII, as well as in Kollman’s embryo Bulle, indicate
that the first plates must belong to the occipital region, and I
have found that there are three occipital muscle plates in
embryo No. II.! Moreover, there is every indication of a
degeneration of the first two plates in XII, so on this account
I am inclined to number them as they are numbered in Fig.
16. Ido not think that any of them ever communicate with
the pericardial cavity as Bonnet found them in the sheep. The
cavities in all of the other plates are small, and they are sep-
arated by a large mass of mesoderm from the coelom. This
all confirms my view.

The chorda extends from Seessel’s pocket to the neurenteric
canal.

There are also a few segmental ducts, some completely and
some partly separated from the ectoderm, as was the case in
Kollman’s embryo. The ducts are small, and extend over one
or two sections only, and occasionally one of them is arising at
several different points between a given two segments. They
are present on both sides between the first and second cervical
segments, second and third segments, third and fourth segments,
fourth and fifth segments, and only on the left side in the region
of the fifth and sixth cervical segments.

The coelom of this embryo is especially instructive. A
sagittal section of the embryo and ovum is given in Fig. 21.
This embryo, when drawn connected with the ovum, is very
similar to Graf Spee’s embryo Gle. as shown in Fig. 15. It is
very easy for us to conceive the von Spee embryo converted
into this embryo, for about all the change that is necessary is
that the embryo grow somewhat and bend upon itself. In so
doing the attachment of the umbilical vesicle becomes smaller
as the amnion encircles the body of the embryo more. The
position of the neurenteric canal, the shape of the allantois,
and the formation of the pericardial cavity, all show that the
curving must be a normal one.

Nearly all other young embryos of this stage, or a little older,
which have been published show a straighter body or even a
curve in the opposite direction. I have also in my collection

1 See also Mall: Journ. of Morph., vol. V.

424 MALL. [VoL. XII.

two embryos of this stage, Nos. Iand XV, which had been taken
out of the chorion and torn from the umbilical vesicle, and both
of them are straight like Kollman’s embryo Bulle and His’s?
embryo L. It is difficult to conceive how my embryo XII
could possibly be torn out of its membranes without straight-
ening it. Weneed only
recall our experience in
hardening embryos of
lower animals to be re-
minded how easily a
curved embryo is
straightened when it is
handled the least bit
roughly before: itis
hardened.

His,.an/ his) (eneat
monograph on human
Gar AA ie embryos, emphasizes a
curve in the back of the
embryo just the reverse
of the one given in Fig.
21. Irefer'to-embryos
sch. BB.) and) Wezas
well as to Minot’s em-
bryo) 195.7). he: sfact

that this inverted bend

Jo Goce ousrm nin sauae (cts a Tun con
coelom ; «7, umbilical vesicle; ad, allantois; A, — stant (His’s Ri., | for
medullary plate; 7 c, neurenteric canal. zi i
instance), and that it

occurs at the time when any tension upon the umbilical ves-
icle could produce it, makes me believe that it is an artifact.
This view was suggested to me a number of years ago, when I
was removing young dogs’ embryos from the uterus, and unwit-
tingly distorted a number of them in this very way before they
were hardened. The middle of the back is the weakest part
of the embryo’s body, and the umbilical vesicle is attached to

1 His: Anat. mensch. Embryonen, Plate VI.
2 Minot: Human Embryology, New York, Fig. 169.

No. 2.] DEVELOPMENT OF THE HUMAN COELOM. 425

it. Under these conditions the simple weight of the vesicle is
sufficient to bend the back of the embryo as pictured by His.

To return to the coelom. At the hinder end of the embryo
the coelom dips into the body overlapping the hind-gut in the
neighborhood of the neurenteric canal, as shown in Fig. 20.
This cavity communicates with its fellow on the opposite side
through an opening between the umbilical vesicle and the
allantois, marked O in Fig. 16. This communication has
already been described by His! for an embryo somewhat older.
If, now, the point O in Fig. 16 is approximated towards NC,
with a flexion of the embryo at the same time, this communi-
cation is easily explained. In other words, as the hind-gut is
being separated from the umbilical vesicle, a groove-like por-
tion of the coelom is also included in the body of the embryo.
At the hinder portion of the embryo, on either side, the
coelomic grooves extend deeper into the body of the embryo,
and communicate with each other around the aboral side of the
stem of the umbilical vesicle. This communication is shown
well by His in Fig. I, B, Plate VI of his AZ/as, as well as in the
same figure, page 299 of Minot’s Embryology. Excellent profile
views showing this point are given in all the embryos figured
on Plate IX of His’s Addas.

I emphasize this point in order to exclude the ventral mesen-
tery for this portion of the embryo. The fact that this
mesentery could never have existed in the human embryo is
also proved by a careful examination of His’s models of human
embryos made by Ziegler.

As we pass towards the head in embryo XII the coelomic
groove communicates freely with the extraembryonic coelom
until the region of the membrana reuniens is reached. This is
shown in Fig. 19, MR, with the membrana reuniens complete
on one side, but not yet united on the other. The membrana
reuniens extends up to the heart, and separates the pericardial
cavity from the extraembryonic coelom, then crosses the ven-
tral median line to return on the opposite side of the embryo.
Throughout the extent of the membrana reuniens there is a
great increase of mesodermal tissue, which encircles completely

1 His: Anat. mensch. Embryonen, I, p. 126.

426 MALL. [Vor XH.

the beginning of the liver, as Fig. 18 shows. A portion of
this mesodermal tissue has been described by His as the septum
transversum.! According to His only that portion of the meso-
dermal tissue is septum transversum which lies between the
posterior part of the pericardial cavity (Parietelhohle), the wall
of the intestine, and the point where the veins enter the heart.
It extends across the body, and has within it the beginning of
the liver. In transverse section this region is shown in Fig.
18. Now the pericardial cavity communicates by means of a
long canal on either side, with the peritoneal cavity, and the
omphalomesenteric vein hangs into this, attached to a kind of
mesentery, as Fig. 18 shows. Lower down, near the commu-
nication (Fig. 19), there is an indication of the beginning of the
umbilical vein, which unites with the omphalomesenteric vein
through the membrana reuniens. The two canals which com-
municate with the extraembryonic coelom are the pleural cav-
ities, and the membrana reuniens aids to separate them from
the peritoneal.

All of the tissues from the diaphragm to the opening of
the liver duct into the duodenum arise from the septum trans-
versum and the membrana reuniens ; the stomach from the
fore-gut, the liver from the liver diverticulum, and the diaphragm
from the septum transversum and the membrana reuniens. The
Cuvierian duct must also have arisen in the membrana reuniens,
in order to pass around the outside of the body cavity to reach
the cardinal and jugular veins, as pictured by His? for the
human embryo.

In the further development of the pleural and _ pericardial
cavities the Cuvierian veins give us our best landmark, as they
define the point where the pleural cavity is to be separated
from the pericardial. And it really seems, as if the greater
portion of the diaphragm is formed from the portion of the
septum transversum on the ventral side of the vein and from
the membrana reuniens, rather than from the portion immedi-
ately in front of the intestine. In other words, there is a

1 His: Anat. mensch. Embryonen, I, p. 126.
2 His: His’s Archiv, 1881, Plate XII, Fig. 9. Also Anat. mensch. Embryonen,
Plate IX, Figs. 10-12, 14.

No. 2.] DEVELOPMENT OF THE HUMAN COELOM. 427

horseshoe-shaped ridge of tissue around the neck of the embryo
to the ventral side of the pericardial and pleural cavities and
parallel to them. The median portion is composed of the
septum transversum, and each wing of the shoe is the mem-
brana reuniens, one on either side of the embryo. Its general
direction in this stage is parallel with the long axis of the
embryo, and within each wing there is an omphalomesenteric
vein.

Origin of Pericardial Cavity. — With the pericardial cavity
opening into the extraembryonic coelom on either side as a
basis, it is possible to trace back the pericardial cavity to its

Fic. 22. Fic. 23. Fic. 24.

Fics. 22-24.— Three Stages to show the Development of the Blastodermic Layers at the Head End
of the Embryo. Fig. 22, Hypothetical Stage. Fig. 23, Embryo No. III. Fig. 24, Embryo
No. XII. V, vein; 4, pharynx; az, amnion; 7z Z, medullary plate; % c, pericardial
cavity ; S, Seessel’s pocket ; #z, mouth; ¢, thyroid ; 7 and 2, first and second branchial pockets ;
~, pleural cavity; # ~, membrana reuniens; wv o #, omphalomesenteric vein, which is
expressed as a dotted line; O, communication between right and left body cavities on the
ventral side of the umbilical vesicle.

origin. Figs. 16 and 24 show that the ventral wall of the peri-
cardial cavity is composed mostly of mesoderm. This is the
portion of the membrana reuniens which is composed of meso-
derm, as the sections, Figs. 18 and 19, show. An earlier stage
is shown in the diagrammatic Fig. 23. It is taken from
embryo No. III. In this specimen, since the ectoderm of the
amnion has not reached completely around the body, as both
the sagittal and transverse sections show (Figs. 23 and 25),
it is evident that the pericardial space is first covered on
the ventral side with mesoderm and later the ectoderm is added
when the amnion begins to close over the head. In embryo
III the canals communicating between the pericardial space
and the extraembryonic coelom are not as long as in embryo

428 MALL. [Vor. XII.

XII, and the ventral walls of the pericardial space are composed
wholly of mesoderm. This indicates that the growth of this
wall was first by a union of the mesoderm, which was followed
by the ectoderm of the amnion to complete the body wall.
The process is shown in Figs. 22-24. Fig. 22 is a hypo-
thetical stage between Graf Spee’s embryo Gle. and my embryo
No. III. As the process from Graf Spee’s embryo continues,
the blood-vessels reach the body to form the heart, as indicated
by the outlines marked v. in Fig. 22. The mesoderm of the
amnion then unites with that of the um-
bilical vesicle, and the first pericardial
space is formed. This is not wholly an
imaginary stage, for it is based upon
Bonnet’s observations upon the sheep,!
as well as Cadiat’s upon the chick.? In
a sagittal section of a sheep’s embryo
of about the same stage (Plate III, Figs.
16-20, c CB) Bonnet gives a similar fold,
and after the pericardial walls are well
BiG, 25 go coca tircughthe Hee eTOrmed he sives.an iivistration ol a,staeic
of Embryo No. III. Enlarged
ss times. P%, pharynx; H,heart. 1n Which it still communicates with the
he arn i the amsotis cavity extraembryonic coelom (Plate IV, Fig.
ure growth of the amnion tocom- 77, KC), With Graf Spee’s embryo
plete the ventral body wall. ; j
Gle. and with Bonnet’s observations
upon the sheep as a starting-point, it is not difficult to interpret
Figs. 22-24.

Extension of the Amnion. — After the stage of embryo XII
is passed the amnion rapidly envelops the whole body and soon
passes out over the cord. The next stage after No. XII which
I have studied is No. XIX. I have very perfect photographs
of this specimen, and the sections are all good, although the
nervous system is macerated. The embryo has rotated in the
amnion, throwing the cord to the right side with the left side
towards the observer. It would have been impossible to obtain
a view of the right side of the embryo without cutting the
cord. The outlines of this embryo and ovum are given in

1 Bonnet: His’s Archiv, 1889.
2 Cadiat: Jour. de l’Anat. et de la Physiol., 1883, Plate V, Figs. 1, 2.

No. 2.] DEVELOPMENT OF THE HUMAN COELOM. 429

Fig. 26. Two sections through the body are given in Figs.
27 and 28.

The amnion has become separated from the body with the
exception of the part about the cord and also that along the
right side of the body, over the heart. The arrow in Fig. 25
shows how the amnion on that side is extended over the ventral
body wall to make the
condition shown in Fig.
28. No doubt the cause _ PS
of this is the rotation of ante he
the body, throwing the ef Uy,
cord to its right side and ides
the amnion with it. In
nearly all young embryos
the cord is on the right
side.1 With the excep-
tion of the four instances
mentioned below, the ro-
tation has always been
so as to throw the left
side of the body away
from the chorion, and in
all of these specimens the
amnion must have swept
ever they body inom leit qe .6.— oven and Eiabryo No: XI enlareedeeaes!
to right, as shown in the Just half of the ovum is shown. A,arm; /, leg; A,

Dri, heart ; z v, umbilical vesicle ; B, branchial arch.
figures. I find a similar
illustration by His in his great monograph.?

Absence of a Ventral Mesentery.— After the septum trans-
versum has been formed as it is in embryo XII, there is on its
ventral side a pretty sharp groove, which indicates that the
umbilical vesicle is being constricted at this point.

It is generally believed that the ventral mesentery of the
intestine extends to the umbilicus, and that ultimately the round
ligament of the liver represents its remnant after most of it has

BS
—
S=\>
fe
ae

1 The exceptions have been published by Waldeyer: Studien des physiol.
Inst. zu Breslau, 1865 ; Janogik: Arch. f. mik. Anat., Bd. 30; His: Anat. mensch.
Embryonen, Plate VIII, Figs. A 1-4; Mall: Journ. of Morph., vol. V.

2 His: Anat. mensch. Embryonen, Plate VI, Fig. 3, No. Io.

430 MALL. [Vou. XII.

disappeared. This theory is expressed by two diagrams in
Minot’s Embryology, page 767. As the liver begins to grow,
and while the heart is being pushed down in front of it, the
ventral end of the septum transversum is turned down to the
umbilicus. While this is taking place the stem of the umbilical
vesicle becomes relatively smaller and smaller, but there is no
union between the umbilical vesicle and the septum transversum
as expressed in Minot’s diagram. The first stage of this

Fics. 27 and 28.— Section through Embryo XIX to show the Attachment of the Amnion to the
Side of the Body. Enlarged 25 times. A, amnion; S, stomach; H, heart; c, cardinal
vein; #, umbilical vein; a, aorta.

process is shown in my Fig. 24, and its successive stages are
shown in His’s A¢las, Plate IX. In all six embryos pictured
on that plate the successive stages are represented, and in none
of them is the umbilical vesicle attached to the septum trans-
versum to form a ventral mesentery. From these embryos of
His we can pass to embryo XIX, in which the umbilical vesicle
communicates by a round canal with the intestine, and the
tube is completely encircled with a space which extends to the
liver, thus cutting off any possible ventral mesentery at that
point. The same thing is shown, but in a later stage, in

No. 2.] DEVELOPMENT OF THE HUMAN COELOM. 431

Fig. 30, O, but a new process has already taken place to com-
plicate matters.

In embryo XII there is just a beginning of an umbilical
vein in the membrana reuniens. In Kollman’s embryo the
vein is more marked.!_ The vein extends out into the somato-
pleure, far away from either the intestine or the median line.
This same position is again shown in His’s embryos BB. and
Lr on Plate IX in his A¢das. The left umbilical vein becomes

Fic. 29. — Embryo No. II Attached to the Chorion. Enlarged 5 times. Just half of the Ovum is
shown. O », otic vesicle ; VU Z, upper extremity ; Z Z, lower extremity; JV, nose; J, //, ///,

branchial arches.

the more prominent, and as the body wall is developed more
and more it moves around towards the ventral median line.
This movement takes place in common with the movement of
the amnion over the body from left to right, as shown in Fig.
28. In embryo No. II, however, the liver has nearly reached
the umbilicus, and the vein has almost moved around to the

1 Kollman: His’s Archiv, 1891, Plate III, Figs. 2, 3, 4. V. umbil.

432 MALL. [VoL. XII.

ventral median line, as shown both in the reconstruction and
the sections (Figs. 30, 37-39). After the vein has moved
around the body to its ventral surface, and after the liver
moves away from the umbilicus up to the permanent diaphragm,
it is easy to explain the formation of the round and broad liga-

SS

mie Si
“7

WN

soy

S

WY

SS SD \WAY

<S
ANS,

Fic. 30. — Reconstruction of Embryo No. II. Enlarged 17 times. V and _X, fifth and tenth cranial
nerves; 7, 2, 3, and 4, cast of the branchial pockets; 7 and 8, first and eighth cervical nerves,
from the fourth the phrenic arises; 72, twelfth dorsal nerve; A, auricle; V7, ventricle; Z,
lung ; S, stomach; P, pancreas ; W” D, Wolffian body ; X, kidney; 17, mesentery ; S 7, septum
transversum ; O, openings which communicate with the peritoneal cavity of the opposite side.
The black line around the heart marks the pericardial cavity.

ments of the liver as a secondary formation, but not as a rem-
nant of a ventral mesentery. It might be called a portion of
the septum transversum, as it is directly continuous withit. A
ventral mesentery does exist between the abdominal walls
and the liver, and only extends slightly below the liver. It is

No. 2.] DEVELOPMENT OF THE HUMAN COELOM. 433

always slightly to the left of the median line, and is in direct
connection with the septum transversum (Fig. 30, O and S 7).’

Coelom of Embryo No. II. — After the body cavity is begin-
ning to separate from the extraembryonic coelom, the next im-
portant stage is the one after the separation is complete, as from
now on the adult body cavities are formed by a simple division
and expansion of the cavities already within the body. This
stage is represented in em-
bryos XVIII, II, and IV.
All of these embryos are
nearly of the same size,
the successive stages being
in the order they are given.
No. XVIII is somewhat
distorted in the middle of
the body, while No. IV is
slightly macerated. No. II
is a perfect specimen, and
has been already described
by me several years ago.1
I shall confine my descrip-
tion of it to the body cavity.

The external form of the
embryo within the ovum is
given in Fig. 29. The Fie. 31.—Cast of the Body Cavity of Embryo No.
Heagon ef the\ umbilical, 7 BUA ee
vesicle, as well as the ex- entery; W B, position of Wolffian body; P,
FeO Het nian ancl the pericardial cavity ; Z, coelom over liver.
relation of the umbilical vesicle and amnion to the chorion, are
all given. The umbilical cord is large and lies on the left side
of the body, while in most embryos already published it is upon
the right side. The cord is short, and midway between the
embryo and its attachment to the chorion it shows a decided
enlargement. The umbilical vesicle is large, measuring 5 X 7
mm., and is located between the head end of the embryo and
the chorion.

The amnion has not grown very much, still leaving a great

1 Mall: Journ. of Morph., vol. V.

434 MALL. [Vor. XII.

space between it and the chorion, the extraembryonic coelom
(compare with Fig. 26). Within it hangs this large umbilical
vesicle, the lumen of which no longer connects with the ali-
mentary canal. The separation is now complete. Around
the stem of the vesicle the extraembryonic coelom communi-
cates freely with the body cavity, as shown in Fig. 30. This
figure is from a reconstruction, and shows the general extent
of the body cavity within the embryo. It encircles the heart,
and then extends to the lungs and over them and to the stomach,
over the intestines, and out into the cord. A cast of the
whole cavity is also given, showing the slit
on the dorsal side for the mesentery of
the intestine, and the grooves on either
side of this for the Wolffian bodies. There
are also grooves in the cast for the veins,
and the place where the Cuvierian duct
enters the heart is marked VY. The sagit-
tal section of the peritoneal cavity is given
in Fig. 32. The striated line indicates
where the cavity crosses the median line
Fic. 32.—Outline of Coelom of the body, while the other lines outline
in Embryo No. Lin Sagittal the cavity beyond. /. outlines the lesser

Section. The striated line
indicates that the coelom peritoneal cavity. Figs. 33-39 give the

crosses the median line. P, . j

pericardial spaces /, pleural. €Xtent of the peritoneal cavity in ‘different

Pai ty of lesser Hortions of the embryo, as indicated by
the lines'in) Fis. 30;

It is not difficult now to imagine the body cavity of embryo
XII converted into the one just described. In that embryo the
heart is high in the neck on the oval and dorsal side of the
septum transversum. In this embryo it is on the ventral and
oral side of the septum transversum, but still above the eighth
cervical nerve. The septum transversum has already received
its nerve supply from the fourth cervical nerve, as pointed out
in the early part of the century by von Bear. This movement
of the septum transversum is accompanied by a movement of
all the other organs on their way into the thorax and abdomen
of the future individual. In the rotation the Cuvierian duct
acts much as the fixed point about which the coelom is bent.

No. 2.] DEVELOPMENT OF THE HUMAN COELOM. 435

The figures all illustrate this beautifully. But as the heart has
rolled over the liver, and the septum transversum has undergone
a quarter-revolution, the Cuvierian ducts and all have moved
away from the head. This is by no means the end of the
excursion of the septum transversum, as its dorsal end must move
down and beyond the twelfth dorsal segment (compare Fig. 30).

The pericardial cavity surrounds the whole heart, as the va-
rious figures show. The cavity is traversed only where the
large veins enter, and where the aorta leaves the heart. The

Fic. 33. Fic. 34.

Fics. 33 and 34.— Sections through Embryo No. II at the Points indicated in Fig. 30. Enlarged
22 times. #&4, brachial plexus; a, aorta; 4 a, bulbus aortae; #2, pharynx; 4%, heart; 7,
trachea; 0, oesophagus ; 7, jugular vein; 2, lung; vc, cardinal vein; C, Cuvierian duct.

cavity completely surrounds the bulbus aortae to its origin
(Figs. 32-35) in the ventricle. On the dorsal side of the
heart the pericardial cavity is separated by a bridge for the
transmission of the veins to the heart. Between the bulbus
aortae and the entrance of the veins into the heart the peri-
cardial cavity crosses the median line as three distinct open-
ings, as expressed by the black areas in front of the trachea in
Fig. 30. On the dorsal side of the heart on either side of the
lungs the pericardial cavity communicates with the pleural
cavities by means of two openings (Fig. 33), each of which is
about .1 X .§ mm. in diameter. Farther on, the pleural cavities

436 DEAE: [VoL. XII.

extend as two slits which encircle the lobes of the liver and
separate them from the alimentary canal on the one hand and
from the body wall on the other (Figs. 34-37). The two

Fics. 35 and 36.— Sections through Embryo No. II. A, aorta; s, stomach; Z, liver; #, umbilical
vein; x, bulbus aortae; 4, heart; 0, omphalomesenteric vein; g, lesser peritoneal cavity ;_/,
foramen of Winslow; c, coeliac axis ; w 4, Wolffian body ; wd, Wolffian duct.

pleural cavities do not communicate with each other around the
lungs, leaving for them both a dorsal and a ventral mesentery.

This appearance of the coelom about the lungs and the liver
can be explained by the lungs and liver both growing into the
two pleural cavities of embryo XII, and this has often made me
think that the membrana reuniens of embryo XII is the main
origin of what is called septum transversum in embryo II. If
this proves to be the case, then the lower end of the mem-
brana reuniens will form the ventral end of the diaphragm, and
not the reverse. A stage between embryos XII and XVIII
(Fig. 41) is required to elucidate this point.

In the neighborhood of the stomach the peritoneal cavity on
either side of it has become asymmetrical, as Fig. 36 shows.
The mesentery has become bent to the left side, leaving a
diverticulum from the right side which extends oralwards to
the tip of the lung (Figs. 34 and 35) to form the beginning of

No. 2.] DEVELOPMENT OF THE HUMAN COELOM. 437

the lesser peritoneal cavity.1_ Further aboralwards the cavities
become symmetrical again (Figs. 37, 38), and then unite
along the ventral median line, as shown in Fig. 39. The ven-
tral mesentery shown in Fig. 38 does not extend more than a
section or two beyond the liver, and is separated by a marked

g

MASS

Fics. 37-39. — Sections through Embryo No. II. A, aorta; wc, cardinal vein; 0, omphalomesenteric
vein ; 4, pancreas ; 2, intestine ; 4, bile duct; Z, liver; Z, heart ; #, umbilical vein ; 7, mesentery ;
w 4, Wolffian body ; @ZZ, allantois.

opening from the stem of the umbilical vesicle in this embryo,

as is shown in Fig. 30, O (see also No. XII, Fig. 16, O). On

the aboral side of the umbilical cord the peritoneal cavities of

the two sides unite in both embryos again, marked O! in both

figures.

Development of Body Cavity in the Chick.— The body
cavity of the chick has been carefully studied by Budge,? who
followed its course by means of injection. With a fine hypo-
dermic syringe he filled the spaces forming the coelom in the
order of their appearance, thus showing their extent in vari-
ous embryos. The splanchnopleure, according to Budge, may
be split into two layers, one dorsal or lymphatic and the other
ventral or vascular. Drasch’s® recent description of the early

1 Mall: Journ. of Morph., vol. V.

2 Budge: His’s Archiv, 1880 and 1887.
3 Drasch: Anatom. Anz., Bd. 9.

438 MALL. [Vox. XII.

formation of the coelom confirms this statement. As the first
blood-vessels are formed, lymph vessels appear on their dorsal
side, which flow together to form a network, and accompany the
primitive veins to the axial part of the germinal area. Here
the lymphatics form two spaces, one on either side of the body,
which soon unite across the body on the ventral side of the
heart. In this way the primitive body cavity of birds appears
at first as an H, the uprights of which are on either side of the
body and the cross-piece on the oral side of the sinus venosus.
In its further development the sinus venosus grows to the dorsal
side of the cross-piece, thus reversing the relation of the vascular
system to the coelom in this portion of the embryo. The
uprights of the H fall to the outside of the body, and are
swallowed up in the formation of the amniotic folds.
According to Budge two diverticula grow from the cross-
piece of the H, one on either side of the chorda, towards the
tail of the embryo, to form the
primitive pleuro-peritoneal
cavities. Budge’s paper was
published from fragmentary
notes after his death, and I
am sure that the above state-
ment is not correct. Profes-
sor His has placed at my
Fi Scot ek or Giese disposal Budge's specimens,
Coelom. Although the embryo has been injected, which J think show conclu-
the injection masses @ and c are not continuous. sively thar iis interrec tion
of this subject is not correct. Most of the injections were
made into the amniotic fold, which is very large in birds.
Cross-sections of chicks at this stage show that the large extra-
embryonic coelom communicates very freely with the body cavity,
and the cross-piece also communicates freely with the cavity at
the anterior end of the embryo. This has already been described
and pictured by Cadiat,! and recently again by Duval.?, Around
the heart, however, the communication is freest between the
extraembryonic coelom and the body cavity, and it is natural

1 Cadiat: Jour. de l’Anat. et de la Physiol., 1883, Plate V, Figs. 1 and 2.
2 Duval: Atlas d’Embryologie, Plate XXII, Fig. 354.

No. 2.] DEVELOPMENT OF THE HUMAN COELOM. 439

that the fluid should find its way along this channel first, and
then extend into the body cavity from this point, giving pic-
tures which in transverse section are like Fig. 40. Surface
views could not decide that these cavities communicate freely ;
and these sections which I have studied were no doubt made
after Budge had written the rough draft of his manuscript, as
they are not referred to in his paper.

Although the body cavity of the bird is formed after the
same manner as it is in the human embryo, there is one marked
difference in the formation of the pericardial space. In the
bird the mesoderm does not extend throughout the head fold
of the blastoderm, leaving a portion of the ectoderm in direct
contact with the entoderm. Later the mesoderm grows into
this region, and at the same time the pericardial cavity extends
to the outside of the body. This condition continues for a long
time, allowing the pericardial cavity to communicate with the
false amnion after the embryo is well formed. Duval’s excel-
lent Atlas shows how the pericardial cavity first communicates
with the exterior of the body, and after the body walls have
united the heart still lies in apposition with the liver, as there is
no septum transversum. The only trace of a septum trans-
versum that I can find in young chicks is at the point the
Cuvierian ducts enter the heart. Here a bridge of tissue passes
transversely to the body. The liver does not grow into it, but
accompanies the single omphalomesenteric vein before it enters
the heart?

Development of the Diaphragm.— Our knowledge of the
development of the diaphragm is based upon the researches of
von Baer,? Cadiat,? His,t Uskow,® and Ravn.6 Each contrib-
uted his portion: von Baer that the diaphragm is at first
located high in the neck and must descend in its development,
and, because of its high position at first, is innervated by a

1 Since the above has been written this subject has been studied thoroughly
by Ravn, whose paper will be found in His’s Archiv, 1896.

2 Von Baer: Entwicklungsgeschichte, 1837.

8 Cadiat: Jour. de l’Anat. et de la Physiol., 1878.

4 His: Anat. mensch. Embryonen, Th. I, 1880.

5 Uskow: Arch. f. mik. Anat., 1883.

® Ravn: His’s Archiv, 1889.

440 MALL. EVOL. XIE

cervical nerve ; Cadiat and His recognized the mass of tissue
in the embryo which is destined to give rise to the diaphragm ;
Uskow and Ravn studied more carefully the separation of the
body cavities from one another; and the wandering of the organs
was emphasized by Uskow.

It is now no great task for me to give the development of
the diaphragm in the human embryo, for I have at my dis-
posal excellent sections, as well as definite knowledge of the
anatomy of the surrounding organs contributed by the above-
mentioned authors.

While the embryo is still straight it is very easy to locate the
various organs and their relations to one another, but through
their shifting and the flexion and extension of the embryo the
relations are constantly changing, and one must not rely too
much upon sections, or else erroneous impressions will often
be obtained. At first the heart is upon the oral and dorsal
side of the septum transversum, then on its ventral side, and
finally again on its dorsal side. At first the lungs are on the
dorsal side of the heart, then on the lateral side, and finally
also on the ventral side of it. At first the liver is on the aboral
side of the septum transversum in the head of the embryo, then
on the dorsal side of it in the cervical region of the embryo,
then as the liver is descending in its excursion it is transferred
to the ventral side of the septum and extends into the sacral
region. At first the Wolffian body extends high into the tho-
racic region of the embryo, but while it is degenerating and the
diaphragm descends, the upper part of the posterior cardinal
vein remains, while the lower part is incorporated with its vena
cava inferior, as shown by Hochstetter.1 As the Cuvierian
ducts and cardinal vein descend into the thorax, the segmental
veins entering the cardinal veins are gradually shifted, so that
veins which originally emptied into the posterior cardinal now
empty into the anterior cardinal. While the whole process is
taking place the arteries arising from the descending aorta also
shift, as I have shown in a previous communication.2 At that
time my collection of human embryos was very limited, and it

1 Hochstetter: Morph. Jahr., Bd. 20, p. 563.
2 Mall: Journ. of Morph., vol. V, p. 472.

No. 2.] DEVELOPMENT OF THE HUMAN COELOM. 441

was necessary to include some observations on lower animals
to prove my point, but now I can give a complete table of
human embryos in which the point of origin of the coeliac axis
is recorded.

TABLE SHOWING POINT OF ORIGIN OF COELIAC AXIS.

LENGTH IN

EMBRYO. MILLIMETERS. ORIGIN OF CoELiaAc Axis.
No. XII. Zar Opposite 4th cervical nerve.!
His’s Embryo M 2.6 < 1st dorsal ta
“ “ B ih “ 2nd “< ‘“ 8
No. II. as oC Athi)‘ Ks
His’s Embryo A eS se 6thy i Ge
No. XLIII. 103} ae rothwames ce
No. IX. 14. ce 11th se “
No. XXII. 18. a Toth ss 6“
No. XVII. 16. cs r2thy «<5 “
No. LVII. 20. Below 12th e «6
Adult. ee r2theee cs 6

1 In the first two embryos the omphalomesenteric artery is noted, and not the
coeliac axis.

2 Compare Fig. 15, Plate VI, His’s Atlas, with M4, Pl. VII.

3 Compare Fig. 35, Plate II, His’s Atlas, with Fig. 1, Plate I.

4 Compare Figs. 79 and 86, His’s Atlas, with Fig. 4, Plate I.

The table shows that the arteries arising from the ventral
side of the aorta to supply the stomach and intestines are con-
stantly shifting until their definite origin is finally reached. In
these specimens the omphalomesenteric artery is shifted ahead
of the coeliac axis. In embryo No. II the omphalomesenteric
artery has a double origin from the aorta, which indicates that
this movement may be brought about by a new anastomosis
forming, which is then followed by an occlusion of the old origin.
At any rate it is impossible that the whole aorta shifts with the
abdominal viscera, for it is bound to the vertebrae and muscle
plates through the segmental arteries.

The various sections and the reconstruction of embryo No.
II show the pleural and pericardial cavities still communicating
freely. The same is true in embryos XIX, XVIII, and IV.

442 MALL. [Vov. XII.

Immediately after this stage there are no embryos in my col-
lection, so I have no specimen in which the communications
between the pleural and pericardial cavities are just closing.

Ue

Eee
i

y
v

\
:

“ISS
UW

EZ

Fic. 41. — Diagram to show the Position of the Diaphragm. The numbers on the blocks indicate
the embryos from which the diaphragms are taken. AO is His’s embryo K O; PP, the out-
line of the opening between the pleural and peritoneal cavity, which is finally closed when the
diaphragm reaches the tenth dorsal segment.

In embryos VIII, V, IX, and XLIII (Figs. 41-43), the pleural
and pericardial cavities are separated, while the pleural and
peritoneal still communicate. In the embryos with a vertex-
breech measurement exceeding 17 mm. the pleural and perito-
neal have been separated completely.

No. 2.] DEVELOPMENT OF THE HUMAN COELOM. 443

The separation of the pleural from the pericardial cavity is
dependent upon the complete development of the diaphragm.
At first the septum transversum and the membrana reuniens are

WS

\

aE! cain ‘iM

ee —<—_——

y
\\
&

|

Pl

Ki
ih

ty
GY

Fic. 42. — Reconstruction of Embryo No. IX. Enlarged 17 times. S 7, septum transversum ; Z,
liver; S, stomach; C, caecum; W, Wolffian body; A, kidney; 7-72, dorsal ganglia; 0,
omphalomesenteric artery. The ventral mesentery of the liver is dotted, as it is only a thin
membrane; S C, suprarenal capsule; A, point of communication between pleural and
peritoneal cavities.

on the ventral side of the pleural cavity, and both are still .
located within the head. As the septum transversum descends
into the body it is next located on the dorsal side of the heart.
In other words, the dorsal end of the septum transversum has
not moved as rapidly as the ventral end, and thus the whole

444 MALL. [Vor. XII.

mass of tissue has turned a quarter revolution. This is accom-
panied by the extreme flexion of the head, as represented in
embryo No. II. At this time the septum transversum has
descended to the lower part of the cervical region. Now the
septum begins to turn in the other direction again, for with the

Fia. 43. — Section through the Point of Communication between the Pleural and Peritoneal Cavities
in Embryo No. IX. Enlarged 15 times. 7, seventh rib; Z, lung; Zz, liver; 1%, ventral
mesentery of liver; a, aorta. The diaphragm is complete on one side, X, while it is incomplete
on the other.

development of the neck the ventral end of the septum becomes
the fixed point and the dorsal end moves more rapidly. The
successive stages in the movement of the septum are best shown
in the diagrammatic Fig. 41.

Fig. 42 shows the septum transversum on the ventral side of
the stomach and the pleural cavity communicating with the
peritoneal at the point X. The Wolffian body and the supra-

No. 2.] DEVELOPMENT OF THE HUMAN COELOM. 445

renal capsule, which is very large, have receded markedly,
and the pleural cavity already forms a pocket on the dorsal
side of them. A sagittal section through this region, somewhat

Fic. 44.— Section through Embryo No. XLIII. Enlarged 8 times. H, hand; A, auricle; V,

,
ventricle; Z, lung; S 7, septum transversum; /, phrenic nerve; U, umbilical vein; S,

stomach ; ”, Wolffian body ; Ov, ovary ; A, amnion; 7-72, ribs.

distant from the median line, is given in Fig. 44. A transverse
section of the embryo pictured in Fig. 42 is given in Fig. 43.
This section is just at the point above the opening, and shows
the communication between its pleural and peritoneal cavities
closed on one side, but open on the other. There is a ridge
on the side of the cavity which projects between the lung and

446 MALL. [Vou. XII.

the liver and continues down to the suprarenal capsule. This
ridge has been well described by Ravn,! who gives an excellent
illustration of the opening with the ridge encircling it.

4
4
4

Fic. 45.— Reconstruction of Embryo No. X. Enlarged 8 times. 2-72, dorsal ganglia; S C, supra-
renal capsule; 1”, Wolffian body; A, kidney; Z, liver; S, stomach; C, caecum. The
dotted area on the ventral side of the liver indicates the extent of the ventral mesentery of
the liver.

In all the embryos in which the pleural and peritoneal cavi-
ties still communicate, the vena cava does not yet communicate
with the posterior cardinal vein.

Fig. 45 is from an embryo slightly larger than the one from
which Fig. 42 was taken. The pleuro-peritoneal communica-

1 Ravn: His’s Archiv, 1889, Plate X, Fig. 16.

No. 2.] DEVELOPMENT OF THE HUMAN COELOM. 447

tion has just closed by the walls of the ridge having grown
together; the extent and shape of the pleural cavity is much as
it is in Fig. 42. The Wolffian body is smaller, and the kidney
and suprarenal capsule have come together.

The story, then, is brief: as the diaphragm descends, its
dorsal end is in apposition with the suprarenal capsule, and
finally, when the capsule approaches the twelfth rib, a ridge of
tissue which also includes the capsule unites with a ridge from
the septum transversum, and the opening is closed. These two
ridges, however, are portions of one and the same ridge, as they
form a circle and in section appear as two ridges. The circle
is closed much after the fashion of tying up a bag.

All of the abdominal organs, with the exception of the kidney,
descend; and the descent is not completed until the pelvis is
formed to admit some of them. In the stages pictured nearly all
the small intestine lies in the umbilical cord, as is the case in
many mammalian embryos. In embryo X (Fig. 45) a large
portion of the liver also projects into the cord. I have also
observed a hernia of the liver in another embryo somewhat
larger. I do not consider the form of embryo X altogether
normal, but this was not noticed until the reconstruction was
complete.

Closely associated with the closing of the pleuro-peritoneal
opening is the development of the coeliac ganglion. In these
young embryos it is extremely large, and can be outlined al-
ready, while the septum transversum is still high in the thorax.
As the septum descends, the various communicating branches
of the nerves are caught up with the coeliac ganglion and
dragged along. This accounts for the high origin of the
splanchnic nerve.

Fig. 46 (embryo VI) shows that all the tissues are becoming
more definitely outlined, and the whole structure is firmer than
in embryo X. The organs of the abdomen are more firmly
clustered together, and the intestine has become more con-
voluted. The lung is much larger, and the pleural cavity ex-
tends to the ventral wall of the embryo, obscuring wholly the
outline of the heart. In general it confirms everything given
in Fig. 45.

448 MALL. (Vox. XII.

Minot! has stated that the pleural cavities are to be con-
sidered a portion of the septum transversum, because they lie
on the dorsal side of it. From what has already been said
above it will be seen that I consider the septum transversum the

=

Fic. 46. — Reconstruction of Embryo No. VI. Enlarged 8 times. 7-72, dorsal ganglia ; S C, supra-
renal capsule ; A, kidney ; W”, Wolffian body ; S, stomach; C, caecum; Z, liver. The dotted
area on the ventral side of the liver indicates the extent of the ventral mesentery.

mass of tissue between the pericardial cavity, the pleural cavities,
and the opening between the two sides of the peritoneal cavity
immediately below the liver, marked O in Figs. 16 and 30.
This tissue includes the membrana reuniens, which is really the
wings of the septum transversum as described by His. In my

1 Minot: Human Embryology, p. 483.

No. 2.] DEVELOPMENT OF THE HUMAN COELOM. 449

account I have employed the term membrana reuniens where-
ever possible to avoid confusion, and have usually employed
the terms septum and primitive diaphragm as synonyms.

There are developed within the region of the septum trans-
versum the whole liver, including its ventral mesentery, the lesser
peritoneal cavity, the stomach, and the suprarenal capsule.
This same region which I have marked out by these three
boundaries as the septum transversum is still sharply defined
in the adult. The point O in Figs. 16 and 30 is still as defin-
ably marked as ever by the round ligament, foramen of Wins-
low, and the duct passing from the liver to the duodenum.
The round ligament is developed by the umbilical vein shift-
ing around the side of the abdominal walls into the ventral
mesentery of the liver, and then when the liver is retracted
from the umbilical cord, the vein and mesentery remain as the
round and broad ligaments respectively.

Lesser Peritoneal Cavity.—I have already discussed the
lesser peritoneal cavity in a separate paper,! and find that I can
confirm all that I have stated at that time. I can only add
that the portion of it extending up under the lung degenerates,
while the omental sac is growing rapidly. I have also found
that it is extremely easy for the omentum to find its way over
the large intestine. At the time this takes place the large in-
testine is in the median line, while the stomach and the omen-
tum are on the left side of the body. After the intestine is
retracted from the cord the caecum falls over to the right side
of the body, while the descending colon is shifted to the left
side, and the omentum then comes to lie on the ventral side of
the transverse colon.

Expansion of the Body Cavity and Obliteration of the Extra-
embryonic Coelom.— After the pleural and pericardial cavities
are separated from each other it is very easy to follow their further
development. In embryo II, Fig. 47, the heart is still upright,
and a transverse section of it is also transverse to the lung.
The pleural cavity lies wholly on the dorsal side of the pericar-
dial, Fig. 32. In the next stage, as the lungs descend more and
more, the heart is tilted over so that its base is towards the

1 Mall: Journ. of Morph., vol. V.

450 MALL. [VoL. XII.

lung and its apex away from it, as in embryo IX, shown in Figs.
42 and 48. The pericardial space has now become separated
completely from the pleural, although both have grown at
about the same pace. From now on the pleural space grows
more rapidly than the pericardial, as shown in Fig. 49. I have
a number of embryos which represent intermediate stages
between embryos IX and XXII, and all of them confirm the
idea that the pleural space develops first and then is followed
by a growth of the lung. Fig. 50, which is a section of
embryo No. XLV, shows a marked increase in the size of the
lung, but the heart and pericardial space are of about the same

_ FR Ky

Fic. 47. Fic. 48. Fic. 49.

Fic. 47-49. — Outlines of the Pleural and Pericardial Cavities to show their Relative Position and
Size. Enlarged 7 times. Fig. 47, Embryo No. II; Fig. 48, Embryo No. IX; Fig. 49,
Embryo No. XXII. H, position of heart ; Z, position of lung.

size as in embryo XXII. A much later stage is shown in

Fig. 51. The scale of enlargement is only half that of Fig. 50,

and when this is considered it is again seen that the heart has

not grown very much but the lung has developed enormously.

It is therefore seen that at first the pericardial cavity is on
the oral side of the pleural, then on the ventral side, and is
finally enclosed by the pleural cavity growing over it.

The growth of the pleural cavity over the pericardial accounts
for the location of the phrenic nerve in the adult. In Fig. 47
the nerve passes to the septum transversum from the lateral
body wall and it is gradually separated from it by the descent
of the septum and by the growth of the pleural cavity between
the nerve and the body wall, thus locating the nerve in a mem-
brane, as Figs. 48 and 49 will readily explain.

The expansion of the peritoneal cavity is by no means as
simple. In it there are many bands and mesenteries, as

No. 2.] DEVELOPMENT OF THE HUMAN COELOM. A5I

well as a marked shifting of the organs. With the descent
of the testis a portion of it is cut off to form the tunica
vaginalis.

In embryo II the peritoneal cavity is extremely simple, as
the figures show, —a simple cavity on each side, communicating
the one with the other by means of two openings, one above and
one below the omphalomesenteric duct. Later, as the diaphragm
descends more and more, the liver
rotates, and its lobes soon fill the
peritoneal cavity, while the intestine
develops out into the cord. The
Wolffian body, sexual glands, and
suprarenal capsule fill the dorsal side
of the cavity and the rudimentary
pelvis. The whole development of
the intestine takes place within the
cord, and finally it is drawn into the

ees : Fic. 50.— Outline of Pleural and Peri-
embryos when it is about 30 mm. in cardial Cavities in Embryo No. XLV.
length. By what process thistakes Enlarsed7 times.
place I am unable to determine, but
it must take place very rapidly, for
I have never seen a human embryo
in which it is only partly retracted.
In the pig’s embryo, however, I have
found the stages in which the in-
testine is in process of retraction.

The liver now fills nearly the Fic. 51.— Outline of Pleural and Peri-

f cardial Cavities in Embryo No.
whole cavity, and extends down to XXXIV, Enlaiged oedues! ee
the pelvis, and in embryo XXII ay oo Deak aE Oona Ce
projects over the ovary and is in
contact with the rectum. As the intestines are retracted
from the cord the liver is relatively higher and higher, for
the expansion of the abdominal walls is now greater below the
umbilical cord than before, giving more space in this region
for the intestine which displaces the liver. In embryos
XXXIV and XLVIII the intestines have been studied, and it
was found that they were still located in the ventral portion of
the peritoneal cavity, as there is no pelvic cavity large enough

452 MALL. [Vou. XII.

to hold any portion of them. This question will be discussed
in the next paper.

The extraembryonic coelom has only a short existence, as it
is already completely obliterated in embryo No. XXII. This
embryo came to me in an unopened ovum, and on this account

Fic. 52. — Embryo No. XXII within the Ovum. Enlarged 3 diameters. The villi of chorion are
outlined on one side of the ovum only. The umbilical vesicle, « v, has become shifted
around to the dorsal and right side of the embryo. The outline is made from a photograph,
and is correct in detail with the exception of the attachment of the chord to the chorion. This
in reality attaches itself to the chorion immediately to the right of the embryo.

is extremely valuable for this purpose. This embryo is about
six weeks old, so, reasoning from it, the union of the amnion
and chorion takes place earlier than is generally believed. In
embryo No. XLIII, which is about five weeks old, the amnion
has expanded over the whole embryo and has nearly reached
the chorion. The earlier stages are given in the sagittal sec-
tions. They show that the amnion is nearest the chorion at
the caudal end of the embryo in the earliest stages, and soon

No: 2.| DEVELOPMENT OF THE HUMAN COELOM. 453

the two unite at this point. As the embryo grows, the union
of amnion and chorion extends. At the end of the fourth
week the extraembryonic coelom is still very large ; at the end
of the fifth week the space between the embryo and chorion is
divided equally between the amnion cavity and the coelom ;
at the end of six weeks the extraembryonic coelom has
disappeared.

BALTIMORE, May 5, 1896.

mM CONTRIBUTION TO THE STUDY OF VARIATION:
(SKELETAL VARIATIONS OF NECTURUS MACULATUS RAF.)

HERMON C. BUMPUS.

“Both heredity and variation are in urgent need of causal explanation.”’— Roux, Wheeler.

In this communication, based upon the comparative exami-
nation of skeletons of Necturus, an effort is made to answer
the following questions:

I. What is the per cent of homeceotic variation ! in the attach-
ment of the pelvic arch to the axial skeleton; is there true
meristic variation, and is homceotic associated with meristic
variation ?

II. Is there a ratio between the absolute length of the
animal and the number of vertebrz?

III. Why does the variation tend towards forward rather
than backward homceosis ?

IV. Can an explanation be given for the frequent occurrence
of oblique or unsymmetrical sacra?

V.-Is the position of the pelvic arch dependent upon the
ordinal position of some one segment (sacrum) of the vertebral
column, or is its position determined by the location of some
topographical point ?

VI. Are there variations in the position of the pectoral arch,
and are these correlated with variations in the pelvic arch?

VII. Are there other skeletal variations associated with
pelvic variation ?

1 Bateson (’94) states that “‘ Homeeotic variation in the spinal column consists

in the assumption by one or more vertebrz of a structure which in the type is
proper to vertebre in a different ordinal position in the series. Examples of this

are seen in the . . . occurrence of a vertebra, normally lumbar, in the likeness of
a sacral vertebra, having its transverse process modified to support the pelvic
girdle. ... In using the expression, Homeeosis, ... we may speak of the

variation as occurring from before backwards or from behind forwards, according
as the segment to whose form an approach is made stands in the normal series
behind or in front of the segment whose variation is being considered. The
formation of a cervical rib on the VII vertebra is thus a backward Homeeosis for
the VII vertebra thus makes an approach to the characters of the VIII,” etc.

456 BUMP US. [Vor. XII.

VIII. Are variations more frequent in males than in
females ?

IX. Are there anatomical grounds for the theory of verte-
bral intercalation ?

METHODS.

Until within a few months the examination of a large series
of skeletons involved the expenditure of considerable time for
their proper preparation, the destruction of valuable anatomical
material, the occasional loss of certain cartilages, and the too
frequent misplacement of disjointed parts. The truly wonder-
ful discovery of Rontgen, however, has placed in the hands of
the anatomist an economical means for the most critical exam-
ination of the bones, zz sztu, without the use of macerating
fluids or the scalpel, and a means readily applicable, without
injury to the most valuable museum specimen or even to the
living animal.

The one hundred alcoholic specimens on the comparative
examination of which this paper is based, belong to the
museums of Brown University and the Boston Society of
Natural History. Each animal, properly numbered, was bound
in a proper position to a thin piece of board, and the board with
the specimen attached was then placed upon an envelope of
black paper containing a common photographic plate. When
large plates were used as many as ten animals were exposed at
the same time.

Immediately over the table on which lay the photographic
plate, and approximately 500 mm. distant, was suspended a
Crooke’s tube of the Elihu Thomson pattern, rendered fluores-
cent by means of a Tesla coil. The exposure was invariably
five minutes. When the tube became dim it was warmed with
a Bunsen flame until the return of the brighter illumination.
The plates were developed with the ordinary “‘ pyro” developer.

The negatives, taken so as to show both the dorsal and
lateral views of the vertebral column, were so clear that the
ultimate joints of the tail could be readily counted. The
amphiccelous structure of the vertebra and the minute pores
of the bones were often exhibited with remarkable detail. A

Ni@s 2:i] THE STUDY OF VARIATION. 457

few specimens showed the opaque fish-hook, and thus betrayed
the mode of their capture.

SeEcTion I.

(1) What is the per cent of homceotic variation in the attach-
ment of the pelvic arch to the axial skeleton ?

This question was considered by G. H. Parker (96), who
based his conclusions on an examination of twenty-seven speci-
mens, cleaned by a preparator, in none of which could the total
number of vertebrze be determined. Of these specimens two
were found with oblique or unsymmetrical sacra. In nineteen
cases the sacrum was developed from the XIX vertebra and in
six from the XX vertebra. Arranged in the form of a per cent
table, Parker’s results were as follows:

The pelvis is attached to the XIX vertebra in 70% of 27 specimens.

“ “ “ “c 66 Ok XEXE “ “c 22% “OE “c
“sc “ “ “ obliquely “cs 7% “c “ “c
“ “ “ “c abnormally “ 29% “6 “

An examination of a larger number of specimens shows that
the number of variable individuals may be considerably increased.
On Plate C the pelvic appendages are represented by short
lines crossing the vertical ordinates on the abscissas of the
XVIII, XIX, and XX vertebra.

The pelvis is attached to the XIX vertebra in 64% of 100 specimens.

“ “ 6“ “ sc OE xX “c “ 28% “cE “cc
“ “ “ “ obliquely “ 8% “ “
“ “c “ 6“ abnormally “ 36% “cS ‘“

The pelvis, then, is attached abnormally in 36% (28 + 8)
of the 100 specimens. This variation should, in the light of
the 127 specimens thus far tabulated by Parker and myself, be
corrected by uniting the two sets of figures. This union gives
the remarkable variation formulated in the following table, vzz.,
an average of 35%, and offers an example of excessive varia-
tion in animals not domestic. |

The pelvis is attached to the XIX vertebra in 65% of 127 specimens.
CPOE SEC us, SER RD ie Uae Pa hfe a «“

““ “ “ “ obliquely “ 8% “6 “c
“ “ “ “ abnormally “ 35% “6 “

458 BUMPUS. [Vou. XII.

(2) Is there true meristic variation ?

Bateson (94, p. 102) states that “numerical change may be
brought about in the series of vertebrae by two different proc-
esses : first, by variation in the total number of segments com-
prising the whole column, in which case the variation is truly
meristic ; and, secondly, by variation in the number or ordinal
position of the vertebrae comprised in one or more regions of
the column, not necessarily involving change in the total num-
ber of segments forming the whole series, and in this case the
variation is homeeotic.’”’ He further states that though the
latter form of variation may be associated with the former it
is rarely possible in any particular case “to distinguish clearly
whether such a change has occurred or not,’ because the
terminal joints of the caudal vertebrze cannot be readily enu-
merated.

The application of the Rontgen rays often so clearly defines
the centers of ossification of even the terminal caudal verte-
bree that I have attempted to plot the several series on Plate C,
though in a few cases, due to the imperfection of the specimens,
tails in process of regeneration being quite frequent among the
larger individuals, the number of the last few vertebrae have
not been determined with certainty. Such specimens, five in
number, are designated with ? over the est¢mated terminal
joint.

An examination of the plotted curve of the vertebral columns
on Plate C will reveal an extremely irregular line passing from
Specimen 1, with 45 vertebra, to 2 with 44, 3 with 47, etc.
(Plate B, Specs. 1 and 2). There is, then, considerable mer-
istic variation, and the second of the questions considered in
this section is answered in the affirmative. Since it is pos-
sible to enumerate the caudal vertebrz, we are in a position
to answer the third question, vzz.:

(3) Is homeeotic associated with meristic variation ?

On Plate C the dotted ordinates indicate the specimens
which are homeeotic, and, taking the specimens in blocks of
twenty, let us see if the homceotic specimens of each block
have a greater variation in the number of vertebrz than do the
normal specimens.

No. 2.] THE STUDY OF VARIATION. 459

In the first block of twenty specimens the average number
of vertebrze for each animal is 45.7. The sum of the depart-
ures from this mean in sixteen normal specimens is 22.4, giving
in this block an average departure from the mean, for each
specimen having the pelvis on the XIX vertebra, of 1.4. The
four homeceotic specimens, however, show an average departure
of 1.6, z.¢., the amplitude of variation among the homeceotic
specimens is greater. The second, third, and fourth groups of
twenty also reveal greater amplitude of variation on the part
of homeeotic individuals, though the last group is exceptional,
the homceotic specimens showing less variation than the nor-
mal. The tabulated results are as follows:

The mean amplitude of meristic variation of the normal specimens
In the first 20 is 1.4 and of the homeotic specimens is 1.60.

In the'second'20) * 94 “ “ « a ¢ SOP sO i):
Imbthethirdy2oy 61.35, sees <c cs SiGe
inthe fourthy2o, 6 1.52) < <6) <6 ‘ “ Ge Gigs.
Imbtheffthy2o) es o1:66) 66 <6 ise Ke ‘ 5 inate,
In I00 oete37e “call the i ss STO.

Though in this table there is considerable irregularity in
the ratios, it seems to the writer that the final figures, 1.37 and
1.61, are sufficient to warrant the conclusion that specimens
having abnormally placed sacra (homeceotic) do present a con-
siderably increased meristic variation, and aside from the fact
that there are twice as many homeeotic specimens on the right
as on the left of Plate C.

Is the converse true? Do specimens presenting extremes
in meristic vertebral variation tend towards homeeosis ?

An examination of the facts in the case brings out most
striking results. There are only two specimens whose
vertebree are reduced to the abnormal number of forty-
three, and both of these are homeeotic. Of the nine
examples bearing forty-four vertebre, only 33% are home-
otic. Of the twenty-one specimens bearing forty-five verte-
bree, only 19% are homceotic. As the vertebrae now leave
the normal and tend towards the other extreme, homceosis
becomes more frequent, as will be observed by reference to
the table.

460 BUMPUS. [Vot. XII.

Of 2 specimens with 43 vertebrz, 100% are homeotic.
“ 9 “ “c 44 7 33% 66 “
“ ay “ “ 45 6c 19% “ “
6“ 18 “ “cs 46 “ 29% “ “
“ 14 “ “ 47 “ 36% “ “

a “ 16 “ “ 48 “ 44% “ “
Tt “ “ 49 ‘“ 45% Cue “
“ 5 “ “ 50 T7 60%, “ 6“
“ 4 “ “ 51 “ 50% “ “

A profile curve, or ‘curve of frequency ’”’ (Galton, ’89), illus-
trating the relative distribution of the several columns accord-
ing to their vertebral enumeration, is drawn in dots at the right
of Plate C, the length of the ordinates being determined by the
number of the specimens having respectively 43, 44, 45, 46, 47,
48, 49, 50, and 51 vertebrae. If now the respective per cents
of homeeotic individuals in each series be represented by a red
curve on the same ordinates, we will observe that as the first
curve ascends to its culminating point the second curve
descends, the ordinate 45 bearing the highest point of the
first curve and the lowest point of the second. To the right
of this ordinate the curves converge in approximately the same
way as they primarily diverged.t

The conclusion that specimens presenting extremes of meris-
tic vertebral variation tend towards homeeosis is irresistible,
and the third question of the first section is answered in the
affirmative.

Section II.

-

Is there a ratio between the absolute length of the animal
and the number of the vertebrz ?

As before mentioned, the animals are arranged on Plate C,
the smaller toward the left and the larger toward the right.
The curve of relative lengths, drawn in red, passes from the
specimen No. 1 at the left, with a length of only 193 mm.,

1 The small dotted curve in red is drawn on the following scale. On abscissa
43 a dot is placed ten squares above the line which forms the right margin,
and these ten squares are taken as representing the 100% of homeeotic individuals

having 43 vertebre. The curve then follows the respective altitudes of 33%, 19%,

29%, 36%, 44%, 45%, 60%, and 50% on the abscissas of 44, 45, 46, 47, 48, 49, 50,
and 51. I regret that the print does not make the course of this curve perfectly
clear.

No. 2.] THE STUDY OF VARIATION. 461

toward the larger specimens at the right, and of course with a
constantly increasing altitude! It will now be profitable for us
to draw upon Plate C a line which shall represent the mean of
the very irregular curve of length in terms of vertebree. If the
specimens having a larger number of vertebre are irregularly
distributed, this line will lie horizontally, but if numerical
vertebral increase tends to occur in larger animals, the line
will tend to rise from left to right. Suppose now we take the
means used in Section I for the several blocks of twenty speci-
mens, and, using them for our ordinates, we plot the curve of
mean numerical vertebral variation. We will find it to lie
between the abscissas of 45 and 48 (see dotted line on Plate C).
The curve shows a vegu/ar rise from left to right, following the
flowing curve of absolute lengths in millimeters, there being an
increase of about three vertebrze in the entire series. The
question at the beginning of this section is answered, then, in
the affirmative.

But the simple addition of three vertebrze, each less than one
millimeter in length, at the caudal end of the animal, cannot
account for the increase of over one hundred millimeters in the
total length of the larger specimens. The increase in length
is not then to be explained by the addition of new seg-
ments behind, after the method of growth of developing
arthropods and worms, but is interstitial, the individual verte-
bre increasing in length with age ; though of course it does
not follow that length is an absolute criterion of age. Under
varying conditions growth may be more or may be less rapid.

Whether the slightly increased number of vertebrze is due
to multiplication of vertebrz after sexual maturity, or is an
index of unusual embryonic vitality, or is predetermined in the
egg, I will not now attempt to answer, though the application
of the X-rays to the developing young will definitely settle the
first of these, vzz., the question of vertebral multiplication
after sexual maturity, and the frequent occurrence of smaller
and average specimens possessed of nearly the maximum
number of vertebrz might indicate early extraordinary vitality,
predetermination, or both.

1 The red figures on the right of the plate indicate millimeters.

462 BUMPUS. [VoL. XII.

It will be observed that the rise in the curve of relative
lengths is quite abrupt until the specimens are 252 mm. long,
when the ascent is gradual until the specimens are 330 mm.
long, from which point to the longest specimen in the series
the ascent is again rapid. The curve then shows that the
larger number of specimens are between 252 and 323 mm. in
length. If the lengths are examined in profile, as shown by the
red curve on the right of Plate C, it will be observed that the
culminating point of constancy is around that specimen which
has a length of 260 mm.

As the sum total of opportunities for death are directly
dependent upon the length of existence, it is not surprising
that the upward trend of the curve is considerably more abrupt
than the descent of senescence. I am inclined to look upon
the absence of specimens measuring between 245 and 255 mm.,
causing the depression in the ascending curve, as accidental,
though it is barely possible that the depression signifies some
change in the rate of growth, consequent upon the approach of
sexual maturity (Minot, ’91); or it may be dependent upon the
season of capture. I am also at a loss to explain the depres-
sion in the curve of descent in the neighborhood of specimens
having a length of 280 mm.

Before leaving this section it should be noted that the pro-
portion of homceotic specimens is not the same among the
shorter as among the longer specimens. Of the first fifty
specimens, but ten are homeeotic, while of the second fifty
twenty-five present cases of forward homeeosis.

It has already been shown that the mere addition of a few
terminal caudal vertebrze cannot alone account for the increased
size of homaeotic specimens. The fact, then, appears, if our
specimens are not exceptional, that this homceotic departure is
a character of increasing frequency among such animals as
have met with success in the struggle for existence. If this
is the case, the present species of Necturus may be considered
as undergoing a process of “mutation,” since the causes of
transformation are acting with considerable “uniformity upon
large numbers of individuals” and “in a definite manner”
(Scott, ’94).

No. 2.] THE STUDY OF VARIATION. 463

Section III.

Why does the variation tend towards forward rather than
backward homeeosis ?

Bateson says ‘the attempt to apply to numerical variations
the conception of variation as an oscillation about oze mean is
not easy, difficulty arising especially in regard to the choice of
a unit for the estimation of divergence; ... to judge from the
scanty indications available, it seems that in cases of numerical
change, variations to numbers greater than the normal number
and to numbers less than it, are not generally of equal fre-
quency. Probably no one would expect that they should be
Soi) (9450571).

It has already been noted that in 28% of the one hundred
salamanders examined, the sacral ribs are borne on the XX
vertebra, while in not a single specimen are they symmetrically
borne on the XVIII. Most certainly numerical variations to-
wards a greater number, than the normal, of pre-sacral vertebrze
are more frequent than variations towards a smaller number.

The positions of the oblique or unsymmetrical sacra also
indicate pre-sacral increase. In only one case among the
eight examples is the territory of the XVIII vertebra invaded
(Plate A, Spec. 62), and this invasion marks the only en-
croachment upon the domain of this vertebra in a total of 127
specimens examined by Parker and myself,—an interesting
fact when it is considered that the XX vertebra, no further
removed morphologically, presents a total of 35 invasions.

Of course the easiest way of disposing of this question is to
look upon the variation as atavistic. The “ancestral type”
was possessed of a larger number of pre-sacral vertebrz, and the
“mud-puppy ” of to-day, by its anatomical variations, kindly
indicates, in approximately 35% of its representatives, the
character of its quasi-progenitor.

But if 35% of the present individuals tend towards pre-sacral
multiplication, z.e., tend to assume ancestral characters, we must
not deny to these potentially progenitorial individuals the same
tendency to vary that is possessed by their young, z.c., 35% of
the thirty-five should have an added increase of the pre-sacral

464 BUMPUS. [Vor. XIL.

region, and thus about twelve specimens should have sacral
ribs on the XXI or XXII vertebra. Such a disposition of the
sacral ribs does not occur, though its absence may be attributed
to a presumed tendency of the parent form towards “ fixity”
or ‘specific stability.”

Rejecting the theory of atavistic variation, I think the for-
ward homeeosis of the sacral vertebra may be explained on
other grounds.

In the first place, the origins of the processes that bear the
ribs, as will be noted by reference to Pl. A, Spec. 87, arise not
from the middle, but from the posterior end of each vertebra,
making it a much shorter distance from the transverse process
of the XIX to the territory of the XX, than from the trans-
verse process of the XIX to the territory of the XVIII trunk
metamere. It will thus seem more probable that variations
occurring in the course of ontogenetic development will fall on
the side of nearer proximity ; and the Anlage of the limb, com-
pounded though it may be by contributions from several trunk
segments, will depart less from its normal position when it
finally falls within the territory of the XX than when it falls
within the territory of the XVIII vertebral segment. This
explanation rests upon the assumption that the territory of
the XX metamere once invaded, the secondary processes of
growth within that segment will arrange for the accommoda-
tion of the sacral rib at its normal place, even though its
Anlage was first located in the anterior portion of the segment.

A third and possibly valid explanation is offered by anatomy
alone. The XVIII, XIX, and XX segments are not, in the
adult at least, of equal length. The Anlage of the limb would
be obliged to vary considerably more, in a linear direction, in
order to influence the XVIII than it would to influence the
XX vertebral segment.

Though these two possible explanations of the more fre-
quent occurrence of forward homceosis are here given, they
are both dependent upon an assumed variation in the position
of the Anlage of the appendage with respect to some one

1 The occurrence in certain fishes of pelvic fins, which lie in front of the pecto-
ral fins, is in this connection of especial interest.

No. 2.] THE STUDY OF VARIATION. 465

“normal” vertebral segment, vzz., the XIX. We assume,
moreover, that when the relative position of the appendage is
once determined in the embryo of Necturus it is determined
for life ; since there is no evidence to show that there is a
“migration’’ of the parts of the pelvic region after the first
rudiments are laiddown. Rosenberg (’76), however, claims a true
pelvic migration for Homo, involving an emancipation of sacral
elements for the production of coccygeal vertebrae, and Credner
(86) claims a distal shifting of the pelvic arch in the fossil
amphibian Branchiosaurus. There are grounds, moreover, as
will be shown in Section V, for the belief that the very process
of local differentiation of the embryonic cells, for the final pro-
duction of the appendage peripheral to the vertebral axis, occurs
in a position in no respect dependent upon the position of any
one vertebral segment, but dependent rather upon the general
proportions of the embryo asa whole. The determination of
the loci of the successive vertebrze and their early differentia-
tion exerts no determining influence on the position of the
appendages. But the Anlage of the appendage once deter-
mined, influence from it will direct the growth of proper sacral
elements in the nearest vertebral segment, be it the XVIII,
DEE On, XX.

It should furthermore be noted that when the Anlage of
the appendage does not fall within the territory of the normal
metamere (XIX), its vertebra, which has presumably been pro-
ducing sacral ribs in two-thirds of the specimens since some
remote geological period, does not show the slightest sign of
sacral differentiation, but is exactly like the other neighboring
trunk vertebrz.

Section IV.

Can an explanation be given for the frequent occurrence of
oblique or unsymmetrical sacra ?

In eight per cent of the specimens examined the sacrum is
not a single vertebra, but is composed of two halves, each
belonging to different metameres (Plate A, Spec. 62). The
legs, moreover, in these specimens, leave the body at points
not directly opposite.

466 BUMPUS. [Vor. XII.

Those who attempt an explanation of this phenomenon on
the principle of intercalation, excalation, or of pelvic migration
meet with most provoking difficulties. Why should only one
lateral half of a vertebra be intercalated, giving rise to a body
segment which bears a single rib on one side and two ribs on
the other? Or why should the process of the formation of
sacral ribs involve portions of two vertebrz rather than one,
and thus produce an asymmetrical sacrum?

The difficulty is partly obviated if we admit that the differ-
entiation of the sacral vertebra is the result of centripetal
influence exerted by the growing Anlage of the appendage. If
the first rudiments of the appendages are not laid down exactly
opposite each other, and there are many reasons, as, for exam-
ple, the curvature of the embryo, the pressure of neighboring
eggs, etc., why their primitive positions might vary, an unsym-
metrical or oblique sacrum would result.

It is a remarkable fact that of the eight unsymmetrical
specimens, all but one, No. 62, have the left half of the sacrum
in advance of the right, z.e., the axis of the sacrum is sinistro-
dextral.

I have been quite unable to find any valid explanation for
this peculiar uniformity, unless it is based upon the curvature
of the embryo. Dr. Whitman has informed me that the young
of Necturus, as it lies upon the surface of the egg, is curved,
but I have not seen the embryos or larve, and I ‘do not know
whether the curve is lateral or dorsal-ventral, or even if it is
fairly constant in its trend. An attempt to correlate the posi-
tion of the sacral axis with the slightly asymmetrical position
occupied by the paired viscera has proved futile.

SEcTION V.

Is the position of the pelvic arch dependent upon the ordinal
position of some one segment (sacrum) of the vertebral column,
or is its position determined by, and is the sacrum the resultant
of, the location of some topographical point ?

I do not know that an attempt has ever been made to com-
pare the distribution of the vertebra and the location of the

No. 2.] THE STUDY OF VARIATION. 467

appendages with the general dimensions of the vertebrate body,
though the radiographs of Necturus, procured by the X-rays,
make the necessary process of measurement extremely simple.

For a standard of measurement I have considered the dis-
tance from the first to the XXX vertebra, no matter what the
actual length of the animal may be, to be represented by 100;
the XXX vertebra being selected because the terminal portions
of the animals are liable to injury. I have taken the inter-
vertebral space between the XIX and XX as the locus of a
variable, and I find by making accurate measurements on the
negatives that the position of this intervertebral space varies
in different animals in a very appreciable way. In certain ani-
mals the trunk, the part lying anteriorly to the selected inter-
vertebral space, is relatively longer, while in others the caudal
portion is relatively longer. The distance from the XX to
the XXX vertebra may in some specimens (Nos. 4, 67, 95)
be only 29% of the entire measured length, while in others it
may be as great as 35% (Nos. 58, 61, 74).

On Plate C these relative measurements of the several
specimens are given, each ordinate representing the length
from the first to the XXX vertebra, and the upper of the two
black lines that cross the plate represents the fluctuating inter-
vertebral septum between the XIX and XX vertebre. In by
far the larger number of cases the caudal region is one-third,
33%, of the length. Specimen No. 1 has a trunk slightly
longer and a caudal region correspondingly shorter than the
average. The trunk of No. 2 is shorter and the tail longer.
No. 3 is like No. 2. No. 4 has an uncommonly long trunk
and short tail, etc.

If this curve is followed across the plate, it will be noted
that the first nineteen vertebraze may suffer regional expansion
and contraction, with correlative contraction and expansion
of the caudal region, and to such an extent as to give an
amplitude of variation to the dividing line between the XIX
and XX vertebre, amounting to 585 (35% minus 29%) of
the average distance between the I and XXX vertebrae. The
amplitude of variation in the column itself, absolutely inde-
pendent of the limbs, is sufficient, then, to include within

468 BUMPUS. [Vor xil

itself ;$, or ;4 of the measured distance. But the verte-
bree in the region of the selected intervertebral spaces are, in
length, considerably less than ;4, the measured distance;
so that the oscillations of the dividing line between the XIX
and XX vertebrz would be such as to carry it back and forth
over a distance greater than one vertebra anteriorly and one
posteriorly.

If we should string a series of thirty short segments of rub-
ber tubing upon a cord, and then, after placing a card between
the XIX and XX, subject the entire series to slight pressure,
we would have what might roughly correspond to the column
under consideration. The card, representing the selected in-
tervertebral space, can be forced forward or backward over a
certain amplitude, according to the compressibility and elas-
ticity of the rubber rings. Suppose, while the card is at rest, we
hold two fingers of the hand in such a position as to represent
the rudiments of the appendages, lying on either side of the
nineteenth ring. Now when the card is forced anteriorly the
fingers lie no longer opposite the XIX, but opposite the XX
vertebra. A forward homeeosis has taken place illustrative
of what happens in a large per cent of the specimens of
Necturus.

But do our data warrant the assumption that with the com-
pression of the anterior vertebrz there is a concomitant varia-
tion in the position of the hind limbs? When specimens tend
toward elongation of the caudal region, do the legs appear on
the XX, and when the reverse obtains do they tend to arise
from the XIX vertebra?

A further examination of the plate will answer in a most
conclusive way.

Specimens No. 58, 61, and 74 have the greatest elongation
of the posterior vertebrze, and in every one of these specimens
the sacral ribs are borne by the XX vertebra.

In other words, in all the specimens presenting extreme
compression of the anterior vertebrz there is concomitant
variation in the position of the hind limb.

Arranged on the abscissa of 34% (Plate C) are thirty-two
specimens, and, if our theory is to be supported, a less number

No. 2.] THE STUDY OF VARIATION. 469

of specimens should present homeeotic variation than was the
case with the three just mentioned, belonging to the abscissa
of 35%. A reference to the plate will show that a less number,
only fifteen specimens (46%), have homeeotic sacra.

A still smaller number of homeceotic specimens should be
found on the abscissa of 35%, and, in fact, of the forty-four
specimens arranged on this line but thirteen present cases of
forward homeeosis, 29%.

An even greater diminution in the number of homceotic
specimens should be found on the abscissa of 32%, where we
count fourteen specimens, only two of which, barely 14%, are
homeeotic.

On the abscissa of 31% among four examples there is only
one that is homeeotic, and among three examples on abscissa
of 30% there is also only a single homeeotic individual.

There are thirty-five specimens below the mean abscissa of
33% and twenty-one specimens above. The former exhibit
eighteen cases of forward homeeosis, 2.¢., 51% are homceotic ;
the latter exhibit only four, z.¢., 19%.

This effort to prove that the sacral ribs are developed as the
result of centripetal influence bears directly upon the proposi-
tion of Bateson “ that individuality should not be attributed to
members of a series which has normally a definite number of
such members,” for it is shown that the normal XIX verte-
bra owes its specialized form to the molding influence of sur-
rounding parts, and not to some inherited directive influence
upon one particular vertebra. It shows that a “redistribution
of differentiation’ may and frequently does take place.

The appearance of the appendage at a definite topographical
point, without respect to the location of certain segments of
the neighboring axial area, is in harmony with the view of Pro-
fessor Whitman ('93) that the real unity, both in development
and in adult stages, is the organism as a whole, “that the
organism dominates cell formation.”

The view here advanced also receives direct support from the
facts of embryology, as shown by the following extract from
Weidersheim’s ‘“Grundriss” ('93): ‘Mit anderen Worten —
und ganz dieselben Gesichtspunkte ergeben sich auch fiir

470 BUMPUS. [VoL. XII.

Reptilien, Vogel und Sauger, und sie gelten auch ebenso gut
bei allen Vertebraten fiir die hintere Extremitat —es handelt
sich um ein festes, die ganze Vertebraten-Reihe beherrschendes
Gesetz, dass der Anstoss zur Entwicklung des Gliedmassen-
skeletes stets von der Peripherie ausgeht, und dass sich die
centralen (Giirtel-) Theile erst secundar unter dem formativen
Einfluss der freien Gliedmasse entwickeln.”’

One of course feels some regret that there is not a suffi-
cient number of examples of backward homeeosis for the basing
of a conclusion in regard to the association of backward homee-
osis with trunk elongation. The only specimen which shows
this unusual homeeotic tendency is No. 62, and on a reference
to Plate C it will be seen that its XIX intervertebral septum
is located on the mean abscissa of 33%.

Section VI.

Are variations in the relative position of the pectoral arch
associated with variations in the relative position of the pelvic
arch?

The pectoral arch, located near the anterior end of the body
and closely associated with the head and visceral skeleton,
would be expected a priorz to show a lesser amplitude of varia-
tion in respect to the segments of the vertebral axis than is
shown by the pelvic arch. In front of it are only three verte-
bree, and the limits of vertebral elasticity, if I may use the term,
are correspondingly restricted.

On the other hand, the shoulder girdle is not articulated to
any one vertebra, and those intimate relations between bones
which we have been taught are so potent in determining their
position and character, do not here exist, and were it not for
the tensile influence of the myotomes the enclosed vertebrze
might slide back and forth through the arch ad /zbztum.

Since the girdle thus lies imbedded only in muscle tissue,
its normal position in relation to the embraced vertebrz is
rather difficult to determine from specimens which have been
killed without special care. My observations, however, con-
vince me that the halves of the arch do not always occupy the
same position with regard to the vertebral segments.

No. 2.] THE SLOUDV OF VARTA TION. 471

The small ossified scapulze are often very clearly shown on
both the lateral and dorsal radiographs, and it is not a difficult
matter to get an approximate idea of their relation to the neigh-
boring vertebree. They ordinarily lie opposite the middle of the
fourth vertebra, though they may lie opposite the intervertebral
space which separates the third vertebra from the fourth, and
less frequently opposite the space which separates the fourth
from the fifth. They are represented on Plate C by the small
cross-lines occurring between abscissas three and four. Let
us examine the extreme cases, and see if there is a correlation
between pelvic and pectoral variations.

Specimens 12, 28, 30, 31, 35, 38, 39, 43, 52, 73*, 79%, 92,
94* all show the scapulze to lie opposite or near the third inter-
Wertebral space, while specimens 23, 24%, 37*,.53*5 58%,00™
71, 77, 78*, 83*, 89*, 90, 95*, 98 show a tendency towards the
approach of the fourth intervertebral space.

Now of the first series of thirteen specimens there are only
three cases (indicated by an *) of forward homeceosis of the
sacrum. In the second series, however, of fourteen specimens,
there are nine cases of forward homceosis.

When one stops to consider that the total number of homee-
otic specimens is approximately only one-half the number of
normal specimens, and that the chances of scapular variation
are hence twice as liable to gather around normal individuals,
the figures in the previous paragraph have an increased
significance.

We can conclude that variations in the relative position of
the pectoral arch ave associated with variations in the relative
position of the pelvic arch, and that both variations are causally
connected with some common factor. I believe that factor to
be the variation of the relative length of different vertebral
regions, as considered in the fifth section.

Section VII.

Are there other skeletal variations associated with pelvic
variation ?

A critical examination of the skeletons of Necturus, made
for the purpose of detecting a possible tendency on the part of

472 BUMPUS. [ VoL. XII.

homeeotic individuals towards variations in other directions,
leads to the conclusion that homeceosis is only an index of
general instability on the part of the skeleton as a whole.

(1) This principle was illustrated in Section VI, where it was
shown that variations in the position of the sacral vertebra are
accompanied by variations, in a definite direction, in the pectoral
region.

(2) The principle was also illustrated in Section I, when it
was shown that homoeotic specimens tend towards numerical
increase in the number of vertebral segments. — Before passing
on to other possible illustrations, let us examine this present
phenomenon more in detail.

According to Bateson’s law, forward homeeosis, involving one
vertebra, should yield a column of oze more than the normal
number of vertebrze. But if we examine the curve of lengths
in terms of vertebra, on Plate C (the lower of the curves
drawn in black), we shall find that whereas the normal speci-
mens average only 45 vertebrz, the homceotic average not one
more, 46, but ¢wo and ¢hree more, 47 and 48, and that the
amplitude of variation, moreover, is much greater among the
latter than among the normal specimens. It is clear, then,
that there is an added increase in the variation of the vertebrze
among homoeotic specimens.

(3) A third illustration of the principle of general variability
was given in Section V, where it was shown that abnormalities
in the relative lengths of the anterior and posterior portions of
the body tend to gather about homceotic individuals.

(4) Dr. Parker ('96) has made use of the first haemal arch in
his comparisons of the vertebral columns of Necturus, showing
that its position is subject to considerable variation. Let us
see if homceotic specimens present greater or less variation in
the position of this arch than do the normal examples.

Of sixty-three normal specimens giving satisfactory radio-
graphs of the first hamal arch,

10 examples (16%) have the arch attached to the XXII vertebra.
51 “c (81%) iT “ “ “ CO DOCU ue “
7 “ 707 74 (73 (73 6c“ 66a ee XOXO, “

a (3%

and of thirty-five homceotic specimens,

Ai

No. 2.] THE STUDY (OF VARIA TION. A738

1 example (3%) has the arch attached to the XXII vertebra.
31 examples (88%) have “ “ se DO DO. L
Supper sms 5 er FONT,

Our figures show that in both normal and homeeotic speci-
mens the point of constancy is the XXIII vertebra, and,
moreover, that the homeceotic specimens are slightly less liable
to vary than the normal, the former presenting 88% of cases of
stability against 81% presented by the normal specimens.

It would seem that the production of the first hamal arch is
then a function of the X XIII vertebra, a function of the axial
portion of the embryo, and that the development of the hamal
arch proceeds entirely independent of the position of the
neighboring appendages, for in 81% of the normal specimens
there are three intervening vertebrz between the sacrum and
the first hamal arch, while in nearly all of the homeeotic speci-
mens there are but two.

Of course if we assume that the normal position of the
first hamal arch is on the fourth vertebra behind the sacrum,
we can then show that the homoeotic specimens present a
greater range of variation than normal individuals. But this
assumption would be unjustified.

(5) Specimens may frequently be found whose caudal verte-
bree are provided with abnormal processes. These abnormalities
may appear as bi-lobed or double neural plates, or double neural
spines. Similar malformations may affect the hamal arch
(Plate:B, Spec:70).

If our belief in the general variability of the skeleton is to be
justified, these minor variations should be of more frequent occur-
rence among the homceotic than among the normal specimens.

Among thirty-six homceotic specimens there are twenty
which present vertebral variations in the caudal region, while
of sixty-four normal specimens there are only twenty-six that
show the same local variability.

It should not be forgotten that imperfect processes of regen-
eration may be responsible for the large number of caudal
malformations, though this does not explain why homeeotic
individuals should be more frequently captured during the
regenerative process.

A74 BUMPUS. iVor. Xai

I wish that alcoholic “ mud-puppies ” presented some striking
coloration, like the spots of Diemyctylus or Amblystoma, so
that one might detect any tendency towards color variation
in specimens presenting skeletal instability ; few animals, how-
ever, are more alike externally than preserved specimens of
Necturus.

Section VIII.

The question may arise: Are variations in the position of
the pelvis correlated with sex, or, if not directly dependent
upon the sex, does one sex show a greater tendency towards
skeletal variation than the other?

It is frequently the case when a large number of individuals
of a single species are collected that the males predominate,
but in the present collection, as will be seen by counting the
indices of sex at the top of Plate C, there are thirty-seven
males and sixty-three females. There should be, then, if varia-
tion is equally distributed among males and females, nearly
twice as many abnormally placed pelves among the females as
among the males. In fact, however, of the thirty-six homeeotic
specimens, only nine occur among the males, while twenty-seven,
three times as many, occur among the females.

It is, I think, quite generally believed that males are much
more subject to variation than are females. Montgomery (96)
in his recent paper on “Organic Variation”’ states that “the
dimensions of birds are more variable in the males than in the
females,’’ and that “it is not impossible that in birds, as in man,
the female may be more conservative and less progressive than
the male.” This of course recalls Brooks ('83) and Geddes and
Thomson ('90), though we should not forget that males are fre-
quently larger among birds and mammals, and thus offer an
increased amplitude of variation, and that males are, moreover,
frequently provided with many and variable secondary sexual
characters whose development is directly dependent upon the
activity of the proper sexual glands, and are hence subject to
considerable variation.

Among Amphibia, however, it has been many times shown
that the male may assume certain of the duties which are

No. 2.] THE STUDY OF VARIATION. A75

generally considered to be attributes of the female, and it is
barely possible that this affectation by the male is only an
expression in habit of a general condition of physical conser-
vatism.

SeEcTIon IX.

Are there anatomical grounds for the theory of vertebral
intercalation ?

If we accept the theory that there is a fixed intimate inter-
relation between the appendage, girdle, and some one support-
ing vertebra ; that all are parts of one developing area, parts
bound together by some intangible law of correlation, the
occurrence of the XX vertebra as the support of the pelvic
appendages can be explained only by the theory of vertebral
intercalation of Albrecht and Baur.

If I understand the theory correctly, it is the function of one
particular vertebra— or where more than one vertebra enters
into the formation of the sacrum, of certain particular vertebrae—
to give attachment to the pelvic arch, and no matter what sec-
ondary alterations may occur before or behind, this vertebra or
these vertebrz will tenaciously hold to their prescribed and
inherited function.

After examining the descriptions of several examples of in-
tercalation given by Baur, in an earlier number of this journal
(91), I fail to see that they necessarily support his theory.

In the first place, Albrecht and Fiirbringer are acknowledged
to differ in their interpretation of the Belgian python, and the
mention of vertebral and costal asymmetry in Pelamis and
Cimoliasaurus does not render the intercalation theory more
probable.

The case of the gavial, described by Baur and considered by
Parker to show “very conclusively that, in place of one verte-
bra, two or parts of two may arise,” is equally inconclusive.
Baur writes: “It is well known that the typical number of the
pre-sacral vertebrz in the living Crocodilia is twenty-four; there
are two sacrals; the first caudal is peculiar by being convex.
In a specimen of Gavialis gangeticus I found twenty-five pre-
sacral vertebrz. As in all living crocodiles, the first caudal

476 BUMPUS. (Vou. XII.

vertebra is bi-convex; but in this case it is the twenty-eighth,
in the other the twenty-seventh. Is it not evident, therefore,
that at some place between the occipital condyle and the first
caudal a new vertebra has been inserted? By careful compari-
son I find that this new vertebra has been intercalated between
the ninth and tenth.”

Dr. Baur does not tell us how he was able to pronounce this
an intercalated segment, leaving it to be presumed, however,
that a new vertebra possesses certain diagnostic features.
Moreover, if perfectly formed sacral ribs may appear through
discontinuous variation on the XX vertebra of the “ mud-
puppy,” I see no reason why the new first caudal of the abnormal
gavial should not produce a bi-convex rather than a proccelous
centrum.

If we could only have the relative trunk and tail measure-
ments of these specimens with abnormal vertebrz, and could
compare them with the normal, we could soon tell whether a
strange vertebra had been wedged into the regular series,
making the trunk abnormally long, or whether the arrest in
the development of a trunk vertebra had not resulted in the
abortion of certain anterior vertebra, drawing into the sacral
region certain primarily caudal elements.

The Helodermas mentioned by Baur, on page 335, would form
excellent material for such a comparative study, since the four
specimens exhibit four variations in the numerical position of
the first caudal vertebra. But in the absence of a large number
of Helodermas, let us examine again our Necturus material
and see if we cannot find vertebre that might be looked upon
as intercalated.

Specimen No. 46 shows a slight reduction in the lengths of
the fifth and sixth vertebra, and specimen No. 61 shows a
similar reduction in the twelfth and thirteenth. In both cases,
the posterior of the affected vertebrz is slightly smaller than
the anterior.

If during the process of embryonic growth there was a dis-
turbance of the regular formation of the body segments, we can
easily see that this disturbance might have resulted in the
interruption of the growth of certain pre-sacral vertebre. The

SS = = —_—
a ee

2S SS

—T

Sa

No. 2.] THE STUDY OF VATA TION. ATT

general proportions of the respective vertebral regions would
thus be altered, and a tendency would result for the XIX
segment to be laid down in a relatively more anterior position.
The sacrum, in other words, would tend towards forward
homeceosis. The two specimens just mentioned should, then,
present forward homeeosis, and both do.

On Baur’s ground, also, both specimens should show forward
homeeosis, but on his ground that portion of the body lying in
front of the normal XIX segment should show, in addition,
an increase in absolute length proportional to the numerical
increase of the vertebra. In the specimens at hand such a
proportional increase in the length of the trunk region does
not take place. Specimen No. 46, though possessed of an
additional pre-sacral vertebra, has its sacrum, so far as the
general proportions of the body are concerned, in the normal
position, z.¢., on abscissa 33%, as will be observed by reference
to Plate C; while No. 61, far from having a longer pre-sacral
region, has this region remarkably short ; indeed, in the one
hundred specimens of which proportional measurements have
been taken, only two show an equal amount of pre-sacral com-
pression.

Again, according to the theory of intercalation, if the first
hzemal arch occurs normally on the XXIII vertebra, four
joints behind the sacrum, on the occasion of the introduction
of an additional pre-sacral vertebra, the pelvis should be simply
forced backward with all its chattels. This, however, does not
occur, —the hzmal arch does not change its position from
the X XIII vertebra.

On the theory of regional compression and expansion, which
we have advanced, the hemal arch, a function of the axial
rather than of the appendicular areas, should tend in forwardly
homoceotic specimens to approach the pelvic arch, z.e., in homee-
otic specimens the number of vertebrz intervening between
the sacrum and the first hamal arch should tend towards
reduction. This is what actually occurs. (See Section VII,
P. 473.)

The assumption that the embryonic pelvic girdle travels
backward and forward over a fixed vertebral column has been

478 BUMPUS. (Von. XII.

questioned both by Bateson and Parker, though believed by
Rosenberg and Credner. With Parker one can agree that the
“sacral region has the power of developing sacral ribs at sev-
eral points on both right and left sides.” But Parker gives us
no reason for the abnormal position often taken by certain
sacral ribs, and does not explain why a particular vertebra (or
vertebrae) produces sacral ribs, and why others, potentially able
to produce sacral ribs, produce ribs of the ordinary sort. If
several vertebrze are endowed with this power, why do we
never find nicely formed sacral ribs passing out into the sur-
rounding tissue as if searching for some ilia with which they
might articulate, and why are the ilia not occasionally attached
to ribs of the ordinary triangular type?

The sacrum, pelvis, and appendages are not intimately asso-
ciated parts of the body that represent one complete whole, and
the definitive location of the sacrum is probably due to cen-
tripetal influence derived from the budding appendage. Inter-
calation in the sense of the introduction of new segments does
not take place, and what have been given as examples of inter-
calation are probably only imperfectly formed body segments.

BROWN UNIVERSITY, PROVIDENCE, R. I.
Nov. 3, 1896.

No.

"76
83
86

89
89
91
91
93

93

94
94

'94

96

'96

2.] THE STUDY OF VARIATION. 479

BIBLIOGRAPHY.

ROSENBERG, E. Ueber die Entwicklung der Wirbelsdule und das
centrale Carpi des Menschen. Morph. Jahrb., Bd. 1, p. 83.

Brooks, W. K. The Law of Heredity. Boston. 1893.

CREDNER, H. Die Stegocephalen aus dem Rothliegenden des Plauen-
schen Grundes bei Dresden, vi. Zeztsch. fiir d. deutsche geol.
Gesellschaft, Bd. 38, p. 576.

GEDDES AND THOMSON. The Evolution of Sex. London. 1889.

GALTON, FRANCIS. Natural Inheritance. London. 1889.

Baur, G. The Intercalation of Vertebre. /ourn. of Morph., vol. iv,
Pp. 331.

Minot, C. S. Senescence and Rejuvenation. Journ. of Physiol., vol.
xil, p. 97.

WHITMAN, C. O. The Inadequacy of the Cell Theory of Develop-
ment. Bvzological Lectures. Woods Holl. 1893.

WIEDERSHEIM, R. Grundriss der vergleichenden Anatomie der Wirbel-
thiere. Jena. 1893.

BATESON, W. Materials for the Study of Variation. London. 1894.

Scott, W. B. On Variations and Mutations. Am. Journ. of Scz.,
vol. xlviii, p. 355.

Roux, W., AND WHEELER, W. M. The Problems, Methods, and Scope
of Developmental Mechanics. Bzological Lectures. Woods Holl.
1894.

MontTGOMERY, T. H., JR. Organic Variation as a Criterion of
Development. /ourn. of Morph., vol. xii, p. 251.

PARKER, G. H. Variations in the Vertebral Column of Necturus.
Anat. Anz., Bd. 11, p. 711.

480 BUMPUS.

EXPLANATION OF PLATE A.

The figures are printed direct/y from the original radiographs, and, so far as
size and general structure are concerned, are faithful reproductions; but the
exquisite details and sharp lines of the originals are unfortunately obscured.

SPECIMEN 62 is peculiar in that the pelvis is oblique. The figure also shows
the location of the scapule. The animal was placed in a supine position upon
the photographic plate.

SPECIMEN 62*, the same animal seen from the side.

SPECIMEN 87. This figure shows a symmetrical pelvis, and indicates the
position of the scapule.

Journal of Morphotlog Va eA

pctmen No, 62.

|
|

. ee Ane
crmen No. 62.

-tmen No. 87%

. Pl. A.
: Journal of Morphology Vol. XI.

Specimen Na. 62.

PE an

LS

~

Na

Specimen No. 62°

oy
Speermen Na 8&7

482 BUMPUS.

EXPLANATION OF PLATE B.

SPECIMENS 1 and 2, the smallest in the series, gave, on the original photo-
graphic plate, the total number of vertebrae with remarkable clearness. They
have greatly suffered in the process of reproduction.

SPECIMEN 79 is especially interesting in that certain of the caudal vertebrz
are provided with double neural and hemal spines.

Journal of Morphology | Fl. B.

\
)
pectmen No,
}
)
}
}

}
j
{
I

pectmen No, 2

pcimen No, 79.

PLB.
al of Morphology Vol. XI.

wrtee

Specimen No 1

Specimen Na 2

=—s

Specimen No. 72

aie ‘i
sh

; a
a ry

ee

eee

ee Te eee ee

484 BUMPUS.

EXPLANATION OF PLATE C.

The vertical lines, ordinates, indicate the several specimens from I to Ioo.
Entire lines represent normal, broken lines represent homceotic specimens. The
numbers at the left, in ascending order from 1 to 55, indicate the number of ver-
tebree and the ratio of the length from the XXX to the XX, and XIX to the I
vertebra.

The irregular line, crossing the plate transversely from 45 on the left, repre-
sents the length of the animals in terms of vertebra, and the more regular dotted
line follows its mean course.

The irregular line extending across the upper part of the plate represents the
relative variation in the ratio of tail-length to body-length.

The red curve, running obliquely across the plate, indicates the variation in
absolute length, and is represented in profile on the right. The figures in red
indicate millimeters ; thus, specimens 71 and 72 are 308 mm. in length.

oO
«©
road

320

310
0

Snes NEE SE EN ROE Uae
Rese eeealed Stele ete pe
DEUS Gees 6oe0r 4 SECU SSeS ae
BBERS Seo e000 Toni oe pee e see ocde
=a ee oe SS aan cae
Eitan tassios BEERS na
rN
i

240
230

Eee ceeaoge
BEER EERE EEE EEE REECE REEL
ca i eo ESO Gamers HbA Tease sesewe
57 et | | a

a a ie]
ERM SSESS TER eae eesesee

peta apn S888

Nae

—

Srysstastsntatt

SHREERSEERaN

EEE EET

BERSREBEEEOS| EaGYs
SISEEGUES
BERS enpea — i
Hebei Boe eee
BESas BES Soe aaoe V0 econ
BEEEEE SES SSSR SReS SRS wee
BE SESee
Baas
BESBSae

Sarat seseree sores a

SE GOG SMe ae ale

i) te asf 7 ea asa

ERESESBeRu4e

tt fe |
ce lad
Bh. SUS BEERS SESSee SUEeRaeee

pee EPEAT ese it [eS fa
Bo EER’ Sees
Poo SERS AN SSD WEARERS eS owe

aa SSS E SE SGRees GRR 2a eee

9 TNT a
05 SRDS Sees
ptt fA DEBBEES se SeSse

SS ee eee sels eset
ERSEERaee

DERE SER (eee es Ose eseoeeee
EEG ARee TAGs Geese Ios Ses deel ae CaS AoE SS aeRO ASSES EL
’ [| sa (| a | | i

Sia eieteieioe

70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100

men

fie

wi

séesnesse

SKesssssltsars

Hnerucvlss

Journal of Mo
dcodcgd9

Ca ee ay
Pea, 18: 9
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

rphology, Vol. XII.

odddogdQgdcggdcQgdodod
9 Pato od edodo 20 Fog Pode Sho ug Cog de vada roo ddo9999cgg999 Crrurreeriririeri
| ‘ 29

HEE

{

“>

E Tf
‘a a al ye
iE a]
|
}_ } }
L a Batata} alepahatspals
ABE zl 1508
HE va ry
e a
ial
HEE
BE
aE
+H
" on is
+
o a lta aid —t

26 27 28 29 30 31 32 33 34 35 36 37 38 39 40

T

41 42 43 44 45 46 47 48 49

SO 51 52 53 54 55

56 57 58 59 60 61

SB BS Ea
ae

62 63 64 65 66 67 68

69 70 71 72 73 74 75 76 77 78 79 80 81 82 &:

i

ea

+

3 84 85 86 87 BB B89 90 91 92 93 94 95 96 97 9B 99

8

390

380

370

350

320.

310

280

230
,

210

200

> ye
. ' “ ‘ =
-
*
\
.
.
~
\
‘
4
« “>
=
\
“
.
< x ”

JOURNAL

dy

“se

MORPHOLOGY.

EDITED BY

C. O. WHITMAN,
Head Professor of Biology, Chicago University.

WITH THE CO-OPERATION OF

EDWARD PHELPS ALLIS.
Milwaukee.

Vor. XII. Marcu, 1897.

BOSTON; “U.S: A. s
GINN & COMPANY.

AGENT FOR GREAT BRITAIN: AGENTS FOR GERMANY: AGENT FOR FRANCE:

EDWARD ARNOLD, R. FRIEDLANDER & SOHN, JULES PEELMAN,

37, Bedford Street, Strand, Berlin, N.W,, 189, Boulevard Saint-
London, W.C. Carilstrasse 11. Germain, Paris. nA

CONTENTS OF No. 3, MARCH, 1897.

: Eee oe ot ay faa “PAGES ah
Peg he Cranial Muscles and Cranial and PORE , i
Spinal Nerves in Amia Calva... . 487-808 ©

EDWARD PHEDES: Bake

Il. The Origin of the Cleagoae Controsomes.. , 809-814

_ KATHARINE Foor.

© he Aebenaum prose Pati EN Da Paes teens
GINN & COMPANY, BOSTON, ‘U.S.A. A at

Volume XII. March, 1897. Number 3.

JOURNAL

OF

MORPHOLOGY

mike (CRANIAL MUSCLES AND CRANIAL, AND

FIRST SPINAL NERVES IN AMIAVCALV A:

EDWARD PHELPS ALLIS.

CONTENTS. PAGE

INTRO DU CILLON feceescessetsscceres sees ce ssneteceucods -os-bo0u2 uct siscsueucbsaceetevenraceseetee teeter eee 489

I. THE ORBIT, THE MUSCLES OF THE EYE, AND THE ASSOCIATED
NERVES AND ARTERIES.

Mem eaye-Wiuscles Cama esc.) ho cctecsccpsensscecs-uess-sncdeonpeccanvcdeoesecee ie anaee tee eee ee 492
2s GAKOUI GMAT ECTICS se: ss ooo suse icsacs dans csscinoasckestcocescesscucdevers tous tres eee ee 497
BMMINIGH VAIS OPEL CUS cee: 2c ces sce scesacecevien2ccscoscoscnsoaeoqaosetevaesothce teen tanta ete eae eee 507
Am NEGUS andreo bus; OlfaActOrius <.ccsccccseccsees seer ces eee ee ee 508
5. Muscles of the Eye, Nervus Oculomotorius, Nervus Trochlearis, and Ner-
VMS PANIC UI CONS) 21 ie22.c onncetescdccecedecel coast en dpe ace baceee ne ae nen ee 516
6. Review and Comparison of the Muscles and Nerves of the Eye.................. 519
7aamusi@phthalmicus, Profundus) Drigeminitsses eee eee 532
SMPNG views anG) © OmPAarisON 2225: .<cc5scc. nc: 2ocege ee tence ee ce nee eae eeed ee 535
II. MuscLtes INNERVATED BY THE TRIGEMINUS AND FACIALIS, AND THE
NERVI TRIGEMINUS, FACIALIS, ACUSTICUS, AND LINEAE LATERALIS VAGI.
1. Sub-Group 1. Muscles innervated by the Trigeminus alone ................. 546
ae Aaductor” Mandsbulae © ara eset ee ee 546
DaZicvator MaxtllaemSuper 20rd ssc cee ee ee 552
CRUE CUALOK, AT CUS RALALIIUE TS eee eee ee ee aE aE 556
d. Dilatator Opercull ............ DEP Pep ne tert oe Paar eee 557

488 ALLIS. (Vou. XII.

PAGE

2. Sub-Group 2. Muscles innervated by both the Trigeminus and Facialis.... 559

ay Laer MAN ALGUL ALAS, See ca ss ans scene naennnd seo cesescncera nents terete eee ee 559
Ds GE7 OLY OLA CUS ececetee neat ecese et eneesatena reals cece snenctenpneeetecn eee tennant eee 559
Ch LLY ONY OFLU CLES yrentne ae nena cent ascetics ee fone eer ea eae arene eee estencenane tase senate re 563
3. Sub-Group 3. Muscles innervated by the Facialis alone.....................-.0-++. 566
an Adductor LLY Oman ALGUL ATES. Lier nig tacecsetesetectenceznentvevaseunensnttteneeeee 566
b. Adductor Operculs and Levator Opercule cn enon cece 566
4. Review and (Comparison '@f Muscles 12:2 ie ornare cesar eet ca tee 567
a. Levator Arcus Palatini and Muscle Ad@y |............-.sees-s0oe0c-s0esee22-2- | 500
b. Levator Maxillae Superioris, Csdy, and AddB ...............0.0-00100 e000 570
CHP ARaUClor AT ANA tO Uae eee rca eee eee eo ace esses va scene eee 578
d. Lntermandibularis, Geniohyoideus, and FyohyotMeus.......22200.0000-- 582
e. Adductor Hyomandibularis, Adductor Opercult, and Levator
Oper culess ne eat Bale eunale sauiccecasaarant so ncnenstscsctaetcat et eeeeaae 590
5. Nervus Trigeminus and Nervus Pactalis sc... ccccsceccse-cosceee-tatsencsecaneeas-tononcaseeees 591
Be Lrasenierto-LAactad (GOR OTe oor a cee ca ssc sates. seencencstnesieszacana sete sancssteetee state 593
b. Ramus Ophthalmicus Superficialis Trigemini and Ramus Oph-
ER QUIUECUS \SSUCDETJECLU LES LI LCLOLES sare neeceet cenean aes oeeenastee ates s eesteteane 598
Coe Ramus Buccales Haciales ee ect 2 es see scpann cones coesenaensees OOS
GST UNCUS MAKI Ar EST PIC ENL IIE oaks csasacskcoenntusare snnnerecacbacuotasidenastesteners 605
@: Ramus Maxillaris SUPCT407, LG ASEWU2Ub. connnce sscnsntennseeescescssansenssevascesnes 607
f. Ramus Maxillaris 11 fe7 20, Late mest ancsta ce-ch scemoceases testoceccassisaszacane 609
g. Truncus Hyoideo-Mandibularis Factadts .......c0ccsessenesccnrecssseeeneneeees 615
6. Ear, Nervus Acusticus, and Nervus Lineae Lateralis Vagi ....................----- 621
CRI 007 NEE ei ae ne ta an oR ee Ae eee PEE RSL aA: 621
DSN e7UISAl CUSTEC US cots Vea ere ee oe ose sei Se ae Pea eee ners kane 623
Ci MLV CPUUSLLTILEDE Le OLEV ALES VEE see eee eden nero ce sane aa ona 625
7. Reéeview.and (Comparison (Of) Nerves. io. scccecr.c:.-.csnz escort senses sconreceecassvenoseeseabteser: 626

III. MuscLes INNERVATED BY THE GLOSSOPHARYNGEUS AND VAGUS, AND
THE NERVI GLOSSOPHARYNGEUS AND VAGUS.

1: (he: Visceral Arches... 73) 50005 00.208 te ee es ee

ETRO 3 U7 Ky TAD D7; 1 HAN Me PPE BS ANSI PRY IED SaASR Oe of MS DALES PN Pee a eM er eh 644
Dei! Baa rec zal Ad OCW eS cc Aue pee cca esas ce eer ee ke a ee eee 646
Cs SAL pot Arche Ae es eee nine), SOR eee ee eet ee 653
d. Operculum and Branchtostegal ROY S.i.:. cc ieccsencte-ccscscenvecentsseeevorsseee 658
Cx andl ar, AG CH kes SIE aA ae Se ee 660
f. Review and Comparison of the Visceral ArcRes ..........:ccseercvencnereees 661
2. Levatores Arcuum Branchialium ................-c.--2-00+-s0ee+-- PEN RPDS ORDERS AIM 667
3. Interarcuales’ Dorsales\ii Sau oe a rd
a. WODlsgege Dor sales: Mera ee eae aces aka tae ae a oe 669
Deseaaese 77st POY SALES £5 evs abe eccccc ora AE Buea ehe ae 670

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 489

PAGE

Ate AAdUCtOres ATCUUM Branchialaum lessee ee een ieee ese eee 671
Ropbniterarcuales : Ventraless 2.22.05... scsctestenscecctel coos ates ncbeeeed eter ene lar eaeet ee 672
Ai ODL tt VCMER ATES) ear aoe Ee re eee ae 672

De Lvs erst Verte Ole crs cons ccctaat ten cement eee 675

c. Pharyngo-Clavicularis Externus and [nter nus .......cecececceeeeececeeeee 676
Gufhkeview and Comparison’ of, Muscless:22.2 tee eee 677
Fe Ne GUS! GlOSSOpharyn Zeus ieee ssc y a eee ere FS LS ok eer 683
SUPINE MIS IN AOUS .incccasccetseve: vedassdcecenactsbatecnscsc ease cee retee soa eee nes em 688

IV. Muscies INNERVATED BY THE POSTVAGAL OR OCCIPITAL NERVES,
AND THE NERVI OCCIPITALES.

PepesT ANC O1OMAN GID ULATES: .ai2.o058s.25-xcxnectmaseucct cece at. at teeta eee coe ee sense 698

PaOLE TONY OMUCUS sot. tescee 2 cept aaee eee si seusrarunsbectretuctunibessuosteeseteltann, suse oueaaees 701

BeOccipitalvemion Of Skul le oi 5 rosea ae cate eee ec 702

4. Occipital Nerves and Corresponding Muscle Segments...............22..::cceeee0- 723

Romie view, ANG: COMPATISON, .°..52.20sessssccns. te cvesstesive erase ere ee 735

(GENERATES U MMAR Vi o22 00255 3.2 ssa cccstaas-cs2, seve sataneceee seosesueeee ore ree ee ee 737
INTRODUCTION.

Mucu the larger part of the purely investigative work on
which the present memoir is based was done in 1886, in con-
nection with a work on which I was at that time particularly
engaged: “The Anatomy and Development of the Lateral
Line System in Amia Calva.’’ It was done simply to better
understand the serial sections on which I was working, and the
notes and sketches once made were put away as of no special
value to any one but myself, it being assumed that the several
memoirs already published on the cranial nerves and cranial
muscles of Amia (van Wijhe, Sagemehl, McMurrich, Wright)
were so full and complete that any further publication on the
subject would be not only unnecessary but superfluous.

In 1889, after the publication of my earlier memoir, the notes
and sketches put away in 1886 were taken up for more careful
examination as a study preliminary to an investigation of the
central origin of the peripheral nerves. A little study and
consideration convinced me that such frequent reference would
have to be made to my own work, rather than to that of others,
that it would be much better for my purpose, even at the risk
of considerable repetition, to make and publish a brief but full
account of my results before beginning the research I especially

490 ALLS: [VoL. XII.

contemplated. I therefore began at once the preparation of the
final dissections needed for illustration in such a publication.
The work was hardly begun when I was obliged to discontinue it
and all other similar work for several years. Mr. Jujiro Nomura,
who was attached to my laboratory as artist, and who was simply
to have made the drawings from my dissections, then undertook
to make the dissections also, under my direction. On this work
he has been engaged continuously since that time, accompany-
ing me from place to place, where considerations of health have
obliged me to go. The length of time he has been engaged on
the work only proves but too well that it has been no simple
matter, even where the number and position of the nerves and
muscles were definitely known, to find them, trace them to their
origins, and lay them bare zz sz¢w, sufficiently clean and clear for
a drawing to be made of them.

As the work has progressed it has repeatedly been found
necessary to study details, and to include in the investigation
whole subjects, not contemplated in the beginning. The work
has, accordingly, gradually lost the preliminary character it
was intended to give it, and has become a somewhat extensive
memoir. It nevertheless, in its general conception and arrange-
ment, still retains many of the characteristics of the prelim-
inary work originally contemplated, for it remains, as it was
intended to be, an accumulation of facts and references grouped
so as to be conveniently used as a basis for further work.

The drawings used in illustration are all by Mr. Nomura.
They all represent actual dissections, made mostly by Mr.
Nomura, but continually controlled by me, not only on the
dissection itself, but also by comparison with the results
obtained in my own earlier work both on the adult and on
serial sections of larvae. Nothing is shown in the adult that
was not controlled in larvae, and everything found in larvae
has been sought for until found, or accounted for, in the adult.

Most of the drawings represent conditions found in a single
fish, but certain of them are combinations of dissections made
on two different heads; such combinations being made only
when the results so given could not possibly be misleading.
Whenever there has been the slightest doubt as to any arrange-

No. 3.] WUSCLES AND NERVES IN AMIA CALVA. 49QI

ment of the parts concerned, another complete dissection has
been made and the drawings changed to agree with it, or dis-
carded and made over again. No pains have been spared to
have them correct. Nevertheless I do not doubt that omissions,
if not errors, will be found in some of them, especially in those
first made.

The numerous serial sections used in the work were pre-
pared for me at my laboratory, the Lake Laboratory, Milwaukee,
at first*by Dr. W. Patten, under the direction’ of Dra; 0:
Whitman, and afterwards by O. S. Strong and A. C. Eycles-
hymer, under the direction of Dr. Howard Ayers. Several of
the dissections used for control, especially those of the occipital
region, have been made by Dr. Julius Dewitz. Dr. Dewitz also
made the final dissections used for Figs. 35 and 58.

In the nomenclature adopted I have followed as closely as
possible van Wijhe, Sagemehl, and Vetter. New names have
been used only when absolutely necessary for descriptive
purposes, for I fully agree with Sagemehl (No. 104, p. 180,
note 1) that radical changes will have to be made, even in
names already in general use, when the development and
anatomy of the vertebrate head is better known.

In the literature referred to I have had to limit myself to
such memoirs and publications as I could readily procure, cir-
cumstances obliging me to live where access to libraries, even
to my own, was impossible. Certain of the references are made
from notes made years ago, others are to abstracts only of the
works concerned. All references to Scomber and Gadus, and
all those to Perca and Esox, where special reference to Vetter
has not been made, refer to work now being done, in my labo-
ratory here, by Dr. Dewitz and Mr. Samuel Mathers.

PALAIS CARNOLES, MENTON.

A492 ALLIS. [Vou. XII.

I. THE ORBIT, THE MUSCLES OF THE EYE, AND THE
ASSOCIATED NERVES AND ARTERIES.

1. Hye-Muscle Canal.

Amia has, as already described by Sagemehl (No. 104, p. 214),
a well-developed “Augenmuskelkanal”’ (emc, Fig. 11, Pl. XX1I),
bounded in front and behind by transverse ridges of the car-
tilaginous base of the skull. The posterior ridge lies near
the hind edge of the petrosals, and bears on its summit the
transverse processes of those bones. The anterior ridge is a
specially developed structure (w, Figs. 9 and 11, Pl. XXI),
called by Sagemehl a transverse ‘“ Wulst,” and by Shufeldt
(No. 115) a transverse “bar” of cartilage. This bar is sharply
marked off from the rest of the basis cranii by the Augenmus-
kelkanal behind, and in front by a narrow but comparatively
deep groove (cz) which extends transversely from orbit to orbit.
At each end of the bar there is usually, but not always, a small
ossification, the basisphenoid of Bridge (B&,S, Figs. 13, 14, and
15, Pl. XXII), which in one large specimen, contrary to Sage-
mehl’s statement (No. 104, p. 215), extended in part through
the cartilage of the base of the cranium, so as to be seen on its
under surface. In the same specimen the basisphenoid on the
left side of the head extended forward across the transverse
interorbital groove, bridging it or being perforated by it. In
all the specimens examined the two ossifications were of unequal
size, and they never touched each other at any point, a median |
region of cartilage always separating them.

At about the middle of the median edge of each basisphenoid
there was in all the specimens examined a deep indentation,
forming, with the adjoining cartilage, a canal (zc, Figs. 9, 10,
14, and 15, Pls. XXI and XXII) through which, and not
through the hypophysial fenestra as stated by Wright (No.
133, p. 495), the internal carotid artery on each side entered
the cranial cavity. From this canal, near its ventral opening,
one, or sometimes two, small canals run forward and laterally
through the bone into the tranverse interorbital groove, or
into the extreme hind end of the orbit near it. From the free
lateral edge of the bone the superior, inferior, and internal recti

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 493

muscles arise, and along its postero-lateral edge, which is slightly
grooved, the external rectus lies in its passage from the orbit to
its origin, close to the middle line of the head near the hind end
of the eye-muscle canal. In none of the specimens examined did
the muscle arise, as stated by Sagemehl, immediately behind
the transverse “ Wulst”’ or bar. It arises from the cartilaginous
floor of the eye-muscle canal, or in part also from the median
rib or ridge of the parasphenoid, which fills the median longi-
tudinal fissure in that floor. This fissure (Zfn, Fig. 10, PI.
XXI) is the hypophysial fenestra of Sagemehl and Wright.
The process of the parasphenoid that fills it, although differing
somewhat in position, may be the homologue of the one which
in Cyprinoids is said by Sagemehl to replace the basisphenoid
of other teleosts (No. 107, p. 575).

In larvae of Amia, even in those 50 mm. in length, the
internal carotid canal is represented by a foramen only which
perforates the basal plate of the future orbital opening of the
eye-muscle canal. The internal carotid artery, having issued
from this foramen, runs upward and forward along the lateral
surface of the base of the transverse cartilaginous bar of the
chondrocranium, traversing an open space between the basal
plate of the skull and the overhanging, upper, lateral edge of the
bar. The anterior corner of this overhanging edge projects
upward and forward, and the artery, having reached the anterior
face of the bar, below and median to its projecting dorso-
anterior corner, turns outward and upward along the dorsal
surface of that corner, not having again entered the cartilage
at any place. While traversing the open space along the lateral
surface of the bar, its sends or receives a communicating
branch from the efferent pseudobranchial artery, the branch
giving rise, undoubtedly, to the small branch canal found in
the basisphenoid of the adult ; though in the adult the commu-
nicating artery itself was not traced. The three recti muscles
which, in the adult, arise from the lateral surface of the basi-
sphenoid, arise, in larvae, from the under surface of the dorsal
edge of the transverse bar. The basisphenoid of Amia is
therefore probably, in large part, a membrane bone developed in
connection with the insertions of certain of the recti musles.

494 ALLIS. [VoL. XII.

The roof of the eye-muscle canal in the adult is formed, as
described by Sagemehl, by the horizontal processes of the petro-
sals, and by a tough, glistening membrane which extends for-
ward from the front edges of those processes to the transverse
bar of cartilage. The membrane, however, in all the specimens
I examined, formed .a much less important part of the roof of
the canal than in the specimens figured and described by Sage-
mehl. The processes of the petrosals, the united front edges
of which have a concave outline, cover nearly, if not quite, one
half of the canal, and the space between their front edges and.
the transverse cartilaginous bar is largely occupied by an
important sack-like depression or pit (pf, Fig. 26, Pl. XXV),
described by Sagemehl as a slight depression only. In this pit
are lodged the hypophysis cerebri and saccus vasculosus. The
rounded front end of the hypophysis projects forward slightly
beyond the anterior edge of the pit, and almost touches the
transverse ‘‘Wulst,’” while the saccus vasculosus projects
backward beyond the hind edge of the opening of the pit and
lies, in large part, under the projecting, overhanging processes
of the petrosals in a backward, pocket-like extension of the pit.

The tough membrane that forms part of the roof of the eye-
muscle canal extends upward, on each side, and forms the
median wall of what Sagemehl has called the upper, lateral
chamber of the canal. In front of the canal the membrane
extends forward across and beyond the transverse bar of car-
tilage, closely attached to its upper surface. Here, on each
side, it extends upward from the lateral edge of the basisphe-
noid and the cartilage of the basis cranii in front of it, to the
lower edge of a thin, projecting plate or fin of the alisphenoid,
and to the lower and posterior edge of the orbitosphenoid in
front of and continuous with that fin, thus filling the optic
fenestra. Immediately in front of the transverse bar the mem-
brane covers and takes part in the formation of a transverse
pad of tough, dense tissue which extends from orbit to orbit
(Fig. 26, Pl. X XV), covering and filling the interorbital groove.
The top of the pad projects backward so as to slightly over-
hang, and the hind edge thus formed has a concave outline.

Under this overhanging, curved hind edge, on each side, at the

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 495

lateral end of the pad, the optic nerve, which runs outward,
forward, and slightly downward, along the upper, shelving sur-
face of the cartilaginous bar, pierces the lining membrane of
the optic fenestra and enters the orbit.

The hind edge of the optic fenestra (ofn, Figs. 9 and 11, PI.
XXI) is formed by the pedicle of the alisphenoid, the bone
being always, in the specimens I have examined, cut out behind
by the large trigeminal foramen, so as to leave a pedicle, and
not simply perforated by it, as described by Sagemehl. The
pedicle thus formed rests externally upon a process of the lat-
eral wing of the parasphenoid, and internally upon a cartilagi-
nous base which forms part of the side wall of the eye-muscle
canal. Between this cartilaginous base and the pedicle exter-
nally, and the lining membrane of the eye-muscle canal and the
thin, projecting plate or fin of the alisphenoid internally, there
is formed a tall and relatively narrow opening by which the
eye-muscle canal communicates with the orbit. Along the bot-
tom of this opening, which may be called the orbital opening
of the canal, the rectus externus enters the main, lower portion
of the eye-muscle canal, and in its upper part the trochlearis,
oculomotorius, and ophthalmicus profundus issue from the
upper, lateral chamber of that canal and enter the orbit.

The upper, lateral portion or chamber of the eye-muscle canal
(emc') lies in a recess in the side wall of the cranium, the
recess extending from the pedicle of the alisphenoid back to
that process of the petrosal that forms the front wall of the
utricular fossa. This chamber has in addition to the tall,
narrow opening just described, which is common to it and the
main eye-muscle canal, five openings leading to the outer sur-
face of the cranium, namely: the large facial foramen (//7)
through the petrosal; the small foramen for the external
carotid (ecfr) immediately below and in front of that foramen
and through the same bone ; the otic canal (ofc) leading to the
top of the skull through the cartilage in front of and above the
petrosal; the large trigeminal foramen (¢/7) through the edge
of the alisphenoid, immediately behind and lateral to the pedicle
of the bone; and the smaller foramen for the ophthalmic
nerves (offr) through the alisphenoid immediately above and

496: ALLIS. [Vou. XII.

in front of the trigeminal foramen and separated from that
foramen by only a thin layer of bone.

Immediately below the upper, lateral portion of the eye-muscle
canal there is, in the side wall of the main canal, an oblong
opening, the palatine foramen (//7), through which the palatine
branch of the facialis leaves the cranium. This foramen is
closed externally by the base of the lateral wing of the paras-
phenoid, and leads into a canal between that bone and the base
of the cranium. This latter canal, which may be called the
palatine canal (fc, Fig. 17, Pl. XXII), begins behind the pala-
tine foramen, at an opening found in the angle formed by the
hind edge of the wing of the parasphenoid and the lateral edge
of the body of that bone, the opening being almost but not
entirely enclosed in bone (zcf/7, Fig. 17, Pl. XXII). Beginning
at this point the canal runs at first medianward and forward
to the hind edge of the palatine foramen, then almost directly
forward across that foramen and onward toward the anterior end
of the head, lying in part between the cartilaginous base of the
skull and the parasphenoid and vomer, and in part in those two
bones. Immediately in front of the anterior edge of the pala-
tine foramen a branch canal is sent forward and outward to the
orbit, the opening of the canal lying in the angle between the
anterior edge of the wing of the parasphenoid and the lateral
edge of the body of the bone in front of the wing. Imme-
diately in front of this canal a second branch canal is given off,
the internal carotid canal, which pierces the basis cranii and
enters the cranial cavity along the median edge of the basi-
sphenoid as already described.

Through the posterior opening of the palatine canal the
internal carotid artery and the pharyngeal branch of the glos-
sopharyngeus enter the canal and join the palatinus facialis
(No. 133, p. 494). The opening may therefore be called the
internal carotid foramen, although it is only indirectly the point
of entrance of that artery to the cranium. Through the first
branch canal the posterior branch of the palatinus facialis enters
the orbit. The external opening of this canal may therefore
be called the foramen of that nerve (ppffr, Fig. 17, Pl. XXII).
Anterior to this point and anterior to the ventral opening of

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 497

the internal carotid canal, the palatine canal lodges the anterior
branch of the palatinus facialis, the pharyngeal branch of the
glossopharyngeus and a small branch only of the internal carotid.

2. Carotid Arteries, Hypophysis Cerebri, and Saccus Vasculosus.

The carotid arteries in the young of Amia have been described
by Wright (No. 133, p. 494). With the results obtained by
him, my work both on the young and on the adult is mainly,
but not entirely, in accord.

The efferent artery of the first branchial arch (ea. I, Fig. 28,
Pl. XXV, and Figs. 61-63, Pls. XXXVI and XXXVII) as it
bends inward and backward to join the dorsal aorta, gives off two
vessels, the common carotid artery (cc) directed forward, and the
hyo-opercularis (Zo) directed forward and outward. From near
the root of this last artery a small branch is sent backward and
upward, the remaining and larger portion of the artery continu-
ing forward and outward until it reaches the facial foramen,
which it crosses under the issuing truncus hyoideo-mandibularis
facialis. It then turns upward, outward, and backward above
the nerve, and running backward above the adductor hyoman-
dibularis, separates into two main parts, the hyoid and opercular
arteries. The former joins and accompanies the truncus hyoi-
deo-mandibularis facialis and its branches ; the latter, destined
to supply the inner surface of the operculum, separates into
two parts, one of which runs backward above, and the other
backward and downward below, the opercular process of the
hyomandibular. As the main hyo-opercularis artery passes
forward across the facial foramen, under the facialis, it sends a
branch upward in front of that nerve, through the foramen into
the upper, lateral chamber of the eye-muscle canal. Branches
are there given to tissues in the canal, and the artery then,
running upward and forward lateral to the trigemino-facial
ganglion, issues through the trigeminal foramen with the ramus
buccalis facialis and truncus maxillaris trigemini, which nerves
it accompanies in their further outward course. No branches
could be definitely traced into the cranial cavity proper or the
auditory labyrinth, as described by Wright (No. 133, p. 495).

4098 ALLIS. [Vou. XII.

The common carotid artery, soon after its origin from the
first aortic arch, and just before reaching the lateral wing of
the parasphenoid, separates into two parts, the internal and
external carotids.

The internal carotid (zc) enters the palatine canal between
the parasphenoid and the basis cranii, and after crossing the
palatine foramen, turns upward into the internal carotid canal,
and enters the cranial cavity under, or median to and in front
of, the optic nerve of its own side. As it leaves the palatine
canal a small branch is sent forward in that canal, and as it
traverses the internal carotid canal another small branch is
sent outward and forward in the small branch canal to join
the efferent pseudobranchial artery, as already described on
page 493.

Having entered the cranial cavity, the internal carotid artery,
in the adult, turns backward and then outward and forward
under the optic nerve. At this point it separates into three
parts, an optic, an anterior or olfactory, and a posterior portion.
The optic portion runs outward with the optic nerve, enclosed
in the membranous sheath enclosing the nerve, and supplies,
according to Wright, the ventral half of the choroid gland.
The olfactory portion runs directly backward to about the level
of the hypophysis, where it separates into two parts, both of
which turn inward and then forward above the optic nerve and
supply the anterior parts of the brain and brain cavity, one
of the branches running forward under the olfactory nerve
toward the nasal sack. The third, or posterior, portion of the
artery runs at first directly backward, median to the olfactory
portion. Reaching the level of the auditory labyrinth, it gives
off a large branch, and, turning inward behind the lobus infe-
rioris, joins its fellow of the opposite side, thus forming a cir-
culus cephalicus, from the median point of which a branch is
sent forward and another backward under the brain.

In embryos the optic chiasma lies in front of the internal
carotid canal; and the internal carotid artery, after issuing from
its canal, runs upward and forward along the lining membrane
of the optic fenestra, behind the optic nerve. In this part of its
course it becomes so flat and thin that I could not definitely

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 499

determine from my sections whether it lay along the inner or
the outer surface of the membrane. When it reaches a point
slightly above and behind the optic nerve, it either perforates the
membrane from its inner to its outer surface, or sends an impor-
tant branch inward through the membrane, and then turns
sharply outward and slightly downward above and behind the
optic nerve, and, lying close to and behind the nerve, along its
open edge, and at no place inside the nerve, as Goronowitsch
(No. 50, p. 443) seems to have found in the adult, enters the
eyeball with it.

The external carotid (ec, Fig. 28, Pl. X XV) enters the upper,
lateral chamber of the eye-muscle canal through the small
external carotid foramen which lies in the petrosal immediately
below and in front of the large facial foramen. In the eye-
muscle canal it runs upward and forward under or in front of
the trigemino-facial ganglion, and near the front end of that
ganglion separates into two parts. One of these parts, the
ophthalmic portion of the artery (ecof), issues from the canal and
enters the orbit through the ophthalmic foramen with the
ophthalmic branches of the facialis and trigeminus; the other
or mandibular portion (ecm) issues through the trigeminal fora-
men with the truncus trigeminus, which it accompanies. Imme-
diately outside the cranium the mandibular portion gives off
the afferent pseudobranchial artery (afs), which runs backward
to the uppper surface of the pseudobranch. From the lower
surface of the pseudobranch the efferent pseudobranchial artery
(eps, Figs. 22-28, Pls. XXIV and XXV) arises and runs for-
ward under the external carotid, or under the mandibular branch
of that artery. Immediately in front of the lateral wing of the
parasphenoid it crosses above the ramus palatinus posterior
facialis, as that nerve issues from the palatine canal by its fora-
men, and, turning inward and forward, passes through a small
foramen (epsfr, Figs. 9-15, Pls. XXI and XXII) in the turned-
up lateral edge of the thin layer of cartilage that forms the
bottom of the passage leading from the eye-muscle canal to the
orbit, and enters that passage opposite the basisphenoid. Here,
in larvae and probably also in the adult, a communicating branch
is sent to the internal carotid, the branch corresponding to and

500 ALETS: Vor sunk

representing the anastomosis of the two arteries described by
Wright in Lepidosteus (No. 133, p. 484). Continuing forward
the efferent pseudobranchial artery crosses under the rectus
externus, turns upward behind and above the rectus inferior
but under the nervus oculomotorius, and reaches the orbit.
It then turns outward and forward, internal to and then above
the inferior branch of the oculomotorius, and lying immediately
behind the optic nerve, enters the eyeball with that nerve, behind
and a little above it, through the large opening in the carti-
laginous sclerotic (0f, Fig. 24, Pl. XXIV). It is the arteria
ophthalmica of Sagemehl (No. 104, p. 203), and gives rise, accord-
ing to Wright (No. 133, p. 495), to the dorsal part of the choroid
gland. During its course through the orbit it is enclosed in a
tough fibrous envelope, which, in the large fish used for illustra-
tion, had become, either naturally or through the action of
reagents, semi-cartilaginous in character and appearance, and
formed a tough, elastic, tube-like structure extending from the
cranium to the eyeball. Sagemehl found no trace of a choroid
gland in Amia (No. 106, p. 116).

Two other vessels (ov and ov’, Figs. 22-27, Pls. XXIV and
X XV) extend from the cranium at the hind end of the orbit to
the eyeball, and, in the one large fish, they were, like the arteria
ophthalmica, semi-cartilaginous in appearance. They are both
of them veins, though they have, even in larvae in sections,
much the appearance of arteries. The smaller of the two, ov’,
issues through a perforation in the sclerotic near its outer edge,
between the insertions of the rectus superior and the rectus
externus, and runs downward, backward, and inward between
those two muscles to join a large vein, the orbital sinus of
Parker in Mustelus, which lies along the side of the skull imme-
diately below and internal to the ophthalmic branch of the
trigeminus. From this orbital vein, distal to the point where
it is joined by the vein ov’, a branch, the anterior cerebral vein
of Parker in Mustelus (No. 83, p. 712), was traced in embryos
through the cartilage of the side of the skull into the brain cav-
ity, where it was distributed to the fore-brain and the olfactory
nerve. In the adult this branch was not traced ; the foramen by
which it leaves the skull was, however, always found, lying some-

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 501

times through the cartilage immediately in front of the alisphe-
noid, and sometimes through the anterior edge of that bone
(acufr, Figs. 9-11, Pl. XXI). In embryos the vein ov! could be
traced inside the sclerotic outward and forward toward the optic
lens, accompanied by the ciliares longi. After it has joined the
orbital vein, the latter enters the upper, lateral chamber of the
eye-muscle canal through the large orbital opening of that canal,
lying external to the profundus ganglion and dorsal to the ciliares
longi.

The second and larger of the two optic veins, ov, issued, in
the large specimen used for illustration, by two main openings
and other smaller ones through a porous portion of the sclerotic
lying immediately below and lateral to the optic perforation. It
arises from the choroid gland and other structures in the interior
of the eye, one branch being traced through the pigment layer
of the retina to its inner surface. Having issued from the eye-
ball it runs inward and backward immediately below the opticus,
and enters the orbital opening of the eye-muscle canal, passing
above and in front of the inferior branch of the oculomotorius,
between the rectus superior and the rectus inferior, and above
the efferent pseudobranchial artery. Its further course was not
traced in the adult. In 40 mm.and 50 mm. specimens it sepa-
rates near the origins of the internal and inferior recti into two
parts, one of which continues backward below the rectus exter-
nus while the other turns upward in front of that muscle and
joins or is joined by the orbital vein. The united veins then
run backward, at first below the trigemino-facial ganglion, then
internal to it, and finally above it, the larger portion of the vein
passing through the ganglion. It then issues through the facial
foramen and joins the jugular vein. The other portion or branch
of the vein ov, the portion continuing backward below the rectus
externus, reaches the external surface of the lining membrane
of the eye-muscle canal and continues inward and backward
along that membrane to the hind end of the hypophysis, where
it turns inward behind that organ, between it and the saccus
vasculosus, and joins the corresponding vein of the opposite side
of the head. That part of the vein that lies in the orbit is
unquestionably the structure considered by Sagemehl (No. 104,

502 ALELS. [VoL. XII.

p. 203) as the homologue of the eyestalk in selachians, and that
part of it that lies under the base of the brain, behind the hypo-
physis, is the “hintere Quer-Anastomose”’ of Gaupp in Anguis
and Lacerta (No. 41, p. 571). The transverse anastomosis in
Anguis and Lacerta lies above the hind end of the hypophysis,
between it and the base of the brain, but as the hypophysis in
Amphibia grows backward from its point of attachment to the
gut (No. 71, Figs. 6 and 7), and in Ganoids forward (No. 69,
Figs. 13-16), this difference of position of the vein is probably
apparent rather than real. In larvae of Acipenser a vessel that
has, in the median vertical plane of the head, exactly the position
of the vein vo in larvae of Amia is considered by von Kupffer
(No. 69, p. 11) as the artery of the mandibular arch. It may
be such in Amia also, the choroid gland representing the de-
generated respiratory organ of that arch, as has been suggested
by Dohrn, I believe, though I do not find the reference. The
artery disappears in older larvae of Acipenser (No. 69, p. 25).

The base of the brain in Amia, above and in front of the
transverse anastomosing vein vo, is, in 50 mm. specimens, flat
and thin, with a large, median, much-plaited fold projecting into
the cavity of the infundibulum. At the middle of its length the
lateral edges of the fold almost touch, leaving a slit-like opening
into its relatively large interior, which is filled in sections with
what is apparently glandular tissue and a dense mass of blood
corpuscles. Immediately below and in front of the fold, and
hence immediately under and in contact with the base of the
brain, is the hypophysis, a flattened oval mass of tissue similar
to and continuous with that that fills the fold. It is as wide
as the underlying hypophysial fenestra, is everywhere separated
from the mouth cavity by the parasphenoid, and is everywhere
limited and defined by membrane excepting toward the brain
and where it comes in contact with the vein vo, at its posterior
end and along the posterior portion of its lateral edge. At
these places the tissue of the organ is apparently directly con-
tinuous with masses of blood corpuscles in the vein. The
saccus vasculosus begins behind the vein, and in its posterior
portion ramifications of the central cavity give to it a glandular
appearance.

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 593

In 12 mm. and 14 mm. specimens the median fold in the
base of the infundibulum is not yet formed, but there is already
a slight thickening of the tissue, indicating its position. The
hypophysis has much the appearance that it has at 50 mm., but
there is less evidence of glandular formation. The vein vo
lies entirely behind the organ, instead of behind it and along
the posterior portion of its lateral edge, as at 50 mm., and is,
as at that age, directly continuous with it. The base of the
brain, above the vein, is formed of a single layer of cells, as is
also the saccus vasculosus, which is, as in Acipenser (No. 69), a
simple sack-like prolongation arising from the lower posterior
limit of the base of the infundibulum and running directly back-
ward, its hind edge being wedged under the front end of the
medulla. It is narrow in front, widens considerably as it extends
backward, and contains a large central cavity, but gives at this
age no indication whatever of a glandular structure or of associa-
tion with such a structure. Immediately dorsal to it there is
another posterior prolongation of the infundibular cavity which
turns upward in front of the front end of the medulla and ends
there blindly. It is the saccus dorsalis infundibuli of von
Kupffer in Acipenser (No. 69, p. 19), and from it arise laterally
the cavities of the lobi inferiores, as in Acipenser.

In the adult (Fig. 64, Pl. XX XVIII) there is a true lobus
infundibulum, not shown by Goronowitsch (No. 50) in his
figures, and there is also a large and long saccus vasculosus (sv)
not shown by him, the part called by him the saccus being the
lobus infundibulum or part of that structure. The lobus infun-
dibulum lies between and below the lobi inferiores (/), the
hypoaria of Gage (No. 40, p. 281), and its posterior portion is
trilobate in outline. The median of the three lobes contains
the posterior prolongation of the infundibular cavity, described
above, and the two lateral lobes contain each a slight lateral
prolongation of that cavity.

In the base of the median lobus there is a median slit-like
opening which leads directly into the anterior end of the
saccus vasculosus, which is here formed of a single layer
of cells lined externally by a large venous sinus of a semi-
circular or horse-shoe shape. At the hind end of the slit

504 ALLIS. [Vou. XII.

the lips of the opening project downward as two columns, which
then turn backward and, as large nerve cords or strands, extend
backward along the upper part of the saccus, lying between the
upper ends of the venous sinus, and forming the upper wall of
the central cavity of the structure. Diverticula of this central
cavity are found along nearly its entire length, extending out-
ward into the venous sinus. They are short in the anterior
portion of the organ, but long and much convoluted in the
posterior portion, where the organ becomes a large glandular
structure. Into this glandular formation nervous strands or
fibres extend from the upper, nervous wall of the structure, and
finally, toward the hind end of the organ, the strands and cen-
tral cavity become entirely lost in the general tissue. No
slightest indication of a canal or canals extending into the
mouth cavity was found, and the organ seems to be devoted
solely to a glandular secretion destined to supply the central
cavity of the brain, as Waldschmidt states to be the case in
Polypterus (No. 126, p. 318). That part of the hypophysis
shown by him in sections seems, however, to be the saccus of
Amia, his ‘“feinere Gefiige’”’ indicated by the letter d being
in part, at least, the nervous portion of the organ. The nerve
cells shown by Bickford in Calamoichthys (No. 11) seem to
correspond in position to the blood corpuscles of the venous
sinus in Amia. In Amia no nerve cells associated with the
organ could be identified.

The hypophysis, in the adult (Zy, Fig. 64, Pl. XX XVIII), is
somewhat heart-shaped, viewed from below, the point of ‘the
heart directed forward. On the sides it extends upward along
the sides of the anterior end of the lobus infundibulum, this
part of the lobus containing the cephalic projection of the cen-
tral cavity of the brain described by Gage. In the floor of
this cephalic projection is the median fold found in the base
of the infundibular cavity in larvae. Canals from the hypophy-
sis open into the infundibular cavity and into the hypophysis
nerve strands and fibres are sent from the brain, the nervous
supply being large and important, as is also that of the glandu-
lar part of the saccus.

The vein vo of larvae is found in the adult (vo, Fig. 64, Pl.

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 505

XXXVIII), crossing under the front end of the hypophysis
immediately internal to the rectus externus and immediately
behind the transverse ‘“‘ Wulst”’ of cartilage that marks the front
end of the eye-muscle canal. From the vein a median branch
is sent backward and upward into the hypophysis. Whether this
branch communicates directly with the venous sinus along the
base and sides of the saccus vasculosus, or passes through the
hypophysis to supply that organ, could not be definitely decided.
Probably the former. The hypophysis and saccus are thus in
all probability both glands secreting, from venous blood, a fluid
destined to fill and supply the central cavities of the brain.

In small larvae the hypophysial fenestra is covered externally
by the parasphenoid as in the adult, and the hypophysis lies
directly upon that bone. The saccus vasculosus lies directly
upon the inner surface of the base of the skull, and the trans-
verse ridge on that base, which, in the adult, forms the posterior
limit of the eye-muscle canal, does not exist. The processes
of the petrosals that form the roof of the canal have also not
yet been formed, and the hind end of the saccus vasculosus is
separated from the base of the brain by membrane only. The
eye-muscle canal of the adult, therefore, does not exist in
embryos or young larvae. It is, however, represented by the
space in which the vein vo traverses the median line of the
skull, and by the space along the sides of the hypophysis and
saccus in which lie, on each side, the vein vo and the rectus
externus.

In larvae 50 mm. in length the saccus begins to present a
glandular appearance and the processes of the petrosals, repre-
sented by cartilages, are well developed; the hind end of the
saccus being separated from the brain by their united edges.
The space enclosed by these processes, the future eye-muscle
canal, is much larger than the saccus, and the space around
that organ is filled with loose tissue such as that described by
Waldschmidt in Polypterus and shown by Bickford in Cala-
moichthys. No trace or indication whatever of a transverse
lymph sinus in this space could be found. The eye-muscle
canal in all probability does not, therefore, as Sagemehl was
led to conclude, owe its origin to such a sinus. It is a space

506 ALES. ivion. xin

formed around the saccus vasculosus into which the rectus
externus on each side has crept, following the vein vo or the
nervus abducens, and the orbital opening of the canal is the
fused foramen of those two structures, or, as in Amia, of those
two foramen and the foramen of the trochlearis, oculomotorius,
and profundus as well. That a muscle should enter the canal
of its nerve certainly seems more natural than that it should
enter one serving simply for the passage of a lymph canal.
In Mustelus, where there is no eye-muscle canal, the rectus
externus arises (No. 123, p. 81) close to the opening of the
canal for the abducens. In larvae of Axolotl (No. 117, p. 16)
the side wall of the skull is pierced, behind the optic foramen,
by a larger opening, which is in part filled with muscle
«Anlagen.”” Through this opening pass also a blood-vessel
and a bundle of nerve fibres.

Although no trace of a transverse lymph sinus was found
in the eye-muscle canal, such a sinus was always found, in
larvae, extending from orbit to orbit between the internal
carotid canals and the optic chiasma, and connecting the
periorbital lymph spaces. This sinus is represented in the
adult by the interorbital groove already described on page 406,
but whether there is a transverse lymph passage at this point
in the adult or not I did not determine, as my attention was
not called to the point in the earlier stages of this work, and
lately I have had no material suitable for its investigation.
The groove is, however, certainly the canalis transversus of
selachians, and Gegenbaur’s description of the canal in Galeus
and Squatina (No. 44, p. 76) could almost be taken for that in
Amia, the “ Sattellehne” in the former being considered as the
posterior edge of the transverse Wulst or bar in the latter.
The carotid canal or canals, however, lie in selachians in front
of instead of behind the canalis transversus. The presence of
this latter canal, together with an eye-muscle canal in Amia,
shows definitely that the latter is not derived from the former
as Gegenbaur (No. 44, p. 78) and Sagemehl (No. 104, p. 217)
were led to suggest.

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 507

3. Nervus Opticus.

The optic nerves (0, Figs. 22-27, Pls. XXIV and XXV) in
the adult have a well-developed chiasma and a long tractus or
pedicle, on each side of the head, connecting the chiasma, along
the side of the brain, with the lower anterior end of the optic
lobe.

Whether the chiasma is entirely free from the base of the
brain or not was not investigated. Each nerve has, as Goro-
nowitsch states (No. 50, p. 443), the appearance of a much-
plaited plate, folded upon itself so as to form a cylindrical or
oval, nearly closed gutter. The open line between the free
edges of the plate lies along the lateral and posterior edge or
surface of the nerve, and, proximally, the free edges of the
plate open and embrace the tractus. The fibres contained in
the upper of these free edges, in part at least, run upward along
the anterior edge of the pedicle and do not enter the chiasma
or cross to the opposite side of the brain. The same is pos-
sibly true of other fibres of the nerve. The fibres along the
median and anterior edge of the nerve run directly into the cor-
responding ones of the nerve of the opposite side.

The nerves in the adult run outward and forward into the
extreme hind end of the orbit, with the recti muscles. In
young specimens they run outward, nearly at right angles to
the axis of the body, a little in front of the place of origin of
the superior, inferior, and internal recti, and not far from the
middle of the orbit, as in adult selachians. They lie imme-
diately behind the tough, dense pad of tissue already described
as extending transversely from orbit to orbit. The position of
this pad, and its appearance in sagittal sections of the young,
correspond closely to that of the mass of fibres, a.c., described
and figured by Balfour and Parker, in Lepidosteus (No. 6, p.
378, and Fig. 45). These fibres in Lepidosteus are considered
by them to be nervous, and to be probably equivalent in part
to the anterior commissure of the brain.

The brain in the young of Amia entirely fills the cranial
cavity ; in the adult it occupies a portion only of the middle
and hind part of that cavity, the remaining and larger portion

508 ALLIS. [Vot. XII.

of the cavity being filled by fatty tissue (No. 105). This marked
change in the relative size of the cranial cavity and brain takes
place mostly in post-larval stages, the unequal development of
skull and brain being most marked anteriorly. By it the floor
of the anterior part of the skull, and its perforations, are carried
forward relatively to the brain, and the optic nerves at their
origin are, on the contrary, pulled relatively backward ; so that,
while in the young they issue in front of the canalis transver-
sus and internal carotid canals, they lie in the adult either
directly above the internal openings of the carotid canals or
even behind them. The olfactory branches of the carotids
must, therefore, first run backward under the optic nerves and
then forward above them to reach their destination.

4. Nervus and Lobus Olfactorius.

The lobus olfactorius, Goronowitsch, or bulbus olfactorius,
Sagemehl, is in the adult Amia a well-rounded mass (/o/, Figs.
25 and 64, Pls. XXV and XXXVIIIJ), lying close to the middle
line of the head, on either side, immediately above the trans-
verse bar of cartilage marking the front end of the eye-muscle
canal, immediately in front of the optic chiasma, and imme-
diately behind the pad of tissue behind and under which the
nervus opticus enters the orbit. Its diameter is about one and
one half times that of the nervus olfactorius, and it contains, as
Goronowitsch has stated, a small lateral extension or diverticu-
lum of the ventricle of the fore-brain. On its upper surface,
in well-preserved specimens, there are three slight furrows, one
running medianward and backward from the lateral surface of
the lobus, near its anterior end, one laterally and backward
from the median surface of the lobus, and the third directly
backward along the median surface, beginning at the point
where the second furrow begins, near the root of the nervus.
These three furrows thus correspond fairly well with the
three furrows shown by Burckhardt on the so-called tuber-
culum olfactorium of Protopterus (No. 16, Figs. 3 and 4).
The first two enclose a slightly marked anterior region of
the lobus.

No.3.] MUSCLES AND NERVES IN AMIA CALVA. 509

The lobus lies partly in front of, and partly under, the median
two thirds of the corresponding hemisphere of the fore-brain,
the upper and lateral surfaces of which, in their posterior por-
tion, show well-marked convolutions, much stronger than Gage’s
figures and reference to them (No. 40, p. 296) would seem to
indicate. Ventrally the lobus extends further backward than
on its dorsal surface, and there is no line of demarcation between
it and the base of the fore-brain beyond it, as Gage has already
stated (No. 40, p. 297). On the sides and dorsal surface, on the
contrary, it is sharply limited and defined by a deep fissure or
furrow, which, on the median side of the lobus, lies just in front
of the lamina terminalis, and separates that structure, exter-
nally, from the lobus. The fissure, therefore, at this point is
the fissura prima. In the remaining portion of its course it
seems to be the sulcus olfactorius of Burckhardt in Protopterus,
but it is evident to the most superficial observation that the fis-
sure and sulcus of Amia have not the position that they have
in Protopterus, and that the lobus of Amia contains elements
not found in the tuberculum of Protopterus. Lee’s conclusion
(No. 72, p. 5) that the latter structure is in reality a lobus,
strictly comparable to the lobus of other animals, may therefore
be questioned. It seems more probable that it is a bulbus
partly separated from the fore-brain.

In embryos of Amia the sulcus olfactorius extends across the
ventral surface of the brain, and a continuous groove is thus
formed, running backward and outward across the dorsal surface
and backward and downward across the lateral surface of the
brain, then across its ventral surface, and then forward and
upward along its median surface between the lobus and the
lamina terminalis.

The olfactory diverticulum is, in 12 mm. and 14 mm. speci-
mens, short and broad, extends outward and forward from the
median or median and upper portion of the basal ganglion of
Goronowitsch, and has at its anterior and lower portion a
pointed prolongation, apparently the true recessus olfactorius,
extending forward into the lobus. In its hinder portion, exactly
where Studnitka (No. 122) describes and figures the cornu pos-
terior of the ventriculus lateralis in Petromyzon, it has a sharply

510 ALLIS. [Vov. XII.

marked angle, or even a slight pointed recessus, directed out-
ward or outward and backward. The whole diverticulum thus
seems to correspond to the ventriculus lateralis of Studnicka
in Petromyzon, with the posterior horn rudimentary, instead of
wanting, as he states it to be in ganoids and teleosts. The
lobus in Amia thus seems to correspond internally, as well as
externally, to the lobus and something more of Petromyzon,
Dipnoi, and Amphibia, — to the prosencephalon, probably, or to
part of it, as Gage has suggested (No. 40, p. 297). The origin
of olfactory nerve fibres at the hind end of the prosencephalon
in reptiles (Edinger, quoted by Gage, No. 40, p. 262) would
seem to support this supposition.

Between the olfactory lobes, and extending downward and
backward to the commissure cw of Gage (No. 40, p. 282), that
is, for the full length of the lamina terminalis, there is in the
bottom of the ventricle of the fore-brain a long and narrow
median prolongation, directed forward and downward. At the
upper, anterior end of this median diverticulum there is a slight
recessus directed forward. in the lamina terminalis at this
point there is no process properly so called, the recessus being
apparently formed, as sagittal sections show, by a thinning of
the substance of the lamina. The lamina here is slightly con-
vex externally, but less so than in its ventral portion. In trans-
verse sections the anterior end of the recessus is often cut so
as to show a slit-like lumen, surrounded or nearly surrounded
by an egg-shaped layer of cells, the small end of the egg directed
upward and wedged in between the lobi olfactorii. The process,
if it can be called such, is the lobus olfactorius impar, and its
cavity the recessus olfactorius impar of von Kupffer (No. 69),
or, as it has been later named by Burckhardt (No. 17), the reces-
sus interolfactorius or recessus neuropricus.

The nervus olfactorius (o/, Figs. 21, 26, and 64, Pls. XXIV,
XXV, and XXXVIII) arises from the front end of the lobus,
and, in the adult, is as long as, or even one half again as long
as, the brain. It runs directly forward, for fully two thirds of
its course in the cranial cavity. It then enters the olfactory
canal (o/c) and, turning slightly outward, issues by the olfactory
foramen (o/fr, Figs. 8 and 9, Pl. X XI) on the floor of the olfac-

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. Ick

tory pit (o/p) beneath the nasal sack. Sagemehl says of it that
it consists of from seven to ten loosely connected bundles of
fibres, between which there is to all appearance no interchange
of fibres. I find it to consist of three principal bundles, two
large ones, one on each side of the nerve, united dorsally for
about one half their length, and a much smaller ventro-median
one (Fig. 64, Pl. XX XVIII), lying in the open triangular space
between the ventral ends of the other two. Between the two
lateral bundles there is dorsally, for one half their length,
frequent interchange of fibres, and the two bundles break up
distally into several bundles. Between the smaller ventro-
median bundle and the lateral ones there is also an inter-
change of fibres, but it is much less important than that
between the larger bundles. This median ventral bundle, in
whole or in part, seems to be the “hitherto undescribed cranial
nerve’ of Pinkus in Protopterus (No. 88).

In the adult the ventro-median bundle of the olfactorius is
distributed, so far as macroscopic observation can show, to the
nasal epithelium at the extreme anterior end of the nose. In
embryos such is also its apparent distribution, though I have
never been able in my preparations to trace it definitely to that
tissue. During the larger part of its course, in 30 mm. to
50 mm. specimens, it is easily distinguished from the lateral
bundles by the presence of the large round cells which Pinkus
describes in Protopterus. Near the anterior end of the ol-
factorius, however, these cells disappear, and the fibres into
which the bundle breaks up cannot be distinguished from the
other terminal branches of the main nerve. All the fibres
of the bundle apparently enter the nasal tissue, and neither
in embryos nor in the adult do I find such a mass of cells
at the end of the nose as Pinkus describes in Protopterus.
They may possibly represent in Protopterus a rudimentary
condition of the intermaxillary gland described by Gage in
Diemyctylus.

Between the ventro-median bundle of the olfactorius and the
lateral bundles there is, in large larvae as in the adult, frequent
interchange of fibres. Toward the root of the nerve in such
larvae the large round cells that characterize the bundle dis-

512 ALLIS. [Vot. XII.

appear, as they do toward its anterior end, and the fibres of the
bundle enter, in large part, with the fibres of the lateral bundles
into the anterior end of the lobus. A very small bundle, how-
ever, continues backward outside the brain membranes to the
hind end of the lobus. There in all my series of sections but
one it was lost. In that one series what seemed to be the nerve
entered the base of the brain lateral to the anterior end of the
recessus interolfactorius, that is, in Amaia, in front of the lamina
terminalis, or sulcus olfactorius, and hence into the lobus olfac-
torius and not into the fore-brain proper. This root of the
olfactorius, or nerve if it be a separate nerve, is found in the
adult in the same position as in the young, but in none of my
specimens could I trace it into the brain or find where or how
it ended proximally. It becomes delicate and disappears in the
membranes of the brain, broken doubtless in preparation. Dis-
tally it could be separated from the ventro-median bundle of the
olfactorius further forward than was possible in sections of lar-
vae, but it finally ran into that bundle, as it did in the young,
and could not be separated from it.

In 12 mm. larvae the large cells found in the ventro-median
bundle of older larvae are collected into a compact mass lying
on the ventral and median surface of the olfactorius at about
the middle of its length. The mass has a well-defined limiting
membrane, and forms a knob-like structure on the outer surface
of the olfactorius, the nerve and mass much resembling in
arrangement and general appearance the nervus oculomotorius
and associated ciliary ganglion. The cells of the mass are
exactly similar in appearance to those of the ciliary ganglion,
and it may therefore be, as that ganglion is said to be in whole
or in part, a sympathetic ganglion, possibly the sphenopalatinum
of higher vertebrates, though I found no connection whatever
between it and the nervus trigeminus. The main olfactorius,
at this age, is often distinctly double at its base, that is, it arises
by two roots, one of which, in sections, is lateral and dorsal and
the other median and ventral. With the median of the two
roots the nerve or bundle “n”’ of Pinkus is associated. In
the adult the separation of the two roots is not distinct. In
sagittal sections of the adult brain a triangular space is, how-

No. 3.]| MUSCLES AND NERVES I[N AMIA CALVA. 513

ever, found near the median ventral surface of the olfactorius
indicating a partial separation into two bundles.

Wiedersheim (No. 128, p. 256) in a figure of the brain of
Rana esculenta shows two roots of the olfactorius, but gives no
special description of them in the text. Lee (No. 72) describes
and figures a similar arrangement in Spelerpes fuscus, and
states, on the authority of Miklucho Maclay and von Rohon,
that the olfactorius is frequently found double in selachians.
Whether the two roots so described are the two main portions
of the root in Amia, or one of them the small root “‘n”’ and the
other the main double root, I am unable to judge.

No ganglion cells other than those described in the ventro-
median bundle were found in any part of the olfactorius in any
of the specimens examined. Whether these cells in the adult
and the ganglionic mass in young larvae are the olfactory gang-
lion of Beard (No. 9, pp. 878 and 884) or not I am unable to
state, not having seen his complete and later work. I am also
unable to judge whether they are, or are not, a remnant of the
ganglionic mass from which the nervus olfactorius is said by
Holm (No. 58) and Chiarugi (No. 18) to develop, my investiga-
tions being confined to embryos already too advanced for the
purpose.

The olfactory canal in Amia (o/c) is considered by Sagemehl
(No. 104, p. 217), it seems to me erroneously, as simply an
anterior extension of the cranial cavity. It is separated from
the anterior end of the orbit by a thin wall of cartilage through
which, at about the middle of the canal, from its side and bottom,
there is a large opening (onfn, Fig. 25, Pl. XXV) leading into
the extreme front end of the upper part of the orbit. This open-
ing in Amia seems to have entirely escaped Sagemehl’s notice,
although in Lepidosteus, in the Characinidae, and in teleosts in
general it and its relation to the nervus olfactorius are described
at considerable length by him (No. 104, p. 219, and No. 106, pp.
68 and 72). In Catostomus he did not find it, and he says of it
that it has no physiological importance whatever (No. 107, p.
566). It lies in Amia internal to and under the antorbital
process, directly above the obliquus superior near its origin, and
through it a large venous vessel passes from the olfactory pit

514 ALLIS. [VoL. XII.

into the orbit. The opening, however, is much larger than is
necessary simply for the passage of the vessel, and it therefore
forms a true fenestration of the side wall of the orbit, compara-
ble in every respect to the larger and more important one
described by Sagemehl in other fishes. Sagemehl gives no
name to the opening, referring to it in the text as the fenestra-
tion ff, letters which do not appear in the figures. I have called
it the orbito-nasal opening or fenestra. That part of the olfac-
tory canal that lies in front of it is formed by the fusion of
two canals, the olfactory canal proper and another canal, quite
probably the homologue of the orbito-nasal canal of selachians
(No. 44, p. 73). That canal in selachians is traversed by what
from Gegenbaur’s description is apparently, and in Callorhyncus
(No. 116, Fig. 1, and No. 59, p. 265) and Mustelus (No. 123,
Fig. 1), judging from the figures alone, by what is certainly the
ramus ophthalmicus profundus; but neither Gegenbaur, Stan-
nius, Hubrecht, nor Tiesing mentions or shows the passage of a
vein with the nerve. In Torpedo and Raja, Tiesing (No. 123,
Figs. 2 and 3) shows the profundus passing with the super-
ficialis through what Gegenbaur calls in the former the pre-
orbital incisur and in the latter the preorbital canal. What
passes through the orbito-nasal canal, the orbital opening of
which is shown by Tiesing in Torpedo but not in Raja, is not
indicated. In Amia there is no ophthalmicus profundus, but
in Polypterus that nerve is found and it passes (No. 93, p. 394)
from the orbit to the cartilaginous nasal capsule through a
“canal in the ectethmoid bone, which it occupies together with
the ophthalmicus superficialis, the obliquus superior muscle and
a considerable vein.” That there must be some canal or
channel in elasmobranchs for the passage of this vein so large
and important in Amia, and that it is probably the orbito-nasal
canal, seems sufficiently evident to warrant the use I have made
of the name.

The development of the olfactory nerves and their canals in
Amia is as follows. In specimens less than 10 mm. in length
the preorbital processes have not yet been fully formed, and the
orbit on each side is directly continuous with the olfactory pit ;
the two regions being separated by an elevation, the beginning

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 515

of the future process, through which there is a large open chan-
nel which, owing to the shape of the base of the brain and basis
cranii, runs decidedly upward as well as forward, almost at right
angles to the olfactory nerve. Following this stage the outer
edge of the open channel near its posterior end grows upward,
then inward, and then inward and downward toward the inner
edge of the channel with which, when the larvae is about 10
mm. in length, it has coalesced, forming a narrow arched bridge
over the posterior end of the channel. Under the bridge thus
formed from its inner wall, the former inner edge of the open
channel, the oblique muscles have their origin, and through the
large opening or canal, formed by the bridge, a large venous
vessel passes from the nasal to the orbital regions of the skull.
Across the anterior, open end of the canal the olfactorius runs,
nearly at right angles to it. The canal and the open channel
in front of it therefore represent the extreme, upper, anterior
end of the orbit, the orbito-nasal fenestra, and that part of the
olfactory canal that in the adult lies in front of that fenestra.
In later stages the preorbital process thickens, grows upward
along the side of the brain and then backward, forming the
beginning of the roof of the orbit. This is the condition in
larvae 12 mm. and 14 mm.in length. The lobus olfactorius at
these stages lies practically in front of the preorbital process,
for the transverse vertical sections that, in 12 mm. specimens,
pass through the process, pass also through or behind the reces-
sus interolfactorius. The side wall of the skull in front of the
process has at these ages not yet been formed, but, immediately
in front of the lobi, a vertical transverse ridge, or a median pier,
marks the anterior end of the cranial cavity. The olfactory
nerve, on each side, arises from the lower, outer, anterior sur-
face of the lobus, and runs downward, outward, and slightly
forward along the shelving lateral edge of the basis cranii.

In larvae 20 mm. in length the preorbital process on each
side has grown forward, enclosing the orbito-nasal canal, and
that canal now opens on the shelving olfactory region of the
basis cranii under or immediately in front of the anterior end of
the lobus olfactorius. Where it opens in front of the lobus a
true olfactory foramen is formed or indicated, lying between the

516 ALLIS. [Vou. XII.

opening of the orbito-nasal canal and the anterior end of the
lobus. The orbito-nasal canal in such specimens, at its anterior
end, turns upward almost at right angles to the olfactory nerve,
and immediately under and lateral to that nerve. In still later
stages the olfactory nerve is enclosed above and in front of the
opening of the orbito-nasal canal, and the anterior part of the
olfactory canal of the adult is formed. The lobus olfactorius
during the same period recedes, with the brain, from the ante-
rior end of the cranial cavity, and the long intracranial portion
of the nervus olfactorius is formed. There is never the
slightest indication of a tractus olfactorius.

The development of the olfactory parts in Amia thus agrees
exactly with that in the trout, as described by Sagemehl (No.
106, p. 76), and in the adult Amia, as well as in embryos, its
nervus olfactorius lies in a certain limited part of its course
exposed in the orbit. Amia, therefore, in this respect belongs
definitely to Sagemehl’s third or teleostean type, and it and all
ganoids and teleosts are in all probability derived directly from
his cyclostome type and not from the selachian type, as he con-
cludes (No. 104, p. 219). In those teleosts that have a tractus
olfactorius I agree with Lee that they owe it, as do the elasmo-
branchs, to special mechanical conditions that have tended to
hold the anterior end of the lobus in its embryonic position near
the nasal membrane. The separate development of the orbito-
nasal and olfactory canals, instead of their fusion as in Amia,
would seem to be either among those conditions or a secondary
arrangement resulting from them.

5. Muscles of the Eye, Nervus Oculomotorius, Nervus Trochlearis,
and Nervus Abducens.

The obliqui muscles (os and oz, Figs. 21-27, Pls. XXIV and
X XV) arise, in the adult, close together from the side wall of
the orbit at its extreme front end immediately below the orbito-
nasal fenestra. A slight depression, occupying the space between
that fenestra and the floor of the orbit, marks the place of origin,
and is unquestionably the small beginning of the anterior eye-
muscle canal of certain teleosts (No. 107, pp. 563 and 566). In

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 5i7

larvae from 12 mm. in length upward the muscles have, as
already described, a similar origin, arising immediately below,
or in the younger larvae immediately posterior to, the fenestra.

The rectus externus (ve) arises, as already stated, near the
hind end of the eye-muscle canal. The three other recti
muscles (7s, 77¢, and vzf) arise from the lateral face of the
basisphenoid, and hence at the external opening of, or partly
within, the orbital opening of the eye-muscle canal. In larvae
the internus has the most anterior origin, the inferior next, and
the superior the most posterior. The rectus inferior is said
by Sagemehl (No. 106, p. 87) to arise in the Characinidae in
the anterior portion of the eye-muscle canal.

The obliquus superior is innervated by the nervus troch-
learis (¢7), which pierces the lining membrane of the upper,
lateral chamber of the eye-muscle canal, and enters that
chamber in front of and slightly above the main trigeminal
root and the root of the ophthalmicus profundus (Fig. 26,
Pl. XXV). It lies, while in the chamber, internal to the
ophthalmic branches of the facialis and trigeminus, and leaves it
through the upper end of the large orbital opening of the canal.
It then runs forward along the inner wall of the orbit, dorsal
to all the recti muscles, and separating into two nearly equal
parts, each of which again separates into two parts, it enters
the obliquus superior on its upper surface, near the middle of
its length. In its course through the orbit it lies below and
internal to the ophthalmic branches of the facialis and tri-
geminus, above and internal to the profundus ganglion, and
above the oculomotorius. No branches of it except those
entering the muscle were found, nor were there any commu-
nicating or connecting branches to or from any of the other
nerves or ganglia. The same is true of the nerve in the
smallest larvae examined ; the nerve, however, often passes in
sections so close to the portio ophthalmici profundi that it is
impossible to say that there is there no interchange of fibres.
No ganglion or ganglion cells were found at any point in the
nerve, either in the adult or in larvae.

The rectus externus is innervated by the nervus abducens
(a6), which runs forward and then forward and outward in the

ALLIS. VoL. XII.
51

cranial cavity, a comparatively long distance from its point of
origin on the under surface of the brain, to the front edge of
the horizontal process of the petrosal. There, immediately
behind and internal to the place of exit of the main facial
nerve, it pierces the membranous roof of the eye-muscle canal,
and lying in a semi-circular foramen (adfr, Fig. 11, Pl. XX1),
at the angle formed by the horizontal process of the petrosal
and the process of that bone that forms the front wall of the
utricular fossa, enters the eye-muscle canal. Turning down-
ward it separates into two main parts which, separating again
into two or more parts, enter the externus while that muscle is
still inside the canal.

The rectus externus is sometimes double through part of its
length, one part lying, relatively to the eye, directly external to
the other.

Schneider (No. 112, p. 10), after describing the origin and
intracranial course of the abducens in Acipenser, says that
that nerve in all ganoids enters the orbit, as it does in Aci-
penser, at the front under corner of the Gasserian ganglion,
and that, in its further course, it lies in all, Amia included,
under the ophthalmicus trigemini, but over all the other
branches that arise from the trigeminal ganglion and have
a forward course. Just what is meant by “over all the
other branches”’ is not entirely evident from this statement
taken alone, but in his Fig. 5@ he shows the arrangement
of all the nerves concerned in Lepidosteus. In that figure
the abducens, after leaving the inner, anterior surface of the
Gasserian ganglion, crosses, from within outward, under the
r. ophthalmicus superficialis trigemini and over, that is
across the upper surface of, the truncus maxillaris trigemini
and r. buccalis facialis. It cannot, therefore, be doubted that
this is the further course ascribed to the abducens in all
ganoids. This, it will be seen, is markedly at variance with
what is found in embryos and in the adult of Amia, and I can
in no way account either for his observation or his statement.
As I differ as markedly and unaccountably from him in certain
other of his statements, I have not been able to accept any of
his results without much reservation.

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 519

The rectus superior, rectus inferior, and rectus internus are
all innervated by the nervus oculomotorius (ocm). This nerve
pierces the lining membrane of the cranial cavity opposite and
above the optic chiasma, and enters the orbit through the
upper end of the large orbital opening of the eye-muscle canal.
Lying dorsal to all the recti muscles at their origin, it separates
into a superior and an inferior portion. The superior portion
(ocms), dividing dichotomously, supplies the rectus superior
alone, one branch going to the upper and the other to the
under surface of the muscle, each branch dividing dichotomously
before reaching the muscle. The inferior, or main portion (ocmz?)
of the nerve runs forward and downward between the rectus
superior and the rectus externus, and behind or external to the
rectus inferior. It then turns forward below the rectus inferior
and rectus internus, and sends two branches, arising close
together, if not from a common root, one to the upper and the
other to the under surface of the former muscle, and two
branches, arising some distance apart, each of them double, to
the latter. The nerve then separates into two parts, which
again divide dichotomously and enter and end in the obliquus
inferior, entering the muscle on its under and upper sur-
faces.

The inferior and internal recti were, in the specimen used
for illustration, connected about midway of their length by two
or three muscle strands, which arose distally in each muscle
—that is, toward their insertions— and formed a festoon con-
necting them. In another specimen the two muscles arose as
a single muscle, which soon separated into two parts.

6. Review and Comparison of the Muscles and Nerves of the Eye.

In Amia the obliquus superior is innervated by the troch-
learis, the rectus externus by the abducens, the rectus superior
by the superior branch of the oculomotorius, and the rectus
inferior, rectus internus, and obliquus inferior, in the order
named, by the inferior branch of the oculomotorius. The
inferior branch of the oculomotorius runs forward and down-
ward behind and below the rectus superior, above and in front

520 ALLIS. PVOE aie

of the rectus externus, behind and below the rectus inferior,
and below the rectus internus.

In all other ganoids, both chondrostean and holostean, the
same arrangement is found ; at least I so interpret the descrip-
tions given by van Wijhe (No. 129) of the muscles and nerves
in Acipenser sturio, Polypterus bichir, and Lepidosteus osseus,
and these three fishes may be assumed to represent all ganoids.
Schneider says (No. 112, p. 12) that in Acipenser, Scaphirhyn-
chus, and Amia the inferior branch of the oculomotorius, after
passing between the rectus superior and rectus externus, con-
tinues forward in the ‘“‘Kegel”’ formed by the recti muscles.
This would seem to mean that the oculomotorius runs forward
above the rectus inferior and rectus internus, but in his Fig. 1,
showing the arrangement in Acipenser, the nerve passes below
the rectus inferior, as it does in van Wijhe’s Fig. 3 of the same
fish. In Amia my work leaves no doubt as to the arrangement ;
the nerve runs below the two muscles.

The same arrangement of the nerves and muscles of the eye
is probably found in all teleosts. The obliquus superior and
rectus externus are certainly innervated, respectively, by the
trochlearis and abducens, as in ganoids, and the remaining
muscles by the oculomotorius, but that this nerve has, in all
teleosts, the same distribution and the same relation to the
muscles, as in ganoids, cannot be definitely determined from
the literature. Most references simply say that the four
muscles are innervated in the usual or well-known way. In
Amiurus, where, according to Wright (No. 132, p. 365), the
inferior branch of the oculomotorius runs over the inferior and
internal recti, it was found in its usual position wxder those
muscles in the one specimen that I examined.

In the Chondropterygii, or Elasmobranchii, a somewhat dif-
ferent arrangement is found. In these fishes the superior
portion of the oculomotorius supplies the rectus internus as
well as the rectus superior, the inferior portion of the nerve
supplying only the rectus inferior and the obliquus inferior.
In Holocephala the rectus internus arises near the front end
of the orbit, at some distance from the other recti muscles,
and the nerve supplying it runs forward above the opticus

No. 3.| MUSCLES AND NERVES IN AMIA CALVA. 52 i

(No; 193, Fig. 12, and: No: 116,,Fig: 1).’- In: Plasiostomata
the internus arises with the other recti at the hind end of the
orbit, but both muscle and nerve lie above the opticus, while
in ganoids and teleosts they both lie below it. In the young
of Scyllium the rectus internus is innervated by an anterior
branch of the superior division of the oculomotorius (No. 77,
p. 88). In the adult Scyllium this anterior branch may
become somewhat widely separated from the branch to the
superior rectus, two superior branches thus arising from the
main nerve, — the proximal one of the two going to the rectus
internus (No. 113, Fig. 10). In Mustelus laevis the oculomo-
torius, after sending branches to the internal and superior recti,

rocms ad rs

Cur 1.— Top view of left eyeball of Galeus canis. oft, ophthalmicus superficialis trigemini ; os,
ebliquus superior; ve, rectus externus ; ~z¢, rectus internus; vocms ad rs, branch of superior
division of oculomotorius to rectus superior ; ~s, rectus superior; 7, trochlearis.

perforates, according to Schwalbe (No. 113, p. 185), the rectus
internus near its upper posterior edge, while according to
Tiesing (No. 123, p. 79) it perforates the rectus superior. It
then runs downward behind the rectus inferior, and forward
below that muscle, and ends in the obliquus inferior. In
Galeus canis I find a similar arrangement (Cut 1). In Scyl-
lium catulus (No. 113, p. 187) the inferior division of the
oculomotorius penetrates the rectus superior near its hind
edge, and in Raja batis (No. 113, p. 188, and No. 123, p. 80)
it runs down behind that muscle, in both cases turning for-
ward below the rectus inferior. In Chimaera, judging from
Schwalbe’s figure alone (No. 113, Fig. 12), it runs down in
front of the rectus superior, while in Callorhynchus (No. 116,
Fig. 1) it apparently lies behind that muscle.

522 ALLIS. [Vou. XII.

These different relations of the oculomotorius to the internal
and superior recti, in elasmobranchs, are due to and are caused
by the gradual shifting from before backward of the origins of
all the recti muscles, and also of the place of exit of the oculo-
motorius from the cranium. As a result of this shifting the
internal and superior recti, at their origins, either traverse, or are
traversed by, the issuing nerve. The nerve does not traverse the
inferior rectus also, by which it may, as in Galeus canis (Cut 2),
be pulled backward and much diverted from its course, simply
because, crossing the muscle as it does midway in its length,
such a process is, or at least would seem to be, impossible
without destroying or disturbing functionally the muscle.

Cut 2.— Same as Cut 1 with rectus superior and obliquus superior muscles removed. ocw2, oculo-
motorius; ocm?, inferior branch of oculomotorius ; 044, ophthalmicus profundus trigemini; rz/,
rectus inferior.

In Petromyzon (No. 37, pp. 38, 58, and 70) the trochlearis
innervates the superior oblique, which arises from the mem-
branous hind wall of the orbit, and the abducens innervates
the inferior as well as the external rectus. The trochlearis, as
it issues from the cranium, lies dorsal to the abducens and
apparently dorsal also to the main trunk of the trigeminus, for
Fiirbringer, speaking of the roots of the three nerves and their
relations to each other as they issue from the cranium, says of
the trochlearis, that it is ‘‘der am oberflachlichsten gelegene.”
If such be its position it must, as it issues from the cranium,
run backward and outward above and across the other two
nerves. It must therefore lie dorsal to the r. ophthalmicus
trigemini, a position confirmed by Ahlborn (No. 1, p. 21). This
ophthalmic nerve in Petromyzon does not, therefore, probably

No.3: MUSCLES AND NERVES IN AMIA CALVA. 523

contain any fibres that are homologous with those that form
the r. ophthalmicus superficialis trigemini of Amia, teleosts,
and selachians, for in all these latter fishes the trochlearis lies
under the nerve. It is more probably the homologue of the r.
ophthalmicus profundus of other fishes, that nerve, when found,
lying always under the trochlearis, as will be shown later.

The abducens in Petromyzon arises (Ahlborn) close in front
of the ventral, motor root of the trigeminus, not ventrally near
the median line as in other fishes, runs upward inside the cranial
wall, and issues (Firbringer) dorsal to — but, as shown in his
Fig. 19, behind — the trigeminus. It must, therefore, acquire
this position by crossing first the ventral and then the lateral
or posterior surface of the trigeminus while the two nerves are
still inside the cranium. That it runs upward internal to, that
is anterior to, the trigeminus, and then crosses the upper sur-
face of that nerve from before backward, seems wholly improb-
able. The descriptions, however, are not concise. It lies below
the trochlearis, sends a short branch to the rectus externus, and
then continues downward and outward behind that muscle to
the rectus inferior,in which it ends. The rectus superior is
innervated by a posterior branch of the oculomotorius, sent out-
ward, according to Fiirbringer, below the ramus ophthalmicus,
a wholly unaccountable position if the ramus ophthalmicus be
the ophthalmicus profundus of other fishes, as its relation to the
nervus trochlearis seems to indicate. The rest of the oculomo-
torius, the inferior division of the nerve, runs forward under
the opticus and there separates into several branches, all of
which enter the obliquus inferior, that muscle being called by
Fiirbringer the obliquus anterior. The branches all enter the
muscle on its outer surface, according to Fiirbringer’s text, but
on its upper or inner surface, as his Fig. 19, giving the arrange-
ment in Petromyzon marinus, plainly shows. Of the several
branches that enter the obliquus, two pass through it to supply
the rectus internus, which lies immediately below the obliquus,
and arises not far from it near the front end of the orbit.

In Dipnoi the muscles of the eye are small and the obliqui
are said to be wanting in certain species. In Protopterus
annectens, where all six muscles are found (No. 88, p. 281),

524 ALLIS. [Von IT:

they are innervated as in elasmobranchs, that is, the rectus
superior and rectus internus both by the dorsal or superior
branch of the oculomotorius.

In Amphibia the muscles of the eye are innervated, accord-
ing to Schwalbe, in Urodela, as they are in elasmobranchs, and
in Anura, as they are in ganoids and teleosts (No. 113, pp. 196,
197, and 200). In Salamandra the trochlearis and the branch
of the oculomotorius to the rectus superior are often more or
less completely fused with the r. nasalis trigemini, and by
Fischer (quoted by Schwalbe) they are described as branches
of that nerve. The branch of the oculomotorius to the rectus
internus is always independent, and lies above the r. nasalis,
according to Schwalbe. According to von Plessen and Rabino-
vicz (No. 92) it would seem to lie in Salamandra maculata, below
that nerve. By these authors both oblique muscles in Sala-
mandra are said to be innervated by branches of the ramus
nasalis, and they failed entirely to find a nervus trochlearis. In
Amblystoma the oculomotorius and all its branches are shown
by Herrick (No. 55) external to, and hence above, the nervus
opticus and ramus ophthalmicus trigemini, in a side view of
the head, but below both these nerves in a top view. In one
of the figures the terminal branch of the nerve innervates the
rectus inferior and obliquus inferior, while in the other the
obliquus is innervated by what seems to be the second branch
of the nerve, this branch innervating also the rectus internus.
From the descriptions of the nerves it cannot be told which
of the figures is correct, or if either of them is. Schwalbe’s
descriptions seem, therefore, to be the most reliable and must
be relied upon for the whole class.

In Apoda an arrangement similar to that in Urodela is indi- —
cated in Fig. 72, Tab. XIX of the Sarasins, thus confirming
what the Sarasins conclude from other reasons; that is, that
Apoda is a subdivision of Salamandrina (No. 108, p. 29).

In both Urodela and Anura a small muscle is found, the
retractor bulbi, not found in fishes. Schwalbe says that this
muscle in Salamandra is innervated by a branch of the inferior
division of the oculomotorius, given off before the branch to
the rectus inferior (No. 113, p. 198). In the sheep, calf, and

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 525

dog, where the muscle is also found, he finds the lower portion
of it innervated by a similar branch given off close to the branch
to the rectus inferior (No. 113, pp. 218 and 220). This inner-
vation of the retractor bulbi does not agree with that given by
others. Wiedersheim says (No. 128, p. 339) that it is inner-
vated by the abducens. Huxley says (No. 60, pp. 67 and 74)
that in all vertebrates “‘the muscles of the nictitating mem-
brane, and the retractor bulbi, or musculus choanoides, when
such muscles exist,” are innervated by the abducens, and that
the musculus choanoides when it exists ‘lies within the four
recti, and is attached to the circumference of the posterior
moiety of the ball of the eye.’” Mivart also (No. 81), in the cat,
gives the innervation of the choanoid muscle by the abducens,
and Hoffmann (No. 56, p. 209) says that in Lacerta the rectus
externus and retractor bulbi develop from the same somite.
Schwalbe’s statement may, therefore, be considered as due to
some error in observation. It, however, finds some support by
what Herrick finds in Amblystoma, for after stating (No. 55, p.
200) that the abducens separates peripherally into two branches,
one of which goes to the retractor bulbi, he adds, “ Although
the two branches of the sixth nerve are indistinguishably blended
when they leave the Gasserian ganglion, yet it would seem that
the branch destined for the m. retractor bulbi ought not to be
assigned to this nerve, but to the trigeminus.”” Herrick further
states that, in Salamandra maculata the retractor is innervated
by the oculomotorius, the statement being probably made on
the authority of von Plessen and Rabinovicz (No. 92, p. 7).

In reptiles, birds, and higher vertebrates the six muscles of
the eyeball are, according to Schwalbe, innervated as in Anura,
for in speaking of the latter he says: ‘‘Das Innervationsschema
der Augenmuskeln der Anuren weicht in nichts von dem
bekannten der hdheren Wirbelthiere ab.” In these higher
forms there may be a musculus choanoides, as already stated,
and there is usually a levator palpebrae superioris, innervated,
according to Schwalbe, when it exists, by a branch of the
superior division of the oculomotorius.

It is thus evident that the muscles of the eye in vertebrates,
although performing similar functions, are not homologous

526 ALLIS. [Vou. XII.

structures. It is also evident that this want of homology is
found entirely in those muscles that are innervated by the
oculomotorius and abducens, that is, in those muscles that are
said to arise in all the Gnathostomata from van Wijhe’s first
and third somites. As the several muscles that operate the
eyeball have been developed from the muscle masses in these
two somites, different arrangements have arisen, which, if it be
assumed that reversions have not taken place, must indicate
totally distinct and different lines of descent. The several
arrangements characteristic of these lines may have arisen
independently and simultaneously from the undifferentiated
condition, or some one or two primary arrangements may have
first so arisen, and from them, by later development, the exist-
ing arrangements. This latter method of development seeming
the more probable, I have attempted to construct a prototype
from which existing arrangements might have been derived.
The superior rectus in ganoids, teleosts, and Anura has
unquestionably been developed from a muscle mass which has,
in Elasmobranchii, Dipnoi, and Urodela, given origin to two
muscles, the rectus superior and rectus internus. As the
rectus internus in Holocephala arises near the front end of the
orbit, at some distance from the place of origin of the rectus
superior, it may be safely assumed that it represents a condition
more primitive than that found in the Plagiostomata, where the
two muscles arise together at the hind end of the orbit; for a
muscle once having arisen with the other recti at the hind end
of the orbit, would not, in all probability, have changed that
place of origin for the apparently less advantageous one near
the front end of it. It may, therefore, be assumed that the
rectus superior of ganoids was formed by the fusion of the two
muscles of elasmobranchs, rather than that it, by splitting,
gave origin to the latter. The elasmobranch type would,
therefore, be more primitive than the ganoid type, and, as a
consequence, the internal and inferior recti of the latter would
be formed by the splitting of the inferior rectus of the former,
instead of by their fusion giving origin to it. We therefore
conclude that in the prototype of all vertebrates, the Cyclosto-
mata excepted, the superior branch of the oculomotorius inner-

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. hay,

vated two muscles, a rectus superior and a rectus internus, and
that the inferior branch of the nerve also innervated two
muscles, a rectus inferior and an obliquus inferior.

Turning now to the Cyclostomata we find that in Petromyzon
the rectus internus arises near the front end of the orbit, and
is innervated by branches of the oculomotorius which first pass
through the obliquus inferior near the middle of its length.
This manner of innervation renders it highly probable that the
muscle could not have been derived from, or have given origin
to, any of the recti muscles known in other fishes, for even if
the muscle should acquire an origin with the other recti, the
inferior branch of the oculomotorius would still lie wholly above
it, a condition not known, so far as I can find, in any other
animal. The position and innervation of the muscle both
indicate its derivation, in Petromyzon, from the outer, under
surface of a muscle or muscle mass giving origin to it and to
the obliquus inferior. The embryonic condition of such a
muscle mass is described by von Kupffer (No. 70) in Ammo-
coetes. It is said by him to be the lateral and lower portion
of the subcerebral part of the trabecular arch, that arch being
the first, or most anterior, of the visceral arches. This same
muscle, or muscle mass, must, in the prototype of the Gnatho-
stomata, have given origin to the obliquus inferior and rectus
inferior, the two muscles innervated by the inferior branch of
the oculomotorius. The outer muscle of the two that so
develop in the Gnathostomata must, however, have become
the obliquus, instead of the inner one, as in Petromyzon ; and
the nerve innervating the two muscles, in the Gnathostomata,
must have run downward along the posterior edge of the mass,
instead of through it, as in Petromyzon. The inner muscle of
the two, the inferior rectus, as it separated from its companion
and acquired an origin with the other recti or the hind end of
the orbit, would then have pushed the inferior branch of the
oculomotorius before it, keeping it always posterior to and
below it, as it is always found.

That the inferior rectus, in the Gnathostomata and higher
vertebrates, arises as above supposed, is strongly indicated by
the early development of the muscles of the eye in elasmo-

528 ALLIS. [VoL. XII.

branchs and in Apoda, as given by van Wijhe, Platt, and the
Sarasins. In Scyllium and Pristiurus, van Wijhe concludes
(No. 130, p. 14) that it arises mainly from the dorsal wall of the
first head cavity, but, in part at least, from an anterior process
of the ventral wall of that cavity, — the obliquus inferior arising
from a second posterior process of this same wall, and the rec-
tus superior and the rectus internus entirely from its dorsal
wall. In Acanthias vulgaris (No. 90, pp. 83, 86, and 103) Platt
says that the inferior oblique arises from the dorsal wall of a
posterior process or constriction of the premandibular or first
head cavity, the inferior rectus from the posterior portion of
the dorsal wall of the remaining portion of the cavity, and the
superior and internal recti from its anterior portion. In her
figures the rectus inferior seems to lie immediately above the
obliquus inferior, and the inferior branch of the oculomotorius
runs outward and downward behind it. In larvae of Ichthyo-
phis the rectus inferior overlaps internally the obliquus inferior
(No. 108, Fig. 72). Whether the head cavities here referred to
belong to the somites or to the visceral arches is unimportant
for my purpose, provided the muscles of the eye in the Cyclo-
stomata and Gnathostomata are of homologous origin, which
no one would, I think, venture to question.

We thus see that, in so far as the muscles innervated by the
inferior branch of the oculomotorius are concerned, the Cyclos-
tomata and Gnathostomata represent two totally distinct and
different lines, and that they must both have descended from a
prototype in which there was but a single muscle, or muscle
mass, innervated by the nerve in question. Assuming a single
muscle to have been first formed it was probably an obliquus,
that muscle being still, in the adult Petromyzon, the largest of
all the muscles of the eye. We accordingly arrive at a proto-
type for all fishes, in which the oculomotorius would innervate
but three muscles, the superior and internal recti by a supe-
rior branch running forward above the optic nerve, and the
inferior oblique by an inferior branch running downward in
front of the rectus superior, and then forward below the opti-
cus. This would leave the inferior and external recti to be
innervated by the abducens, as they are in Petromyzon, and

No: 3:] J7ZOSCLES. AND NERVES [N AMIA CALVA. 529

the superior oblique by the trochlearis, as it is in all fishes.
This last muscle is developed from van Wijhe’s second myo-
tome, and accordingly lies, primarily, posterior to all the other
muscles. Its position, therefore, in Petromyzon, where it arises
from the membranous hind wall of the orbit and is inserted on
the under surface of the bulbus, probably represents a primitive
condition. The position toward the front end of the orbit,
found in all other fishes, must then have been acquired, as
Firbringer suggests, by the gradual shifting of the origin of
the muscle along the upper edge of the orbit. Nothing in the
arrangement of the nerves of the orbit would have interfered
with this change of position zf the r. ophthalmicus in Petromy-
zon be the r. ophthalmicus profundus of other fishes, nor would
the muscle, during the process, have been ever functionally at
a disadvantage. What is apparently an intermediate stage
in this shifting process is shown in Callorhynchus, where the
muscle still arises from the upper edge of the orbit dorsal to
the ophthalmicus superficialis, — the trochlearis, which inner-
vates the muscle, lying, as it should, below that nerve. Other
intermediate stages are probably presented by the Monotre-
mata, for Goppert (No. 46) finds, in Echidna, fibres of the
obliquus superior arising from the frontal near the Trochlea,
that is, from about the place where the whole muscle arises in
Callorhynchus ; and in both Ornithorhynchus and Echidna he
finds the muscle arising less deep in the orbit than in other
mammals.

In Fig. 12, Pl. XXII, the arrangements of the muscles
and nerves of the eyeball found in the several orders of the
Ichthyopsida, and assumed to be found in the several proto-
types of those orders, is represented diagrammatically. From
the prototype of all vertebrates there are two lines of descent,
the impulse leading to each being the splitting of the large
inferior oblique. In one line, the outer of the two muscles
thus formed becomes a rectus internus, and the earlier muscle
of that name, innervated by the superior branch of the oculo-
motorius, disappears, or fuses with the rectus superior, giving
rise to the arrangement found in Petromyzon. In the other
line, the inner of the two muscles formed becomes a rectus

530° ALLIS. [Vov. XII.

inferior, and the earlier inferior rectus, innervated by the
abducens, either fuses with the rectus externus, thus giving
rise to a holocephaloid arrangement of muscles and nerves,
or, if Huxley’s statement as to the innervation of that muscle
be accepted, persists as the retractor bulbi, thus giving rise
to the arrangement found in Urodela. Assuming this last
arrangement to be the one first formed, there would arise what
may be called a proto-urodelian type from which there are
three lines of descent. One of these lines leads to the Holo-
cephala and Plagiostomata, in both of which the retractor bulbi
disappears, and in the latter of which the rectus internus
acquires an origin with the other recti at the hind end of the
orbit. A second line leads to the Ganoidei and Teleostei,
where the retractor bulbi likewise disappears, probably by fusion
with the rectus externus, as the frequently double muscle and
double nerve in Amia indicate, and where a still further
change is brought about by the formation of a new internus
by the splitting of the rectus inferior. As this new muscle
develops, the old internus disappears ; thus leaving but one
muscle innervated by the superior branch of the oculomoto-
rius. A third line leads to Amphibia and the higher verte-
brates, the Dipnoi either arising from this line or from the line
leading to elasmobranchs. The amphibian part of this third
line separates into two parts, one leading directly to Urodela
with but little change, and the other to Anura, where a new
internus is formed, as in Ganoidei and Teleostei, the old inter-
nus disappearing as in those fishes. In the higher vertebrates
the arrangement is as in Anura, but there is often a levator
palpebrae superioris, innervated, when found, by a terminal
branch of the superior branch of the oculomotorius. This
muscle thus seems to be developed from the disappearing
rectus internus of the prototype, and the line leading to the
higher vertebrates must accordingly lie between those leading
to the two divisions of Amphibia, just as the line leading to
Amphibia lies between those leading to the two great divisions
of Pisces.

In Ichthyophis a retractor tentaculi has been formed from
one of the muscles of the eye (No. 108, p. 200), probably from

Noa3:|) JZUSCLES AND NERVES IN AMTA CALVA. 531

the rectus internus, and not from the retractor bulbi, as the
Sarasins suggest. Ichthyophis, therefore, presents a condition
intermediate between Urodela and Anura, and may represent
the beginning of the line leading from Amphibia to Sauropsida.
Burckhardt (No. 16) also places Ichthyophis intermediate
between Anura and Urodela, but derives it from Protopterus,
the Sauropsida being derived from Ceratodus and representing
a totally different line. In a later work I believe he has stated
that it represents, in the arrangement of the different parts of
the brain, a type leading directly to Sauropsida and the higher
vertebrates. I do not, however, find the reference.

In the higher vertebrates the retractor bulbi may persist
or it may disappear, probably by fusion with the rectus externus,
as indicated by the occasional double muscle in man, and also
by its innervation in man, the abducens entering the orbit
between the two heads of the muscle (No. 100, vol. II, pp. 290
and 291).

In this ancestral tree of the muscles and nerves of the
eyeball the several lines of descent are sharply and positively
defined, but they are based on most insufficient and perhaps
inaccurate data. The important assertion that reversions have
not occurred must also be granted. If that be allowed, the
other assumptions, or rather deductions, are not important or
improbable, and the schema becomes an ancestral tree of ver-
tebrates. In this tree it will be noticed that the lines leading
to the higher types of each class resemble each other in that
the superior branch of the oculomotorius in such lines inner-
vates but one of the muscles of the eye; that in the lines
leading to the lower types it always innervates two ; and that
the line leading from one class to the next higher arises be-
tween the lines leading to the two types of the class ; that is,
from the so-called proto-urodele type upward there has been
in the development of the muscles of the eyeball but one
impulse, if it may be so called, leading to the formation of
the arrangements found in higher types, and not repeated ones.

The forms indicated in the schema as intermediate ones are
the Holocephala, Polypterus, Ichthyophis, and the Monotre-
mata, and the grouping of the several orders is exactly the same

532 PA TEI ES Yo ba. @ hE

as that given by Hasse (No. 52, p. 84, and No. 53, p. 296),
based on the development and structure of the vertebral
column, and by Maurer (No. 78, p. 476), based on the develop-
ment of the muscle cells and muscle fibres. It differs from
that recently given by Dollo (No. 23), in that he derives the
Dipnoi and Amphibia directly from the “ Ganoides crosso-
pterygiens,”’ instead of from some form intermediate between
the teleostomes and elasmobranchs.

7. Ramus Ophthalmicus Profundus Trigemini.

The root of the ophthalmicus profundus in the adult Amia
arises from the median or anterior side of the root of the tri-
geminus, at a considerable distance from the base of that root,
but still inside the cranial cavity (7f, Figs. 26 and 27, Pl. XXV,
and Figs. 38 and 39, Pl. XXX). It pierces the lining mem-
brane of the lateral chamber of the eye-muscle canal, and
enters that chamber in front of the root of the trigeminus, but
behind the trochlearis and oculomotorius. At the front end
of the chamber, or in the passage leading from it into the
orbit, it enters the profundus ganglion (g/) which lies near the
upper end of the passage, internal to the pedicle of the ali-
sphenoid, and external to the thin projecting plate of that bone
to which the lining membrane of the eye-muscle canal and of
the optic foramen is attached. The ganglion, in the adult, is
somewhat arrowhead in shape, with the point of the arrow
directed backward where the radix profundi joins it. It lies
immediately below the trochlearis and above the oculomo-
torius, but neither in the adult nor in larvae could any connec-
tion with either of those nerves be found. In larvae the
ganglion lies so close to the oculomotorius that it is difficult to
determine whether there is or is not any connection between
them. None could be established in any of my sections.

From the upper, anterior corner of the ganglion the portio
ophthalmici profundi of van Wijhe (fofp), which may be single
or double, arises ; from the lower corner the radix longa (77) ;
between these two, from the front edge of the ganglion, close
to the base of the portio ophthalmici profundi two ciliary

» Ne

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 33

nerves, the rami ciliares longi (c/) ; and between them and the
radix longa a nerve which was not satisfactorily traced in any
of the dissections made. When this last nerve was found, it
always accompanied the ciliary nerves as they ran forward and
outward between the external and superior recti. Beyond that
point it was always lost, appearing sometimes to fuse with the
ciliary nerves, and at others to disappear in the general tissues.

The portio ophthalmici profundi runs upward and forward
above all the muscles of the eye, and joins the ophthalmic
branch of the trigeminus while that nerve still lies under the
overhanging cartilaginous roof of the orbit. The two nerves
then fuse, the fusion being so complete that the two could not
be separated in dissection.

The two rami ciliares longi run outward, upward, and for-
ward above the rectus externus, and below the rectus superior,
and then join and become attached to the vein ov’, which runs
from the orbital sinus to the outer edge of the eyeball. They
lie along the upper, posterior side of the vein, and pierce
the cartilaginous capsule of the eye internal to and below the
point where the vein leaves it ; that is, near the insertion of
the rectus superior, between it and the rectus externus. In
larvae the nerves could be traced outward, along the inner
surface of the sclerotic, towards the lens and cornea.

The radix longa runs downward and forward toward the
inferior branch of the oculomotorius, which it accompanies,
closely applied to it, in its course between the superior and
external recti. Immediately beyond those muscles it enters
a small ganglionic swelling, the ciliary ganglion (gc), which lies
directly in its course and, like it, is closely applied to the
oculomotorius. Between the radix longa and the oculomo-
torius no connecting or communicating fibres were found ;
between the ciliary ganglion and the nerve such fibres, rep-
resenting the radix brevis, were always found. From the
ganglion, and as a direct continuation of the radix longa, a
large nerve, the ciliaris brevis (cd), runs outward and forward
along the outer surface of the envelope enclosing the efferent
pseudobranchial artery, and enters the capsule of the eye with
that artery through the opening for the optic nerve. No sym-

534 ALLIS. [VoL. XII.

pathetic fibres could be traced to the ganglion, nor could I
find any indication of a branch running forward along the
oculomotorius to and beyond the inferior oblique muscle, such
as Ewart describes in sharks and rays (No. 30).

In larvae, the branch of the profundus ganglion found in the
adult between the ciliares longi and the radix longa was not
found, a third portio ophthalmici profundi connecting the pro-
fundus ganglion with the ophthalmicus superficialis trigemini
was, however, often found ; always much shorter than the other
portions, and always running directly outward from the ganglion
to the nerve, instead of outward and forward. This third
branch was never found in the adult, owing doubtless to the
fact that the dissection at this point is exceedingly difficult.
A tough membranous tissue surrounds all the parts, and six
dissections, at least, of this particular region had to be made
before one sufficiently complete for illustration had been
obtained.

The ciliary ganglion in larvae always appears in transverse
sections as a well-rounded mass attached to the upper, outer
surface of the inferior branch of the oculomotorius, relatively
much larger than in the adult, and enclosed in a thick fibrous-
like envelope. It contains small cells, similar in appearance
to those found in the ganglion on the under surface of the
olfactorius, and not at all like those found in the profundus
and the other cerebral ganglia. No ganglion cells were found
in any part of the oculomotorius itself, and the ganglion cells
described in that nerve in the adult of other forms, and con-
sidered by different investigators as the ganglion or ganglions
of the oculomotorius, are in all probability simply parts of the
ciliary ganglion of Amia scattered along the ciliary nerves or
roots, as Ewart finds them in sharks and rays (No. 27), instead
of being collected into a distinct and definite mass, as in Amia.
I do not mean by this to express any opinion as to the origin
of the cells, for the disposition of the parts in the youngest
stages of Amia that I have examined might be taken to
indicate an origin from or in connection with the profundus
ganglion alone, or from that ganglion and the oculomotorius
combined.

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 535

8. Review and Comparison.

In Laemargus (No. 26) the profundus arises by a root which
is fairly separate and distinct from the root of the trigeminus,
but connected with it by several small communicating branches.
In Raja (No. 26) the roots of the two nerves are more inti-
mately connected. In both Laemargus and Raja, and in Tor-
pedo also (No. 28), the ganglion of the profundus is apparently
wholly separate from and unconnected with the trigeminus and
its ganglion, for Ewart does not mention any communicating
branches in any of those fishes. He, however, says, in
another publication (No. 27), that in skates the profundus
ganglion is frequently connected by a communicating branch
with the Gasserian ganglion. In sharks other than Laemargus,
Ewart, in his preliminary communication, does not state defi-
nitely whether there is or is not a separate profundus ganglion;
in his complete work, in a diagram indicating the distribution
of the cranial nerves in selachians in general (No. 29, Fig. 2)
he, however, shows an entirely separate ganglion and root.
Beard says (No. 8, p. 570) that in sharks the ganglion (meso-
cephalic) and its root are found distinct in early stages of
development, but in later ones both become fused with the
ganglion and root of the trigeminus.

Ewart, in his preliminary publication, describes but one nerve
arising from the profundus ganglion in Laemargus. It is the r.
ophthalmicus profundus. From it a small branch passes out-
ward above the rectus superior, in the place, therefore, where
the portio ophthalmici profundi is found in Amia, and then,
from the main nerve, two or three ciliary nerves, presumably
the ciliares longi and ciliares brevi combined, arise and, running
forward between the rectus superior and rectus externus, enter
the eyeball. No ciliary ganglion is described, but delicate
communicating fibres are said to pass from the inferior branch
of the oculomotorius to the ciliary nerves. The main nerve
is then said to bend inward between the obliquus superior and
the rectus internus, and end in the cutaneous and subcutaneous
tissues in the snout. In its forward course small branches are
sent outward, in front of the eyeball, and others to the skin

536 ALLIS. [Vou. XII.

and tissues of the snout. In his later publication (No. 29, p.
75), which seems to give the distribution of the nerves in sharks
in general rather than in Laemargus in particular, Ewart
describes a ciliary ganglion lying at the junction of the pro-
fundus and oculomotorius fibres. From this ganglion the
ciliares brevi arise, the ciliares longi arising from the main r.
ophthalmicus profundus, as in Laemargus in particular. The
profundus is here said to pass under the rectus internus instead
of over it, between it and the obliquus superior, as in the pre-
liminary publication. This earlier statement is, therefore,
probably an error. It apparently never fuses with the r.
ophthalmicus superficialis trigemini, and Ewart says it has
a distribution similar to that in Raja and in Amia. No
reference is given, in this particular instance, indicating his
authority for the statement regarding Amia. It is, however,
probably based on my earlier work (No. 3), as definite reference
is made to that work in describing the nerve in Raja (No. 32,
p. 90). The nerve referred to in Amia is, therefore, probably
the portio ophthalmici profundi, and not the ramus ophthalmicus
profundus.

In Scyllium, Schwalbe (No. 113) says that the profundus
perforates the rectus superior near its hind edge and then runs
forward under the rectus internus and obliquus superior, pass-
ing, however, in its course through the sclerotic, a fact con-
firmed by Marshall and Spencer (No. 77) in the embryo. In
Hexanchus it apparently has a similar course (Gegenbaur).
In Galeus (see Cut 1) it lies above the rectus superior, running
downward in front of that muscle, through the rectus internus,
near its hind edge, and then forward below the rectus internus
and obliquus superior. In Mustellus, Schwalbe gives it a
similar course, but Tiesing (No. 123) says that it issues below
the rectus superior and lies below that muscle, below the
internus, and below the obliquus superior. Whatever its rela-
tions to the muscle of the eye, it always lies behind, or external
to, and then below, the superior branch of the oculomotorius,
and above the inferior branch of that nerve; that is, it crosses
the upper surface of the oculomotorius distal to its superior
branch, and then runs forward always under the trochlearis.

Nox3:] J7CUSCLES AND NERVES IN AMIA CALVA. Seyi

In embryos of Acanthias vulgaris, Mitrophanon (No. 80, pp.
178 and 179) describes both a ramus ophthalmicus superficialis
trigemini and a ramus ophthalmicus profundus trigemini. The
former arises from the anterior root of the trigeminal Anlage,
in connection with the Gasserian ganglion. It enters distally
into connection with the nervus trochlearis. The profundus
arises from an ophthalmic prolongation of the posterior root of
the trigeminal Anlage, sends a delicate branch, the portio oph-
thalmici profundi, to the epidermis, and then enters the ciliary,
or mesocephalic, ganglion. From this latter ganglion a gan-
glionic mass separates the ganglion of the oculomotorius, and
from this ganglion the primitive nervus oculomotorius is de-
veloped. The ciliary ganglion, here described, is certainly the
profundus ganglion of Amia, as the name mesocephalic indi-
cates, but its origin from the posterior root of the trigeminal
Anlage seems to differ from the origin of the ganglion in Amia,
as will be shown later. The ganglion of the oculomotorius
seems, with equal certainty, to be the ciliary ganglion of Amia.
The presence of both a superficialis and a profundus trigemini,
and of a portio profundi also, is of interest. The relations of
the three nerves to the muscles of the eyeball and to the
oculomotorius are not given. The nervus trochlearis is, how-
ever, said to pass above the ramus ophthalmicus superficialis
facialis, which is exceptional.

In Raja, taken by Ewart as a type for skates in general, the
ciliary nerves arise in part directly from the profundus ganglion
and in part from the ciliary ganglion (No. 26, p. 535 and No.
27, p. 289). The ophthalmicus profundus in the same fish
lies, according to Schwalbe, below the usual two recti muscles
and the obliquus superior, and always below the trochlearis and
below the superior branch of the oculomotorius. Tiesing gives
it a similar course in the Batoidei examined by him, adding to
what Schwalbe has stated, the fact that it lies dorsal to the
eyestalk. No ciliary nerves arise, according to him, either from
or in connection with it; they arise from the oculomotorius.

In Callorhynchus the ciliary ganglion and ciliaris brevis
are not given by Stannius; the ciliaris longus is said by him
to arise from the ophthalmicus profundus, as in sharks. In

538 ALLTS. [VoL. XII.

Chimaera, Schwalbe does not give either the ciliaris longus or
brevis, but he describes a ciliary ganglion on the inferior branch
of the oculomotorius. In both Chimaera and Callorhynchus
the r. ophthalmicus profundus crosses the orbit below the
superior and internal recti and below the superior oblique, and
it may even lie below the inferior oblique, if Stannius’ figure is
correct. It always lies, as it does in the Plagiostomata, below
the trochlearis and below the superior branch of the oculo-
motorius, and Schwalbe even shows it in his Fig. 12 as lying
below the inferior branch of that nerve, unquestionably an
error.

In elasmobranchs the ophthalmicus profundus, its ganglion,
and its ciliary branches are thus seen to present several some-
what different arrangements. The differences in these arrange-
ments are, however, only apparent; and they are all easily
explained by the more or less complete fusion of the profun-
dus ganglion and its root with the ganglion and root of the
trigeminus, by the adhesion, or more or less complete fusion,
of the ciliares longi with the ophthalmicus profundus, or even
with the ophthalmicus superficialis trigemini at its base, and
by the more or less complete fusion of the ciliary ganglion
with the inferior branch of the oculomotorius.

In most teleosts (No. 116, p. 39) a ciliary ganglion with two
distinct roots, a radix longa and a radix brevis, is found, and
from the ganglion the ciliaris brevis, single or double, arises.
The ciliaris longus, as in elasmobranchs, varies in its origin.
It may arise from the profundus ganglion, as in Trigla; from
the so-called r. ciliaris, a nerve which arises from the Gasserian
ganglion close to the r. ophthalmicus superficialis trigemini ;
or from the latter nerve itself, as in Salmo.

In no teleost, with the single recorded exception, so far as I
can find, of Trigla (No. 116, p. 25), is there a separate profundus
ganglion and root. Both ganglion and root are apparently always
completely fused with the ganglion and root of the trigeminus.
In Trigla the profundus root arises from the trigeminal root, as
it does in Amia, and its ganglion is separate and distinct from
the trigeminal ganglion, also as in Amia. From the ganglion
arise a ramus ciliaris longus and a radix longa. No other

No. 3.]| MUSCLES AND NERVES IN AMIA CALVA. 539

branches arising from it are given, and there is accordingly no
ramus ophthalmicus profundus in Trigla. That nerve, as a
separate nerve, is also not described, so far as I can find, in
any teleostean forms excepting Trichomycterus and Clarias,
in which fishes it is described by Pollard (No. 97, pp. 392 and
398). As Pollard describes no ramus ophthalmicus superficialis
trigemini in any siluroid, it is exceedingly probable that the
nerve described by him as a profundus is in reality the super-
ficialis, its position in Trichomycterus being in that case most
unusual. Pollard’s nomenclature or observation, in this instance
as in some others, seems to vary somewhat from my own,
and to lead him to somewhat different conclusions. In all
siluroids, excepting the two above named, he says that “the
ophthalmicus profundus may only be represented by some fibres
running along with the ophthalmicus superficialis of the Facial”’
(No. 97, p. 398). By most other writers it is assumed that
the ramus ophthalmicus profundus trigemini has fused com-
pletely with the superficialis trigemini, which, so far as the
course and position of the nerve are concerned, would amount
to the same thing as the statement made by Pollard. Regard-
ing such a fusion in individual cases and in general, several
authors have expressed some doubt. To me it seems impos-
sible. How a nerve that lies below the trochlearis and below
the superior branch of the oculomotorius could fuse with a
nerve lying above both those nerves without inclosing the
two nerves in the single nerve so formed, it is difficult to
imagine. It seems much more probable that the ophthalmicus
profundus of elasmobranchs is entirely wanting in teleosts, and
that the profundus elements in the latter are represented in the
former by some branch or branches of the nervus profundus,
its ganglion or its root, which lie above the two nerves con-
cerned. Such a branch is found in Laemargus in the small
branch which runs upward and forward above the rectus supe-
rior (No. 26, p. 527). In teleosts and ganoids this small
branch must undergo a special development as the rest of the
nerve disappears, becoming the portio ophthalmici profundi,
which may be found as a partly separate nerve, as in Amia, or
as fibres completely fused with the ophthalmicus superficialis, as

540 ALLIS. [Vor. XII.

in teleosts. A confirmation of this supposition is found in the
arrangements presented by ganoids other than Amia.

In Polypterus, van Wijhe (No. 129, p. 260) describes two com-
missural branches connecting the profundus ganglion (called by
him the ciliary) with the ophthalmicus superficialis, and one con-
necting it with the oculomotorius. He also describes a large
nerve, the r. ophthalmicus profundus, which, arising from the
profundus ganglion, runs forward under the rectus superior and
obliquus superior, and joins the ophthalmicus superficialis at the
front end of the orbit. Pollard (No. 93, p. 394) practically con-
firms van Wijhe’s observations, for he describes, (1) two commis-
sural branches to the ophthalmicus superficialis, considered by
him as a branch of the facialis; (2) the main nerve, which has
the general course given by van Wijhe, but passes through the
obliquus superior; and (3) a branch which joins the “motor
nerves,” these latter nerves being presumably the oculomotorius,
though he does not so state. The main nerve in Polypterus
thus corresponds exactly in position with the nerve of the
same name in elasmobranchs, and with the small nerve which,
arising from the profundus ganglion in Amia, runs forward
under the superior rectus and is lost in the general tissues.
The commissural branches to the ophthalmicus superficialis
correspond with the portiones ophthalmici profundi in Amia,
and the commissure to the oculomotorius to the radix longa.

In Lepidosteus there is a profundus ganglion and a portio
ophthalmici profundi (No. 112 4 and 4, Fig. 5a) as in Amia,
but in Lepidosteus the portio profundi and its ganglion lie,
most unaccountably, if Schneider has made no error, imme-
diately below and not above the main trunk of the oculomotorius
(No. 112, p. 18). There is in Lepidosteus no r. ophthalmicus
profundus, unless that nerve is represented in one of the three
nerves called by Schneider “ciliary,’’ which have become fused
with, and have an apparent origin from, the superior and infe-
rior branches of the oculomotorius.

In Spatularia, according to van Wijhe (No. 129, p. 249),
the ramus profundus “issues under the origin of the rectus
superior,’ and then, running forward above the other recti
muscles, separates into two parts, one of which joins the

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 541

ophthalmicus superficialis, and the other enters a canal in the
front wall of the orbit median to the origin of the obliquus
superior. The relations of these two nerves to the other nerves
of the orbit is not given by van Wijhe, and no ophthalmic
branch of the facialis is described. Collinge also (No. 109, p.
516) gives no ophthalmic branch of the facialis, but, as he
describes three dendritic systems along the supraorbital canal,
and gives no branch of the facialis that could in any probability
innervate the sense organs that must be associated with them,
it seems probable that the nerve called by him and van Wijhe
the r. ophthalmicus superficialis trigemini is in part the ophthal-
micus facialis. The origin of the r. oticus from the superficialis,
as given by Collinge, sufficiently proves that the latter nerve
contains elements destined to supply the sense organs of the
lateral canals, and those elements must belong to the facialis.
The sensory canal system of Spatularia does not, therefore,
in its innervation present such exceptional interest as Collinge
has ascribed to it.

In Acipenser, van Wijhe figures and describes a ramus _pro-
fundus lying dorsal to the trochlearis, and dorsal to all the
muscles of the eye. Ina later footnote (No. 129, p. 230) he
says that the nerve so described is probably the ramus super-
ficialis trigemini, and as nothing more is said of the profundus,
the existence or non-existence of that nerve is left somewhat
in doubt. Branches are sent from the nerve so described to
the glands of the “Augenmuskelschlauche,” and the nerve does
not join his ramus ophthalmicus superficialis, which is doubtless
the ophthalmicus facialis. Goronowitsch, who seems to have
published without having seen the later footnote of van Wijhe,
rather confirms than otherwise the non-existence of a true
profundus. According to him, the trigeminus I arises by two
roots which never entirely fuse (No. 50, pp. 478 and 481).
From these two roots three nerves arise, the rami maxillares
inferior and superior and the ramus ophthalmicus profundus.
This last nerve runs forward dorsal to the trochlearis and
above the muscles of the eye, as described by van Wijhe.
Schneider states that the ciliary nerve arises from. the oculo-
motorius, a nerve which never has, according to van Wijhe

542 ALES: [Vor. XII.

(No. 129, p. 229), any connection whatever with any other
nerve.

It is evident from the above references that the work on
ganoids needs some revision. It is also evident that Polypterus,
and probably Spatularia also, represent, in the arrangement of
the ophthalmic nerves, a condition intermediate between that
found in elasmobranchs and that in Amia and the teleosts. In
selachians there is a relatively small ramus superficialis trige-
mini, or no such nerve at all, as in Torpedo (No. 22, p. 62); a
large ramus profundus ; and in Acanthias embryos and skates
a portio profundi, represented in the former by a separate
nerve, and in the latter by communicating branches from the
profundus to the Gasserian ganglion. In Amia there is a large
ramus superficialis, a remnant only of a ramus profundus, but
a portio profundi well developed. In teleosts the remnant
even of a ramus profundus has disappeared, and the portio
profundi is completely fused with the ophthalmicus superficialis.
The superficialis and profundus seem to vary in relative impor-
tance directly as the number of terminal buds found on the
top of the snout and head. In bony ganoids and teleosts
those buds are there found in great quantity (No. 79, p. 69),
while in elasmobranchs they are much less abundant. In
Amia they are innervated by the superficialis trigemini (No. 3,
p. 513) and in elasmobranchs they must be innervated in the
same manner, for Ewart says that the profundus takes no part
whatever in their innervation. The profundus in elasmobranchs
is distributed, in part, to the tubes extending from the ampullae
to the skin (No. 26, p. 527). The disappearance of these organs
in Amia and the teleosts may account for the great reduction
of the profundus in these fishes, while the retention of a closely
related organ, the nerve sack, in cartilaginous ganoids (No. 79,
p. 38) may account for the possible retention of the nerve by
them.

In Protopterus (No. 88, p. 296) there is a large ophthalmicus
profundus and no ophthalmicus superficialis trigemini. From
the profundus three important branches are sent upward and
forward to the top of the head and snout, where they become
associated with the ophthalmicus facialis. They, therefore,

No. 3.] MUSCLES AND NERVES [IN AMIA CALVA. 543

undoubtedly represent the portio ophthalmici profundi of Amia,
or that nerve and the superficialis trigemini also. Under
these branches, but, so far as can be judged from Pinkus’
Fig. 1, above the remaining main portion of the nerve, the
nervus oculomotorius and nervus abducens run, from within
outward, to the muscles of the eye. In the text Pinkus says
that the oculomotorius leaves the profundus on its ventro-lateral
side. The position of the main profundus relative to the supe-
rior and inferior branches of the oculomotorius is, therefore, left
somewhat in doubt, as is also its relation to the abducens.

In Petromyzon the ramus ophthalmicus trigemini lies under
the nervus trochlearis, but, if Firbringer’s figure is correct,
over the superior branch of the oculomotorius. There is no
other ophthalmic nerve, and that the terminal buds in Petro-
myzon are homologous with those in Amia seems from Merkel’s
description an open question. The “Nervenhiigel’’ that he
describes in Petromyzon seems to resemble the “ Nervensicke”’
rather than the “ Nervenhiigel”’ of ganoids, and the nerve in
that case in Petromyzon would probably be the profundus, as
its position below the trochlearis indicates, rather than the
superficialis, which, in all other fishes, lies always above that
nerve.

In Salamandra and Rana (Schwalbe) the ramus nasalis trige-
mini lies under the trochlearis, under the rectus superior, and
under the nerve supplying that muscle. In Salamandra it lies
also under the branch to the rectus internus. It, therefore, cor-
responds exactly in position to the ophthalmicus profundus of
selachians, as Schwalbe has himself pointed out (No. 113, p.
200). In both Salamandra and Rana a ciliary ganglion is
described by Schwalbe as a ganglion of the oculomotorius,
connected in Rana with the ramus nasalis by a very delicate
radix longa, and probably so connected in Salamandra also,
although Schwalbe did not find the nerve. In Salamandra,
and also in Menobranchus and Siredon, ciliary branches arising
from the ramus nasalis are described in addition to those aris-
ing from the ciliary ganglion. In Rana and Salamandra rami
frontales are described, arising from the base of the ramus
nasalis and considered by Schwalbe as the equivalents of the

544 ATSIELS: [VoL. XII.

ophthalmicus superficialis trigemini in fishes. This latter nerve
as a separate and distinct nerve is not described, and in
Amphibia there are no terminal buds on the top of the head
(No. 79, p. 76). Strong (No. 121, p. 108), in larvae of Rana,
finds the ramus ophthalmicus trigemini arising from a partially
separate anterior, ventral and median portion of the Gasserian
ganglion. The nerve runs forward between the two divisions
of the oculomotorius, and gives off its first branches after pass-
ing that nerve. It lies apparently below the abducens, but the
description is not precise (No. 121, p. 134). Strong does not
describe a ciliary ganglion, a profundus ganglion, or a radix
longa. He, however, says that sympathetic fibres are con-
nected with certain of the trigeminal branches (No. 121, p.
119).

In the adults of Amphibia and higher vertebrates a profundus
ganglion is not described, so far as I can find. It is, however,
described in certain embryos. In late embryonic stages of
Lacerta, Hoffmann (No. 56, p. 205) describes it as a more or
less separate and distinct ganglion, called by him the ganglion
ophthalmicum. From it the ciliary ganglion is developed.
From the profundus ganglion proper, after this differentiation,
there arise a ramus frontalis and a ramus nasociliaris, and from
the latter a ramus ciliaris connecting the nerve, and hence its
ganglion, with the ciliary ganglion. The ramus nasociliaris is
thus the ophthalmicus profundus of Ichthyopsida ; the ramus
ciliaris is the radix longa, and the ramus frontalis is the portio
ophthalmici profundi of ganoids and teleosts, or branch 1 of
Ewart in Laemargus. The ciliary ganglion is connected with
the inferior branch of the oculomotorius bya short radix brevis.
As in Amphibia, so in Lacerta, there is no ramus ophthalmicus
superficialis trigemini, for from the posterior of the two
trigeminal ganglia described by Hoffmann, that is, from the
entire ganglion of the trigeminal outgrowth minus the pro-
fundus ganglion, there arise only the so-called second and third
branches of the trigeminus, that is the superior and inferior
maxillary nerves.

In embryos of Torpidonotus natrix the oculomotorius and
trochlearis are shown by Grosser and Brezina (No. 51) in one

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 545

reconstruction, Fig. 8, median to, that is ventral to, the ramus
nasociliaris trigemini and its ramus frontalis. In another
reconstruction of an older embryo, Fig. 9, they are shown
lateral to, that is dorsal to, the same nerves. One or the other
must be wrong.

In the guinea pig the ophthalmic nerve arises, according to
Chiarugi (No. 18, p. 507), from a distant ophthalmic ganglion
which soon fuses with the ganglion of the fifth nerve. This
ganglion is unquestionably the profundus ganglion of Amia, and
the ophthalmic nerve, therefore, the ophthalmicus profundus.

In a five-months human embryo Ewart (No. 27, p. 290) finds
vestiges of a profundus ganglion, lying under cover of the inner
portion of the Gasserian ganglion. After describing it he states
his conviction that ‘the ophthalmicus profundus of the elasmo-
branch is represented in man by the so-called nasal branch of
the ophthalmic division of the fifth nerve.”

The arrangement of the ophthalmic nerves and their asso-
ciated ganglia in reptiles and the higher vertebrates seems,
therefore, to be exactly the same as in Amphibia, if a certain
allowance be made for the indefiniteness or insufficiency of the
observations and descriptions. In these higher forms, and in
certain fishes as well, sympathetic branches are traced to the
ciliary ganglion, and several investigators state that the gan-
glion is in all probability a sympathetic, and not a spinal, or
cerebro-spinal, ganglion. Retzius (No. 103) states positively
that it is such in mammals, and as it is known as the ganglion
ophthalmicum in man, that name should certainly not be given,
as proposed by Hoffmann, to the ganglion of the profundus
where that ganglion is found as a separate ganglion, as in
fishes and in Lacerta.

In the schema (Fig. 12, Pl. XXII) showing diagrammati-
cally the relations of the oculomotorius to the muscles of the
eye, I have also endeavored to show the relations, as above de-
scribed, of the ophthalmicus profundus and ophthalmicus super-
ficialis trigemini to those muscles and to the oculomotorius.
From it it will be seen that in the disposition of these nerves,
as well as in the manner of innervation of the muscles of the
eye, the arrangement found in Amphibia and higher vertebrates

546 ALETS, [VoL. XII.

is naturally derived from some such arrangement as that shown
in the hypothetical proto-urodele type which lies intermediate
to and below the two great divisions of Pisces.

II. MUSCLES INNERVATED BY THE TRIGEMINUS AND
FACIALIS, AND THE NERVI TRIGEMINUS, FACIALIS,
ACUSTICUS, AND LINEAE LATERALIS VAGI.

The muscles included in this group are in part innervated by
the trigeminus alone, in part by the facialis alone, and in part
by nerves formed by the fusion of branches of both those
nerves. Whether, in the latter case, the muscles concerned
are innervated by the facial elements, or by the trigeminal, I am
unable to determine ; the muscles of the group are, therefore,
placed for convenience in three separate sub-groups.

1. Sub-group 1. Muscles innervated by the Trigeminus alone.
a. Adductor Mandibulae.

The adductor mandibulae (Figs. 29-42, Pls. XXVI-X XX)
is a large and complex muscle having an extended surface of
origin and several different insertions. It has three main divi-
sions: a superficial one Az, an inner or deeper one A3, and a
mandibular one Ao.

The superficial portion Az is much the largest of the three.
It lies immediately underneath the postorbital bones, and is
exposed when those bones and the thick tough dermis extend-
ing from their hind margins to the outer edge of the preoper-
culum are removed. Its outermost fibres present a fan-like
appearance, radiating approximately from the coronoid process
of the mandible and spreading out through somewhat more than
aright angle. The lower, posterior fibres run, from their origin,
forward and slightly upward, and the upper, anterior ones for-
ward and downward, the most anterior ones turning inward at
the front edge of the muscle and running downward and slightly
backward along its inner surface.

The muscle has an extended surface of origin. Its super-
ficial fibres arise, above, from the postorbital process and

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 547

from the under surface and outer edge of the squamosal ;
behind and below they arise from the entire outer, anterior
face of the preoperculum, excepting only its outermost edge
which gives attachment to the dermis covering the side of
the head. A few fibres often arise from the under surface of
the upper postorbital bone along its upper edge. Its deeper
fibres arise from the outer surfaces of the hyomandibular,
quadrate, and symplectic, the surface of origin lying immediately
in front of the preoperculum and extending somewhat under
the lower end of that bone and forward onto the hind edge of
the mandible. This inner, deeper portion of Az is intimately
connected, particularly below, with the inner, deeper portion, A,
of the adductor. A part of Az, often found as a somewhat
separate bundle, arises from a slightly raised portion of the
outer surface of the hyomandibular immediately behind the
middle portion of that bone. This surface is well defined above
by the lower, nearly horizontal, border of the dilatator operculi ;
in front by the line of attachment of the strong metapterygoid
membrane, which extends across the open space between the
hyomandibular and the median process of the metapterygoid
and covers the insertion of the posterior portion of the levator
arcus palatini ; behind by the front edge of the preoperculum ;
and below by the external opening of the facial canal through
the hyomandibular, and by the slight groove extending from
this canal downward and backward to the point where the hyoid
branch of the facial nerve passes under the preoperculum. It
is below and in front of this groove that the superficial muscle
Az is, at its origin, so intimately connected with 43.

Although continuous and single at its origin, the superficial
muscle, 42, always shows, toward and at its insertion, indications
of a separation into three portions, a lower or posterior one, A2’,
a middle one, 42”, and an upper or anterior one, A2'’. The
middle portion, A2!’, sometimes shows a further but much less
marked separation into two parts, as shown in Fig. 40. Between
the three main portions, at their insertion, the surfaces of sepa-
ration are clean and distinct, the actual separation of the mus-
cle into three parts is, however, wholly artificial. The surfaces
of separation lie in the direction of two planes intersecting at

548 ALIAS: Vor, X11

the top of the coronoid process of the mandible, the planes
extending obliquely into the muscle from its outer surface, but
not passing entirely through it excepting at or near its inser-
tion. The separation, therefore, on the external surface of the
muscle extends much nearer its origin than in its deeper por-
tions. The plane between the upper and middle portions of
the muscle extends downward, backward, and inward; that
between the middle and lower portion almost directly down-
ward and inward; so that the line separating these two parts
on the outer surface of the muscle runs almost directly back-
ward from the tip of the coronoid process. The lower or pos-
terior division of the muscle overlaps the middle one and so lies
partly superficial to it, and the middle portion in the same way,
but to a much greater extent, overlaps and lies superficial to the
upper or anterior one. The separation between the lower and
middle divisions is much more complete than that between the
middle and upper ones.

The outer and upper fibres of the lower portion, A2’, have
their insertion on the outer surface of the supra-angular, near
its hind edge, and on that hind edge from the tip of the coro-
noid process to the lower end of the bone. At the top of the
coronoid process they extend almost onto a strong tendon or
ligament (No. 76, p. 126), which has its origin there and is
inserted on the inner surface of the maxilla at about its ante-
rior third. The connective tissue sheath covering this portion
of the muscle at its insertion is continuous with that covering
the ligament, so that, before it is removed, the ligament has
strongly the appearance of being the tendon of the muscle.
The remaining deeper and lower fibres of Az’ separate from the
outer ones and are inserted on the inner surfaces of the artic-
ular, the supra-angular, and the coronary cartilage. This por-
tion of the muscle is continuous at its lower, ventral margin, and
toward its origin, with the lower fibres of 3, the two muscles
sometimes enclosing between them, intimately connected with
them, a special bundle of fibres (Fig. 37, Pl. X XIX), which arise,
partly tendinous, from among their fibres, and, joining the man-
dibular portion of the adductor, Ao, at its hind edge, are in-
serted continuous with that muscle on the inner surface of the

No. 3.] MUSCLES AND NERVES [IN AMIA CALVA. 549

articular in its lower, anterior part. This is the only connec-
tion A2’ has with Aa, for although its deeper portion, where it
passes to its insertion on the inner surface of the coronary
cartilage, lies closely upon the outer surface of Ao, there is no
interchange of fibres between the two muscles, and the r. max-
illaris inferior trigemini lies between them, as shown in Fig.
31, where A2’ has been entirely removed.

Az", the middle portion of the muscle, is larger than either
of the other divisions. Some of its outer, upper fibres are often
inserted on the inner surface of the dentary, at the coronoid proc-
ess and in front of it. The remaining fibres pass in large part
directly into the ramus of the mandible, and there form the
outer portion of the outer division, Aw’, of the mandibular
muscle. A part of them, however, including the deepest and
lowermost fibres, are inserted along the lower, ventral edge, and
on the outer surface, of a broad tendon or fascia, 42 Aw’, which
is formed on the inner surface of Az, where it contracts to pass
into the mandible (Figs. 41 and 42, Pl. XXX).

Az'", the upper, anterior division of Az, consists mainly of
those fibres of the adductor that arise from the postorbital
process, the surface of origin on this process being the deeper,
lower, portion of the cap-like piece that forms the upper, outer
angle of the ossification (No. 3, p. 480). It also contains those
fibres of Az that arise from the under surface of the squamosal,
immediately behind this cap-like piece, and some or all of those
that arise from the under surface of the upper postorbital.
The muscle is inserted entirely along the upper, dorsal edge,
or on the outer surface of the upper half of the tendon Az Aa’.
Although continuous at its origin with the upper, outer fibres
of A", and at its insertion on Az Ao! with the deeper fibres of
that muscle, it is, nevertheless, wholly distinct and separate
anteriorly where there is no interchange of fibres whatever, the
fibres of Az!” going entirely to the tendon Az Aa’, and those
of A2" entirely to the inner surface of the dentary.

In larvae of from 20 mm. to 50 mm. in length, 42!” arises
entirely from the cartilage of the postorbital process, the sur-
face of origin lying below the postfrontal bone, in front of the
plane of the dorsal opening of the spiracular canal, and imme-

550 ALLIS. [Vor. XII.

diately dorsal to the origin of the levator arcus palatini, the sur-
face of origin of which muscle extends a little in front of it.
The muscle lies on the inner surface of Az, does not come to
the level of the outer surface in any part, is nearly vertical in
position, and the fibres of 42" run downward and forward across
it at a considerable angle. Its fibres, which are all inserted
along the upper, anterior edge of the tendon, or fascia, on the
inner surface of Az, are long in front and short behind where
they gradually vanish, the fascia here extending upward to the
origin of the muscle. At this point A2” begins, arising as a
thin sheet, partly tendinous in front and wholly so behind, where
it lies superficial to the dilatator operculi. This fascia, rather
than muscle, is attached in front to the upper, outer edge of the
cartilaginous rib that forms the lateral boundary of the anterior
diverticulum of the temporal groove, which rib is, in the adult,
inclosed in a V-shaped process on the under surface and outer
edge of the squamosal (No. 104, p. 188). As this V-shaped
process has not, at this age, begun to form anteriorly, the origin
of the fascia in that part is entirely from the cartilage. Poste-
riorly, where the process is represented by a single bony fin or
plate along the outer edge of the squamosal, the line of origin
of the fascia extends onto that fin, that is, onto the lateral edge
of the squamosal. In the adult this fascia has become muscu-
lar, and the surface of origin of the muscle is a narrow line
extending along the outer edge, and along the under surface of
the squamosal, back as far as the upper end of the preopercu-
lum. This uppermost portion of the muscle is in the adult thin
and partly tendinous. It lies immediately superficial to the
levator arcus palatini and dilatator operculi, and passes almost
abruptly, at the lower edge of the dilatator, into the thick fleshy
lower portion of the muscle.

The inner, deeper division of the adductor, A3, lies imme-
diately underneath Az. Although a relatively small muscle, it
has a large surface of origin, arising from the hyomandibular
below the external opening of the facial canal, from the outer
surface of the quadrate down to the articular swelling at its dis-
tal end, from the outer surface of that part of the metaptery-
goid that lies behind the front edge of its median, dorsal pro-

No. 3.] MUSCLES AND NERVES [IN AMIA CALVA. Sys

cess, from the entire outer surface of that process, and also by
a few fibres from the outer surface of the metapterygoid mem-
brane, some of the outermost of these last fibres overlapping the
membrane and arising from the outer surface of the hyoman-
dibular beyond it. A; is continuous below, or ventrally, with
A2',as already described. Its fibres converge strongly and are
all inserted on a broad tendon or fascia, 4; Aw", formed where
the muscle passes into the ramus of the mandible.

The mandibular portion of the adductor, Aw, is that part of
the muscle that lies inside the hollow of the ramus of the man-
dible. It can be separated into two parts, Aw! and Aw", which
are distinct and separate in origin and insertion. There is,
however, no smooth and even surface of separation between
these two parts, as there is between Az and 43. On the con-
trary, the surfaces are rough and uneven, muscle fibres, or
bundles of one division, projecting into and interlocking with
those of the other, so that they must be carefully pulled apart
to effect a separation of the muscles. There is, however, no
interchange of fibres between the two divisions, each being
confined strictly to its own region of origin and insertion.

The larger, outer portion, Ao’, has its origin from the broad
tendon Az Aw’. Its fibres arise from the outer surface of the
tendon, near the hind end of the mandible, and pass directly,
with the outer fibres of A2”, to their common insertion on the
entire inner surface of the dentary and the anterior part of the
angular. The arrangement is such as would arise if a single
broad muscle, arising at or near the preoperculum and having
its insertion inside the ramus of the mandible, thus represent-
ing Aw! and the greater part of Az, should become contracted,
and in part tendinous, near the middle of its length.

The inner, deeper portion, A", of the mandibular muscle is
much smaller than the outer one. It arises entirely from the
tendon or fascia formed on the outer surface of A 3, and is inserted,
as a flat, thin muscle, along the upper surface of the posterior half
of the longitudinal cartilaginous rib which, as described by van
Wijhe (No. 129, p. 281), projects horizontally from the lower
edge of Meckel’s cartilage into the hollow of the ramus of the
mandible. Considered together with 4, it forms in outline, on

552 ALLIS. [Vox. XII.

a smaller scale, a fair copy or reduplication of the outer muscles
Ao! and Az taken together. The tendon A, Ao” has its front
edge posterior to that of Az Ao! and, although lying closely
against the inner surface of that tendon, is wholly separate fror
and disconnected with it, except at one point near the hind edge
of the splenial. There the front edges of the two tendons are
connected by a short tendinous band, which extends between the
points where they are joined, 42 Ao! by the tendon of the first
division of the levator maxillae superioris, and A; Aw! by the
tendons of the second and third divisions of that muscle. The
tendon A; Aw", where it passes across the hind edge of the
splenial, is strongly attached to a tough band of dermis formed
along that edge, so that this becomes in a measure a point of
attachment for the two divisions of the adductor Az and A3, and
also for the first three divisions of the levator maxillae superioris.

b. Levator Maxillae Superioris.

In addition to the adductor mandibulae, properly so called and
above described, there are four muscles intimately associated
with it, and in teleosts and selachians considered, in part, as parts
of it. They have been called by McMurrich, in Amia (No. 76,
p. 122), the second, third, fourth, and fifth divisions of the leva-
tor arcus palatini. They are, however, so evidently derived from
the levator maxillae superioris of selachians, and from a muscle
that Vetter (No. 124) has called AddB, a part of the adductor
mandibulae, in Acanthias and Scymnus, that it has seemed best
to reject the names used by McMurrich, and to call the muscles
the first, second, third, and fourth divisions of the levator max-
illae superioris. It seems best, also, to change slightly the order
of numbering used by McMurrich, and to consider his third divi-
sion of the levator arcus palatini as the first one of the levator
maxillae superioris ; his second division as the second ; his fourth
division as the third; and his fifth division as the fourth.

The fourth division of the muscle (Lws‘+, Figs. 21-42) is short,
stout, and conical in shape, with the base of the cone below
at its insertion on the upper surface of the palatine bone near
its outer, hinder edge. At its apex or origin the muscle was,

No.3.] ZUSCLES AND NERVES IN AMIA CALVA. 553

in most of the specimens examined, double, arising by one head
from the under surface of the posterior extremity of the antorbi-
tal bone, and by the other from the anterior and under face of
the prefrontal process this process being pyramidal in shape, as
described by Sagemehl (No. 104). Where the muscle was single
in its origin it arose entirely from the antorbital, and in speci-
mens of about 20 mm. in length it had apparently this origin
only. Where it has a partial insertion on the antorbital, it
must, in contracting, depress that bone and so diminish the
size of the nasal cavity, thus having some functional connection
with the nose.

Immediately posterior to the insertion of this muscle a
strong ligament arises from the outer edge of the ectoptery-
goid. It runs backward closely attached to the inner surface of
the loose dermal fold that extends from the edge of the pala-
tine arch to the maxilla, and which may be called the supra-
maxillary fold (No. 3, p. 489, and Fig. 20), and disappears on
the under surface of the dermis behind the superficial fold or
crease that marks the hind edge of the mandible, its fibres
running downward and backward somewhat parallel to that
crease. Another ligament arises from the under surface of
the anterior of the two infraorbital bones, and running back-
ward and downward closely attached, as the preceding one is,
to the dermis between the maxilla and the palatine arch, dis-
appears on the under surface of the dermis behind the hind
margin of the mandible and behind the first ligament.

The third division of the muscle, Zms3, arises in full-grown
fishes immediately behind the antorbital process from the upper
surface of the palatine bone near its hind edge. In nearly all
the fishes examined some of the lateral and upper fibres of the
muscle continued forward above and across the ligament that
binds the palatine arch to the prefrontal ossification, and had
their origin on the posterior face of that bone. The muscle is
a long one, of flat oval section. It runs downward, outward,
and backward along the upper surface of the palatine arch
until it reaches the angle or bend in the upper surface of the
metapterygoid, where it turns downward and contracts abruptly
into a long, slender tendon. This tendon continues downward

554 ALES, [VoL. XII.

and backward, in the direction of the muscle, to the anterior
edge of the large tendon A;Ao”, which it joins near the hind
edge of the splenial. Here it joins and usually fuses com-
pletely with the tendon of the second division of the muscle,
ILms?. The united tendons, as a broad flat band, continue
backward and then downward in a curved line along the outer
surface of A; Aw", closely attached to that tendon, but in no
way fused with it. They have their own particular insertion
on a small ossification which lies immediately behind the hori-
zontal projecting rib of Meckel’s cartilage, and has been called
by Bridge ossicle ¢ (No. 15, p.618). This is the only insertion
of the muscle evident in larvae. In the adult, however, what
seemed to be a second tendon of the muscle was always found.
It lay on the inner surface of A; Aw", as shown in the
figures (Figs. 41 and 42, Pl. XXX), and had its insertion pos-
terior to the insertion of the united tendons of Laws3 and Lms?,
apparently on the same bone, ossicle ¢, at its extreme edge, or
possibly on ossicle 6. This could not be definitely determined.
The tendon always had much the character of a string of con-
nective tissue, but it was always found well defined through
the greater part of its length. At its insertion it was always
joined by parts of the posterior portion of tendon 4; Ao”.

In larvae, Zs3 arose entirely from the upper surface of
the palatine cartilage, and in no instance had it acquired any
attachment to the antorbital process. In one instance it was
double throughout its length, arising, one part a little in front
of the other, on the upper surface of the palatine, the two
divisions lying side by side and contracting separately into
separate tendons, which united to form a single tendon just
before reaching the large tendon of the adductor. In several
other specimens there were indications of such a division into
two parts, but there was no actual separation of the fibres.

The second and first divisions of ws arise, in the adult, as a
single muscle, mainly from the front edge of the hyomandibular
at its extreme upper end, but the line of origin extends forward
from the hyomandibular, across the upper end of the lateral
wing of the parasphenoid, to the front edge of that bone.
In young specimens also, of from 20 mm. to 40 mm. in

No. 3.| MUSCLES AND NERVES IN AMIA CALVA. 555

length, these two muscles arise as a single muscle, but they
have their origin entirely from the front edge of the hyoman-
dibular, the attachment to the parasphenoid being along the
upper edge of the muscle and by connective tissue only ;
that is, the attachment is from the side of the muscle fibres
and not at their ends. This part of the muscle in these young
specimens is a flat band lying close against the inner surface
of the levator arcus palatini, and immediately external to the
pseudobranch. It runs almost directly forward from its origin
on the hyomandibular to the level of the front edge of the
levator arcus palatini, where it turns downward and outward
along the outer shelving surface of the metapterygoid, and there
separates into two parts, a lower, posterior one, Ls?, and
an upper, anterior one, Zms!. The former, Lms?, contracts
near the front edge of A; to a slender tendon, and joins the
large tendon A; Ao” at the point where that tendon is joined
by the tendon of Zs3, with which it usually fuses and has the
course and insertion already described. A secondary semi-
tendinous connection of this tendon, similar to that of tendon
Lms3, was sometimes found along the inner surface of A; do”,

Lms*t, the upper, anterior division of the levator maxillae
superioris, contracts near the front edge of Az to a flat, ter-
minal edge and is inserted by a short, thin, fascia-like tendon
on the inner surface of the tendon Az Aa! at its upper, anterior
edge, internal to the coronoid process of the mandible and
posterior to the hind edge of the splenial. Near its insertion
the muscle is moulded against the front, inner edge of 42'”, the
fibres of the two muscles ending at the same level, and the
short, fascia-like tendon of Zs! appearing simply as a branch
of, or slip from, the main tendon Az Ao’. It cannot be traced
across this tendon as the tendons of Zs? and Lms3 were across
the tendon A; do”. At its outer, anterior edge it is strongly
attached by connective tissue to the dermis at the outer corner
of the mouth, and at its inner, posterior edge it is connected,
as already described, by a tendinous or connective tissue band
with tendon A; Ao", the point of attachment to that tendon
being at the angle formed by the crossing of the two tendons
Lms3 and Lms?, where they join 4; Ao”.

556 AL TELS. [VoL. XII.

In specimens of from 10 mm. to 12 mm. in length the four
divisions of the levator maxillae superioris are directly con-
tinuous with the adductor mandibulae, their tendons, found in
the adult, not yet having been formed, or at least not being
distinguishable. The first and second divisions appear as a
single muscle, connected with both the inner and outer divi-
sions of the adductor, and arising from the front edge of the
median process of the metapterygoid or from the upper sur-
face of that bone near the process, the attachment to the
hyomandibular and the position internal to the levator arcus
palatini not yet having been acquired. The third division of
the levator is well developed and arises in connection with the
first and second divisions, but from the inner division only of
the adductor. It extends forward under the eye and is either
attached to the upper surface of the palatine cartilage or
vanishes in the general tissues above it, having no visible
attachment. No trace was found at this age of the fourth
division as a separate muscle.

ce. Levator Arcus Palatini.

This muscle (Lag, Figs. 30, 31, and 36, Pls. XXVI, XXVII,
and X XIX), called by McMurrich the first division only of
the levator arcus palatini, is a stout muscle lying in origin
and insertion internal to the plane of the median, dorsal
process of the metapterygoid. It has in its upper part a
strong, median, longitudinal aponeurosis, which extends into
the muscle from its origin and front edge for about two thirds
the length and breadth of the muscle. From this aponeurosis
all the fibres of the inner portion of the muscle arise, and those
of the outer in great part.

The most anterior fibres of the muscle belong entirely to
the outer of these two portions. They are distinctly marked
off on the surface of the muscle (Fig. 36) from the remain-
ing fibres of the outer portion, and have their origin from
the lateral and front edge of the postorbital process below the
dermal postfrontal bone and anterior to the lateral wing of the
parasphenoid. They run almost vertically downward, bellying

Non3.| “OSCLES AND NERVES TN: AMIA CALVA. SS

forward a little, and are inserted along the front edge of the
median, dorsal process of the metapterygoid, and on the upper
surface of the bone in front of that process. The fibres of the
outer portion of the muscle immediately behind this anterior
part arise partly from the postorbital process, but mostly from
the outer surface of the median aponeurosis of the muscle.

The inner portion of the muscle begins about on a level
with the front edge of the metapterygoid process; it arises
entirely from the inner surface of the median aponeurosis
which, in front of this point, has been merely a strong mem-
brane lining the inner surface of the outer portion of the
muscle. Its fibres arise at a considerable angle to the aponeu-
rosis, curving inward and downward and then a little outward,
thus forming a strong belly on the inner side of the muscle.
The muscle has its insertion on the entire inner surface of the
median process of metapterygoid, on the entire inner surface
of the strong metapterygoid membrane, and on the outer
surface of that small portion of the hyomandibular that les
internal to that membrane, and is marked off from the rest
of the bone by the line of attachment of the membrane to it.
This small portion of the hyomandibular is much thinner than
the rest of the bone and lies depressed below the level of the
rest of its outer surface. It extends from the front, upper
corner of the bone down along its front edge to the point
where the metapterygoid joins or overlaps it, at about two
thirds its length.

dad. Dilatator Operculi.

The dilatator operculi (Do, Figs. 30, 31, and 36, Pls. XXVI,
XXVII, and XXIX) is described by McMurrich as a part of
the first division of the levator arcus palatini, his description
of it being: “Those [fibres] lying most posteriorly are directed
backwards, passing under the upper extremity of the preoper-
culum, thus fulfilling the function of the separate dilatator
operculi of the Teleostei.”

In all the adult and larval specimens that I have examined,
without exception, the dilatator has invariably been found as a

558 ALLIS. [VoL. XII.

wholly separate muscle, lying immediately above and in close
contact with the levator, but having no connection whatever
with it except at its origin. It arises from the entire lateral
edge of the squamosal back as far as the upper end of the pre-
operculum, and from the postorbital process behind and above
the origin of the levator. The surface of origin of the two
muscles at this place is continuous, but the muscles them-
selves are, from the beginning, markedly distinct and separate.
The dilatator runs backward across the upper end of the
hyomandibular, along its outer surface. It passes through
the narrow opening between the hyomandibular and the upper
end of the preoperculum, and is inserted by a tendon, which
forms on its outer surface, on the inner surface of the oper-
culum above and in front of the facet forming the articulation
of that bone with the hyomandibular.

In specimens of 12 mm. in length and under, the dilatator
is, at its origin, united to and a part of the levator, the two
muscles arising, as a single muscle, from the upper edge of
the cartilaginous postorbital process immediately in front of
the blind, upper end of the spiracular canal, which, at this age,
lies entirely outside the cartilage of the cranium. The united
muscles pass backward, external to the spiracular canal, and
then separate. The levator is inserted on the inner surface of
the vertical process of the metapterygoid and on the inner sur-
face of the metapterygoid membrane behind it. Both mem-
brane and muscle extend backward beyond the front edge of
the hyomandibular, lying external to that bone and having no
apparent connection with it. The muscle here is inserted
entirely on the membrane, and the membrane shows no trace
of attachment to the hyomandibular, but vanishes, as the
muscle does, in the general tissues superficial to it. The outer
division of the adductor mandibulae, A2, has its origin, at this
age, from the outer surface of this same membrane, there being
no attachment whatever to the cranium. The dilatator ex-
‘tends further backward than either the levator or the adductor,
and vanishes apparently closely attached to the under surface
of the dermis.

No. 3.) JZOSCLES AND NERVES IN AMIA CALVA. 559

2. Sub-Group 2. Muscles innervated by both the Trigeminus and

Facialis.
a. Intermandibularis.

The intermandibularis (/m, Figs. 43 and 44, Pl. XXXI) isa
short, stout muscle extending transversely from one ramus of
the mandible to the other, immediately internal to the front
end of the gular plate. Its fibres arise from the dentary of one
side and are inserted on that of the other, the lines of insertion
extending close up to the symphysis of the bones. Although
to all appearance a single muscle, it can be easily separated
along its front edge into two parts, each of which extends
across the inter-ramal space and is thicker at its origin from
the dentary of its own side than at its insertion on that of the
other. The two portions are completely united posteriorly and
overlap as the hyohyoidei do, the muscle of the left side lying
superficial to or below that of the right.

The superficial or inferior portion of the geniohyoideus, in
passing from its origin on the mandible backward and inward
toward the middle line of the head, lies partly superficial to the
posterior portion of the intermandibularis; and the tendons
of the deeper superior portion of the same muscle pass imme-
diately dorsal to the intermandibularis closely attached to it.
The branchio-mandibularis is either firmly attached to the
hind margin of the intermandibularis or lies immediately dor-
sal to it between it and the integument of the floor of the
mouth.

b. Geniohyoideus.

The larger part of the geniohyoideus (Giz and Ghs, Figs.
43 and 44, Pl. XXXI) lies immediately internal to the gular
plate, which plate covers also the intermandibularis and the
anterior portions of the sternohyoideus and hyohyoideus.

The gular plate is a large and slightly convex dermal ossifica-
tion about three quarters as long as the mandible (No. 3, G.
Fig. 47, Pl. XLI). It is pointed in front, where it is firmly
attached by ligament to the symphysis of the mandible, and

560 ALLIS. [Vov. XII.

rounded behind, where, lying in a dermal fold, it overlaps the
first branchiostegal rays and covers and protects the isthmus.
It nearly fills the space between the rami of the mandible, but
is small enough to pass freely up and down between them, when
the motions of the mouth require it. The external surface of
the bone has the sculptured markings characteristic of the
dermal bones in Amia, the ridges proceeding ray-like from a
higher central point near the middle of the bone. Immediately
internal to this point there is a dense connective tissue, almost
cartilaginous in character, in one of the specimens examined ;
and on a level with it, along the lateral edges of the bone, there
is a thick, rigid dermis forming a rim projecting upward and
laterally. The gular bone itself does not give direct insertion
to any muscle fibres.

The geniohyoideus has two well-marked portions, a superior,
deeper one (Gs) and an inferior, superficial one (G/z). The
deeper portion is the larger, and is a broad, flat, thick muscle,
thinning out anteriorly, where it lies immediately below the
integument of the floor of the mouth. It arises from the outer
under surface of the ceratohyal, behind the origin of the hyo-
hyoideus. It is, at its origin, V-shaped in outline, the two
edges of the muscle extending much further back than the
central portion. The median edge extends as far as the
eighth or ninth branchiostegal ray, the muscle in this part of
its course lying close along the bases of the rays. It is directed
forward and inward, its median fibres meeting those of the
muscle of the opposite side of the head in a vertical aponeurosis
which lies in the middle line of the body and extends forward
to the hind edge of the intermandibularis. The dense connec-
tive tissue mass under the central point of the gular plate lies
in or is a part of this aponeurosis, lying near its hind end. It
gives insertion posteriorly to a part of the most posterior or
median fibres of the muscle, which fibres form a somewhat
separate bundle, rising above the rest of the deeper muscle to
the level of the superficial one, the posterior fibres of which
also have their insertion here. Posterior to this point the
fibres of the deeper muscle are more or less continuous with
those of the opposite side, and sometimes entirely so, as shown

No. 3-| MUSCLES AND NERVES IN AMIA CALVA. 561

in Fig. 43. The fibres of the deeper muscle immediately lat-
eral to the median bundle pass forward beyond the hind edge
of the superficial muscle, dorsal to it, and have their insertion
on the median aponeurosis, or on the hind edge of the inter-
mandibularis. The fibres lying still more toward the lateral
edge of the deeper muscle become tendinous, after passing dor-
sal to the superficial muscle, and are continued beyond that
muscle, dorsal to the intermandibularis, as a sheet or series of
flat tendons which lie immediately beneath the integumental
lining of the floor of the mouth. They are closely attached to
that integument, and have their insertion partly in it, and partly
on the inner surface of the mandible near the symphysis. The
extreme lateral fibres of the muscle run almost directly forward,
from their origin on the ceratohyal, parallel to the ramus of the
mandible, and have their insertion in the integument of the floor
of the mouth along the median edge of the tough layer or fold
of dermal tissue that extends from the lateral edge of the
muscle downward and medianward to the lower, inner edge
of the mandible. Some of the fibres of the muscle are also
often inserted along the dermal rim formed at the middle of
the lateral edge of the gular plate.

The superficial, inferior portion of the geniohyoideus arises,
in young specimens, in a narrow horizontal line from an exposed
portion of Meckel’s cartilage on the inner surface of the man-
dible immediately below the lower edge of the splenial and
between that bone and the dentary (Fig. 6, Pl. XX). In full-
grown fishes the surface of origin includes also the edge of the
dentary immediately below this cartilage. The line of origin
extends approximately from the hind edge of the intermandib-
ularis to the level of the front end of the mandibular portion
of the adductor mandibulae ; that is, through about one quarter
of the length of the mandible. The muscle is flat and broad
and runs backward and inward from its origin to the middle
line of the head, where it is inserted, with its fellow of the oppo-
site side, on the median, vertical aponeurosis common to it and
to the deeper portion of the muscle. Its insertion is below, or
superficial to, that of the deeper portion, and only extends back
to the point under the middle of the gular plate where, as

562 MELTS. [Vor XI.

already described, it meets the posterior bundle of fibres of
the deeper muscle.

The two divisions of the geniohyoideus are intimately con-
nected at and near their insertion on the median aponeurosis,
but the fibres of the two portions lie at a considerable angle to
each other, and the continuity described by McMurrich was not
found in any specimen, old or young.

In larvae 10 mm. in length some of the fibres of the super-
ficial portion of the muscle are continuous with those of the
intermandibularis, the fibres of the left side being continuous
with the ventral layer of the intermandibularis, and those on
the right side with the dorsal one. In these young larvae the
geniohyoideus is also apparently continuous with the hyohyoi-
deus.

The geniohyoideus acts either as an adductor of the hyoid
arch or as a retractor of the mandible according as the one or
the other of those structures is stationary, or is so considered.
The hyoid arch is loosely connected, in front, with the inner
surface of the mandible by a fold of the tough integument of
the floor of the mouth, which, when the parts are at rest, is
folded under the tip of the hyoid, thus forming the covering of
its under surface and the floor of the anterior part of the mouth.
Back of this, along each side, the hyoid is connected with the
lower, inner edge of the mandible by a strong fold formed by
the closely united external dermis and internal lining membrane
of the mouth, the two membranes being so closely united that
they are with difficulty separated. A strong membranous fold
is thus formed, which, when the mouth parts are at rest, lies
directed upward, from the lower edge of the mandible, parallel
to and against its inner surface. By this arrangement the
hyoid arch, lying between the rami of the mandible, is con-
nected with them, on either side, by a long and flexible fold, so
that its anterior end, which is smaller than the space between
the rami, can be freely elevated and depressed between those
rami to the full extent permitted by the fold. A deep groove
(No. 3, Fig. 16, Pl. XXXV and Fig. 22, Pl. XXXVI) on the
under surface of the head marks the position of this fold,
the groove extending from near the symphysis backward along

No. 3:] A7TOSCLES AND NERVES IN AMIA CALVA. 563

the inner edge of the mandible to the preopercular articulation,
and then upward and backward between the two upper poste-
rior branchiostegal rays, onto the external surface of the gill-
cover behind and beyond them, where it disappears. The fold
may be called the hyoideo-mandibular fold.

In its action as a retractor of the mandible the geniohyoi-
deus is, as Vetter and others have stated, accessory to the
sternohyoideus, which is the main retractor of the hyoid and
through it of the mandible. The sternohyoideus is inserted
on the under surface of the extreme anterior end of the hyoid
arch. In its action it pulls this end downward and backward
in a curved line, so placed, relatively to the curve described
by the tip of the mandible in opening, that the hyoid not only
pushes the parts below it downward between the rami of the
mandible, but, having slipped backward over those parts to the
full extent allowed by its integumental connection with them
and with the mandible, has its motion transferred to the man-
dible, and thus acts as its retractor. The motion so given
by the sternohyoideus is accelerated and increased by the
action of the geniohyoideus, both divisions of which muscle,
by their contraction, tend to hold the mandible down over the
anterior end of the hyoid so that it shall follow more promptly
the motion given to the latter. The geniohyoideus also draws
the hyoid forward, this being its simplest and most direct
action, but it can do this, and so act as an adductor of the
hyoid, only when the mandible is fixed by the action of its
adductor.

ce. Hyohyoideus.

The hyohyoideus (Zz and hs, Figs. 43-46, Pls. XXXI
and XXXII) arises from the median edge of the flat, spreading
blade of the ceratohyal, from a surface of depression begin-
ning at the base of the first or second branchiostegal rays and
extending forward along the ventral side or edge of the shank of
the ceratohyal, and from the inner surfaces of the branchiostegal
rays, mainly from the inner surfaces or front edges of the first
and secondrays. These first two rays do not generally come into
contact, at their bases, with the ceratohyal, the first ray being

564 ALLIS. [VoL. XII.

often separated from that bone by a considerable interval. The
remaining rays are firmly bound by their bases to the under sur-
face of the blade of the ceratohyal, a little beyond its median
edge, so that this edge is left free throughout its entire length.
From this free edge the hyohyoideus in part arises.

Anterior to the branchiostegal rays, the hyohyoideus runs
forward and inward, a broad, continuous band, thick at the lateral
edge, but very thin at the median one, where, in its anterior por-
tion, it passes beyond the middle line of the body, and overlaps
the muscle of the opposite side to such an extent that at their
insertions the two muscles are often nearly but not entirely
superimposed. The muscle of the left side lies superficial to that
of the right side, and the two end in front in a common, curved,
tendinous line, attached at each end by strong tendons to the
ventral surface of the hypohyals. Each of the two muscles has
its own tendon at either end of the common anterior margin,
but the tendons at each end unite more or less completely as
they approach their insertion, so that a single tendon is formed,
which lies in the extended line of the curved front edge of the
muscles. This single tendon then separates into two smaller ten-
dons, one of which continues in the direction of the main tendon,
and is inserted on the under anterior surface of the hypohyal.
The other turns sharply forward and medianward, and splits up
into numerous, smaller, slender tendons, which spread out finger-
like over the under surface of the hypohyal, and are inserted on
it, and in the tough tissue in front of it toward the tip of the
tongue, the tendons of the one side overlapping and crossing
beyond those of the other. No muscle fibres extend onto or
into the tongue ; but the presence of such a network of inter-
crossing tendons, onto which muscle fibres or their tendons are
inserted, seems to indicate an earlier stage in, or perhaps a dif-
ferent manner for, the “muscularization”’ of the tongue than
that given by Gegenbaur (No. 45). The tongue in Amia can
certainly not be considered as wanting a direct and definite con-
nection with muscle fibres which are not in any way associated
with a glandular system.

The two hyohyoidei are not continuous anteriorly; they
merely end in curved, tendinous edges, which are superimposed

No. 3.] WUSCLES AND NERVES IN AMIA CALVA. 565

and bound together. The curved line of these tendinous edges
lies a little behind the symphysis of the hypohyals, and is concave
forward, leaving a space in front of it, between it and the hypo-
hyals, through which the branchiomandibularis passes.

The anterior or inferior part of the hyohyoideus is somewhat
divided into two portions or bundles, one arising from the bran-
chiostegal rays, and the other from the shank of the ceratohyal,
the division between the two portions being faintly indicated
by surface lines. The fibres arising from the branchiostegal
rays are the ones that pass across the middle line of the body,
and enter into the tendon that lies on the opposite side of the
head, while those arising from the ceratohyal usually pass
directly forward into the tendon of their own side. The fibres
arising from the branchiostegal rays are more or less separated
into parallel bundles, closely held together by connective tissue.
This condition is most apparent near the front edge of the first
ray, and there gives to the muscle a stringy or striped appear-
ance. These bundles have their origin in part from the front
edge and under surface of the first ray, and in part they are
continuations of more or less continuous cord-like muscles,
which have their origin from parts beyond this ray, and form
what Vetter has called in other fishes the superior part of the
muscle. These muscular cords are wholly, and in places some-
what widely, separated; they branch somewhat, and are alter-
nately tendinous and muscular, the muscular portions lying, in
every case, across the interspaces between successive rays.
Throughout their course they are firmly held in connective
tissue, which is closely attached to the inner or dorsal surfaces
of the rays. The little muscular swellings, or bellies, lie most
frequently directly in the course of the more or less continuous
cord, but they may extend from one cord to another, or arise
from the side of a cord in front, and be inserted on the inner
surface of the next ray behind. Along the median or free
edges of the rays, each cord forms a continuous line, but
toward the edge of the ceratohyal the other arrangements
prevail, the cords becoming somewhat crossed and interlaced ;
while from the edge of the ceratohyal larger muscle bundles
arise, which, running forward and outward along the inner,

566 ALELS. (VoL. XII.

dorsal surface of the rays, unite with one or more of the cords
encountered there. Some of these cords, alternately muscular
and tendinous, extend as far upward and backward as the inner
surface of the interoperculum; and other small, thread-like
cords, found in the integument beyond the free edges of the
rays, can be traced nearly, if not quite, to the upper edge of the
operculum. In young fishes the fibres of the hyohyoideus
extended up into the region of the levator operculi, but the
continuity of the fibres, and hence of the two muscles, could
not be established.

3. Sub-Group 3. Muscles innervated by the Facialis alone.
a. Adductor Hyomandibularis.

This muscle (Az, Figs. 25 and 37, Pls. XXV and XXIX) is
broad and short, extending from the auditory region of the lat-
eral wall of the skull to the inner surface of the hyomandibular.
Its line of origin begins on the petrosal, at the hind margin of
the facial foramen, and extends backward and upward across
that bone, and across the intercalar, nearly to the end of its
posterior process. It is inserted on the inner surface of the
hyomandibular, the line of insertion extending from the front
edge of the bone, backward and upward, nearly to the hind end
of the opercular process. It lies mostly below the internal
opening of the facial canal, through the hyomandibular, but
overlaps that opening somewhat, so that the facial nerve, in
its passage from its cranial foramen to the internal opening of
the canal, lies partly imbedded in the fibres of the muscle.
The muscle runs outward, backward, and downward, lying, in
its anterior portion, more nearly horizontal, and being there
much shorter than in its posterior part.

b. Adductor Operculi and Levator Operculi.

These two muscles (Ao and Lo, Figs. 20-30, Pls. XXIII-
X XVI) are continuous at their insertion, which is on the upper
portion of the inner surface of the operculum, behind the artic-
ulation with the hyomandibular, the surface of insertion occu-

No. 3.] WUSCLES AND NERVES IN AMIA CALVA. 567

pying a little more than one third of the inner surface of the
bone.

The levator has its insertion posterior to that of the adductor,
and is double at its origin, arising behind directly from the under
surface of the suprascapular, and in front by tendinous attach-
ment to a tendinous formation, which extends backward under
the extrascapular, from the hind edges of the squamosal and
parietal. A part of the tendon of the muscle can usually be
traced through the tendinous formation to the under surface of
the parietal, near its hind edge, and at about the middle line of
the temporal groove. This long and slender part of the ten-
don lies immediately above the anterior extension of the trunk
muscles that fills the temporal groove.

The adductor lies anterior to the levator. Its superficial fibres
form a thin, narrow band, which crosses the hindermost fibres of
the adductor hyomandibularis at a considerable angle to them,
lying on their upper, outer surface, and arising, just above their
origin, from the lateral wall of the skull, the surfaces of origin
of the two muscles being continuous. The deeper fibres arise,
partly by numerous delicate tendons, partly by direct attach-
ment, from the posterior process of the intercalar, from the
postero-lateral angle or edge of the cranium immediately above
that process, and from the hindermost lateral extremity of the
squamosal, this last attachment not being an important one.
The muscle in its deeper part is continuous at its insertion
with the levator operculi behind, and with the adductor hyo-
mandibularis in front.

4. Review and Comparison of Muscles of Group II.

The muscles described under this group belong to the hyoid
and mandibular arches, or perhaps also, in part, to one or more
preoral arches (No. 124, p. 448). They form a large, compli-
cated, and very variable group, and their definite innervation
is but little known. Any attempt to compare the muscles in
Amia with those described in other fishes can, therefore, be of
but little value in so far as the possible homologies established
for the several muscles or the different parts of those muscles

568 ALETS. [Vov. XII.

are concerned. It seems, however, proper to attempt to bring
my results into some sort of accord with, or relation to, those
of others. In doing this, I shall frequently refer to dissections
made of several fishes, more particularly of Carcharinus littor-
alis and Galeus canis, in the earlier stages of this investigation.
They were made most superficially, and simply to determine,
approximately, certain points that I thought obscure but of
importance in Vetter’s descriptions. Notes and sketches were
made at the time, and, as I have only them to refer to, the
statements now made regarding the several fishes may not be
entirely reliable.

a. Levator Arcus Palatini, Dilatator Operculi, and Muscle
Addy.

The levator arcus palatini and dilatator operculi are certainly
parts of a single muscle. They are considered by Vetter (No.
125, pp. 484 and 531) as the homologue of the protractor hyo-
mandibularis in Acipenser, and of the levator maxillae superioris
in selachians. He derives them, as he does those muscles, from
the dorsal half of the superficial layer of the general constrictor
of the mandibular arch (No. 124, p. 407). McMurrich, on the
contrary (No. 76, p. 127), derives them from some part of the
adductor mandibulae of selachians, that is, from a muscle that
is developed, according to Vetter and Gegenbaur (No. 124, p.
446), from the deeper layer of the general constrictor of the
mandibular arch. Whatever their derivation may be, they seem
to me to be the homologue of the muscle called by Vetter in
selachians Addy. It seems impossible that they should be the
homologue of the levator maxillae superioris, as Vetter sug-
gests, or be derived from that muscle if, as I shall attempt to
show, the muscle Zms!, in Amia, is the homologue of one or
more of the spiracle muscles of selachians, the muscles called
by Tiesing (No. 123, p. 92) the levator palpebrae nictitantis,
retractor palpebrae superioris, and constrictor superficialis dor-
salis /y. These muscles in selachians cross the outer surface
of the levator maxillae superioris, while Zs!, in Amia, crosses
the inner surface of the levator arcus palatini.

No.3.) WUSCLES AND NERVES IN AMIA CALVA. 569

The muscle Addy, in Scymnus and Acanthias (No. 124, p.
448), arises partly tendinous from the under surface of the skin,
immediately behind the eye, and immediately in front of the
postorbital process. It runs downward and backward, and is
inserted on the outer surface of the adductor mandibulae, being
there more muscular than in its upper portion, and being in part
a direct continuation of the fibres of Csv2, which Vetter con-
siders as the ventral half of the superficial layer of the con-
strictor of the hyoid arch. In Heptanchus Addy seems, from
Vetter’s figure, to arise from the lower rather than the upper
corner of the orbit, as in Scymnus; it runs more nearly back-
ward, its hindermost fibres are inserted on the cartilage at the
hind corner of the upper jaw, and it has no connection with the
muscle fibres of the ventral half of the constrictor. The inner-
vation of the muscle is not given in either fish described. In
Mustelus and the rays, Tiesing does not describe this muscle.
That it is not peculiar to the fishes described by Vetter is
shown by Carcharinus, where a muscle is found closely resem-
bling that of Heptanchus. It arises near the upper corner
of the eye, is inserted in part on the lower end of the hyoman-
dibular, and its posterior portion is crossed by many branches
of the facial, which go entirely or in large part to cutaneous or
subcutaneous tissues. The muscle is apparently innervated by
branches of a nerve which arises inside the cranial cavity, from
the main truncus or ganglion of the trigeminus, the nerve issuing
immediately behind the eye, around the front edge of the levator
maxillae superioris, and then turning backward along the outer
surface of Addy.

Addy is considered by Vetter as the last remnant of the super-
ficial constrictor of some one or more preoral arches (No. 124,
p. 448). Its position, however, in Scymnus and Acanthias, its
connection with the adductor, and its continuity with the
ventral half of the constrictor, seem to indicate that it belongs
to the mandibular arch, and that it is derived from the dorsal
half of the superficial constrictor of that arch. Its innervation,
if the innervation found in Carcharinus be correct, agrees
closely with that of the levator arcus palatini in Amia, and such
slight changes only would be needed in its origin and insertion,

570 ALLIS. [Vou. XII.

to derive that muscle from it, that it may safely be considered
as its homologue. The strong aponeurosis in the upper part
of the muscle in Amia corresponds to the tendinous, upper
part of the muscle in Scymnus, and in the young of Amia the
muscle is connected or associated with the adductor mandibulae
much more than with the palatine arch, as it is in the adult.
In Galeus canis Addy was not found. It is, however, probably
represented in one of the two spiracle muscles.

In Chimaera and Acipenser Addy cannot with certainty be
identified in the descriptions given by Vetter, but it may be
represented in Chimaera by Cs4, and in Acipenser by the upper
tendinous part of Cs: and Csz. In the one specimen of Aci-
penser that I examined, a young Acipenser ruthenus about one
foot in length, this upper part of Cs: and Csz seemed to bea
tendinous formation into which the muscle fibres of the lower
portion were inserted, rather than a tendon of that lower por-
tion. Its position does not differ greatly from that of Addy in
Heptanchus, and if it be that muscle, it would account for the
unusual insertion, below and in front of the eye, of Cs: and
Cs2, the lower portions of which represent, in Acipenser, that
part of the superficial constrictor from which the geniohyoideus
in other fishes arises. This part of the constrictor is, in Acan-
thias (No. 124, Fig. 3, Pl. XIV), inserted at the hind edge of
the lower jaw, mainly into a fascia that covers and arises from
the lower portion of the outer surface of the adductor mandi-
bulae, just as Addy arises, in the same fish, from the upper
portion of that surface. Slight changes only in the insertions
of these muscles in Acanthias would bring about the arrange-
ment found in Acipenser.

b. Levator Maxillae Superioris, Csd:, and AddB.

The levator maxillae superioris, in Heptanchus, arises from the
occipital part of the skull under the projecting and overhanging
postorbital process. Immediately behind it, and partly contin-
uous with it, is another muscle, called by Vetter Csd1. Both
muscles run downward and forward, and are inserted on the
inner side of the upper jaw, the hind muscle covering with its

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. S72

inner, posterior surface the front wall of the spiracular canal.
The two muscles represent, according to Vetter, the dorsal half
of the superficial constrictor of the mandibular arch.

In Carcharinus and in Acanthias much the same relations
exist, but in Scymnus the hind muscle, Csd;, has become
entirely independent. It arises from the hind, upper corner of
the skull, and running forward and downward, unquestionably
along the outer surface of the levator although it is not so
stated, it encircles the front side of the outer ends of the two
spiracular cartilages and is inserted on the inner side of the
quadrate part of the upper jaw. It is called by Vetter the
muscle of the spiracular cartilage.

In Galeus a still further differentiation of this muscle has
taken place, for it is here inserted near the hind edge of the
lower eyelid, some few fibres only, from its inner surface, leav-
ing the main muscle to be inserted on the inner surface of the
upper jaw. A second spiracular muscle is also found. It arises
immediately behind the spiracular canal from the outer surface
of, or from tissues superficial to, Csd2. Its lower fibres run
upward and then forward, partly encircling the spiracle, its
upper fibres directly forward. Beyond the spiracle the muscle
divides, allows the larger spiracle muscle to pass through it,
and then, reuniting, is inserted, according to my notes, near
the edge of the lower eyelid, at the corner of the eye, above
and behind the insertion of the larger muscle. In Mustelus,
also, there are two spiracle muscles. The larger of the two
is described by Tiesing as the levator palpebrae nictitantes, and
is unquestionably derived directly from the muscle called by
Vetter Csd;. The other, smaller, muscle lies internal to the
larger one, and is inserted on the hind end of the upper eyelid.
Its insertion is thus higher than that I have noted for the
corresponding muscle in Galeus. It is called by Tiesing the
retractor palpebrae superioris, and is undoubtedly, as he con-
cludes, also derived from Csd1, or perhaps from that muscle
and Addy also.

In Heptanchus the anterior part of the adductor mandibulae
extends almost to the ‘“ palatobasal”’ articular process of the
upper jaw. In Carcharinus a part of this anterior portion

5/2 ALLIS. [Vou. XII.

forms a distinctly separate muscle lying in front of and over-
lapping externally the main muscle. It is tendinous above, the
long, somewhat slender tendon running forward and becoming
muscular again before its insertion, as a short, stout muscle body,
on the anterior wall of the orbit immediately below the inser-
tion of the obliqui muscles. In Galeus the tendon connecting
these two portions of the muscle is short, and the stout, orbital
muscle-body lies imbedded in the remaining lower portion of
the muscle, with which its fibres are in part directly contin-
‘uous. That part of the lower portion of the muscle that lies
immediately posterior to the tendon of the upper portion, has
its origin from the upper edge of the cartilage of the upper
jaw ; that part that lies in front of the tendon has its origin, in
part, from the front wall of the orbit and, in part, it extends for-
ward on the under side of the head towards the middle line of
the body, having its origin there. The outer fibres of the lower
portion of the muscle become tendinous posteriorly, the tendon
being in part continuous with a fascia covering the outer sur-
face of the adductor and extending to the hind edge of the
lower jaw, and in part continuous with a tendinous formation
extending into the adductor from its outer surface. The deeper
fibres of the muscle are continuous with the fibres of the
deeper part of the adductor. A somewhat similar arrangement
is found in Acanthias and Scymnus, but the separation of the
muscle from the main adductor seems to be more complete
than in Galeus, and the tendon in its lower posterior portion
more developed. The muscle is, however, considered by Vetter
a part of the adductor and is called by him Add. The corre-
sponding muscle in Mustelus is called by Tiesing the levator
labii superioris.

In Heptanchus the special innervation of that part of the
adductor that corresponds to AddB is not given by Vetter. In
Acanthias and Scymnus, he says, that it is innervated by a
branch of the r. maxillaris superior trigemini, but Stannius
(No. 116, p. 46) finds the corresponding muscle in Spinax, and
Tiesing (No. 123, pp. 86 and 96) the corresponding muscle in
Mustelus innervated by a branch of the r. maxillaris inferior
trigemini. In the three fishes described by Vetter the levator

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 573

maxillae superioris, Csd1, and the spiracle muscle are all said by
him to be innervated by a single branch of the r. maxillaris
inferior trigemini, the branch being given off immediately after
the main nerve issues from its foramen. The branch runs
outward and backward under the postorbital process along
the front or outer surface of the levator, and then apparently
backward along the lateral surface of the levator and the outer
surface of Csd:. In Mustelus (Tiesing) the corresponding
muscles have a similar innervation, the nerve innervating them
being sometimes double.

In Galeus the larger spiracle muscle and the levator maxillae
superioris are innervated by a branch of a small double nerve,
which arises inside the cranial cavity, to all appearance from
the upper surface of the truncus trigeminus, or of its ganglion.
As it issues through the main trigeminal foramen this double
nerve lies along the upper surface of the inferior maxillary
nerve, crossing that nerve from its inner to its outer edge.
It then turns under the nerve, to its lower surface, and con-
tinues forward in this position to the upper edge of the upper
jaw. Here, as the inferior maxillary turns downward and back-
ward, the double nerve issues from beneath it, at its inner edge,
and, running forward, supplies the muscle Add8. The nerve
is so closely applied to the inferior maxillary, and, where it
issues from beneath it, appears so exactly like a branch of that
nerve, that it would not have been recognized as a separate
nerve if the arrangement of the nerves and muscles in Amia had
not indicated that it must be there as such. It is unquestion-
ably the r. ad musc. levator maxillae superioris of Amia, but the
nerve in Galeus is always double, as it was found to be in one
specimen only of Amia. From the outer strand of the nerve
in Galeus, soon after it leaves the cranium, a branch is sent to
the levator maxillae superioris and the larger spiracle muscle,
which two muscles therefore unquestionably correspond respec-
tively to the second and first divisions of the levator maxillae
superioris in Amia. The same strand then innervates the
lower and interior portions of Add, which accordingly corre-
spond to the third division of the levator in Amia, while the
orbital part of the muscle, which is innervated entirely by the

574 ALLIS: [VoL. XII.

inner strand of the nerve, corresponds to the fourth division.
This nerve was traced in Galeus on one side only of a single
specimen. Its being double suggests an explanation of an
apparently abnormal innervation of Lms* by the superior max-
illary nerve, found in one specimen of Amia. The strand
destined for the innervation of this muscle in this particular
specimen had doubtless become separated from the rest of the
nerve and attached secondarily to the superior maxillary which
it so closely accompanies. It arises in Galeus from the upper
surface of the truncus trigeminus, while in Amia it arises from
the under surface of that truncus, an important difference.
In Galeus it partly encircles the inferior maxillary, holding that
nerve as ina loop. If this position of the nerves represents
an earlier condition than that found in Amia, the inferior maxil-
lary, as it moved backward to the position found in Amia, must
have caused an apparent shifting of the origin of the nerve
from the upper to the under surface of the truncus.

The innervation of the smaller spiracle muscle in Galeus was
not determined. The muscle may, from its position, be de-
rived either from Csd; or from Addy. The innervation of the
spiracle muscles in Mustelus by two nerves suggests that they
are of similar, but perhaps not of identical, origin. The same is
true of the muscles innervated by the double nerve in Galeus.
It therefore seems probable that Csdi1, Lms, AddB and Addy
in Heptanchus, Carcharinus, and other selachians, represent
parts of the constrictor superficialis dorsalis of the mandibular
arch, or that they represent parts of that constrictor and parts
also of the corresponding portion of the constrictor of one or
more preoral arches. From Addy the levator arcus palatini
and dilatator operculi of Amia are derived, and possibly also
one of the spiracle muscles in Galeus and Mustelus ; from Csdr
is derived the larger spiracle muscle of Galeus, or perhaps both
the spiracle muscles; from the larger spiracle muscle of Galeus
is derived the muscle Zs* of Amia; from the levator maxillae
superioris of selachians is derived the muscle Zs? of Amia; and
from Add (the levator labii superioris of Tiesing) the muscles
Lms3 and Lms‘. In rays this last muscle splits up, according
to Tiesing, into several more or less independent portions.

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 575

In Chimaera and in Callorynchus there is a muscle, called by
Vetter the levator anguli oris, which lies below and in front of
the eye, immediately superficial to the adductor mandibulae.
The anterior part of this muscle arises as a thin, broad tendon
above and in front of the eye. It contracts immediately to a
smaller tendon, and then, becoming muscular, separates into
two portions and is inserted by round tendons on the adjoin-
ing ends of the maxillary and mandibular cartilages (No. 125,
p. 441). The posterior part of the muscle arises from the
lower edge of the orbit, and is inserted by a very slender ten-
don on the mandibular cartilage and in the skin at the corner
of the mouth. Vetter gives the innervation of both divisions
in Chimaera by branches of the inferior maxillary, but his
figure seems to show that the anterior division of the muscle is
innervated by a branch of the superior maxillary, which nerve
supplies also the labialis anterior. In Callorynchus both divi-
sions of the muscle, and the labialis anterior also, are said by
Stannius to be innervated by the superior maxillary (No. 116,
p. 46 and Fig. 1, Pl. 1). This want of agreement in the two
descriptions strongly suggests the error to which Vetter has
himself called especial attention (No. 125, p. 495, Note 1):
that of not tracing a nerve or the branch of a nerve to its
origin, and leads one to infer that the three muscles are inner-
vated by a nerve corresponding to the r. ad musc. levator max-
illae superioris in Amia, and hence that they correspond to the
four divisions of the musculus levator maxillae superioris of
Amia ; the labialis anterior, probably, to the fourth division of
that muscle; the anterior division of the levator anguli oris,
with its double insertion, to the third, which is frequently
double through part of its length in Amia; and the posterior
division, which is inserted in part in the skin at the corner
of the mouth, to the second and first divisions in Amia, the
former of which in Amia has in part this same attachment.
The position of the levator anguli oris superficial to the adduc-
tor mandibulae is naturally suggested by the arrangement found
in Galeus, where AddB lies in part superficial to Add.

In Acipenser the large and powerful protractor hyomandibu-
laris is considered by Vetter the homologue of the levator max-

576 ALLIS. [Vor. XII.

illae superioris of selachians (No. 125, p. 484). He finds the
muscle, in Acipenser, innervated by a branch of the r. maxil-
laris inferior trigemini arising from the still undivided trunk of
the nerve (No. 125, p. 474). Stannius gives the same innerva-
tion, but adds that the muscle contracts when the second root
of the trigeminus is irritated, thus indicating that the nerve is
a branch of the superior maxillary. In the one specimen that
I examined the nerve arose from the upper surface of the
truncus or ganglion of the trigeminus, and as there was no
branch to other muscles or tissues, the muscle it innervated
may correspond either to the first and second divisions of the
levator maxillae superioris of Amia, or to the levator arcus
palatini. The third and fourth divisions of the former muscle
were represented by a tendinous, fatty, and degenerate muscle
tissue extending from the cartilage below and in front of the
eye to the under surface of the skin near the lower end of
the hyomandibular. The r. maxillaris superior trigemini and
the r. buccalis facialis of van Wijhe ran forward across the
upper end of this tissue; the r. palatinus anterior facialis
internal to it.

In teleosts the levator arcus palatini and dilatator operculi
do not differ greatly from the corresponding muscles in Amia.
The levator maxillae superioris, on the contrary, has undergone
great changes. In most teleosts all four divisions of it have
been largely absorbed by the adductor mandibulae, but in
being absorbed they have determined in many ways the
arrangement and disposition of that muscle. In Esox part
of the levator muscle is undoubtedly represented in the
aberrant bundle A438, of Vetter. The tendon of another part
has undoubtedly determined in Esox, and in Perca, Cyprinus,
and Barbus as well, the insertion of A; at the hind end of
Meckel’s cartilage, and in Perca another tendon persists as
tendon Azt of Vetter.

In Amiurus there are two tentacle muscles, the adductor
and the abductor tentaculi (No. 75, p. 314). The adductor is
either simply the first and second divisions of the levator max-
illae superioris of Amia, or some or all of the divisions of that
muscle and the muscle 43 combined. The abductor is described

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 577

by McMurrich as the anterior portion of the adductor arcus
palatini, and is said by him to be innervated, as the posterior
portion of that muscle is, by a branch of the facialis. But for
this innervation this muscle would certainly seem to be the
third or fourth division of the levator maxillae superioris of
Amia, for both in its origin and insertion it differs but little
from and could easily be derived from that of those muscles.
Pollard says (No. 97, p. 390) that the motor nerves to the cor-
responding muscle in Auchenapsis ‘run along with the pala-
tine nerve,” the palatine nerve referred to being described as a
branch of the trigeminus. Pollard further states, as a general
principle, that the motor supply of all the tentacle muscles in
siluroids must be by nerves that “ proceed out with the sensory
nerves”’ of the tentacles (No. 97, p. 382). This general state-
ment shows, what Pollard does not otherwise definitely state,
that he found all the tentacle muscles in all the fishes he exam-
ined supplied, as the tentacles themselves are said to be, by
branches of the trigeminus. The innervation of the muscle
in Amiurus may therefore be by the trigeminus instead of by
the facialis, as McMurrich states, in which case it is certainly
the homologue of the third and fourth divisions of the levator
maxillae superioris.

In the Marsipobranchii Add is represented probably in the
group of muscles that in Mixine are innervated by the so-called
r. ophthalmicus trigemini, and to the muscles that in Petromy-
zon are associated with the “ Vorknorpel”’ and innervated by
the r. maxillaris trigemini (No. 37, pp. 34 and 40). The r. oph-
thalmicus trigemini in Mixine, according to Fiirbringer (No. 37,
p. 71), gives off a sensory branch which runs forward above the
optic nerve and then itself runs under that nerve, giving off
both sensory and motor branches, the motor branches going to
the muscles of the tentacular circle and the nasal tube. In
Petromyzon the r. ophthalmicus trigemini is said to be entirely
sensory, and runs over the optic nerve, as does the first sensory
branch of the nerve in Mixine (No. 37, p. 62). In Bdello-
stoma also the main trunk of the nerve runs over the opticus,
but whether it contains motor fibres or not Fiirbringer does
not state (No. 37, p. 31). The ophthalmic nerve in Mixine

578 ALLIS. [VoL. XII.

thus finds no homologue in Petromyzon, Bdellostoma, or other
fishes, and it seems safe to assume that that part of it that lies
under the opticus corresponds to the ramus ad musc. levator
maxillae superioris and parts of the ramus maxillaris superior
trigemini combined of Amia, while the branch above the opti-
cus is the true ramus ophthalmicus. Whether these deduc-
tions be correct or not, I certainly should not accept existing
descriptions of these nerves as supporting Pollard’s dictum,
based in part upon them, that “the topographical position and
course of nerves is not of great importance”’ (No. 97, p. 397).

c. Adductor Mandibulae.

The adductor mandibulae proper in Heptanchus, Acanthias,
and Scymnus, as described by Vetter, is an extremely simple
muscle ; practically an uninterrupted and undivided layer of
muscle fibres extending from near the upper edge of the pala-
tine arch to the lower edge of the mandible. In Carcharinus
and Galeus it is not so simple. In both these fishes the r. max-
illaris inferior trigemini penetrates the muscle between a super-
ficial and a deeper portion, which are entirely separate at and
near the surface. A surface line extending upward and forward
across the muscle, from, or from near, the corner of the mouth,
marks the front edge of the superficial portion. The fibres of
this superficial portion are inserted into a transverse tendinous
formation, or aponeurosis, which extends inward and upward into
the muscle from a curved surface line which extends from the
corner of the mouth backward, outward, and upward toward the
hind corner of the upper jaw. This tendon does not extend
entirely through the muscle. It is inserted behind and above
on the hind corner of the upper jaw, and in front it is attached
to the dermis at the corner of the mouth. It is pierced by the
r. maxillaris inferior trigemini, in Carcharinus at a considerable
depth below the outer surface, in Galeus near that surface, the
nerve in the latter fish coming outward toward the surface along
the front edge and outer surface of the tendon. In Carcharinus
the nerve, having pierced the tendon, or having even passed
entirely beneath it near its front edge, continues for some dis-

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 579

tance below the surface of the muscle, and then comes to its
outer surface ; in Galeus it comes immediately to the outer
surface of the muscle and in its further course lies on that sur-
face. In both fishes it is joined near the front end of the
muscle by a branch of the facialis, apparently the r. mandibu-
laris externus facialis of Amia.

The tendon above described gives insertion at the corner of
the mouth to the tendon of Add8. From its lower, inner sur-
face the fibres of the superficial portion of the mandibular part
of the adductor arise. It seems, therefore, to represent the
tendon Az Ao! of Amia, and the muscle fibres arising from it
to represent Az and Aa’, the deeper portions of the adductor
muscle corresponding to A; and do’. At the point where
the r. maxillaris inferior trigemini pierces or meets the tendon
several branches are sent into the adductor muscle, supplying
its deeper as well as its outer portions. These branches run
backward between the two portions of the muscle, thus corre-
sponding in position to the branches that, in Amia, run down-
ward and backward between Az and A3, innervating those
muscles. That part of the muscle that arises from the upper
surface of the tendon, and hence lies directly superficial to the
main inferior maxillary nerve, therefore, corresponds to the
superficial layer Az! of Amia. The rest of the superficial por-
tion, or possibly the tendon alone, represents the deeper layers
A." and Az!" of Amia, which, if they are represented by the
tendon alone, must have acquired their relative importance
later.

With the several portions of the adductor found in rays
and described by Tiesing I am unable to make any satisfactory
comparison whatever.

In teleosts the adductor presents markedly different arrange-
ments in different species. The innervation of the different
parts of the muscle not being definitely known, a definite com-
parison with the muscles found in Amiais not possible. Certain
probabilities can, however, be arrived at.

The third and fourth divisions of the levator maxillae supe-
rioris of Amia are, as already stated, probably found as special
insertions of the adductor in teleosts, or as aberrant bundles

ALLIS. Vor xt
580 [

of that muscle. Their further consideration is not necessary.
The first and second divisions of the levator either disappear
entirely or fuse with the adductor proper of Amia to form the
main portion of that muscle in teleosts. This McMurrich
(No. 76, p. 127) has already suggested, the adductor in teleosts
being, in his opinion, equivalent to that of Amia plus the leva-
tor maxillae superioris of selachians, that is, plus the second,
or the first and second, divisions of the muscle of Amia. This
is indicated by certain features in the innervation of the
muscle, such as the innervation of the superficial portion of
the adductor in Amiurus by a special branch of the trigeminus,
the branch arising proximal to the branch that supplies the
deeper portion of the muscle (No. 75, p. 313), and the innerva-
tion of a part of the adductor in Silurus, Salmo, and Gadus by
branches arising directly from the ganglion of the nervus tri-
geminus (No. 116, p. 45). The innervation, in Amiurus (No.
132, p. 368), of the levator arcus palatini and dilatator operculi
by certain fibres arising from a superficial branch of the r. ad
musc. adductorem mandibulae, has no parallel in Amia.

The deeper part, 43, of the adductor in Cyprinus is inserted
mainly by tendon on the inner side of the dentary, but in part
it passes into a short, flat muscle Aa, and is inserted at the hind
end of Meckel’s cartilage and on the adjoining parts of the den-
tary. In Barbus, A; is inserted mainly by tendon at the hind
end of Meckel’s cartilage, but in part it becomes muscular again
and is inserted farther forward on the inner surface of the den-
tary. In Perca and in Esox the muscle Aw” has entirely dis-
appeared, and A; is inserted wholly by tendon at the hind end
of Meckel’s cartilage, this manner of insertion being doubtless
determined by the preéxisting tendons of Zsm? and Lsm3. In
Amiurus also the fibres of A, have followed apparently the
tendons of Lsz3, but in the opposite direction, that is, toward
the origin of the muscle, A; being inserted by tendon at the
base of the maxilla. The tendon is split near its insertion
and the rr. maxillaris sup. trigemini and buccalis facialis run
forward between the two ends (No. 75, p. 314 and No. 132,
p. 368), as they do between the two heads of Zms3, or between
Lms3 and Lms* in Amia.

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 581

The superficial muscle Az, of Amia, has in teleosts under-
gone much modification, due mainly to the marked development
of the superficial portion of the muscle, 42’, and its differentia-.
tion in some fishes as a wholly separate muscle, the deeper
parts of the muscle 42" and Az!” undergoing at the same time
an actual or relative reduction. In Amia the outer fibres of
A,’ extend at their insertion almost onto a strong ligament,
which extends from the coronoid process of the mandible to
the inner surface of the maxilla. A slight change in the inser-
tion of these fibres would give rise to the condition found in
Perca, where the muscle A1, of Vetter, is inserted in part
by tendon on the inner surface of the maxilla, and in part has
retained its connection with the deeper fibres of the adductor,
and is inserted with them on the anterior, lower corner of the
articular. In Barbus this last connection has disappeared, and
the muscle is inserted entirely on the maxilla along its lower
edge. In Cyprinus the muscle has separated into two portions,
wholly distinct at their insertions on the maxilla, but still
united at their origins, while in Moxostoma, which I examined,
the separation at the origin has become complete, and each of
the two muscles is again partly separated into two portions. In
Moxostoma the deeper muscle Az of Vetter, which is equiva-
lent to 42” and A2'” in Amia, is greatly reduced, and is simply
a thick muscle band lying along the upper edge of A3. That
this band is the muscle A2 of Vetter, or a part of it, and not
a part of A, which it has every appearance of being, is evident
from the position of the r. ad musc. adductor mandibulae,
which runs under the band and issues between it and 43, the
remaining lower portion of the muscle ; for it is not probable
that this nerve, which lies so constantly between Az and A; in
other fishes, should have so radically shifted its position in this
one as to penetrate A; from its under surface.

In Esox and Amiurus more primitive arrangements are pre-
sented than that found in Amia, for in neither of them is there
any separation of a superficial layer, although there are in both
indications either of the absorption, or of the preéxistence, of
such a layer. In determining the existence of this layer, the
course and position of the inferior maxillary nerve is of first

582 ALLIS. (VoL. XII.

importance, for the nerve seems to lie always between A; and
Az, a position derived directly from that found in selachians.
Apparent departures from this position are often given, but in
the three cases that I have examined, namely, Esox, Amiurus,
and Moxostoma, it seems due to a misconception of the dif.
ferent parts or layers of the muscle, and not to a variation
in the position of the nerve. In Esox no part of the muscle
Az of Vetter is inserted on the hind edge or outer surface of
the mandible, the entire muscle joining, or passing into, the
well-developed mandibular muscle do, and having its insertion
with that muscle inside the ramus of the mandible. A super-
ficial portion is, however, indicated, and it is along and slightly
under the front edge of this portion, between it and the bundle
A38, and then on the outer surface of Aa, that the inferior
maxillary nerve enters the mandible. Vetter’s description,
which leads one to infer that the nerve enters the mandible
between Az and A3, that is, internal to the tendon Az Aa,
which it must therefore first pierce to reach its later position
on the outer surface of Aw (No. 125, p. 495), is anerror. In
Amiurus some of the superficial fibres of Az, the AM of
McMurrich, are inserted on the hind edge of the mandible,
and as the inferior maxillary nerve enters the mandible along
their front edge, on the outer surface of the remaining, deeper
portion of the muscle, these fibres undoubtedly represent Ar.
In Moxostoma, where A; is found well developed, the inferior
maxillary nerve lies along the front or upper edge of Az and
not internal to it, that is, between it and A3, as Vetter
describes it in Cyprinus and Barbus.

d. Intermandibularis, Geniohyoideus, and Hyohyoideus.

These three muscles are derived by Vetter from the ventral
portion of that part of the general constrictor that, in selachians,
lies in front of the first gill slit. In the earlier part of his
work Vetter states (No. 124, pp. 411, 417, and 426) that this
part of the constrictor is innervated by the facial alone, and
that it accordingly belongs entirely to the hyoid arch. Tiesing’s
results agree with this earlier work of Vetter’s. Vetter, how-

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 583

ever, in his later publication (No. 125, p. 471), questions the
correctness of his earlier determinations, and states positively
that, in Sphyrna malleus, Prionodon glaucus, Scyllium canicula,
and Galeus canis, that part of the muscle that lies immediately
behind the angle of the chin is innervated by the trigeminus,
and hence belongs to the mandibular arch. The nerve by
which this small anterior portion of the muscle is innervated
is said by Vetter to be a branch of the r. maxillaris inferior
trigemini, given off close to the symphysis of the mandible.
It runs backward and inward over the cartilage into the
muscle, and there forms numerous anastomoses with the deli-
cate terminal branches of the r. hyoideus facialis. In the
American Galeus canis I could not with certainty identify this
nerve. In the specimen examined of that fish a branch was
given off by the r. maxillaris inferior trigemini while that
nerve still lay on the outer surface of the adductor mandi-
bulae, at some little distance from the end of the muscle, and
hence at a still greater distance from the symphysis of the
mandible. The branch ran medianward along the outer surface
of the adductor, and, crossing the narrow line of cartilage which
alone separates the adductor from the ventral constrictor, entered
the latter muscle at some distance from its front end. It seems
hardly possible that this can be the branch described by Vetter,
but it seems still less possible that in this careful work he could
have failed entirely to find so large a nerve. In Carcharinus
the corresponding branch was a large and important one given
off before the main nerve reached the outer surface of the
adductor. It separated into two fairly equal portions before
reaching the constrictor, one of which ran forward and the other
backward in that muscle. A small branch was sent backward
and medianward from, or from near, the base of the nerve. In
both Galeus and Carcharinus the main inferior maxillary, after
leaving the adductor at its front end, continued forward along
the outer surface of the cartilage of the lower jaw toward the
symphysis, giving off several branches. One of these branches
may be the one described by Vetter as going to the extreme
front end of the constrictor. I did not so find it, but I did not
make careful search, and cannot say that it does not exist. If

584 ALLIS. [Vou. XII.

there be such a branch, it must correspond to the nerve that I
shall describe in Amia as r. ghi, and the important branch in
Galeus and Carcharinus to the nerve r. ghs ; if there be no such
branch, and if there be no other branch sent to the constrictor
proximal to the one that I have described in Galeus and Car-
charinus, then this one branch alone must give rise to the two
branches in Amia, one of its branches splitting off from the
main nerve as the regions innervated by the branches became
distinct and separate.

Whatever the homologies of these several branches of the
trigeminus may be, their distribution in selachians shows
that a large and important part of the ventral constrictor in
those fishes may belong to the mandibular arch. Vetter’s
muscle Csvz, therefore corresponds, in its innervation partly by
the trigeminus and partly by the facialis, to the three muscles
described under this subheading in Amia. As the muscle in
selachians, in its superficial portion, forms an unbroken layer
extending from near the symphysis to and beyond the hind end
of the lower jaw, the determination, even in a general way, of
which part of it gives rise to each of the three muscles in Amia
is largely a matter of supposition.

In Galeus and Carcharinus there is no superficial bundle
of Csvz inserted on the outer surface of the adductor mandi-
bulae, as described by Vetter in Scymnus and Acanthias. The
hindermost fibres, however, of that part of the muscle that is
inserted along the lower edge of the mandible, and which may
be called the anterior portion of the muscle, form a somewhat
separate bundle, continuous with the rest of the muscle in its
median ventral portion, but distinctly separate from it at its
insertion. This bundle is also distinctly separate at its inser-
tion from that part of the constrictor that lies immediately
posterior to it. When its fibres are separated with the scalpel
from the rest of the muscle, an inner, deeper layer is disclosed
lying immediately internal to it. This deeper layer has its
origin entirely from the hyoid arch, and is continuous behind
with the posterior portion of the superficial constrictor. It ex-
tends forward about one half the length of the mandible, runs
almost directly medianward to the lateral edge of the m. coraco-

Nong) J7USCLES AND NERVES IN AMIA CAEVA. 585

mandibularis, which it overlaps slightly ventrally, and is inserted
in a fascia or aponeurosis which seems to extend across the
ventral as well as the dorsal surface of that muscle.

The facial nerves in both Galeus and Carcharinus come to the
outer surface of the head along the front edge of Csd2. Sev-
eral branches are sent backward to that muscle as the nerve
reaches the outer surface, or even before it reaches the surface.
A large branch, corresponding apparently to the ramus mandib-
ularis facialis in Amia, is then sent downward in front of the
main truncus. After giving off this important mandibular
branch, the remaining, larger part of the facial nerve continues
downward and backward, immediately behind the upper or hind
edge of the upper jaw, to the point where the hindmost bundle
of the anterior part of the ventral constrictor is attached to the
hind edge of the mandible. There a branch, apparently sensory,
is sent medianward and forward across that bundle. After
giving off this branch, and after having sent several branches
to the posterior portion of the constrictor, the remainder of the
facial nerve passes internal to the hindermost bundle of the
anterior portion of the muscle, and, lying along the outer sur-
face of the inner, deeper layer of the muscle near its lateral
edge, runs medianward and forward, sending branches to the
deeper muscle and also to the superficial one. A large and
important branch, if not the main nerve itself, enters the super-
ficial muscle near the front edge of the inner, deeper one, and
not far from the point where the superficial muscle is entered
by the branch already described of the r. maxillaris inferior
trigemini. This part of the facial nerve is unquestionably the
r. hyoideus of Amia, and the inner, deeper muscle, along the
outer surface of which it lies, the only part of the ventral con-
strictor of the hyoid and mandibular arches that has become
an even fairly independent muscle, is the hyohyoideus either
wholly or in part.

This muscle, the hyohyoideus, is apparently much more differ-
entiated in Galeus and Carcharinus than in either of the three
selachians described by Vetter, for in them the muscle is said
to arise from the inner surface of a tendinous streak or aponeu-
rosis that forms the median, ventral portion of the superficial

556 ALLIS. [Vou. XII.

muscle. It is, in all three, continuous behind with the posterior
portion of the superficial muscle, and the facial nerve in all pen-
etrates the muscle at this point ; the point of penetration in
Heptanchus being at about one third the length of the upper
jaw, and in Scymnus and Acanthias at its hind end, as in Galeus
and Carcharinus. In all three fishes the nerve separates at
once into two parts: a r. hyoideus which runs forward along the
outer surface of the hyohyoideus, supplying it, mainly (No. 124,
p. 418), but sending branches also to the superficial muscle, and
a branch which runs forward along the inner surface of the lower
jaw and is said, in Scymnus and Acanthias, to supply in part the
most anterior portion of the constrictor.

That part of the superficial constrictor that lies in front of
the point where the muscle is penetrated by the facial nerve
thus forms an anterior portion, innervated in its anterior part,
probably, by the trigeminus mainly, but by the facialis in part,
and in its hindermost portion by the facialis mainly or alone.
From the anterior part of the muscle the intermandibularis and
the inferior division of the geniohyoideus probably arise; from
the next following and hindermost portions the superior division
of the geniohyoideus. In the several selachians here consid-
ered and the other fishes described by Vetter a regular transla-
tion of this last portion of the muscle from before backward is
shown. In Heptanchus its hindermost fibres extend only to a
point about two thirds the length of the mandible; in Scymnus
and Acanthias they have reached the hind end of the lower
jaw; and in Acipenser they have passed beyond the mandible,
the muscle arising in part (Cs: and Csz) from a fascia below
and in front of the eye, an origin most naturally arising from
the insertion of its superficial bundle in Scymnus, and in part
(Cs3) from the upper end of the ceratohyal above and behind
the surface of origin of the hyoideus inferior. In Amia this
last attachment only is found. In Acipenser the r. maxillaris
inferior trigemini, after entering the lower jaw, sends two
branches perpendicularly across the outer surface of Meckel’s
cartilage (No. 120, p. 233). The anterior of these two branches,
in the specimen I examined, went to the muscle Cs6 of Vetter,
and the posterior one backward along the outer surface of Csr

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 587

and Cs2. Whether or not it extended to Cs3 and Cs4 I did
not attempt to determine, nor did I attempt to trace the course
of the facial nerve. It was sufficient for my purpose at the time
to know that Csr and Csz represent, in all probability, the supe-
rior division of the geniohyoideus of Amia, and that Cs3 either
represents also a part of that muscle or that, continuing its
backward translation, it has finally lost its attachment to the
upper end of the ceratohyal and become associated with Cs4,
probably as part of the superior division of the hyohyoideus,
Css representing, in part at least, the inferior division of that
muscle.

As the lateral ends of that part of the constrictor that gives
rise to the superior division of the geniohyoideus thus travelled
backward relatively to the mandible, the median ventral portion
of the entire muscle seems to have travelled in the same direc-
tion, and relatively at a still more rapid rate, thus giving rise
to the backward and inward direction of the fibres seen in the
entire constrictor in Scymnus and Acanthias, in the mylohyoi-
deus in Acipenser, and in the inferior geniohyoideus in Amia.
The backward translation of this median part of the muscle
then apparently ceased, and a translation relatively forward of
the median end of the superior geniohyoideus alone began,
giving rise to the condition in Acipenser, where the median
portion of the superior muscle, Csi and Csz2, overlaps externally
the inferior muscle, Cs6, and to the condition in Amia where
the superior division of the muscle extends in part internal to
the inferior division and is inserted in the median aponeurosis,
on the hind edge of the intermandibularis, on the floor of the
mouth, or on the mandible near the symphysis.

In teleosts the inferior geniohyoideus is gradually absorbed
by the superior muscle and finally disappears entirely. The
first stage in this process is shown in Esox, where the fibres of
the inferior division, although retaining their special origin on
the mandible, have become at their insertion directly contin-
uous with fibres corresponding to the median, superficial portion
of the superior muscle in Amia. In Perca and Cyprinus the
origin of the fibres of the inferior muscle on the mandible has
disappeared, but its insertion is still in a measure retained, for

588 ALLIS. [Vox. XII.

the geniohyoidei of opposite sides are still somewhat united at
the middle line of the head. In Barbus even this indication of
an inferior muscle has disappeared, the two geniohyoidei simply
touching each other in the middle line without any connection
whatever. In all four fishes the muscle runs forward either
below or above the intermandibularis, or in Barbus between the
fibres of that muscle, and is inserted, in all, on the mandible, or
in the floor of the mouth immediately behind it.

The inferior hyohyoideus of Amia, and probably only the ante-
rior or lateral part of that muscle, is represented in Heptanchus,
Scymnus and Acanthias by the inner, deeper layer that lies inter-
nal to the posterior part of the anterior division of the ventral
constrictor. It is still connected at its ventral end in these
fishes with the general median aponeurosis, but in Galeus and
Carcharinus this connection has disappeared and the muscle is
inserted, along the edge of the coraco-mandibularis, in a fascia
which is most strongly developed along the dorsal, inner surface
of that muscle. This part of the hyohyoideus is probably repre-
sented in Acipenser by Css, which is still inserted on the inner
surface of Cs2, as it is in Heptanchus and Scymnus. In Chi-
maera it is represented by the hyoideus inferior of Vetter which
shows a separation into two parts. The anterior part is
inserted on the lower jaw of its own side, while the other, run-
ning internal to the coraco-mandibularis between it and the
coraco-hyoideus, passes beyond the middle line of the head and
is inserted on the lower jaw of the opposite side of the head,
overlapping at its insertion the muscle of that side. In Amia
the lateral part of the inferior portion of the hyohyoideus corre-
sponds closely in origin and insertion with this muscle in Chi-
maera. It arises from the ceratohyal immediately beneath and
distal to the geniohyoideus, and is inserted on the under surface
of the hypohyal and in the integument of the floor of the mouth
beyond it, its tendons passing internal to the branchiomandibu-
laris, between it and the sternohyoideus, and, beyond the middle
line of the head, overlapping the tendons of the muscle of the
opposite side and extending toward the inner surface of the
mandible. If this part only of the muscle in Amia is repre-
sented by the entire muscle in selachians, Acipenser and Chi-

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 589

maera, the remainder of the muscle in Amia, that is, the superior
division and the median part of the inferior division, must be
represented by that part of the ventral constrictor which in sela-
chians lies behind the point where the facial nerve penetrates
the muscle, with the addition perhaps of some part of the super-
ficial layer immediately in front of it. This muscle is repre-
sented in Acipenser by Cs4, with or without the addition of
Cs3, and in Chimaera by Cs: or Cs1 and Cs2; Css, although lying
internal to the branchiomandibularis, probably representing
some part of the geniohyoideus. In both these fishes the
muscle is still inserted in a median aponeurosis and lies, as it
does in Amia, external to the coraco-mandibularis, the homo-
logue of the branchiomandibularis. In Amia the insertion has
shifted to the under surface of the hypohyal, the muscle of
either side crossing the middle line of the head and overlapping
its fellow of the opposite side.

In the teleosts described by Vetter an apparent reversion or
retrogression in the hyohyoideus has taken place. The bran-
chiomandibularis, which in Amia lies between the two parts of
the muscle at their insertion, has in the teleosts disappeared,
and the two parts of the hyohyoideus, no longer kept apart by
it, tend to unite; either that or the anterior and lateral part of
the muscle gradually disappears. In Esox the tendinous front
edge of the muscle, found in Amia, has disappeared, and the
muscle shows at its insertion a distinct separation into two
parts. In Perca the anterior of these two parts has been
reduced to a few fibres only, and in Cyprinus and Barbus not
only have these disappeared, but the attachment to the
hypohyal, still retained in Perca, has been lost, and the muscle
is inserted entirely in a median aponeurosis with its fellow of
the opposite side of the head, or has become continuous with
that muscle.

In selachians the hyohyoideus and the hyoid branch of the
facial that supplies it both lie external to the cartilaginous gill-
rakers, while in Amia, even in young larvae, and in teleosts,
both muscle and nerve lie internal to the branchiostegal rays.
Vetter suggests, in explanation (No. 125, p. §37), that the mus-
cle has passed from the outer surface of the gill-rakers to the

590 ALLIS. [VoL. XII.

interspaces between them and then to their inner surfaces,
dragging the nerve along with them. It seems much simpler
to suppose that the branchiostegal rays and gill-rakers are not
homologous structures, and that the former have been devel-
oped in their present position as simple dermal bones.

e. Adductor Hyomandibularis, Adductor Operculi, and Levator
Operculi.

These three muscles arise from the muscle Csd@2 of Vetter,
that is, from a muscle which he derives from the dorsal half of
the superficial constrictor of the hyoid arch. Csdz arises in its
anterior portion, in Heptanchus, Acanthias, and Scymnus, from
the upper, outer corner of the occipital region of the skull. In
its posterior portion it arises in Heptanchus mainly from the
lateral edge of a superficial dorsal fascia, and in Acanthias and
Scymnus from the edge of a corresponding fascia, and from a
tendinous streak which marks the place where the long free
edge of the first gill-cover or wall in Heptanchus has in part
disappeared by fusion with the next following partition (No. 124,
Note 1, p. 414). In Acanthias this fusion incloses and marks
the position of a part of the lateral canal of the head.

The posterior part of the muscle in all three fishes is either
directly continuous with the ventral half of the constrictor, or
is inserted with that muscle in an aponeurotic formation which
lies between them. The anterior part of the muscle is, in
Heptanchus, inserted mainly along the hind, upper edge of the
upper jaw, a thin layer only, which separates from the inner
surface of the muscle, beingyinserted on the outer edge of the
upper member of the hyoid arch. In Acanthias the muscle is
inserted mainly on the front, lateral edge of the hyomandibular,
a small portion only, forming the anterior bundle of the muscle,
being inserted on the upper, hind corner of the upper jaw. In
Scymnus even this slight attachment to the upper jaw is not
found, the anterior fibres of the muscle, corresponding to
the anterior bundle in Acanthias, being inserted instead on a
strong ligamentous band which extends from the upper edge
of the hyomandibular to a projecting corner of the hind edge
of the upper jaw.

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 591

In Galeus Csd2 arises, as it does in Scymnus and Acanthias,
from the upper, outer corner of the skull, from the line mark-
ing the position of the lateral canal of the head, and from the
tendinous streak which extends upward and forward from the
upper end of the first gill slit. That part of the muscle that
arises from the skull has, however, been differentiated as a
wholly separate muscle and lies immediately external to the
anterior part of the rest of the muscle, the front edges of the
two exactly coinciding. It is a thin, delicate, band-like muscle,
nearly as broad as it is long, inserted entirely on the hyoman-
dibular and lying immediately external to the large spiracle
muscle. The facialis, as it runs outward around the front edge
of the muscle below the spiracle muscle, sends a branch out-
ward and then backward along its outer surface to supply it.
It thus corresponds in its innervation to the adductor hyoman-
dibularis in Amia, and but slight changes in origin and inser-
tion would be needed to produce that muscle. The muscle
in Galeus has simply to travel backward at its origin beyond
the origin of the spiracle muscle and then downward along the
side of the skull. If this little muscle in Galeus is the homo-
logue of the adductor hyomandibularis in Amia, the remaining
larger portion of Csd@2 must either disappear or, more probably,
give rise to the opercularis in Acipenser, and to the adductor
operculi and levator operculi in Amia, which muscles arise
wholly or in part from the dermal bones of the head, as does
the muscle Csd@2 in Galeus.

In Chimaera the adductor operculi and the levator operculi
are probably represented by Csz ; the retractor hyomandibularis
of Acipenser and the adductor hyomandibularis of Amia being
represented by the muscle which Vetter has called the hyoideus
superior, and which he considers the homologue of the superior
division of the hyohyoideus in teleosts.

5. Nervus Trigeminus and Nervus Facialis.

The several branches of the nervus trigeminus and nervus
facialis arise from what is commonly called in fishes the Gas-
serian ganglion, or, more properly, the trigemino-facial gangli-
onic complex. This ganglion, in the adult Amia, lies in the

592 AVEIELS: [VoL. XII.

upper, lateral chamber of the eye-muscle canal, and hence,
although lying inside the cartilaginous cranium, lies outside
the cranial cavity, properly so-called. From the ganglion arise
the nine large nerves that form the trigemino-facial complex,
as generally described, and several smaller nerves which may
be considered as belonging to one or other of the nine principal
ones. Of the nine principal nerves, three are commonly con-
sidered to belong to the trigeminus, two to the dorsal root of
the facialis, and four to the ventral root of the facialis. The
three nerves that are assigned to the trigeminus are, the r.
ophthalmicus superficialis trigemini plus the portio ophthal-
mici profundi, or trigeminus I; the r. maxillaris superior trige-
mini, or trigeminus II; and the r. maxillaris inferior trigemini,
or trigeminus III. The two nerves that are assigned to the
dorsal root of the facialis are, the r. ophthalmicus superficialis
facialis, and the r. buccalis facialis, of which the r. oticus is
properly a branch (No. 3, p. 516). The four nerves that are
assigned to the ventral root of the facialis are, the r. palatinus
facialis, the r. hyoideus facialis, the r. mandibularis internus
facialis, and the r. mandibularis externus facialis. As the
latter nerve innervates only the canal organs and the surface
pit organs of the operculo-mandibular portion of the lateral
canal system (No. 3, p. 518), and as it arises in Amia (Allis)
and in Rana (Strong) as a branch of the so-called dorsal root
of the facialis, and in sharks (Ewart) as a part of that root
or as a separate dorsal root, it should be added to the dorsal
branches. The rami maxillaris superior and maxillaris inferior
trigemini together form the truncus maxillaris trigemini, which
is in Amia very short; and the rami hyoideus, mandibularis
internus, and mandibularis externus facialis together form the
truncus hyoideo-mandibularis facialis, which is somewhat
longer. The mandibularis externus and mandibularis internus,
although arising from different roots, are at first intimately
associated with each other, and after separating from the
truncus hyoideo-mandibularis they together form the truncus
mandibularis facialis. The ophthalmicus superficialis trigemini
and the ophthalmicus superficialis facialis always arise separately
from the ganglionic complex and are, the portio minor and the

No. 3.] MUSCLES AND NERVES [N AMIA CALVA. 593

portio major respectively of the ophthalmicus superficialis tri-
gemini of Schwalbe and Van Wijhe.

a. Trigemino-Facial Ganglion.

The trigemino-facial ganglionic complex, as commonly con-
sidered, usually includes the profundus ganglion. As that
ganglion is found, in Amia, as a wholly separate ganglion, the
complex, in Amia, consists of a profundus ganglion and what
will be called, in these descriptions, the trigemino-facial ganglion
proper. As this last ganglion contains the so-called lateral
elements of the facialis, it should, perhaps, not be assigned
entirely to the trigeminus and facialis.

In 12 mm. larvae the profundus ganglion lies in front
of and internal to the anterior, upper end of the trigemino-
facial ganglion proper, separated from that ganglion by a de-
scending portion of the orbital vein or venous sinus. It has
a long and delicate root running almost directly backward to
the extreme lateral portion of the anterior surface of the
medulla oblongata, where the fibres turn inward and backward
and enter the brain as a distinct bundle having a somewhat
deep origin. The root is entirely separate from the anterior,
or trigeminal, root of the trigemino-facial ganglion, lying in
front of, above, and internal to it, and separated from it by
the orbital vein.

The trigemino-facial ganglion proper consists of two dis-
tinct or nearly distinct parts, the main ganglion and the
ganglion of the ophthalmic and buccal branches of the facialis.
The former, in the earliest larval stages examined, is a large,
irregular mass of cells, separated imperfectly into three parts
or regions, an anterior, a median, and a posterior, the anterior
and posterior portions each being traversed by a large bundle
of fibres. The anterior end of the ganglion has two projections,
an outer one giving origin to the truncus maxillaris trigemini,
and an inner one to the ramus ophthalmicus superficialis trige-
mini. The latter, or inner end lies at a higher level than the
outer one, running upward and forward in front of, and in a
measure internal to the ganglion of the ophthalmic and buccal

594 ALLIS. [Vor. XII.

branches of the facialis. From its base the anterior root of
the ganglion arises. It isa short, stout root, running inward
and backward to the anterior and lateral corner of the medulla,
where it has its origin. It is here distinctly double, having a
_dorsal posterior portion and a ventral anterior one. The former
arises by fibres that spread and extend backward in the brain,
parallel to the outer surface of the brain, almost to the level of
the origin of the facialis. The other, or ventral portion, arises
in the brain as two bundles, which extend, as bundles, backward
and medianward into the brain. The anterior bundle arises, as
seen in horizontal sections, in front of the ventricular cavity of
the brain, from the anterior, upturned end of the tract from
which the ventral motor component of the facialis arises, and
which I take to be the posterior longitudinal fasciculus of
Osborne in cryptobranchs. This bundle is the motor V, or V
minor, of Strong in Rana.

The fibres of the antero-ventral part of the anterior root of
the main ganglion form the larger part, if not all, of the anterior
of the two bundles of fibres that traverse the ganglion. From
this transverse bundle the truncus maxillaris in large part arises,
the bundle turning downward in that part of the ganglion that
lies external to the orbital vein, and then outward and forward
into the truncus. In one 20 mm. larva the transverse bundle,
or commissure, was distinctly double, the fibres of both parts
arising from the antero-ventral part of the anterior root of the
main ganglion. Both parts entered the truncus maxillaris.
Whether they represent the two bundles into which their root
separates on entering the brain, or only one of those bundles,
was not determined. The fibres arising from the postero-
dorsal portion of the main root disappeared gradually as they
approached the truncus maxillaris. The transverse bundle in
the ganglion is called by Strong in Rana (No. 121, p. 135) the
motor bundle of the portio minor of the trigeminus.

The orbital vein, which lies at first anterior to, and internal
to, the upper, anterior part of the ganglion, runs outward,
downward, and backward under that portion and its root, and
issues on the dorsal surface of the median and posterior portions
of the ganglion. It thus pierces the main ganglion under its

No. 3.] WUSCLES AND NERVES IN AMIA CALVA., 595

anterior root and between its anterior and posterior portions.
While on the dorsal surface of the ganglion it lies between it
and the ganglion of the ophthalmic and buccal branches of the
facialis. It there receives a large branch from the direction of
the ramus buccalis and truncus maxillaris, and, continuing down-
ward and backward along the lateral surface of the ganglion,
joins or becomes the jugular vein. Under the vein the anterior,
or trigeminal, part of the ganglion becomes continuous with
the posterior or facial portion.

The ramus ophthalmicus superficialis trigemini arises from
the extreme upper, anterior end or horn of the main ganglion,
and the fibres that enter it do not apparently arise at all from
the antero-ventral portion of the anterior root, or from the ante-
rior transverse commissure of the ganglion. They arise, in
large part, from the superficial portions of the median portion
of the main ganglion, and run forward above and below the
anterior commissure along the outer surface of the ganglion.
A large bundle of these same fibres, arising from the ventral
surface of the ganglion, joins the truncus maxillaris.

The ganglion of the ophthalmic and buccal branches of the
facialis lies, as already stated, immediately behind the upper,
anterior and inner end or horn of the main ganglion. It
- reaches to a somewhat higher level than that ganglion and is
Y-shaped in form, as already described in my earlier work (No.
3, p. 516), the anterior end of the main ganglion lying between
and below the arms of the Y. Its ophthalmic branch is, in
young larvae, much larger than the ophthalmic branch of the
trigeminus, and at these ages the terminal buds on the top of
the snout are not yet largely developed. In later stages, when
the buds become numerous, the two nerves have about the
same size.

No fibres connecting the ganglion of the ophthalmicus and
buccalis facialis with the main portion of the trigemino-facial
ganglion were found in larvae or in the adult, but I cannot posi-
tively say that they do not exist. The ganglion has a large,
long root which runs at first inward and backward, and then
directly backward along the side of the medulla, passing
between it and the inner edge of the anterior wall of the ear

596 AELLTS. EVOL. 2k

capsule into that capsule. It enters the brain immediately in
front of the root of the acusticus, almost as a part of that root,
the two roots issuing each by two or more rootlets from the
summit of a slight swelling or eminence on the side of the
medulla. From its under surface, close to its origin, or pos-
sibly as a separate root, a large branch is sent downward,
outward, and forward, to the posterior portion of the main
trigemino-facial ganglion, where it turns outward and then out-
ward and backward, traversing the ganglion and issuing as
part of the truncus hyomandibularis facialis. It forms a large
part of the posterior bundle already spoken of as traversing the
ganglion. These two roots, or two branches of a single root,
together form the sense-organ, or, as it has been called, lateral
component of the trigemino-facial ganglionic complex.

Close to and intimately associated with this lateral root arise
two other bundles which together form the true posterior or
facial root of the main ganglion. The anterior of the two arises,
as a bundle, at a high level in the brain, probably from the
fasciculus communis of Osborne and Strong, and issues in
front of and internal to the sense-organ root. It enters mainly
into the median portion of the main ganglion, and from this
portion arise the ramus palatinus facialis as well as the fibres
already spoken of that go to forma large part, if not all, of the
ophthalmicus superficialis trigemini and a considerable part of
the truncus maxillaris. The other root arises also as a distinct
bundle deep in the brain, but nearly on the same level as its
point of exit, which is in front of and below the root of the
acusticus and below the lateral root of the complex. It is prob-
ably the motor root VII ab of Strong in Rana, although its
position is not exactly the same as that given by him. It
enters the posterior portion of the main ganglion immediately
in front of the hyomandibular branch of the lateral root, and
apparently traverses the ganglion with that root, forming
_ with it the posterior transverse bundle or commissure of the
ganglion.

The abducens in these larvae arises immediately behind the
posterior end of the ganglion of the acusticus, and in my sec-
tions is at its origin distinguished with difficulty from that

No. 3:)) J7ZUSCLES AND NERVES IN AMIA CALVA. 597

ganglion. Whether it arises in part from the ganglion, or
wholly from the brain, it certainly has a much more lateral and
higher origin than in the adult, agreeing in this with the origin
of the nerve in the adult of Petromyzon (No. 1, p. 21), but
lying behind instead of in front of the so-called motor root of
the trigeminus.

In the adult, as in the young, the trigemino-facial ganglion
(gg, Figs. 38, 39, and 64, Pls. XXX and XXXVIII) has two
apparent roots, an anterior (v¢fa) and a posterior one, the latter
consisting of two parts, the true posterior root of the ganglion
(riff) and the sense-organ root or roots (voéf). The anterior root
is the smaller of the two. It runs forward and outward in the
cranial cavity, from its place of origin, pierces the lining mem-
brane of the eye-muscle canal in front of the utricular fossa,
and enters the anterior and upper part of the main ganglion.
From its inner, anterior surface, while still inside the cranial
cavity, the much smaller root of the ophthalmicus profundus (rp)
arises.

The two portions of the posterior root arise close together,
or as a single root, immediately in front of and above the root
of the acusticus. The part destined for the ganglion of the
ophthalmicus and buccalis facialis separates at once from the
rest of the root, but runs forward and outward with it, closely
applied to its upper surface, and escapes with it from the cranial
cavity through the inner, membranous edge of the front wall of
the utricular fossa, separated by that edge from the anterior
root at its exit. The front wall of the fossa, which is also the
hind wall of the upper, lateral chamber of the eye-muscle canal,
is formed by a process of the petrosal and a cartilaginous ridge
continuous with that process immediately above it. The inner
membranous edge of this wall and not the wall itself is what is
pierced by the root of the facialis as it issues from the cranial
cavity. Having entered the upper chamber of the eye-muscle
canal, the upper portion of the root enters its own ganglion
and the lower portion enters the posterior portion of the main
ganglion.

598 ALLIS. [VoL. XII.

b. Ramus Ophthalmicus Superficialis Trigemini and Ramus
Ophthalmicus Superficialis Facialis.

The ophthalmicus superficialis trigemini (of¢ in many figures)
arises in the adult, as in larvae, from the median and upper
portion of the trigemino-facial ganglion immediately under the
ophthalmicus facialis (off). It leaves the upper, lateral cham-
ber of the eye-muscle canal and issues from the cranium with
the ophthalmicus facialis by the ophthalmic foramen through
the alisphenoid (offr, Figs. 9 and 10, Pl. X XI), immediately
above the large trigeminal foramen. The two ophthalmic
nerves are, from the very beginning of their extra-cranial
course, closely and intimately associated. They run upward
and forward along the side of the cranium, and not at some
distance from it, as Schneider states, lying at first under the
overhanging, cartilaginous roof of the orbit, and then under the
overhanging frontal, which, opposite the eye, where the roof
of the cnondrocranium is relatively narrow, projects so much
beyond that roof that the nerves lie approximately under the
middle line of the bone, and hence under that part of the supra-
orbital sensory canal that lies in that bone.

About opposite the front edge of the eye, the two ophthalmic
nerves leave the orbit, and reach the upper surface of the chon-
drocranium, passing as they do so through a notch in the edge
of the cartilaginous cranium. This notch (z, Figs. 8, 9, and
10, Pl. X XI) lies a little behind the prefrontal ossification, and
a little in front of the middle point of the frontal bone, and is
the preorbital incisure of Pristiurus, Prionodon, and Zygaena
(No. 44, p. 71). It marks, as in those fishes, the posterior limit
of the preorbital process (fv). The anterior limit of the proc-
ess 1s defined, as in those fishes, by an ethmoid incisure (ze).
Having reached the upper surface of the cranium, the ophthal-
mic nerves continue directly forward, soon entering a groove on
the under side of the frontal bone, in which they pass above the
projecting hind end of the premaxillary. Leaving the groove
at the front end of the frontal, they pass through the narrow
strip of dermal, or subdermal, tissue that lies between that
bone and the nasal, and then pass under the nasal, lying, in

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 599

this part of their course, in the thin membranous covering of
the nasal sack. There is no ethmoid canal in Amia, and as that
canal is found in Polypterus (No. 93) it would seem to depend
on the presence of a ramus ophthalmicus profundus, and on the
partial fusion of the canal for that nerve with the canalis pre-
orbitalis. If, then, the canal for the profundus in Polypterus is,
in its proximal portion, the orbito-nasal canal of selachians, as
seems probable, that canal must sometimes in ganoids fuse with
the preorbital canal, as in Polypterus, and sometimes with the
olfactory canal, as in Amia. In selachians the distal portion
of the canal seems to fuse with or become the ethmoid canal,
for Gegenbaur, in all his ventral and lateral views, never gives
but one anterior opening, calling it sometimes the distal open-
ing of the ethmoid canal and sometimes that of the orbito-nasal.
There may even be two canals for the profundus in selachians :
one, the ethmoid canal, for the dorsal branch of the nerve,
Ewart’s nerve on! (No. 29, Fig. 2) ; and one, the orbito-nasal,
for its orbital branch, Ewart’s nerve ov. Ewart unfortunately
does not describe the course of the latter. It would seem,
however, from its name to remain in the orbit.

Immediately after issuing from its foramen, the ophthalmi-
cus trigemini gives off two branches one of which arises from
the outer and the other from the inner surface of the nerve.
These branches are flat and broad at their bases, and closely
embrace the ophthalmicus facialis. They run upward and
backward, one on either side of the facialis, pierce the alisphe-
noid, and, entering the cartilage of the cranium by the same or
by closely adjoining canals, issue on its upper surface either by
the same or by adjoining foramen. They are the first two, or
first pair, of frontal branches of the nerve, and the course and
position of the canal or canals traversed by them, as also that
of the other small canals traversed by other frontal branches
of the nerve, vary greatly in different specimens. There are
frequently both in the young and in the adult, small branch
canals connecting these canals and giving passage to delicate
anastomosing branches of the nerve.

One of these first two branches of the ophthalmicus tri-
gemini is accompanied by the first branch of the ophthalmicus

600 ALLIS. [Vou. XII.

facialis. This branch of the facialis (offa/) arises inside the
cranium from the upper surface of the nerve, near its origin
from its ganglion, and issues with it through the ophthalmic
foramen, lying along its upper surface. It then turns upward
and backward, with the two frontal branches of the trigeminus,
and having reached the upper surface of the chondrocranium
turns medianward and backward under the frontal and parietal
bones and supplies the anterior dorsal line of pit organs (No. 3,
p. 513). In Ammocoetes the corresponding branch arises,
according to Hatschek, from a small anterior portion of the
facial ganglion, and is the nervus temporalis facialis or cuta-
neus dorsalis facialis (No. 54, p. 156). It is, in Amia, always
closely accompanied, after reaching the upper surface of the
cranium, by the inner of the first two frontal branches of the
ophthalmicus trigemini.

The other or outer nerve of the first pair of branches of the
ophthalmicus trigemini usually separates into two parts, both
of which run almost directly backward along or near the outer
edge of the temporal groove, and either unite again to form a
single nerve or are connected by anastomosing branches.
From the median one of its two parts an anastomosing branch
is usually sent to the inner nerve of the pair. The lateral one
is always directly continuous with a nerve formed usually by
the fusion of two branches of the vagus. From the compound
nerve thus formed several branches are sent to the dermal
bones and underlying tissues of this part of the head, but
which of these branches belonged to the trigeminus and which
to the vagus could not be determined.

Immediately in front of the inner one of the first two
branches of the ophthalmicus trigeminus there were, on each
side of the head in the specimen used for illustration, two
branches arising close together and forming what may be for
convenience called the second pair of frontal branches. About
on a level with them, or a little in front of them, the branch of
the facialis destined to supply organ 7 of the supraorbital sen-
sory canal arises (off.so?). These three nerves all run upward
and backward through the cranial cartilage and issue on the
upper surface of the chondrocranium, but their course and

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 601

their relations to each other vary greatly in different specimens.
On the left side of the head, in the specimen here described,
the facialis branch to organ 7 was accompanied by the anterior
of the two trigeminal branches that form the second pair, the
other trigeminal branch of the pair issuing by a separate and
slightly lateral canal. On the right side the facial branch to
organ 7 joined the inner one of the first pair of trigeminal
branches, those forming the second pair issuing separately by
two somewhat anterior openings. On both sides of the head
there were, associated with these first two pairs of frontal
nerves, one or two branches of a nerve which arose from the
trigemino-facial ganglion under and lateral to the root of the
ophthalmicus trigemini and which will be later described as
nerve ‘a.’ On the left side of the head these branches of nerve
‘a’ pierced the cranial cartilage near its lateral edge and issued
on the top of the cranium lateral to and slightly in front of
the two pair of trigeminal branches; on the right side, where
two pair of frontal nerves issued at some distance apart, the
branches of nerve ‘a’ issued lateral to, but between them.

The next branch of the ophthalmicus facialis is the one to
organ 6 supraorbital. It runs upward and backward external
to the cranial cartilage, lying in or just in front of a strong
notch in the cartilaginous roof of the orbit. Ona level with
this nerve, in the specimen used, four branches arose from the
ophthalmicus trigemini on the left side of the head, and three
on the other. One of the branches on the left side, and two of
those on the right, lay internal to, or above, and the others
external to, or below, the ophthalmicus facialis. Those lying
internal to the facialis ran upward immediately behind the
branch to organ 6, either lying in the notch already described,
or passing through special canals in the cartilage. Having
reached the upper surface of the cranium, they turned median-
ward and were distributed to tissues toward the middle line of
the head. The other branches, issuing below the ophthalmicus
facialis, ran outward and forward into the fatty tissues above
the eye. On the right side of the head there were no branches
from the ophthalmicus trigemini between this group of nerves
associated with the facialis branch to organ 6 and those asso-

602 ALLIS. (Vou. XII.

ciated with the branch to organ 7. On the left side there were,
however, between these two groups, two pair of trigeminal
branches, one branch in each pair lying above, or internal to,
and the other below, or external to, the facialis. Some of these
nerves pierced the cranial cartilage and issued on the top of
the cranium, others were distributed to fatty and other tissues
above the eye. One of the nerves in each pair was double at
its origin. Intimately associated with an internal branch in
the anterior pair there was an external branch which arose
toward the base of the ophthalmicus trigemini, ran forward
closely applied to the external surface of the ophthalmicus
facialis, and, turning inward across the upper surface of that
nerve, joined the internal branch referred to. So closely and
intimately was this branch associated with the facialis that
Mr. Nomura for some time stoutly maintained that it was a
branch of that nerve, although repeatedly assured that, there
being no supraorbital organs between organs 6 and 7, there
could be no branches of the ophthalmicus facialis between the
branches to those organs. This marked instance of misinter-
pretation by an exceedingly careful workman is given simply
to enforce the statement already made that many so-called
anastomoses may be simply instances of intimate juxtaposi-
tion, and that similar misinterpretations made by others have
undoubtedly given rise to much of the confusion in the litera-
ture of the subject.

The branch of the ophthalmicus facialis to organ 5 supra-
orbital is given off as the main nerves pass to the upper sur-
face of the chondrocranium, and it had on the left side no
trigeminal branch associated with it. On the right side two
branches arose from the trigeminus near it, both passing out-
ward and forward below the facialis, and both distributed to
fatty and other tissues above and in front of the eye.

The branch of the facialis to organ 4 is given off after the
nerve has entered the groove in the frontal, and has passed, or
is about to pass, above the hind end of the premaxillary. The
branches to organs 3, 2, and I are given off after the nerve
passes under the nasal, the nerve ending in organ I.

On a level with the facial branch to organ 4, a branch is

mt.

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 603

given off by the ophthalmicus trigemini on both sides of the
head, and the nerve then breaks up into several terminal
branches all of which lie under the nasal bone, in the mem-
brane covering the nasal sack, and run forward on both sides
of the facialis toward the end of the snout.

The portio ophthalmici profundi (off) joins the ophthalmi-
cus superficialis trigemini a little proximal to the branches of
that nerve that accompany the branch of the facialis to organ
6 supraorbital. On the right side of the head, in the one
specimen, the superficialis had separated, opposite the point
of union with the portio profundi, into two nearly equal
strands. No other instance of such a separation was found.

c. Ramus Buccalis Facialis.

The buccalis facialis (f) arises from the outer side of the
ganglion common to it and the ophthalmicus facialis. It issues
from the upper, lateral chamber of the eye-muscle canal, through
the large trigeminal foramen, with the superior and inferior
maxillary nerves, lying immediately above them. It then
turns downward and runs downward, outward, and forward
behind and under the eye, and then forward and inward toward
the end of the snout, lying immediately external to and above
the maxillaris superior. It is distributed entirely to the organs
of the infraorbital sensory canal (No. 3, p. 514). On the left
side of the head there were, in the specimen used for illus-
tration, fifteen of these organs, on the right side the normal
number, fourteen.

The oticus facialis (ef), which is simply a branch of the
buccalis, arose on both sides close to the base of that nerve.
It ran upward and outward through the cranial cartilage and,
as in the embryo, supplied organs 15 and 16 infraorbital and
the organ of the spiracular canal. Close to its base the branch
to organ 14 infraorbital arose, a little further forward the
branch to organ 13, and close to this last branch, on the left
side of the head, two branches, and on the right, a single one,
double at the end. There were beyond this group of nerves
eleven buccal branches on each side of the head, the first

604. ALLIS. [Vov. XII.

branch arising some little distance beyond the most distal
one of the group. There must accordingly have been fifteen
organs or groups of organs in the infraorbital canal on the left
side of the head, and either fifteen organs or fourteen organs,
one of them double, on the right. Unfortunately no attention
had been given at the beginning of the dissection, which was
begun on the left side, to determine, either the number of den-
dritic systems or the number of sense organs in the canals,
it being assumed, from earlier embryological work, that the
branches of the nerves concerned would everywhere fully
indicate the number of sense organs and the arrangement
of the dendritic systems. After the dissection of the left side
had been nearly completed and fifteen branches of the buccal
instead of fourteen had been found, the advantage of knowing
the number and position of the sense organs in the different
canals became apparent, and on the right side the outer sur-
face of all the bones concerned was carefully removed, so as to
lay open the canals that traversed them and thus expose, zx
situ, the sense organs, and the ends of the nerves that supplied
them. The number and position of the organs, as indicated by
the projecting nerve ends, being everywhere normal on this
side, no further thought or attention was given to the matter.
Later it was found that one of the branches of the buccalis was
double at the end, but as the dermal bones had all been removed
it was too late to determine which organ or organs had been
supplied by it. In my earlier work on larvae, groups 11 and
12 infraorbital were always supplied, either by branches of a
single nerve or by nerves arising close together from the buc-
calis, and there should therefore normally be, both in the adult
and in the embryo, but ten groups of organs beyond group I1,
and but ten nerves beyond the double nerve supplying groups
It and 12. In the present instance, in this specimen used for
illustration, there were, however, eleven such nerves on each
side of the head. There must, therefore, have been an extra
system, 10’, on both sides, or, assuming no error in observation,
the nerve to group 11 must have become somewhat widely sep-
arated at its origin from the nerve to group 12, and that group
must have become entirely double on the left side, and partly

NOsga)) JIOSCLES AND NERVES, INVAMIA CAEVA. 605

so on the right, as indicated by the projecting nerve end in the
figure (4/2012, Fig. 19, Pl. XXIII).

The branches to organs 10 to 7 infraorbital, inclusive, arose
successively from the buccalis at fairly equal distances apart.
The branches to organs 6 and 5 arose close together, that to
organ 6 passing outward above, that is, behind and lateral to the
fourth division of the levator maxillae superioris as in 40 mm.
specimens, but that to organ 5 passing through that muscle at
its origin instead of under it, that is, in front of it, as in the
young. The buccalis itself also passed through this fourth
division of the muscle instead of between it and the third divi-
sion asinthe young. Beyond this point the branches to organs
4, 3, 2, and 1 were given off regularly in succession, the nerve
ending in the last-named organ.

da. Truncus Maxillaris Trigemini.

The truncus maxillaris separates immediately, in the adult,
into its two main portions, the r. maxillaris superior (#s¢) and
the r. maxillaris inferior (#zt). From these two nerves close
to their origin, from the truncus maxillaris itself, from the
upper surface of the main trigemino-facial ganglion, and perhaps
also in part from the r. ophthalmicus superficialis, several small
nerves arise. They vary in number and arrangement in every
specimen and are the homologues of the three accessory tri-
geminal branches described by Strong in Rana (No. 121, p.
111). In the specimen used for illustration one of these
nerves issued from the cranium with the ophthalmic nerves
through the ophthalmic foramen ; the others all issued with the
superior and inferior maxillary nerves through the main tri-
geminal foramen. They formed numerous anastomoses with
each other and were distributed almost entirely to the fatty
and other tissues above and behind the eye and to the dermal
and subdermal tissues of the cheek. On the left side of the
head there were five of these nerves, a, 6, c,d, ande. They
are shown much enlarged in Figs. 38 and 39, Pl. XXX. In
these figures the nerves are all pulled somewhat apart and
hence somewhat displaced from their natural positions, and in

606 ALLELES: [VoL. XII.

Fig. 39 the ophthalmicus facialis and the buccalis facialis are
turned back for better illustration. The mandibularis facialis
is not shown.

Nerve ‘a’ (v.a) arose under the bundle of fibres that go from
the posterior root of the trigeminus to the r. ophthalmicus trige-
mini. It lay at first under the ophthalmic nerve, closely applied
to it, issued with it through its foramen, and had as it issued
every appearance of being a branch of it. One of its branches,
running upward between the ophthalmicus and buccalis facialis,
pierced the cartilaginous roof of the orbit and, reaching the top
of the chondrocranium, was distributed with those branches of
the ophthalmicus trigemini that are associated with the first,
or first two branches of the ophthalmicus facialis, as already
described. Other shorter branches entered the tissues above
and behind the eye. One long branch joined a branch of nerve
‘c,’ and ran outward with it (ac), in front of, dorsal to, and ex-
ternal to the buccalis, around the front edge and then backward
along the outer surface of the adductor mandibulae, lying near
the upper edge of that muscle. It lay dorsal to and external
to nerve 13 buccalis facialis, between it and nerve 14. Nerve
‘b’ (7.6) arose near nerve ‘a,’ ran forward below it, formed no
anastomoses with the other nerves, and was distributed entirely
to tissues above the eye. Nerve ‘c’ (7c) arose immediately
lateral to nerve ‘b.’ It separated into two parts, one of which
ran outward in front of the buccalis below or internal to nerve 12
buccalis facialis, between it and nerve 11, and then, after being
joined by a branch from nerve ‘e,’ ran (7ce) outward and back-
ward along the outer surface of the adductor mandibulae. It
lay in this part of its course immediately dorsal to nerve ‘ac.’
The remainder of nerve ‘c’ separated into two parts, one of
which joined the long branch of nerve ‘a,’ as already described,
while the other ran downward and joining a branch of nerve
‘d’ was distributed to tissues behind and below the eye (cd).
Nerve ‘e’ (”¢) arose immediately lateral to nerve ‘c.’ It sep-
arated into two parts one of which issued behind, that is, inter-
nal to the buccalis, while the other ran outward and downward
in front of and external to that nerve. The ramus buccalis
thus lay between the two main branches of nerve ‘e,’ and as

No.3.] MUSCLES AND NERVES IN AMIA CALVA. 607

terminal portions of those branches fuse, the buccalis, in a way,
can be said to perforate the nerve so formed, as it does one of
the corresponding nerves in Rana (No. 121, p. 117). The inter-
nal branch of the nerve followed closely the maxillaris inferior
and was distributed to dermal and subdermal tissues along its
course, one long branch, ‘r.e,’ being sent downward behind
the eye and another backward to join a branch of nerve ‘c,’ as
already stated. The external portion of the nerve fused at once
with nerve ‘d’ and the nerve so formed having crossed the
buccalis between nerves 10 and 11, followed the course of the
internal part of nerve ‘e’ lying immediately above it and fusing
in part with it. It also sent an anastomosing branch to that
branch of nerve ‘e’ that was distributed to tissues behind and
below the eye.

The four nerves a, 4, c, and e arose apparently from that part
of the trigemino-facial ganglion that gives origin to the maxil-
laris inferior. Nerve ‘d’ (7d) on the contrary arose by two
roots from that portion of the ganglion that gives origin to the
maxillaris superior. Both roots joined the external portion of
Metvie, ec),

e. Ramus Maxillaris Superior Trigemini.

This nerve (mst) accompanies the buccalis in its course
through the orbit, lying at first immediately under it and then
immediately internal to it. The two nerves run downward and
forward internal to, above, and in front of the first and second
divisions of the levator maxillae superioris, and then dorsal to
the third and fourth divisions of that muscle. They lie while
in the orbit immediately external to, that is, behind or below,
the periorbital lymph space. The membranes lining that space
form a sort of sack which encloses the eye and eye muscles, as
shown in Fig. 20, Pl. XXIII. In this figure the eyeball, which
lies within the sack, and the fatty tissues which lie above and
behind it, have been removed and the ventral nerves are seen
through the membrane.

As the maxillaris passes through the orbit it gives off several
branches. ‘In the specimen used for illustration there were on
each side of the head eight of these branches, seven of them

608 ALLIS. Vion. XI tk

running outward and forward above the buccalis and one run-
ning outward and forward below it. The seven branches that
lie above the buccalis form several anastomoses with each other
and accompany approximately the branches of the buccalis to
organs 5 to 10, inclusive. In this relation to the buccalis they
agree with the branches of nerves a, c, d, and e, which accompany
approximately the branches of the buccalis to organs II to 14.
The one branch that lies below the buccalis (#st.mx) leaves the
maxillaris about opposite the buccal branch to organ 8, and is
by far the largest branch of the nerve. It runs outward and
downward and then forward across the upper surface of the
third and fourth divisions of the levator maxillae superioris and,
near the front edge of the fourth division of that muscle, breaks
up into several branches. In the specimen used for illustration
there were four of these terminal branches on each side of the
head. Three of them turned outward and backward to supply
the inner and outer surfaces of the maxilla, the other continuing
forward under the front articular end of the maxillary bone
toward the front end of the head, and breaking up into several
branches all of which lay above the palatine and vomer and
below the premaxillary. At the point where the main nerve
breaks up into its four terminal branches it lies external to and
close to the r. palatinus posterior facialis with which there is an
interchange of fibres or partial anastomosis. This anastomosis
is undoubtedly the one which in Cyprinidae, according to Sage-
mehl, represents unquestionably the ganglion sphenopalatinum
of higher animals (No. 107, p. 558).

The main maxillaris superior leaves the orbit under the pre-
orbital process and between the third and fourth divisions of
the levator maxillae superioris close to their origins. On the
left side of the head it passed with the buccalis through Lms4
and a branch was sent outward and forward between the two
heads of that muscle. The main nerve then runs forward
above the articular head of the maxillary and separates into two
main branches and other smaller ones. One main branch runs
forward and downward, sending one branch backward and down-
ward along the anterior part of the lower edge of the maxilla,
and others forward, and forward and inward to the tissues

No.3.] MUSCLES AND NERVES IN AMIA CALVA. 609

between the premaxillary and the external dermal bones, and
along the edge of the upper lip. The other main branch of
the nerve turns inward and forward above the premaxillary
and crosses the depression in that bone that forms the bottom
of the front end of the nasal pit. The nerve here lies ona
level with the outer edges of the bone and hence somewhat
above the bottom of the pit, extending across it immediately
under the front end of the olfactory tissue. In this part of its
course several branches are given off, all of them running for-
ward or forward and inward under the buccalis and above the
premaxillary to the dermal and subdermal tissues toward the
end of the snout. The branches form numerous anastomoses
with each other and with other terminal branches of the nerve
and also with a terminal branch of the palatinus anterior facialis
which, having traversed the cartilaginous base of the skull,
issues by a foramen between the inner edge of the septomaxil-
lary and the inner edge of the olfactory perforation of the pre-
maxillary, and reaches the upper surface of that last bone (af,
Figs. 21 and 25, Pls. XXIV and XXV, and paf-fr, Figs. 8 and
10}, PSX XT).

One small branch of this branch of the palatinus facialis was
traced into one of the teeth of the premaxillary.

ft. Ramus Maxillaris Inferior Trigemini.

The maxillaris inferior (#72?) is the largest of the three main
divisions of the trigeminus. It runs at first downward, outward,
and forward, and then downward, outward, and backward around
the front edges of the muscles Lms!, Lms2, Do, and Am. It
then runs downward and backward along the outer surfaces of
A! and Az" to the upper edge of Az’, at about one third the
distance between the insertion of the muscle and its origin.
Here it turns downward and then downward and forward
between A2! and A2", and entering the hollow of the mandible
reaches the inner surface of the coronoid process of Meckel’s
cartilage near its anterior edge and at the posterior end of the
surface of insertion of Aw’. In its course up to this point it
has been a broad, thin, ribbon-like nerve, the flat surface of the

610 ALLIS. [Vou. XII-

nerve lying on the outer surface of Az” and Ao’. Here, how-
ever, it is twisted inward and downward and, although still broad
and thin, lies imbedded in and between the fibres of Ao! at
their insertion, so that the thin edge of the nerve is presented
superficially. Continuing downward and forward in this posi-
tion, along the inner surface of the articular and dentary, it grad-
ually becomes rounded in outline and issues from the hollow of
the mandible, at its extreme front end, through a long longitu-
dinal opening (mzéfr., Figs. 6 and 7, Pl. XX) between the den-
tary, below, and Meckel’s cartilage and the lower edge of the
splenial, above. Beyond this point it extends to the tip of the
mandible lying along the inner surface of the dentary median
to Meckel’s cartilage and immediately below the lower edge of
the splenial. It is in this part of its course closely associated
with the r. buccalis facialis which lies immediately median to it,
the two often appearing as asingle nerve. Between them there
is, however, no interchange of fibres, and they are easily separ-
ated in dissection.

The first branch given off by the maxillaris inferior, ~./af.do,
leaves the nerve from its outer surface before it has separated
from the maxillaris superior, that is, it arises from the truncus
trigeminus, but from that portion of the truncus that contains
the fibres destined for the inferior division of the nerve. It is
given off immediately after the truncus enters the orbit and
runs outward and forward to the inner surface of the levator
arcus palatini near its front edge. There it separates into two
parts, an inner, posterior portion and an outer, anterior one.
The inner, posterior portion runs backward and slightly upward
along the inner surface of the muscle, sending branches into
it and lying between it and the lateral wall of the cranium
immediately above the united first and second divisions of the
levator maxillae superioris. It reaches the front edge of the
hyomandibular near its upper end, and, running along the outer
surface of that bone, along the line separating the dilatator
operculi and the posterior portion of the levator arcus palatini,
sends branches to those two muscles. The other or outer por-
tion of the nerve runs outward around the front edge of the
levator arcus palatini, and then backward along the outer sur-

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 611

face of that muscle, between it and the adductor mandibulae.
On the left side of the head it here separated into two parts,
one of which sent branches into the levator arcus palatini and
finally entered that muscle, while the other continued back-
ward along the outer surface of the dilatator operculi. Whether
this second branch sent branches into either of the two muscles
or not was not determined. On the right side of the head the
outer portion of the nerve was found as a single branch extend-
ing across the outer surface of the levator arcus palatini and
backward onto the outer surface of the dilatator operculi.

Not far from the origin of this first branch of the maxillaris
inferior, but from the under surface of the main nerve, a second
branch, 7./ms, is given off. In young specimens it arises at
almost the same level as the first branch, close to the root of
the truncus trigeminus if not from the ganglion itself, for gang-
lionic cells are found beyond both nerves, extending even as
far as the point where the truncus separates into its superior
and inferior portions. This second branch is often double
through a large part of its course, the two portions being
closely applied to each other. It runs at first downward and
forward immediately internal to and in front of the first and
second divisions of the levator maxillae superioris, and then
forward under the eye, lying at first a little below and lateral
to the united buccalis facialis and maxillaris trigemini, and then
immediately beneath them. At the front end of the orbit it
runs diagonally across the upper surface of the third division
of the levator maxillae superioris, sending branches into that
muscle, and then forward between the third and fourth divi-
sions of the muscle, sending one branch to the outer surface of
the latter muscle, and itself entering it on its inner surface.
As the nerve ran downward and outward internal to the first
and second divisions of the levator maxillae superioris it sent,
in different dissections that were made, one, two, or three
branches into those muscles; and a little further forward, in
one specimen, a branch, unfortunately cut in dissection, was
found entering the tissue near the corner of the mouth, and
in another a loop was found in the main nerve, as if its two
portions had become partially separated.

612 ELIS. [Vou. XII.

The third branch, 7.@m, of the maxillaris inferior arises some
distance beyond the points of origin of the first two branches.
It is given off from the inner surface of the main nerve as it
turns outward, downward, and backward around the front edges
of the levator arcus palatini and adductor mandibulae, and is
destined to supply the two divisions of the latter muscle. It
separates at once into two main portions, both of which run
downward and backward, one entering at once the deeper por-
tion of Az, and the other lying along the outer surface of A3,
and sending branches backward and downward into it.

The next or fourth branch of the maxillaris inferior is given
off from the posterior margin of the ribbon-like nerve just
before it passes under the upper edge of Az’. It runs backward
and downward and then downward along the outer surface of
A.', and is, apparently, destined entirely to the supply of cuta-
neous and subcutaneous tissues. It separates into three or
more portions, all of which continue downward along the outer
surface of A2’, to the extreme lower edge, or corner, of that
muscle. Here one of the three main branches of the nerve
turns forward and sends branches upward and downward along
the outer surface of the hind edge of the mandible; the second
and largest branch turns inward and forward around the lower
edges of Az and A3, and then runs forward along the inner
surface of Ao, lying a little above the upper edge of Meckel’s
cartilage; the third or posterior branch turns inward and back-
ward between the lower end of the preoperculum and the sym-
plectic, and then runs backward and upward along the inner
surface of the preoperculum, following approximately the line of
the hind edge of that bone. In larvae branches of this third
branch were traced directly to the terminal buds on the inner
surface of the mouth cavity near the upper end of the hyoid.

Near the point of origin of this fourth branch of the maxil-
laris inferior a small branch is always found arising from the
outer surface of the nerve and running forward along the ten-
don that extends from the origin of Az! to the inner surface of
the maxilla. It is usually imbedded between Az! and Az”, and
is not seen without separating slightly those muscles.

The fifth branch, ~g/s, of the maxillaris inferior is always

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 613

given off at the point where the main nerve reaches the inner
surface of the coronoid process of Meckel’s cartilage and is
twisted inward and downward. It runs downward along the
inner surface of the cartilage and along the inner surface of
the articular in front of the cartilage, across the hind end of
the surface of insertion of Aw’, to the lower edge of that
muscle. There it crosses the anterior edge of ossicle ‘c,’
then turns inward between the articular and the lower hori-
zontal part of Meckel’s cartilage, and, having reached the
inner ventral angle of the cartilage, issues from the mandible
between the articular and the lower edge of the splenial, lying
immediately dorsal to and in contact with the r. mandibularis
facialis externus. It then turns upward and backward and
enters the thin fold of skin that extends from the inner, lower
edge of the mandible to the outer edge of the geniohyoideus
muscle, and which I have already described as the hyoideo-
mandibular fold. Traversing this fold upward and backward it
reaches the outer edge of the geniohyoideus near its origin
from the ceratohyal, and there turns downward and forward
along the outer surface of that muscle. It crosses the muscle,
reaching its median edge opposite the base of the third or
fourth branchiostegal ray, and passes under that ray onto the
outer, ventral surface of the inferior division of the hyohyoideus.
There it anastomoses completely with the r. hyoideus facialis,
the two nerves running directly into each other and forming a
complete circuit, so that it is impossible to tell where one ends
and the other begins. One long branch, apparently belonging
to the nerve gis, is sent forward along the outer surface of
the superior division of the geniohyoideus ; it sends branches
into that muscle, and runs directly into a terminal branch of
nerve 7,g/i, described below, forming with it a complete circuit,
similar to that formed by the main nerve and the r. hyoideus
facialis. Numerous other smaller branches are sent from the
united main nerves forward onto the outer surface of the supe-
rior division of the geniohyoideus, and forward, medianward,
and backward, internal to the branchiostegal rays onto the
outer surface of the hyohyoideus, being distributed mainly to
the inferior portion of that muscle.

614 ALETS, [Vou. XII.

Beyond its fifth branch, 7.g/s, and while still inside the
hollow of the mandible, the maxillaris inferior gives off several
branches, the arrangement of which differs in almost every
dissection. There is always one important branch, and often
two or more smaller ones, sent upward and forward along the
inner surface of the dentary, imbedded in the fibres of Ao!
at their insertion, toward the upper edge of the mandible; and
one important branch sent downward and forward. This last
branch may reunite with the main nerve just before the latter
issues from the hollow of the mandible, as was the case on both
sides of the head in the fish used for the drawings of the under
surface of the head.

Immediately after issuing from the hollow of the mandible,
the maxillaris inferior often sends one or more small anasto-
mosing branches (Fig. 44, Pl. XX XI) upward along the inner
surface of the splenial to the r. mandibularis facialis internus,
and then gives off an important branch, ~g/z, which, like ~ghs,
runs downward and then medianward and forward ventral to
the horizontal part of Meckel’s cartilage, between it and the
dentary, and dorsal to and in immediate contact with the r.
mandibularis facialis externus. It then turns upward and for-
ward, and having traversed the anterior end of the hyoideo-
mandibular fold reaches the outer surface of the inferior
division of the geniohyoideus, where it continues its course
medianward and forward, sending numerous branches forward
and backward along the outer surface of that muscle. One
large branch directed backward passes the hind edge of the
inferior division of the muscle, and, continuing along the outer
surface of the superior division of the muscle, anastomoses with
a branch of ~ghs as already described. A branch, 7.2, is
always sent forward and medianward to the outer, ventral sur-
face of the intermandibularis, which it apparently innervates.
McMurrich concluded (No. 76, p. 131) that both the inter-
mandibularis and the anterior division of the geniohyoideus
were innervated by the trigeminus.

Beyond the branch ~,g/z the maxillaris inferior, as it continues
forward to the symphysis, sends many branches upward and for-
ward in the dentary toward the upper edge of the mandible.

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 615

g. Truncus Hyoideo-Mandibularis Facialis.

The large truncus hyoideo-mandibularis facialis (4m) arises,
in 35 mm. specimens, from the ventral, outer, and hinder part of
the trigemino-facial ganglion. It runs forward and outward,
then outward, and finally outward and backward through the
large facial foramen which lies in the petrosal near the front
edge of that bone and a little behind the lateral wing of the
parasphenoid.

In the adult the truncus arises from the lateral surface of its
ganglion, nearly at right angles to that ganglion and to the line
of its posterior root. It runs directly outward, or outward and
backward, through its foramen, and then slightly upward along
the upper surface of the adductor hyomandibularis, where it
sends a large branch, the r. opercularis facialis (of7f) backward
and upward along the upper surface of the adductor hyoman-
dibularis, and beyond it onto the outer surfaces of the adductor
and levator operculi, all three muscles being supplied by the
nerve. After the nerve reaches the outer surface of the
adductor operculi it breaks up into numerous branches, which
spread out over the surfaces of that muscle and the levator
operculi, lying immediately internal to a similar network of
branches belonging to the glossopharyngeus and first vagus
nervus and destined to supply the cutaneous and subcutaneous
tissues of the operculum.

The r. opercularis facialis, on the left side of the fish used
for this dissection (Fig. 25, Pl. XXV), was easily separated
from the truncus facialis as a thin, flat layer closely applied to
the posterior surface of that nerve, the layer separating even
beyond the ganglion onto the main root of the nerve, and being
connected distally by fibres with the main truncus.

After giving off the r. opercularis the truncus turns down-
ward, and running downward, outward, and backward, somewhat
imbedded in the fibres of the adductor hyomandibularis, enters
and passes through the facial canal in the hyomandibular (/,
Figs. 1, 2, and 5, Pl. XX). After issuing from that canal the
truncus facialis continues downward, outward, and backward,
lying in the depression described on page 000 on the outer

616 ALLIS. [ior ll.

surface of the hyomandibular, and separates almost immediately
into its two main portions, the truncus mandibularis facialis (m/f)
and the r. hyoideus facialis (Zf). The former lies anterior to
the latter, turns downward, and soon separates into its two
portions the r. mandibularis externus facialis (#ef) and the r.
mandibularis internus facialis (sz/).

The r. mandibularis externus facialis supplies only the sense
organs of the operculo-mandibular canal and those of the pit
lines of the cheek and mandible (No. 3, p. 518). The first
branch of the nerve is given off before the main truncus has
separated into its mandibular and hyoidean portions. It arises
from the mandibular part of the truncus, while this truncus is
still inside the canal of the hyomandibular, or immediately after
it issues from that canal. It runs upward and backward, enters
the preoperculum, and then separates into two parts, one going
to organ 15 of the operculo-mandibular line, and the other to
organ 16. This branch and the others also that go to organs in
the preoperculum enter the bone along its inner, anterior edge.

The second branch of the main nerve, in 35 mm. specimens,
goes to the horizontal cheek-line of pit organs, and the third
one to organ 14 operculo-mandibular. In the adult these two
branches arise together, after the truncus has separated into
its hyoidean and mandibular portion, but before the latter has
separated into an externus and internus. The branch to the
cheek-line (#ef:h/) passes outward through the adductor man-
dibulae, near its origin from the preoperculum, and then runs
forward along the outer surface of that muscle; the branch to
organ 14 runs directly backward to that organ. In the adult
specimen used for illustration a small communicating branch
was sent from the branch to organs 15 and 16, to the one for
organ 14; and in one 35 mm. specimen a similar branch was
sent from the branch to the cheek-line to that for organ 14.
Whether these communicating branches belonged to the exter-
nus or were simply fibres of the internus, accompanying the
branches of the externus, it was not possible to determine. In
no other part of the lateral-line system were connecting branches
of the kind found, and it is to be noted that in both dissections
the connecting branch ran to organ 14 operculo-mandibular.

No. 3.] WUSCLES AND NERVES IN AMIA CALVA. 617

The next or third branch of the mandibularis externus, in
the adult used for illustration, was given off just before the
internus separated from that nerve. It ran almost directly
downward and entered the preoperculum opposite organ 13,
sending a branch to that organ and then continuing downward
in the bone to organs 12 and 11. Before entering the bone a
branch was sent outward and downward through the adductor
mandibulae and then outward and downward along the outer
surface of that muscle. It separated into two portions and
supplied the pit organs of the vertical cheek-line and mandib-
ular line, the branch to the latter crossing superficially the
branches of the fourth branch of the trigeminus, where they
turn inward at the lower corner of the adductor. In 35 mm.
specimens, and in most adult specimens also, there were two
branches arising from the externus after the internus had sepa-
rated from it, one branch going to organ 13 and one to organs
12 and 11, the branch to the cheek-line arising between the
other two, or as a branch of the one to organs 12 and 11.

After giving off the branch to organs 12 and 11 the externus
passes under the cartilaginous posterior corner of the palato-
quadrate arch, and then under the preoperculum into the canal
between that bone and the symplectic. From that canal it
issues at about the middle of the hind edge of the symplectic,
between that bone and the preoperculum, and, turning slightly
backward, passes external to the ligamentum mandibulo-hyoi-
deum, then downward and forward around the ventral surface
of that ligament into a groove on the median surface of ossicle
a.’ Tt then continues forward in that groove, and in a pro-
longation of the groove on the median surface of the ventral
thickened part of the articular, the nerve being here covered
internally by the lower edge of the splenial.

While passing downward along the external surface of the
ligamentum mandibulo-hyoideum a branch is sent forward to
organ 10 operculo-mandibular, the branch entering the articular
along its hind edge lateral to the articular process for the sym-
plectic. This branch seems to have been mistaken by van
Wijhe for the main r. mandibularis internus (No. 129, p. 292).
The next branch of the externus, that to organ No. 9, is given

618 ALIAS. VoL. XhE

off while the nerve lies in the groove on the median surface of
the articular. The nerve then comes to the upper surface of
the thickened lower edge of the articular at the point where
the branch ~ghs of the maxillaris inferior trigemini issues .
from the hollow of the mandible. In that position, lying im-
mediately ventral to the median edge of Meckel’s cartilage, it
runs forward, sending branches to organs 8, 7, 6, 5, and 4.
Beyond this last branch, at the point where the maxillaris
inferior trigemini issues from the hollow of the mandible, the
externus issues from beneath Meckel’s cartilage, and, passing
immediately ventral to r. ghi trigemini, continues forward along
the inner surface of the dentary, sending branches to organs 3,
2, and a.

The r. mandibularis facialis internus enters the canal between
the preoperculum and symplectic with the externus, lying in
front of that nerve, but it issues along the anterior instead of
along the posterior edge of the latter bone, between it and the
posterior edge of the quadrate, and reaches the inner surface of
the palato-quadrate arch. It then turns forward, crosses inter-
nally the lower end of the quadrate, enters the mandible oppo-
site thé articular cup of the quadrate, and, lying between the
splenial and the vertical part of Meckel’s cartilage, near the
upper edge of the latter, continues forward to the tip of the
mandible. It lies dorsal to the origins of the geniohyoideus
inferior and intermandibularis, and may receive, near the hind
edge of the former, an anastomosing branch or branches from
the maxillaris inferior trigemini, as already described.

The r. hyoideus facialis, after leaving the main truncus facia-
lis, runs downward and backward along the outer surface of the
hyomandibularis, passes through a canal between that bone and
the preoperculum, and issues on the inner surface of the gill-
cover beyond the hind edge of the cartilaginous interspace that
lies between the ventral edge of the hyomandibular and the
dorsal edge of the symplectic. The opening of the canal lies
immediately ventral to the ventro-posterior process, or corner,
of the hyomandibular, and immediately dorsal to the articular
cup for the hyoid, and the nerve, continuing downward and
backward internal to the suboperculum and interoperculum,

No. 3.] WZUSCLES AND NERVES IN AMIA CALVA. 619

passes dorsal to and behind the interhyal, dorsal to or across
the upper end of the ceratohyal, and then downward and for-
ward along the inner surface of the branchiostegal rays, lying
immediately behind the posterior edge of the ceratohyal. On
reaching the fourth or third ray from in front, it anastomoses
completely with the r. ghs trigemini as already described.
During its course along the inner surface of the gill-cover a
large branch is sent backward along the lower edge of the
operculum, another along the lower edge of the sub-operculum,
and others, irregularly distributed, along the inner surface of
the interoperculum, the inner surface of the branchiostegal
rays, and the outer surface of both divisions of the hyohyoi-
deus, which muscle they supply.

The r. palatinus facialis (ff) arises from the ventral surface
of the trigemino-facial ganglion toward its median edge. It
runs downward and forward in the upper, lateral chamber of
the eye-muscle canal, and, passing through the palatine foramen,
enters the palatine canal between the parasphenoid and the
ventral surface of the cartilaginous cranium. There it turns
forward and separates into two parts, the r. palatinus anterior
facialis (faf) and the r. palatinus posterior facialis (pff). The
latter turns outward and forward through the foramen palatini
posterioris facialis, immediately in front of the anterior edge of
the lateral wing of the parasphenoid, passes immediately ven-
tral to the efferent pseudobranchial artery, crosses the space
between the lateral edge of the parasphenoid and the median,
dorsal edge of the palato-quadrate arch and reaches the dorsal
surface of the entopterygoid (Z/VP, Fig. 2, Pl. XX). Along
the outer, dorsal surface of that bone it runs downward, out-
ward, and forward to the edge of the cartilaginous palato-quad-
rate (fg), under which it passes, then under the posterior edge
of what van Wijhe calls the autopalatinum (4 UP), and onward
to the lateral edge of that bone, either coming to the upper
surface of the bone through a foramen near its hind edge and
then lying on that upper surface, or continuing downward to
the edge of the bone between it and the dermopalatinum (DP),
The nerve then passes beyond the lateral edge of the palatine,
and lying in the thick upper lip, turns forward, sends anasto-

620 ALLIS. (Vou. XII.

mosing branches to the r. maxillaris superior trigemini, passes
under the articular head of the maxillary, and is distributed to
the lip and to tissues ventral to the premaxillary. During its
course several branches are sent forward and backward along
the inner surface of the palato-quadrate arch, the branches,
like the main nerve, lying dorsal and external to the entoptery-
goid and the other dermal bones of the inner surface of the
arch, but ventral and median to the cartilage of the arch and
the bones formed in that cartilage. The main nerve may
come to the upper surface of the autopalatinum, and hence to
the upper or external surface of the arch, near the lateral edge
of the autopalatinum, and hence of the arch, but it is for a
very short distance only, and the branch or branches given
off during this part of its course reénter the autopalatinum
from its upper surface. In a 40mm. specimen in which the
nerve was traced it lay between the cartilage of the arch and
the forming dermopalatinum ; in a 44mm. specimen it pierced
the cartilage near its edge. In one specimen it was double on
both sides of the head as it issued from its foramen, the two
branches or parts soon uniting again.

The r. palatinus anterior facialis (paf) after its separation
from the posterior nerve continues forward in the palatine canal
accompanied by the r. pharyngeus glossopharyngei and a branch
of the internal carotid artery. It soon gives off a large branch,
which comes to the outer surface of the base of the skull and
runs forward along the edge of the skull, ventral to the vomer,
toward the end of the head. Somewhat further forward the
glossopharyngeus leaves the palatine canal, and with or with-
out a branch of the facialis, comes to the ventral surface of the
vomer and continues forward in that position toward the end
of that bone, lying median to the first branch of the facialis.
The remaining main portion of the palatinus continues forward
in the vomer, or between it and the chondrocranium, sending
one important branch, if it be not the main nerve itself, upward
through the base of the skull in a relatively long canal which
issues at the median edge of the septomaxillary, between it
and the median edge of the olfactory perforation of the pre-
maxillary, as already described. This last branch anastomoses

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 621

with the terminal branches of the r. maxillaris superior trige-
mini, and sends branches along the ventral and along the
dorsal surface of the premaxillary, one branch of the nerve
being traced into one of the large teeth of that bone.

6. Har, Nervus Acusticus, and Nervus Lineae Lateralis Vagi.
a. Far.

In 14mm. larvae of Amia the ear (Fig. 57, Pl. XXXV, of
adult ear) already shows the well-known divisions of the teleos-
tean ear. The utricular chamber is large, and has a central, an
anterior, a posterior, and a vertical portion. The two former,
together, are the utriculus proper, the two latter the sinus
utriculi posterior and sinus utriculi superior respectively of
Wiedersheim’s figures (No. 128, p. 351). The utriculus and
the sinus posterior lie in vertical planes approximately at right
angles to each other, and at 45° to a vertical longitudinal plane
of the body. From the anterior, outer end of the utriculus the
canalis anterior (csa) arises, the lower end of the canal being
already slightly differentiated as the ampulla anterior (aa).
The crista acustica amp. ant. extends transversely across the
lower anterior wall of this ampulla.

Immediately proximal to the ampulla anterior, from the outer
side of the utriculus, the canalis externus (cse) arises, running
at first almost directly backward, and then turning sharply
inward and forward, against the anterior surface of the sinus
posterior, to join the lateral surface of the central portion of
the utricular chamber, that is, the posterior portion of the
utriculus proper. The ampulla externa (ae) is slightly indi-
cated, the crista acustica amp. ext. forming a transverse, ver-
tical, ridge-like projection at the boundary line between it and
the canal. The crista lies nearly in the middle of the lateral
portion of the canal.

The sinus posterior runs outward and backward at a little
more than 45° to the axis of the body, and lies close against
the inner and posterior wall of the posterior portion of the
canalis externus. From its outer end the canalis posterior
(csp) arises, the ampulla posterior (af) being slightly indicated

622 ALLIS. [VoEn.AXie

at its lower end, and containing, as a transverse ridge across
its lower surface, the crista acustica amp. post.

The sinus superior arises from the central portion of the
utricular chamber, median to the opening of the canalis
externus, runs directly upward, and receives at its upper end
the median, upper ends of the anterior and posterior canals.

The sacculus (sc) begins immediately under the central por-
tion of the utricular chamber, connected directly with the latter
by a large opening without the intervention of what could be
properly called a canalis utriculo-saccularis. The front wall of
the sacculus is nearly vertical, and lies on the level of the ante-
rior edge of the opening connecting it with the utriculus.
Posteriorly, on the contrary, the sacculus extends directly
backward a considerable distance, at first immediately under
the sinus posterior, and then immediately below but median to
that sinus. Its upper wall lies close to the ventral wall of the
utriculus, and there is, as yet, no marked constriction indicating
that this portion is the future lagena. On the inner and upper
wall of this portion, immediately behind the posterior edge of
the opening connecting the sacculus with the utriculus, is the
large papilla acustica lagenae, and immediately above it, in the
ventral and median wall of the sinus utriculi posterior, the
smaller macula fundi utriculi, or macula acustica neglecta,
whichever it may be. From Wiedersheim’s description and
figures these two maculae seem to occupy exactly the same
position in fishes and reptiles, and to be apparently identical
(No. 128, p. 347). In Amia the organ is connected by a line
of thickened epithelium with the papilla lagenae.

The macula acustica sacculi lies in the inner and anterior
wall of the sacculus, and the macula acustica utriculi in the
ventral and slightly lateral wall of the utriculus in front of the
opening between the utriculus and sacculus.

In the adult, as the figure shows, the main divisions of the
ear differ but little from those of larvae. The cristae and
maculae acusticae have, however, acquired somewhat different
positions, but no special description of them seems necessary.

Neither in the adult nor in any of the larvae did I find any
satisfactory indication of a ductus or saccus endolymphaticus.

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 623

Why these structures, found so generally in fishes, should be
wanting in Amia I cannot imagine. I can only suppose that
my sections were imperfect, so far as the ear is concerned, or
that the ear not belonging especially to my subject, and but
little attention having been given to it, I failed to recognize
the endolymphatic parts of it. The membranous parts of the
ear were often much collapsed and distorted in my sections,
but in the adult they seemed perfect.

Along the median surface of the sinus superior, in the adult,
there was always a broad band of delicate, strongly pigmented
tissue, running upward from about where the ductus endo-
lymphaticus would naturally begin. It was always closely
attached by its edges to the wall of the sinus, and in that wall
there was a white line, due apparently to a thickening of the
tissue of the wall.

b. Nervus Acusticus.

The nervus acusticus, in young larvae, arises from the summit
of the lobus acusticus by three distinct roots, two of them
often so fused that it is difficult to distinguish them in sec-
tions. In different specimens they seem, as they issue, to
have slightly different positions relative to each other. In one
14mm. specimen, cut in horizontal sections, they were particu-
larly distinct, one root lying slightly above the other two,
which two were, however, cut in the same section. They all
had their apparent origins immediately below the utriculus
and immediately in front of the sacculus, and each of them
entered a definite part of the large irregular acoustic ganglion.
The upper root, which is the main posterior root of the nervus,
gave origin to the ramus cochlearis, which turned directly back-
ward internal to the sacculus, supplied at once the macula
sacculi, then the papilla lagenae and macula fundi utriculi by
branches arising close together, and then, turning outward and
backward above the lagena, between it and the ventral surface
of the sinus utriculi posterior, and always separating into two
terminal branches, supplied the crista ampullae posterioris.

The other two roots together formed the main anterior root
of the nervus and, together, gave origin to the ramus vestibu-

624 ALLIS. [Vou. XII.

laris. The fibres of the anterior of the two roots ran forward
and outward, and supplied first the macula utriculi, which lay
immediately above the ganglion of the root, and then the crista
ampullae anterioris. The fibres of the posterior of the two
roots ran outward and slightly forward and supplied the crista
ampullae externae.

The main posterior root, or ramus cochlearis, and its gang-
lion were always much more distinctly separated from the other
two roots and their ganglia than those two roots and ganglia
were from each other.

As the ramus ampullae posterioris passes outward and back-
ward between the lagena and the sinus posterior it passes
immediately above the root of the nervus glossopharyngeus,
which runs at this point almost directly outward. The two
nerves here come into intimate relations and there is always
an interchange of fibres, the exact nature of which I have been
unable to satisfactorily determine. To the best of my observa-
tion some of the fibres of the acusticus join the dorsal or sense-
organ root of the glossopharyngeus, but whether they run
distally or proximally in that root I am unable to say. Distal
to this point the dorsal root of the glossopharyngeus is sepa-
rate and distinct from the remaining much larger portion of
the root of that nerve, lying immediately behind it and having
an entirely separate ganglion, as already described in my earlier
work (No. 3, p. 516). While passing inward under the acusti-
cus, through the narrow space between the lagena and utriculus,
the two roots of the glossopharyngeus are much flattened, and
the large fibres of the dorsal root were always lost in my sec-
tions. Median to the acusticus, after passing under it, they,
however, reappear in the sections and receive an important bun-
dle of fibres from the acusticus, the bundle being apparently as
large as the dorsal root of the glossopharyngeus itself. Proxi-
mal to this anastomosis the dorsal root still persists, lying in its
accustomed position immediately behind and in contact with
the main root of the glossopharyngeus. This latter root, in a
14 mm. and 20 mm. specimen of which I had excellent sections,
then passed through the root of the nervus lineae lateralis,
near its ventral edge, and there was again an interchange of

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 625

sense-organ fibres. In the 35 mm. and 40 mm. specimens
examined in my earlier work the ventral portion of the root of
the nervus lineae lateralis seemed to pass through the dorsal
part of the main root of the glossopharyngeus, receiving there
an important addition to its fibres (No. 3, p. 517). In all the
adult specimens that I have examined such is also the apparent
relation and arrangement of the nerves, for I have never found
a ventral bundle of the nervus lateralis lying ventral to the root
of the glossopharyngeus. I therefore conclude from my pres-
ent work that the relations of the nerves must vary in different
specimens, or at different ages. I am also convinced that it is
the so-called dorsal root of the glossopharyngeus that receives
fibres from the nerve of the lateral line, and not the reverse, as
I formerly supposed. Whether that dorsal root is formed
entirely of fibres derived first from the root of the nerve of
the lateral line and later from the ramulus ampullae posterioris
acustici or not, I have been unable to positively determine.
Such seems to me, however, to be the case. As the communi-
cating branch between the acusticus and the dorsal root of the
glossopharyngeus lies proximal to the crista amp. post., there
is thus something in Amia to support the view that the nerve
to the posterior ampullae is sometimes derived from the glosso-
pharyngeal nerve (No. 5, p. 232).

In the adult, in the four specimens examined after my atten-
tion was called to the matter, two small anastomosing branches
were always found, one running proximally from the posterior
surface of the glossopharyngeus distally into the outer and
lower surface of the nervus lineae lateralis, and the other
distally from the anterior surface of the glossopharyngeus
proximally into the inner surface of the ramulus amp. post.
acustici. The nature of these branches could not be deter-
mined from macroscopic examination.

c. Nervus Lineae Lateralis Vagi.

This nerve in 14 mm. specimens arises close to the root of
the glossopharyngeus, slightly above and in front of that root
and slightly behind the root of the acusticus. It seems to

626 ALLIS. [Vou. XII.

arise, however, immediately behind the slight swelling of the
tuberculum acusticum, and not from that swelling, as Strong
states to be the case in all fishes (No. 121, p. 179). It runs
almost directly backward, as described in my earlier work, an-
astomoses with the root of the glossopharyngeus, as described
above, and then continues backward, lying at a higher level
than the ganglion on the posterior root of the acusticus, and
at a lower level than the issuing roots of the vagus. On these
latter roots an intracranial ganglion is formed, the root of the
lateralis lying along its lower, outer surface. The lateralis
then issues with the vagus through the vagus foramen, enters
its ganglion and has the distribution already given (No. 3,
p. 517). In the 14 mm. and 20 mm. larvae in which a ventral
bundle of the root was found passing ventral to the root of the
glossopharyngeus it could be distinguished as a separate bundle
almost to the exit of the nerve through its foramen.

7. Review and Comparison of Nerves.

There are, as is well known, several kinds of more or less
specialized sense-organs of ectodermal origin, found in greater
or less number in all vertebrates. In the Ichthyopsida, where
these organs are found in the greatest number and variety, the
innervation of certain of them in certain forms is now definitely
known, that of others is still largely unknown, as is also the
development and interrelationships of all the organs. Some
intimation of the relationships of certain of them can, however,
be deduced from their innervation as already known, and the
innervation, and hence relationship, of others is in a measure
indicated by the relative size and extension of certain nerves,
or their branches, in forms where certain of the organs undergo
unusual development.

In Amia three kinds or varieties of these organs are found,
namely, canal organs, pit organs, and terminal buds, —the organs
of the eye, the ear, and the nose being left out of consideration.
The canal organs are always found, in the adult, inclosed in
the canals of the lateral line system, and the pit organs at the
bottom of slight pit-like depressions that are arranged in lines

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 627

having definite and constant relations to those canals or canal
organs. The terminal buds are, on the contrary, found scat-
tered irregularly over the external surface, mainly of the head
and fins, and on the inner surface of the mouth and branchial
cavities (No. 3).

The canal organs and pit organs in Amia are all innervated
by what are called the dorsal, suprabranchial, or lateral branches
of the facialis, glossopharyngeus, and vagus nerves, and they all
belong to what are called by Merkel nerve hillocks. In early
stages of development they closely resemble each other (No. 3,
p. 502), but in later stages they differ greatly, the pit organ
retaining its embryological form and individuality, while the
canal organ first increases greatly in size, and then, by the
independent growth of other similar organs immediately adjoin-
ing it, gives rise to a large sensory patch, or nerve ridge (Mer-
kel), in which the separate organs lose to a greater or less extent
their individuality.

Portions of the canal lines in Amia are represented in cer-
tain teleosts by lines of surface organs, and certain of the pit
lines in Amia are apparently represented in certain teleosts
and bony ganoids by canals containing organs. To the exam-
ples already given of this in my earlier work (No. 3, pp. 471
and 521) must probably be added the mucous canal at the base
of the dorsal fin in Polypterus, mentioned, but not described,
by Pollard in his work on the lateral line system in siluroids.
He there states (No. 94, p. 545) that this canal is represented
in Amia by the dorsal body line of pit organs. As I do not
find a description of the canal and its innervation, I am unable
to judge whether this be so or not. The canal may, however,
be of the kind described by Emery as the accessory lateral
lines of Fierasfer (No. 25, p. 209), and hence not comparable
at all with the canals of the lateral line system as found in
Amia and teleosts. As there is in Polypterus a supratemporal
crosscommissure, the canal in question cannot be that canal, as
is probably the case with the canal described by Pollard in the
same place, that is along the base of the dorsal fin, in Clarias
and Auchenapsis (No. 94, pp, 530 and 533). In these two
fishes the innervation of the canal is given as by a great recur-

628 ALLIS. MViOL. XT.

rent dorsal branch of the facialis; but that recurrent branch is
joined, in both fishes, by an important branch from the first
branch of the lateralis vagi, and as this branch of the lateralis
vagi supplies organ No. 8 of the main lateral canal, in both
fishes, it is in all probability this branch, and not the recurrent
facialis, that supplies the organs of the canal in question.
Organ No. 8 les in the squamosal near its hind edge,
and as there is, according to Pollard, no extrascapular and
no supratemporal crosscommissure in siluroids, the organ No.
8 of Clarias and Auchenapsis probably corresponds to organs
18 and 19 of the main canal in Amia. The commissural or
recurrent branch that connects the nerve supplying it with the
recurrent dorsal branch of the facialis would, therefore, corre-
spond to the supratemporal extension of the nerve supplying
organs 18 and I9 in Amia. The canal along the base of the
dorsal fin in Clarias and Auchenapsis would then be the supra-
temporal crosscommissure of Amia, somewhat changed in posi-
tion and direction, a change that could easily arise since the
bone traversed by the canal in the one form is wanting in the
other.

The great recurrent branch of the facialis described by Pol-
lard seems to be the ramus lateralis trigemini of other authors.
In Silurus glanis this nerve is distributed to the dorsal fin, as
in Clarias and Auchenapsis. In Trichomycterus and Chaetos-
tomus a large branch is sent from it to the pectoral fin (No.
94, pp. 536 and 539), and in Gadus morrhua I find this branch
distributed to the breast fin also. In elasmobranchs and
Amphibia the nerve is wanting, so far as I can find. Its dis-
tribution indicates that it is destined largely or entirely to the
supply of terminal buds, for these buds are not found on the
body in elasmobranchs and Amphibia, and are found in great
quantity, but with a greatly varied distribution, on the body,
and more especially on the fins, in teleosts. The nerve in
Gadus lies immediately beneath the skin, but crosses the later-
alis vagi internal to that nerve. It arises as two nerves or
bundles from the deeper portions of the trigemino-facial gan-
glion, the two bundles embracing the root of the buccal and
ophthalmic branches of the facialis exactly as the first pair of

No.3.) J7USCLES AND NERVES IN AMIA CALYVA. 629

branches of the ophthalmicus superficialis trigemini in Amia
embrace the ophthalmicus facialis. The nerve in Gadus has,
contrary to the arrangement of the branches found in Amia,
an intracranial course, as it has in siluroids, issuing on the
top of the skull near its hind end. As the trigemino-facial
ganglionic mass lies, in Clarias at least (No. 94, p. §29), inside
the skull beside the brain, this difference of course is probably
of no importance.

In Clarias Pollard states that tube § of the infraorbital or
main canal “must be regarded as the last rudiment of a canal
represented in Amia by the vertical line of pit organs”’; that
tube 6 ‘appears also to be the rudiment of a canal represented
in Amia by the horizontal line of pit organs’’; and that tube 8
‘corresponds to the rudimentary canal represented in Amia by
the dorsal line of pit organs.” These conclusions are certainly
wrong, for the canals of all the lateral lines are always formed
in sections, each section inclosing an organ, and the tubes lead-
ing to the surface from the developed canals always lie between
two adjacent organs, and represent simply the fused open ends
of two such sections (No. 3, p. 523). A sense organ is pri-
marily never found in such a tube, and the tube can in no way
represent a line of surface organs. The terminal tubes of a
canal line may lie in the direction of such a line of surface
organs, and may even inclose secondarily one of those organs,
as is the case with tube 8 supraorbital in Amia (No. 3, p. 506).
The tube has, however, in its development nothing whatever to
do with the line and cannot represent it. A line of pit organs
represents apparently the possibility of a canal; with the dis-
appearance of the organs the possibility of the canal, even in
rudiment, would certainly disappear.

Pollard did not find any lines of pit organs in siluroids, and
Collinge (No. 20) does not mention any. I, however, find
them both in Amiurus catus and in Silurus glanis. In both
these fishes there is an anterior crosscommissural line over the
top of the end of the snout, and, two or possibly three lines on
the top of the head; and in Silurus there are three lines on
the side of the head. One of the lines on the top of the head
is a direct continuation of the supraorbital canal, and in larvae

630 ALLIS. [Vou. XII.

of Amiurus, as in Amia, it is a direct continuation of the line
of the supraorbital canal organs. Those on the side of the
head in Silurus are one vertical, one horizontal, and one on the
hind end of the mandible; the lines do not, however, have the
same positions relatively to each other that they have in Amia.

The arrangement of the canals in larvae of Amiurus shows
beyond question that Pollard’s pore 6 infraorbital in Clarias is
formed by the fusion of pores 6 infraorbital and 6 supraorbital,
his pore 6 supraorbital being therefore in reality pore 7 of that
line, and the terminal pore of the line. That it is the terminal
pore of the line is shown sufficiently by the presence of a
sense organ in what Pollard considers as its tube, and by the
absence of such an organ in what he considers as the continua-
tion of the main canal. Pore 7 of the main canal in Clarias
has fused with the terminal pore of the operculo-mandibular
canal and is found on that line, but not shown in Pollard’s
figure, and pore 7 in his figure is in reality pore 8. The same
is true of Auchenapsis and Chaetostomus. In each of these
fishes the last pore and tube of the ascending portion of the
suborbital canal, behind the eye, is a double system, formed by
the fusion of a suborbital and supraorbital pore and tube; and
the last pore and tube of the operculo-mandibular canal is also
a double system, formed by the fusion of a tube and pore of
the main line with the terminal pore and tube of the operculo-
mandibular line.

The sense organs of all the canals in all the fishes described
by Pollard are, therefore, with a single exception, regularly
placed one between each two consecutive tubes of the line to
which it belongs, instead of being irregularly placed, as the
figures and numbering used by Pollard would indicate. The
single exception, or apparent exception, is in the operculo-
mandibular line in Chaetostomus, where one organ seems to
be wanting, but the development, if it were known, would
undoubtedly show that the irregularity even here is only
apparent. In Chaetostomus the operculo-mandibular canal
turns backward and downward into the interoperculum and con-
tains an organ in that bone. This may be another instance of
a pit line in one form becoming a canal in another, for in Gadus

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 631

there is a line of surface organs on the outer surface of the
opercular bones immediately behind the preoperculum. They
are all innervated bya special branch of the mandibularis exter-
nus facialis, and are therefore of the character of pit organs, and
unquestionably represent in Gadus one of the cheek-lines in
Amia, or a similar line not found in Amia. One of these
organs has possibly become inclosed in a canal or tube in
Chaetostomus. The organ may, however, belong to the main
canal line, for Collinge states (No. 20, p. 277) that the
operculo-mandibular canal in Clarias nienhofii traverses the
inter-operculum. Pollard, unfortunately, does not describe
the operculo-mandibular canal in Clarias. Collinge’s state-
ment cannot therefore be controlled, and his statements seem
sometimes to need considerable control or confirmation.

In Esox lucius, for example, Collinge says (No. 20, p. 288),
“there is a distinct and very large lateral canal’ extending
the full length of the body. I find only a lymph canal lying
under the dermis. Other investigators have doubtless only
found this same canal, as, according to Collinge’s own state-
ment, they find no lateral canal. In Salmo salar he describes
(No. 20, p. 296) ‘‘a series of functionless canals passing through
the substance of the bone”’ and not communicating with the
surface ; and a series “‘of small, drainpipe-like ossifications ”’
from which have arisen a “‘more superficial series of canals,”’
“which gradually replaced altogether those traversing the
deeply seated cranial elements.” In Salmo namaycush I find
the canals traversing the same bones they do in Amia, except-
ing only the antorbital and ethmoid. The extrascapular of
Amia is, however, represented in Salmo namaycush by a series
of small, tube-like bones, such as Collinge describes in Salmo
salar, and there is a small bone of a similar character between
the upper end of the preoperculum and the lateral edge of the
squamosal. As the canals in Coregonus (white fish) also lie
in the main cranial bones, Collinge’s statement regarding
Salmo salar seems most exceptional, the process of develop-
ment he assumes still more so. Why organs once found in
canals resembling those of other fishes should afterwards
migrate into a secondary, superficial series of canals, certainly

632 ALLIS. [Vor. XII.

needs some explanation. Collinge further repeatedly states
that the rami buccalis, oticus, and ophthalmicus superficialis are
branches of the trigeminus, and then, based on this assertion,
arrives at the conclusion (No. 19, p. 515) that, “The fact of
the trigeminal actually innervating a part of the sensory canal
system is of special interest.” Had he assigned the nerves to
the facialis, or to the lateralis, as is usually done, the recorded
facts would lose their special interest.

Collinge, in referring to my earlier work, corrects a state-
ment made by me regarding the operculo-mandibular canal in
Amiurus catus. As I find the canal in Silurus glanis as he
gives it in Amiurus, he is undoubtedly correct. In further
reference to my work, he says (No. 109, p. 504) that I gave the
name ‘peripheral organs”’ to “a system of fine dermal canals”’
“opening by a series of fine branches to the surface by isolated
pores.” This is as evidently an error. For these canals and
their openings in Polyodon, Collinge proposes the name “ clus-
ter pores: Water aim the text (No. 10, p).513),;cluster pores
are said to be a “series of organs” agreeing “almost in every
detail with the Spaltpapillen of Fritsch,” and containing, like
them, sense cells, into which fine terminal nerve fibres pass.
«Primitive pores” in Polyodon are said to be “fine pore-like
openings, spoken of as pin-hole pores by many authors.” These
primitive pores connect with the small branches of the main
canals (No. 19, p. 509), apparently as their external openings,
but each pore contains a sensory organ at its base (No. 19,
p. 514), and in section (No. 19, Fig. 5) closely resembles a
pit organ of Amia. As the branches of the dendritic canal
systems open on the external surface by these “pores,” each
of which contains an organ, and as no organs of any kind are
definitely described at any place in the canals themselves, it is
evident that Polyodon, as the author states, “exhibits a number
of interesting features at present not known to occur in any
other ganoid.”

In Gadus a line of surface organs is found along the lower
edge of the mandible, parallel to the mandibular canal, and
it is innervated by a long branch of the externus facialis,
which first runs forward through the adductor mandibulae, to

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 633

the hind edge of the infraorbital canal behind the eye, and
there turns downward and reaches the mandible. A nerve in
Esox corresponding in position to this nerve in Gadus inner-
vates a line of surface organs lying on the upper jaw imme-
diately below the infraorbital canal. In Polyodon the maxillary
branch of the hyomandibular canal (No. 19, p. 512) may repre-
sent this line.

Although certain portions of the canal lines of Amia are thus
represented in teleosts by lines of surface organs, and vice versa,
I find nothing in the literature at my disposal that tells defi-
nitely whether the organs in such cases change in character ;
that is, whether the organs found on the external surface in
teleosts, in lines that are represented in Amia by canals, have
the character of canal organs, or of pit organs, or of some inter-
mediate or different form of organ. Indirectly one is led to
conclude that they retain or acquire an intermediate character,
for Leydig includes them with pit organs and terminal buds,
under the general term “ Becherorgane.”’ He also states, defi-
nitely (No. 74, p. 125), that “freie Sinneshiigel zu solchen
werden kénnen welche in Schuppencandlen liegen”’ (No.
74, p. 90), that the “ Becherorgane”’ tend to differentiate into
several distinct groups, and that this differentiation, or tendency
to differentiate, is more strongly marked in Amphibia and in
reptiles than in fishes. He, however, places together, under the
general term “lateral organs,” the canal organs, and the canal
organs only, in fishes, and the surface or exposed nerve hillocks
of the lateral lines in Amphibia (No. 74, p. 106), thus making
a distinction between these exposed lateral organs in Amphibia
and similar ones found in fishes. He says, however, as does
Wiedersheim (No. 128, p. 298), that between the several kinds
of organs, canal organs included, there are so many interme-
diate forms that it is impossible to draw a sharp limiting line
between any of them.

Wiedersheim says (No. 113, p. 298) that nerve hillocks, in
their development, pass through the stage represented by the
terminal bud, and that the latter is, therefore, the oldest phylo-
genetically of all these organs. That being so, what has been
established for the development of nerve hillocks and the nerves

634 ALLIS. [Vot. XII.

innervating them, must be, in general, true for the terminal buds
and the nerves that innervate them. The nerve hillocks, or
sense organs of the lateral line, are said by Beard (No. 10,
p. 209) and Wilson (No. 131, p. 244) to arise separately and
independently along lines of sensory epithelium that either
differentiate in one or more directions from certain central
points, or grow directly from such points by cell division. From
the deeper layers of this sensory epithelium the nerve supply-
ing the organs of the line arises. Terminal buds and the nerves
innervating them should therefore arise in this same way.

The only nerves in Amia from which I have been able to
trace branches definitely to terminal buds are the ophthalmicus
superficialis trigemini and the maxillaris inferior trigemini.
The former of these two nerves in 14 mm. specimens derives
the larger part, if not all, of its fibres from the median part of
the trigemino-facial ganglion, that is, from that part of the
ganglion that is formed on, or in connection with, the fascicu-
lus communis root. From this same part of the ganglion a
large bundle of fibres is sent to the truncus maxillaris trigem-
ini; from it arises also the ramus palatinus facialis, which is
distributed in Amia to a region covered with terminal buds,
and which in Rana innervates such buds (No. 121, pp. 121 and
123); and from it also a bundle of fibres is possibly sent to the
truncus hyoideo-mandibularis facialis, as is said to be the case
in Rana and Amblystoma, in which animals it gives origin,
according to Strong, to the ramus mandibularis internus fac-
ialis, which nerve in Rana innervates terminal buds (No. 121,
pp. 130, 132, and 195). I was unable to definitely trace this
bundle of fibres in Amia into the truncus hyoideo-mandibularis,
and the mandibularis internus in larvae of Amia, through the
full length of its course, never passes near a region where ter-
minal buds are numerous. In the lining membrane of the
mouth cavity, internal to the hyomandibular, and hence not
far from the truncus hyoideo-mandibularis, there are many
buds, but the pretrematic branch of the glossopharyngeus runs
near them, and branches of that nerve are found extending
toward them. It is therefore, in all probability, that nerve
and not the mandibularis internus that innervates them.

No. 3.]} MUSCLES AND NERVES IN AMIA CALVA. 635

Terminal buds in Amia are just beginning to appear on the
outer surface of the head in 10 mm. larvae. Their distribution
is shown in Figs. 4 and 5 of my earlier work; Figs. 6 and 7
showing a slightly more advanced condition. It will be noticed
that the organs are, in many places, arranged in lines, and that
well-marked lines are found on each side of the line of the supra-
orbital canal line, and on one side only of the infraorbital line,
that is, on its dorsal side, between it and the eye. Other lines
of organs are found on the maxilla and at its base, below and
in front of the line of the infraorbital canal. If, now, the rami
-ophthalmicus superficialis and maxillaris superior trigemini of
the adult be considered, it is seen that branches arise from the
ophthalmicus trigemini in pairs, one branch issuing on the
median side and one on the lateral side of the ophthalmicus
facialis ; that on the maxillaris superior all the small branches
below and behind the eye, excepting only the accessory branch
‘““r.c,”’ issue on the dorsal or median side of the buccalis, and that
the one large branch, the one destined to supply the maxilla,
issues on the ventral side of that nerve. The branches of these
two trigeminal nerves in the adult thus have the same relations
to the nerves innervating the canal lines that the lines of termi-
nal buds in 10 mm. larvae have to the lines of the canals. As
the trigeminal nerves both lie deeper than the corresponding
facial nerves, and as their branches issue on both sides of the
latter, it is evident that*they must first have been split off from
the ectoderm before the facial nerves, split off apparently from
exactly the same lines, could have arisen. This is exactly what
Platt finds for the ophthalmic nerves in Necturnus (No. 91, p.
949).

The branch of the trigeminus that goes to the maxilla in
Amia is the branch which, according to Pollard, furnishes the
sensory supply of the maxillary tentacle in siluroids, and a part
of that of the coronoid tentacle. Where there is but a single
tentacle at the angle of the mouth, the maxillo-coronoid tentacle,
the larger part of its sensory supply is from this same nerve.
The nerve is called by him the ramus maxillaris. His ramus
premaxillaris is represented in Amia by those branches of the
maxillary nerve of Amia that continue forward toward the pre-

636 ALLIS. [VOL xi.

maxillary, instead of turning backward into the maxilla ; and his
ramus palatinus trigemini (No. 97, p. 400) is either what I have
considered as the main terminal portion of the maxillaris superior
trigemini, or that nerve plus the palatinus posterior facialis.
The name and distribution given by Pollard seem to indicate
that it is the two nerves fused, but as he does not describe the
branches of the nervus facialis one cannot be sure. In the fishes
described by Pollard the anterior end of the palato-quadrate is
found as a distinct and separate piece, which is called by him the
prepalatine, and is said to be the homologue of the autopalatine
of van Wijhe in Polypterus (No. 97, p. 403). The autopalatine
is found in Amia also (AUP, Fig. 2, Pl. XX), and the ramus pal-
atinus posterior facialis runs under it, as Pollard says his pala-
tinus trigemini does in Misgurnus. This nerve in Misgurnus
is thus certainly the r. palatinus posterior facialis of Amia. As
this nerve and the r. maxillaris superior trigemini, in Amia, are
both sensory nerves distributed to the same or to adjoining
regions, and as they partially anastomose in Amia at the prox-
imal end of the maxillary bone, there seems no reason why
they might not fuse entirely if the bone or cartilage that nor-
mally separates them should disappear, or become reduced and
displaced, as it has in siluroids. That the palatinus of Mis-
gurnus, lying below the palato-quadrate cartilage should be
the homologue of a nerve lying above that cartilage seems
wholly inadmissible.

In addition to the terminal buds described above there are,
in 10 mm. larvae, indications of the beginning of similar organs
above organ 16 infraorbital and on the side of the cheek above
the horizontal line of pit organs (No. 3, Fig. 4). InatIi mm.
larva (No. 3, Fig. 6) the organs on the cheek have increased,
and those above the horizontal pit line are seen to be arranged
in horizontal lines. Buds have also appeared on the mandible
on both sides of the mandibular canal line, and a spot on the
operculum at its upper anterior end indicates the beginning of
similar organs there. In 14 mm. specimens (No. 3, Fig. 8)
the buds have increased in number, those on the cheek below
the horizontal pit line being arranged in lines that run back-
ward, or downward, or downward and forward. In 18 mm.

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 637

larvae (No. 3, Fig. 12) the buds have increased at the upper end
of the operculum, are beginning to appear along the branchio-
stegal rays, spreading backward from the mandible, and are
also found above organ 17 infraorbital.

The buds on the cheek as they first appear in larvae have
approximately the position and direction of the accessory trige-
minal nerves in the adult, and these accessory nerves have the
same relation to the branches of the buccalis, behind the eye,
that those branches of the maxillaris superior, that arise below
the eye, have to the buccal branches there. These branches
seem, therefore, to innervate the terminal buds, and in that
case they would belong to the fasciculus communis root, and
not to the general cutaneous root of the trigeminus, as Strong
finds in Rana. The only other supposition is that the buds on
the cheek are innervated by the mandibularis internus facialis,
which seems improbable.

The buds on the mandible lie in the region supplied by
the anterior branches of branch 4 of the maxillaris inferior
trigemini. These branches have the course and position of
the ramus coronoideus of Pollard, a nerve which, in siluroids,
innervates part of the sense organs of the coronoid tentacle.
Accompanying this nerve in Silurus glanis I find a large branch
of the mandibularis facialis. This nerve in Silurus may be a
branch corresponding to the one which in Amia innervates the
mandibular line of pit organs, and in Gadus innervates a man-
dibular line of the slit-like organs peculiar to that fish. If this
be the case, pit organs or other organs of a similar character
should be found in Silurus in the region innervated by the
nerve. Of this I have not yet attempted any investigation.
The nerve corresponds almost exactly to a branch of the
facialis in Carcharinus.

The buds above organ 17 infraorbital, and hence posterior to
the middle pit line of the head, are innervated either by the
first pair of branches from the ophthalmicus superficialis trige-
mini, or by the branches of the first vagus that anastomose
with that nerve. The branch of the glossopharyngeus that
innervates the pit organs of the middle head line lies super-
ficial to the united trigeminal and vagus nerves, as it should if

628 ALETS. [Vou. XII.

o

nerves to terminal buds in any region are split off from the
ectoderm before the similar formation, in the same region, of
nerves that supply the sense organs of the canal and pit lines.
The branch of the glossopharyngeus to reach its destination
runs first outward below the branches of the trigeminus and
vagus and then medianward above them.

The buds on the upper end of the operculum must be inner-
vated either by the first dorsal branch of the vagus, by the
hyoideus facialis, or by terminal extensions of the accessory
branches of the trigeminus. Those on the gular plate and on
the lower end of the gill cover are almost certainly innervated
by those branches of the trigeminus that I have called ~ghs
and v.ghi. The growth of the buds upward and backward from
the lower part of the gill cover, corresponding to the direction
of the branches of these two nerves, strongly indicates their
innervation by them. If their innervation be by these nerves,
the mandibularis internus facialis would probably take no part
in the innervation of any of the buds on the outer surface of
the head, and the nerve must be, in Amia, either a general cuta-
neous nerve or a nerve like the palatinus facialis destined to
supply terminal buds in the mouth cavity, as Strong supposes
it to be. The almost total absence of buds along its course in
Amia is, however, against this last supposition, and its course
and position indicate strongly a nerve comparable to the one
called by Pinkus branch 4 of the hyomandibularis in Protop-
terus (No. 89, p. 304). As the nerve, in Amia, lies behind the
spiracular canal it is a posttrematic branch of the facialis, and
cannot, therefore, be the chorda tympani, for the course of that
nerve in man through the upper portion of the tympanic cavity
and then downward anterior to that cavity certainly indicates
that it is a prespiracular nerve. That this nerve, in Amia, is
the homologue of the nerve of the same name described by
Ewart, Pollard, and Strong in other Ichthyopsida, and consid-
ered by them as the homologue of the chorda tympani, is
hardly open to question. The nerve in Amia is probably to
be compared to the branch which, on each of the branchial
arches, runs downward over the anterior face of the arch onto
the inner surface of its ventral portion. Its position in Amia,

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 639

along the inner surface of the mandible, could be easily derived
from that in selachians as given by Vetter. In Heptanchus
what seems to be the nerve is shown lying along the posterior
edge of the mandible ; from this position, as the hyoideo-man-
dibular fold of Amia was formed, the nerve could as naturally
come to lie along the inner surface of the mandible as along
the lateral surface of the hyoid. The more superficial nerves
r.ghs and r.ghi of Amia would at the same time acquire their
somewhat peculiar course.

The nerves ~ghs and 7.ghi vary greatly in importance and
in position in different fishes. In Gadus they, and part also of
branch 4 of the maxillaris inferior of Amia, seem to be repre-
sented by a single nerve, the larger portion of which lies along
the lower edge of the mandible, small branches only passing
into or toward the hyoideo-mandibular fold. In siluroids they
are in all probability the nerves of the submandibular and
mental tentacles, the nerves called by Pollard the ramus sub-
mandibularis and ramus mentalis trigemini. Pollard states
(No. 97, p. 411) that these nerves in siluroids either contain,
or are accompanied by, motor fibres, destined to supply the
muscles that move the tentacles. The muscles themselves
he does not describe. As all tentacle muscles in siluroids are
said to belong (No. 97, p. 382) to a special system, not homolo-
gous with the metameric body muscles, they must either have
entirely disappeared in Amia or have become absorbed in other
muscles.

In Auchenapsis and Callichthys Pollard describes a ramus
mandibularis externus trigemini, which he says (No. 97, p. 410)
may be a dissociated branch of his ramus mentalis trigemini.
The mentalis lies internal to the coronoid process of Meckel’s
cartilage, the externus external to that cartilage and external
to the mentalis (No. 97, p. 391). In Callicthys the externus
supplies the anterior face of the mental tentacle. Whether
this nerve finds its homologue in one of the branches of branch
4 of the maxillaris inferior trigemini of Amia, in some part of
nerves 7.ghs and r.ghi, or in a branch of the facialis, seems
open to question. It is to be noted that in Callicthys there is
no lateral line canal in the mandible (No. 94, p. 534), and that

640 ALLIS. [Vou. XII.

in Auchenapsis the mandibularis facialis does not divide into
an externus and an internus (No. 94, p. 533).

If these several nerves are destined largely or entirely to the
supply of terminal buds, as seems probable, the variation in
their importance and position in different fishes is easily ex-
plained, for the distribution of terminal buds and their number
in any particular region varies continually. The nerves in
Amia strongly indicate that in this variation certain of the
terminal buds have wandered from the outer surface of the
head into the mouth cavity. No other explanation can be
given of the course and terminal distribution of the two inter-
nal branches of the fourth branch of the maxillaris inferior
trigemini. That sense organs can pass from the outer surface
of the head to its inner surface is seen in the organ of the
spiracular canal in Amia (No. 3, p. 501). Furthermore, Beard
states (No. 10, p. 191, footnote 1) that he has evidence indicat-
ing or proving that the end organs of taste arise from epl-
blastic thickenings which have wandered through certain gill
clefts into the buccal cavity.

The two internal branches of the fourth branch of the maxil-
laris inferior trigemini, in Amia, with or without the branches
r.ghs and ght, seem to be represented in Protopterus by the
inferior branch of the palatinus facialis, which Pinkus consid-
ers the homologue of the chorda tympani (No. 89, p. 310).
This inferior branch of the palatinus as a separate nerve is not
found in Amia. In teleosts also it is not given by Stan-
nius, but it is described by him in selachians. Van Wijhe
does not give it in Acipenser or in the other ganoids described
by him, but Collinge probably describes it, in Polyodon, as the
mandibularis internus facialis, for that nerve, as shown in his
Fig. 12, lies in front of the spiracle, and he says that it inner-
vates ‘‘the primitive pores of the sides of the head and mouth”’
(No. 19, p. 518). It lies immediately in front of and closely
accompanies the mandibular branch of the trigeminus which
innervates “primitive pores” in the region of the maxilla.
These two nerves if fused would seem to represent the maxil-
laris inferior trigemini of Amia, provided the “ primitive pores ”’
described by Collinge are terminal buds or their derivatives.

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 641

In Protopterus the inferior branch of the palatinus facialis
sends a branch to the pharyngeal branch of the glossopharyn-
geus, and another to the mucous membrane in the hind part of
the mouth cavity (No. 89, p. 309). In Rana bundles of fibres
from the fasciculus communis tract join and issue with the
united glossopharyngeal and vagus roots, and form in part the
pretrematic and pharyngeal branches of those nerves.

The fibres of the fasciculus communis tract thus seem des-
tined to form in part or in whole the pretrematic and pharyn-
geal branches of the nerves with which they are associated, and
it is always in the regions which, whether on the outside or the
inside of the body, are, from their relation to these nerves,
presumably innervated by them, that terminal buds are found.
The fibres of the tract that join and form part of the facialis
enter into, or form entirely, the inferior and superior branches
of the palatinus facialis, and those two branches are respec-
tively the pretrematic and pharyngeal branches of the nervus.
The fibres that join the glossopharyngeus enter, in Rana
(Strong), into the ramus lingualis and the ramus pharyngeus,
which are, respectively, the pretrematic and pharyngeal branches
of that nerve. As there are no communicating branches in
Amia from the palatinus facialis to the glossopharyngeus, the
fasciculus communis component of the latter nerve must be a
postspiracular nerve, as must, therefore, probably also be both
parts of Pinkus’ branch of the palatinus facialis to the nerve
in Protopterus.

Whether there is or is not a prefacial component of the
fasciculus communis tract seems problematical. Brandis (No.
14, p. 539) says there is none in birds. The ophthalmicus
superficialis trigemini may, however, be in part such a nerve,
as may also be the ciliaris brevis. The ophthalmicus trigemini
innervates terminal buds, and is therefore functionally to be
classed with the palatinus facialis and the ramus anterior of
the glossopharyngeus, both of which are pretrematic or pharyn-
geal nerves. The pretrematic and posttrematic nerves in
Petromyzon all arise, according to Kupffer (No. 67, p. 45), in
connection with epibranchial ectodermal thickenings, and the
pretrematic nerves, all, from and including the trigeminus

642 ALLIS. FVOL. Xt:

backward, become connected with a second, pretrematic, series
of ectodermal thickenings. From both these thickenings
ganglia arise, the first or epibranchial ganglion being common
to the post- and pretrematic branches of the nerve, while the
second or pretrematic ganglion is connected with the pretre-
matic nerve alone. In Petromyzon no ectodermal sense organs
of any kind arise in connection with the epibranchial ganglia
(No. 67, p. 30). If this be true for other fishes, terminal buds
must arise in connection with the pretrematic ganglia and
ectodermal thickenings, if their origin and development is sim-
ilar to that generally accepted for the organs of the lateral
line. They would, therefore, seem to be the rudimentary
sense organs found by Froriep on the facialis, glossopharyn-
geus, and vagus in calf embryos, and as the eye belongs,
according to Kupffer (No. 67, p. 50), to the line of epibranchial
ganglia, the lens and its nerve, the ciliaris brevis, may belong
to the pretrematic series. Sedgwick says (No. 114, p. 97)
that the profundus ganglion, from which the ciliaris brevis and
other ciliary nerves arise, is, when it is first laid down, in
contact with the ectoderm.

In short, the nerve fibres arising from the fasciculus com-
munis tract seem destined in large part, if not in whole, to the
supply of terminal buds, as Strong (No. 121) has suggested
might be the case. The fibres so arising may issue from the
brain as a separate and distinct root, on which a separate and
distinct ganglion is found; they may issue in part as com-
ponents of certain nerves with the roots of those nerves, and
in part as a separate root which becomes immediately more or
less fused, it and its ganglion, with other roots and ganglia ;
or they may apparently issue entirely as components of certain
nerves.

To the first category belongs, apparently, Protopterus ; to
the second, Amia, Rana, and many other fishes and amphibia ;
to the third, birds, judging from Brandis’ descriptions (No. 12,
p. 539, and-No. 13, p..647), for he finds the fibres of the
funiculus solitarius issuing with the facialis and glossopharyn-
geus, and possibly also with the vagus, and the funiculus soli-
tarius of higher vertebrates corresponds, according to Strong

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 643

(No. 121, p. 186), in every detail with the fasciculus communis
of fishes and Amphibia.

III. MUSCLES INNERVATED BY THE GLOSSOPHARYNGEUS
AND VAGUS, AND THE NERVI GLOSSO-
PHARYNGEUS AND VAGUS.

1. The Visceral Arches.

The visceral arches in Amia have been described by Bridge
(No. 15) and by van Wijhe (No. 129). There are seven of
them in all, the mandibular, or perhaps more properly palato-
mandibular, arch, the hyoid arch, and five branchial arches.
Of the latter, four are complete arches bearing each a double
row of gill filaments, while the fifth is a half arch without gill
filaments.

Between the fifth arch and the next preceding, or fourth arch,
the gill opening lies entirely on the ventral aspect of the arches,
and does not extend to their hinder, outer angle ; that is, it lies
between the ceratobranchials only, of the two arches, and does
not extend upward onto the dorsal aspect of the arches
between the epibranchials, as is the case with the other open-
ings. Between the fifth arch and the next following, or pec-
toral arch, there is a relatively large space, closed externally by
a thin layer of integument which extends upward and forward,
beyond the upper end of the fifth arch, along the posterior edge
of the dorsal portion of the fourth arch, and then forward,
across the dorsal ends of that arch and the other anterior
arches, median or proximal to the gill filaments, to the inner
upper angle of the operculum.

Immediately internal to the dorsal portion of this thin mem-
brane, and attached to it, there is, in the adult, a much degen-
erated glandular formation, extending from near the upper
anterior corner of the opercular opening back to the front
edge of the supraclavicular. It is evidently the thymus, but
apparently so degenerated that it disintegrates and floats away
when stirred with a scalpel under water, or even when washed
with a pipette. In embryos it is found as a well-defined, rela-
tively large mass, not at all degenerated in appearance, and not

644 ALLENS. [Vou. XII.

apparently attached to, or in any way directly connected with,
the epidermis. It extends, in embryos, approximately from the
level of the dorsal end of the second arch backward, above and
beyond the dorsal ends of the other arches, onto the outer sur-
face of the front edge of the supraclavicular. A large venous
vessel lies at first immediately dorsal to it, and then immedi-
ately internal to it, but in the purely superficial examination
that was made no branches were found leading into it.

There is no hyoidean demibranch or opercular gill, but there
is a well-developed pseudobranch, lying in what Wright has
called the pseudobranchial canal (No. 133, p. 493).

a. Basal Line.

Between the distal ends of the branchial arches there is a
median line of basal elements, three in number in the adult
(BB'-3, Figs. 49-51, Pl. XX XIII).

The first of these elements is as long as, or a little longer
than, the other two together, and is, in its middle portion,
ossified, the ossified portion lying between the ventral ends of
the second and third arches. The other two elements are
entirely cartilaginous, and it is perhaps worthy of note that
the single piece of bone in the entire line gives no direct
attachment either to the arches themselves or to the muscles
of the arches. In Lepidosteus the corresponding portion of
the basal line is also osseus (No. 129, p. 272).

The anterior cartilaginous portion of the first element is,
roughly speaking, square in section. It is longer than either
of the other two portions, and has a deep transverse fissure
across its ventral surface near its front end. The part in front
of this fissure is considered by van Wijhe as the basihyal partly
fused with the first basibranchial. On its anterior face, extend-
ing close to the median line, it has two deep depressions in
which it receives the articular ends of the hypohyals. Imme-
diately behind the fissure, beginning in it and extending
upward and backward transversely across the element, there
is on either side a deep articular cup for the articular end of
the first hypobranchial ; and immediately in front of the

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 645

middle bony portion of the element there is a similar articular
cup for the articular end of the second hypobranchial. The
bony portion of the element is much constricted in its middle
portion, both dorso-ventrally and laterally, the dorsal surface of
the piece being, however, left nearly its full and even width, so
that the piece in section is V-shaped, or even T-shaped. At
the hind end of its ventral edge there is, on either side, a roll-
like projection directed backward, outward, and downward.
The posterior cartilaginous portion of the element is shorter
than either of the other portions, is V-shaped or T-shaped in
section, the broad surface above, and on either side, at the
extreme front end of its ventral edge, it projects outward and
forward, and caps the ends of the roll-like projections on the
hind end of the bony portion of the element. From each of
these processes arises a short, tough, conical ligament which is
inserted on the end of a cylindrical process of the third hypo-
branchial, the body of that hypobranchial articulating with the
basibranchial in a very slight depression at the anterior end
of its dorsal edge.

The second basibranchial has somewhat the shape of a short,
slightly curved T-rail, the flat surface of the rail presented
dorsally on the outside of the curve, and the ventral edge
thickened slightly at its front end, where it gives attachment
to a short, tough ligament similar to the one found on the pre-
ceding element. This ligament connects with a process of the
fourth hypobranchial, the body of that hypobranchial articulat-
ing with the basibranchial near its dorsal edge, in the same
way and manner that the third hypobranchial articulates with
the preceding element.

The third basibranchial lies in a continuation downward and
backward of the curved line of the second basibranchial. It is
strongly T-shaped in section, and ends in a long, prow-like
point, directed backward and lying about on a level with the
ventral edge of the first basibranchial. This gives to the
element a keel-shaped appearance, the edge of the keel below,
with a strong angle at about the middle of its length.

The first and second basibranchials are strongly connected
by ligament, the ligament continuing backward along the ven-

646 ALLIS. AViOL. xT:

tral edges of the second and third basibranchials to the ventral
angle of the latter piece. The dorsal surface of the entire line
has a fairly even width.

In a 20 mm. specimen, examined in sagittal sections, the basal
line consisted of three elements, as in the adult; in two 14 mm.
specimens, examined in sagittal and horizontal sections, the
arrangement was somewhat different. In both these latter
specimens the second element was distinctly continuous with
the posterior end of the first element, the two pieces being
separated ventrally by a deep fissure, but connected dorsally by
a thin bridge of cartilage, exactly as the portion called by van
Wijhe the basihyal was connected with the anterior end of the
element. This latter connection was as distinctly marked as in
the adult. The third element seemed to be in process of separa-
tion from the second, the outlines of the abutting ends of the
two pieces being indicated and evident, but the space between
them filled with semi-cartilaginous tissue. These early stages in
Amia, and the continuous basal line in embryos of Lepidosteus
and Acipenser (Parker), therefore indicate that the basal line
is, In ganoids as it is in teleosts (No. 118), originally a
continuous mass, which later separates into several members,
rather than that it is formed by the fusion of several originally
independent members, Amia representing an intermediate con-
dition in the process, as van Wijhe’s statement that the basihyal
is partly fused with the first basibranchial would seem to in-
dicate (No. 120, p. 284).

No indication of a separate basihyal was found in Amia,
either in the adult or in larvae.

b. Branchial Arches.

The first and second branchial arches (Figs. 49-55, Pls.
XXXIII-XX XV) consist each of a hypobranchial, a cerato-
branchial, an epibranchial, and one or two pharyngobranchials ;
the third and fourth each of a hypobranchial, a ceratobranchial,
and an epibranchial, and a single large piece which serves as a
pharyngobranchial for the two arches, but is considered by van
Wijhe as the pharyngobranchial of the third arch only. The

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 647

fifth arch consists of a single large piece generally considered
as a ceratobranchial, and a very small piece of cartilage, usually
but perhaps not always found attached to its outer dorsal end.
The piece that I consider as the epibranchial of the fourth arch
is considered by van Wijhe as the infrapharyngobranchial of
that arch. Bridge states that there are both an epibranchial
and a pharyngobranchial in the fourth arch, but he does not
describe them.

The hypobranchials (#B./—/V) of the first four arches dimin-
ish regularly in length from the first to the fourth, the first
being nearly as long as its ceratobranchial, the fourth only about
one half as long. The first two hypobranchials closely resemble
each other in shape, as do also the second two. The first two
bend sharply inward at their lower ends, and are compressed
laterally, much as if the end of the bone had been flattened
dorso-ventrally, and then, after being twisted through some-
what more than a right angle, had been sharply bent backward
through about forty-five degrees. This flattened end of the
element is triangular or wedge-like in shape, and is strongly
capped with cartilage. The shank or upper portion of the
element is semi-cylindrical in shape, with its ventral surface
strongly grooved, the lower end or opening of the groove lying
in front of the bent and flattened end of the piece. Between
the shank and the bent end there is, on the anterior and ventral
edge of each element, a slight process, and on the posterior side
a slight ridge for the attachment of the strong interarcual
ligaments that connect each arch with the arch on either side
of it.

The most anterior of these ligamenta interarcualia ventralia
(/2v.h) arises from the dorsal surface of the hypohyal, and forms
with the others (v./-/V) a nearly straight, though disconnected
line, running backward and inward from the hypohyal to the
distal end of the fifth ceratobranchial. The anterior ligament
is described by van Wijhe in Amia, and by Vetter in Esox, but
in both cases it is described as passing from the first hypobran-
chial to the ceratohyal, instead of to the hypohyal as I have
found it. It is considered by Vetter as a part of the obliquus
ventralis of the first arch. A similar ligament is found between

—

648 ALLIS. [Vou. XII.

the first and second arches in Esox, and is considered by Vet-
ter as part of the obliquus of the second arch.

The hypobranchials of the third and fourth arches are much
flatter than those of the first and second. Their lower ends are
bent backward and upward slightly, but there is no twist or
lateral compression to the bone, and at the bend or angle on
the front edge of each there is a small process directed down-
ward and forward, approximately in the line of the anterior
edge of the piece, to form a second articulation or connection
with the basibranchial. Both distal ends of the element are
capped with cartilage, as is also the proximal end, the proximal
cap of the fourth hypobranchial and that on the larger of its
distal ends being often continuous along the hind edge of the
element, the bony portion of the element having in such cases
a semicircular form, the arc of the circle directed backward.
The two hypobranchials are appreciably grooved, on their ven-
tral surfaces, only at their outer ends and mainly in the cartilagi-
nous cap of the piece. The lower end of the groove lies or
opens between the two articular ends of the element. At the
base of the anterior articular end of each element, on its ante-
rior edge, there is a slight process for the attachment of the
interarcual ligament, and there is a similar process, or more
properly eminence, on the posterior edge of the third hypobran-
chial, where the proximal cartilaginous cap joins the bony part
of the element, and on the fourth hypobranchial in the proximal
cartilaginous cap itself, the process on this last piece being
sometimes so pronounced that it forms an angular projection
overlapping ventrally the end of the fifth ceratobranchial.
Similar processes are indicated in the proximal cartilaginous
caps of the third and second hypobranchials. The bony portion
of the fourth hypobranchial is formed concentrically around the
point of attachment of the interarcual ligament between that
arch and the third arch, and in the bone of the third hypo-
branchial is also indicated a similar concentrical arrangement
around the point of attachment of the ligament between that
arch and the second. The ligament between the fourth and
fifth arches is inserted along the entire anterior edge of the
distal cartilaginous cap of the fifth ceratobranchial, and extends

No. 3.] WUSCLES AND NERVES IN AMIA CALVA. 649

also slightly onto the edge of the bone immediately beyond
the cap. This interarcual attachment, on this arch, is stronger
even than that of the ceratobranchial to the basal line.

The ceratobranchials (CB./-/V) of the first four arches are
semi-cylindrical pieces, slightly curved upward and deeply
grooved on their ventral surfaces. They are thicker at their
inner, or distal, than at their outer or proximal ends, this being
especially marked in the third and fourth arches, and they are
capped at both ends with cartilage. The distal cartilaginous
caps have each a prominent swelling on the anterior half of
their ventral surface, and beyond, that is distal to, the swell-
ing, the ventral edge of the cartilage is bevelled, so that the
articulation with the hypobranchial lies at a higher level than
the ventral surface of the piece, thus leaving a hollow for the
reception of the outer ends of the muscle of the arch. The
proximal cartilaginous caps bend slightly upward, and that of
the fourth arch has on its posterior edge a large, flat, thin pro-
jection, giving to the cap a triangular shape. The ceratobran-
chial of each arch is bound to its hypobranchial by strong liga-
mentous or connective tissue, which passes in between the two
pieces and forms a tough, relatively thick, bi-concave pad be-
tween them.

The ceratobranchial of the fifth arch (CB.V) is shorter and
more slender than those of the other arches, is only faintly, or
not at all grooved on its ventral surface, and is triangular in
section. The cartilaginous cap at its distal end is relatively
large and flat, that at its proximal end much like those on the
other arches. Attached to the extreme posterior and outer
corner of the proximal cap is a small piece of cartilage (EB.V),
already referred to, which represents either the epibranchial or
one of the pharyngobranchials of the arch. Because of its rela-
tion to the ceratobranchial I have considered it as an epibran-
chial, although in its relation to the muscle of the arch it
strikingly resembles the suprapharyngobranchial of the second
arch.

The epibranchial of the first arch (Z4./) is triangular in
shape, the base of the triangle directed inward and forward,
that is, proximally, with its posterior corner lying at a much

650 ALLIS. (Vou. XII.

higher level than the anterior one, and its ventral surface
deeply grooved through its entire length. Both ends of the
piece are capped with cartilage, the cap on its outer, distal end
curving downward and diminishing gradually in thickness
toward the end, where it meets and is bound to the upturned
cartilaginous end of the ceratobranchial of its arch, a space
being thus left in the angle between the two pieces. The cap
on the inner, proximal end of the piece extends backward a little
along the posterior edge of the piece, and has on the distal end
of this part, just before it unites with the bone, a slight em1-
nence, which in one specimen had on its summit a small tit-like
process. The cartilage along the middle portion of the cap
seems sometimes to be wanting, thus leaving two separate caps
on the proximal end of the piece.

The posterior, upper corner of the proximal end of the
first epibranchial articulates with the infrapharyngobranchial
(JPB.L1) of the next following, or second arch, at about the
middle of its anterior, upper edge, the articulation being with a
prolongation of the cartilaginous proximal cap of that element.
The ligamentous articular attachment of the two elements,
however, extends outward, that is, distally, for a short distance
along the anterior edge of the bony portion of the infrapha-
ryngobranchial, and is continuous both at its origin and inser-
tion with the ligamentum interarculae dorsale of the first arch.
This ligament (/zd./) arises from the posterior edge of the
proximal cartilaginous cap of the first epibranchial, and is
inserted along the anterior edge of the second infrapharyngo-
branchial, from the middle of its length to the outer end of the
piece, the insertion even passing slightly onto the adjacent
proximal end of the second epibranchial. Between the two
ligaments, or more properly through the continuous united liga-
ments, there is a sharply defined circular aperture for the pas-
sage of the pharyngeal branch of the first vagus nerve.

The anterior and lower corner of the proximal end of the
first epibranchial articulates with the infrapharyngobranchial
of its arch. This latter element (/P8./) is a long slender bone
capped at each end with cartilage and slightly grooved on its
posterior or postero-dorsal surface, the groove turning forward

No. 3.} MUSCLES AND NERVES IN AMIA CALVA. 651

under the proximal end of the piece onto and across its ven-
tral surface. Nothing at all corresponding to the supra-
pharyngobranchial of van Wijhe was found, unless it be the
little tit-like process found on the posterior edge of the proxi-
mal cartilaginous cap of the first epibranchial, and already
described. In Scomber, however, such a piece is always
found lying between, and forming the connection between, the
first epibranchial and the second pharyngobranchial.

The epibranchial of the second arch (£4.//) is shorter than
that of the first. Like the latter it is somewhat triangular in
shape, the base of the triangle directed proximally, and so
placed that its posterior edge is at a much higher level than
the anterior one. Both ends of the element are capped with
cartilage, and both resemble the corresponding ends of the first
epibranchial ; the proximal cap on this element is, however, much
more developed than that of the first epibranchial, extends
further outward along the posterior edge of the element, and
has a well developed posterior process at its outer distal end.
On this process there was, in the specimen used for illustration,
a small piece of cartilage which lay against the outer edge of
the levator of the arch, and is undoubtedly the suprapharyngo-
branchial of the arch (SP&.//). Whether this element is
always present in all specimens or not I cannot say, as it was
found only when especial attention was directed to it in the
last of several dissections. The posterior angle of the proxi-
mal end of the epibranchial articulates with the outer end of
the anterior portion of the cartilaginous rim of the third infra-
pharyngobranchial, the articular ligament connecting the two
pieces extending slightly onto the adjoining bony portion of
that piece. The posterior process of the element is connected
by ligament, the ligamentum interarcuale dorsale II (/zad.//),
with the outer edge of the bony portion of the third infra-
pharyngobranchial, with the anterior end of the posterior, outer
cartilaginous edge of that piece beyond the bone, and with the
cartilaginous cap on the proximal end of the third epibranchial.
Between this ligament and the articular ligament there is a
large oval opening through which descends the pharyngeal
branch of the second vagus nerve.

652 ALLIS. [Vor. XII.

With the anterior and larger portion of the proximal end of
the second epibranchial the infrapharyngobranchial (JPB.//.)
of the arch articulates. It is a large, somewhat rectangular
piece, as long as, or slightly longer than, the epibranchial of the
arch, and lies inclined in the opposite direction to that piece ;
that is, with its anterior edge at a much higher level than
the posterior one. It is capped on both ends with cartilage,
the cartilage on the proximal end extending outward along the
anterior edge of the piece for about one half its length. At
this outer or distal end it articulates with the first epibranchial,
its edge overlapping dorsally the edge of that piece.

The third epibranchial (2B.///) is a short, nearly straight
piece deeply grooved on its dorsal surface and with a rounded
proximal end by which it articulates with the outer, posterior
cartilaginous edge of the third infrapharyngobranchial, and
with the anterior corner of the fourth epibranchial. Distal to
this last articulation the proximal cartilaginous cap of the
piece extends outward along the posterior edge of the element,
for about one half its length, and ends in a strong, upturned
process which much resembles the corresponding process on
the second epibranchial plus the suprapharyngobranchial of
that arch. This cartilaginous posterior process, in some speci-
mens, becomes independent of, and not connected by cartilage
with the proximal cartilaginous cap. From its posterior edge
at its base the ligamentum interarcuale dorsale III (“d.///)
arises and is inserted along the anterior edge of the bony por-
tion of the fourth epibranchial. Between this ligament and
the articular ligament there is, as on the other arches, an
opening through which descends the pharyngeal branch of the
third vagus.

The upper pharyngeal bone (/P8.///) is a large piece lying
perpendicular to the direction of the arches, rounded in front
and pointed behind, and may be considered, with van Wijhe, as
the third infrapharyngobranchial alone, or, and perhaps more
properly, as the fused infrapharyngobranchials of the third and
fourth arches, and possibly of the fifth arch also. It is mainly
cartilaginous, the bony portion of the piece being a circular
ossification lying at its anterior, outer edge and entirely sur-

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 653

rounded by cartilage excepting’ in that portion that lies between
the points where the element articulates with the second and
third epibranchials. This exposed bony edge thus corresponds
to the distal half of the anterior edge of the second infra-
pharyngobranchial. At about the middle of the postero-lateral
edge of the element there is a slight thickening on its dorsal
surface marking the point where it articulates with the fourth
epibranchial. In front of and behind this thickening, near the
edge of the piece, there are one or two perforations for the
passage of pharyngeal branches of the vagus.

The fourth epibranchial (£4./1V’) is a flat piece, not grooved
on its dorsal surface, and about as long as the second epi-
branchial. It is capped at both ends with cartilage, the proxi-
mal cap having three strong processes, an anterior, a median,
and a posterior, the two latter passing beyond the outer edge
of the third infrapharyngobranchial dorsal to that piece. By
the anterior process it articulates with the third epibranchial,
and by the posterior with the third infrapharyngobranchial, the
attachment to the latter piece being very strong, the ligament
passing between the two pieces and forming there a tough pad
of tissue. The piece is considered by van Wijhe as the fourth
infrapharyngobranchial because of the articulation in front with
the third epibranchial; but as it articulates also, and much
more strongly, with the posterior portion of the large and per-
haps compound third infrapharyngobranchial, there is equally
good reason for considering it as an epibranchial. Between
the posterior articulation, with the third infrapharyngobranchial,
behind, the two articulations at the proximal end of the third
epibranchial in front, the third infrapharyngobranchial below,
and the fourth epibranchial above, there is a large open passage,
but no vessel or structure of any apparent importance was found
passing through it.

The fifth epibranchial (H4.V) has been described with the
ceratobranchial of its arch.

c. Hyoid Arch.

The hyoid arch, as ordinarily considered, consists of three
parts. These three parts are, in Amia, the hypohyal, the

654 ALLIS. [Vor. XII.

cerato-epihyal, and the interhyal of Bridge. No pharyngeal
element is ordinarily given, as such, but the hyomandibular and
symplectic are generally considered to represent the upper end
of the arch. I consider them the pharyngeal elements. The
cerato-epihyal and interhyal of Bridge then become the cera-
tohyal and epihyal, respectively, as van Wijhe has already sug-
gested that they must be (No. 129, p. 308).

The hypohyal (#77), which is more than half cartilage, is the
distal member of the arch. To its anterior and ventral portion
the large tendon of the sternohyoideus is attached, and around
this point as a centre the piece is partly ossified, the ossifica-
tion extending only about half way through the element so
that its dorsal surface is left entirely cartilaginous. From the
distal, dorsal, and inner corner of the piece a large cartilaginous
process is directed medianward and backward, and articulates
with the front end of the first basibranchial. There is no per-
foration for the passage of an arteria hyoidea as in Scomber,
Perca, and Gadus.

The ceratohyal (CH) is the next following member of the
arch, and is in larvae a single continuous piece of cartilage. In
the adult it contains two ossifications. The lower and much
larger one is the bone considered by Bridge and van Wijhe as
the ceratal member of the arch. It is a long, strong, curved
piece having an upper, or proximal, flattened or blade-like por-
tion, the thin, outer edge of which is deeply grooved for nearly
its entire length, and a lower, or distal, rounded portion or shank
which is not grooved. The thin, outer edge of the proximal
portion is directed ventrally and backward, and corresponds to
the ventral surfaces of the ceratobranchials. The nerve of the
arch, the ramus hyoideus facialis, lies, as on the other arches,
immediately external to this groove, but no associated artery
was found.

The ceratohyal is capped with cartilage below, and articu-
lates with the hypohyal, a thick pad of connective tissue lying

| between and separating the two elements. Above, or proxi-
mally, the lower bone ends abruptly in a nearly straight edge
at right angles to the longitudinal edges of the element.
Beyond this upper edge of the lower ossification there is a

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 655

large triangular terminal piece which to all appearances is
simply the greatly enlarged cartilaginous proximal cap of the
element. In it is found the second, and perhaps secondary,
ossification of the piece, the relatively small semicircular bone
considered by Bridge as an epihyal. The free straight edge of
this ossification forms the larger part of the dorsal side of the
triangular cap of the element, but it does not extend quite to
the upper tip or quite to the lower edge of the cap. Its semi-
circular side is thus left entirely surrounded by cartilage, and
the bone touches and is connected with the large lower ossifi-
cation only by means of a small thin splint-like process, which
projects downward from its outer surface at its lower corner
and dove-tails into a corresponding depression on the outer
surface of the dorsal edge of the lower bone. Near the upper
end of its straight, free edge, and directed at right angles to
that edge, there is a strong condylar process always strongly
capped with cartilage, the cap being sometimes connected with
the main cap by a narrow line of cartilage along the upper,
dorsal edge of the element. On the flat outer surface of the
ossification there is a large depression in which the upper pos-
terior end of the ligamentum mandibulo-hyoideum (/wh) is
inserted, and it is around this point as a center that the ossifi-
cation is formed. A bone having the shape of the one shown
in van Wijhe’s Fig. 13 I did not find in any of the specimens
examined, but I do not doubt that, in specially developed
specimens, it may exist.

The epihyal (2/7), or next following member of the arch, is
the interhyal of Bridge and van Wijhe. It is a small rhom-
boidal piece of cartilage articulating with the condylar process
of the ceratohyal, and directed upward, forward, and slightly
inward in the line of the long axis of the hyomandibular, at
right angles to the lower edge of that bone, and at right angles
to the axis of the upper end of the ceratohyal. Its flat surface
is perpendicular to that of the hyomandibular and perpendicular
also to that of the blade of the ceratohyal, the edge of the
piece being, therefore, presented laterally. Its inner edge is
longer than the outer one, and lies in the plane of the inner
surfaces of the hyomandibular and the blade of the ceratohyal.

656 ALLIS. [Vou. XII.

Its outer edge projects externally beyond those bones. On
its lower or distal edge, at the inner, lower corner of the piece,
there is a deep facet which receives the condylar end of the
ceratohyal, and at its outer, distal corner there is a slight
enlargement which gives attachment to the articular ligament,
which binds the piece to the cartilaginous cap or to the sec-
ondary ossification of the ceratohyal. A slight depression on
the upper surface of the piece between these two corners indi-
cates possibly the longitudinal groove found on the dorsal
surfaces of the epibranchials. At its broader, anterior end the
epihyal articulates with the cartilaginous interspace between
the hyomandibular and symplectic. A strong double ligament
(id.z), arising from the inner surface of the hyomandibular,
extends along the inner edge of the epihyal, and is inserted,
one half on the posterior end or edge of the cartilaginous cap
of the ceratohyal, and the other on the inner surface of that
cap near its posterior or postero-ventral edge, and beyond the
cap on the inner surface of the upper end of the lower ossifica-
tion, the ligament turning abruptly through a right angle at
about the middle of its course. These ligaments are not the
articular ligaments of the arch, they resemble much more, one
or both of them, the interarcual ligaments of the branchial
arches, and I consider them as such.

The pharyngohyal, or proximal member of the hyoid arch,
is undoubtedly the single piece formed by the hyomandibular,
the symplectic and the interspace of cartilage connecting those
two bones. The hyomandibular (/7JZD, Figs. 2 and 5, Pl. XX)
is much the larger of the two bones, and is, roughly speaking,
rectangular in shape with a large process, the opercular process,
capped with cartilage and directed upward and backward from
the middle of its posterior edge. At the middle of the outer
surface of the bone there is a large opening, the external open-
ing of the canal by which the facial nerve traverses the bone,
and immediately below that opening, marking the further
course of the nerve and its three main branches, there is a
slight depression on the outer surface of the bone extending to
its lower edge and widening or separating into two parts as it
proceeds. Beyond the lower edge of the bone the two depres-

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 657

sions are continued across the cartilaginous interspace, one
passing dorsal to and behind the articulation with the epihyal,
and the other ventral to and in front of it. The upper end of
the bone is capped with cartilage, and articulates with the side
of the skull in a deep facet lying immediately under its over-
hanging upper, outer edge and extending forward, downward,
and inward from its posterior, outer corner to the hind edge of
the postorbital process. The facet is deepest at its anterior
end, and there lies immediately behind the spiracular canal.
The anterior edge of the hyomandibular, in front of the facial
canal, is thin, and is, as van Wijhe concluded (No. 129, p. 269),
of secondary origin, for even in fishes 40mm. in length it is
found as membrane only. The facial canal, however, lies
entirely in that portion of the bone that is of cartilaginous
origin, and in the smallest specimens examined, those 12 mm.
in length, it apparently lay relatively further from the front
edge of the element than in the older ones.

The symplectic (SY, Fig. 4, Pl. XX) is an irregularly shaped
bone directed downward and forward and ending below ina
large facet lined with cartilage, by which it articulates with
a large process at the lower end of the hind edge of the
coronoid process of the mandible, the process containing and
being strengthened by ossicle d of Bridge. It lies entirely
below the hyomandibular, with its upper, posterior edge on a
level with the corresponding edge of the quadrate, and its
anterior half covered externally by the posterior and upper
portion of that bone. A depressed portion on the inner sur-
face of the quadrate receives the symplectic, and the two bones
are firmly bound together by tissue. The rest of the symplec-
tic, excepting only the lower end of its articular head, is cov-
ered externally by the lower end of the preoperculum, which
abuts with its front edge against the thickened hind edge of
the quadrate. Between the three bones, the preoperculum
and the quadrate externally, and the symplectic internally, a
space or canal is left through which the mandibular branches
of the facialis pass. The three bones are firmly bound together
by tissue, as are also the upper end of the preoperculum and
the hyomandibular, the former bone crossing, and being closely

658 ALLIS. [Vov. XII.

applied externally to the lower posterior portion of the latter
and to the base of its opercular process. The thin hind edge
of the deeper portion of this part of the preoperculum overlies,
and corresponds with, the thin hind edge of the hyomandibular
between the opercular process and the lower corner of that
bone. A groove on the inner surface of the preoperculum,
corresponding with the posterior or hyoidean portion of the
facial depression on the outer surface of the hyomandibular,
marks where the hyoideus facialis passes downward and
backward between the two bones.

The symplectic lies, as van Wijhe has stated, nearly at right
angles to the hyomandibular, and the two bones are separated
and connected by a triangular interspace of cartilage. The
anterior corner of this piece of cartilage is overlapped and
covered externally by the posterior, cartilaginous corner of the
palato-quadrate arch. The rest of the piece, excepting only a
very small portion, is covered externally by the preoperculum,
and the hind edge of the piece is turned outward, upward and
forward against the inclined hinder and inner surface of the pre-
operculum, and firmly attached to that surface by tissue. On
this turned-up portion lies the oblong articular facet for the
epihyal. The long axis of the facet lies, as does the epihyal, in
a plane perpendicular to the lower edge and to the outer sur-
face of the hyomandibular, but the axis of the facet is inclined
at a considerable angle to the plane of the hyomandibular,
instead of being at right angles to it, as might have been ex-
pected. The facet therefore lies on the inner surface of the
cartilaginous interspace, instead of on its hind edge, and the
relatively long front edge of the epihyal is inclined to the axis
of that piece, instead of being at right angles to it.

d. Operculum and Branchiostegal Rays.

The three opercular bones, the operculum (OP), suboperculum
(SOP), and interoperculum (JOP), fit with their anterior edges
under and against the outer, posterior edge of the preoperculum
(POP), separated from it on the outer surface of the undissected
head by a fold or crease in the dermis, and separated also from
the first branchiostegal ray by a similar fold or crease. The

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 659

three bones are bound tightly together by connective tissue,
extend nearly the entire length of the preoperculum, and form
the upper portion or piece of the gill cover. The operculum
articulates, by a large facet on its inner surface, with the end
of the opercular process or the hyomandibular. The facet is not
lined with cartilage, and even in the youngest larvae examined
there is not the slightest indication of cartilage at or in any
part or portion of the bone. The bone in Amia is strictly of
dermal origin.

The suboperculum is attached by ligament to, but does not
articulate by facet with, the lower, posterior, cartilaginous corner
of the hyomandibular; and the interoperculum is attached by
strong connective tissue to the upper end of the upper ossifi-
cation of the ceratohyal. The interoperculum is also attached
by ligament (4mz) and by strong, tough, dermal tissue to ossicle
a, and to the hind corner of the mandible immediately external
to the insertion of the ligamentum mandibulo-hyoideum. The
operculum overlaps considerably at its lower edge the upper
edge of the suboperculum, and the latter overlaps slightly the
interoperculum.

The first branchiostegal ray (BRG) extends from the lower
end of the interoperculum almost to the upper, posterior end of
the suboperculum. It is overlapped by both these bones and
is strongly attached by connective tissue to their under surface.
It is also attached by strong connective tissue to the upper ossi-
fication of the ceratohyal, near its lower edge, and by strong,
tough, dermal tissue to the hind end of the mandible. The
hyoideo-mandibular fold or crease extends upward and _ back-
ward between it and the second ray, so that on the undissected
head the first ray has more the appearance of belonging to the
opercular bones than to the branchiostegal rays. The next
following rays (indicated in Fig. 43, Pl. XXXI) are loosely
attached to the outer surface of the lower ossification of the
ceratohyal, near its hind edge, the most anterior or distal rays
lying beyond the edge of the ceratohyal upon the outer surface
of the hyohyoideus. There are usually ten or eleven rays in all,
and in none of the larval stage examined was there the slightest
indication of cartilage, either in them or connected with them.

660 ALLIS. [Vou. XII.

e. Mandibular Arch.

The mandibular arch as generally considered consists of two
parts, the palatine arch, or palato-pterygo-quadrate apparatus,
and the mandible; but the mandible and pterygo-quadrate
represent, quite possibly, the entire mandibular arch, the pala-
tine part of the apparatus representing or belonging to a pre-
mandibular arch. To the description of the parts as given by
Bridge and by van Wijhe I have but little to add.

At the hind end of Meckel’s cartilage (1/7, Figs. 2, 6, and 7,
Pl. XX) four ossifications, ossicles a, 4, c, and d of Bridge, were
sometimes found, and sometimes but three, ossicles 6 and ¢
being fused, a faint line almost always indicating the line of
fusion. The outer, lower corner of ossicle @ had usually a
slightly different color from the rest of the bone, and there was
sometimes a faint line between the two portions; but, like van
Wijhe, I never found the separate dermal scale described by
Bridge. The canal (mefc) for the ramus mandibularis externus
facialis, on the inner surface of the ossicle, was always con-
tinued forward some little distance on the inner surface of the
articular, and opened underneath the horizontal part of Meckel’s
cartilage, beyond ossicle 6. Ossicle @ was always capped with
cartilage, and between the mento-meckelian ossicles (A7J/) of
the two sides of the head there was always a median, tough,
semi-cartilaginous piece, intimately connected with the ossicles,
as if it were the fused, cartilaginous tips of those bones. It may,
however, be in part a basimandibular, such as White describes
in Hexanchus and Laemargus (No. 127, p. 60). The hind edge
of the cartilage of the coronoid process was always seen as a
line behind the hind edge of the supra-angular, and the process
itself was always found well developed as a process and not as
a separate piece, even in specimens 12 mm. in length. In the
palatine arch the anterior process of the metapterygoid was
always capped with cartilage, as van Wijhe found it, and there
was always on the hind edge of the quadrate, between it and
the symplectic, a small piece of cartilage that escaped his
notice.

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 661

f. Review and Comparison of the Visceral Arches.

A complete visceral arch in fishes consists normally, accord-
ing to generally accepted views, of four pieces or elements:
a pharyngeal element above, then an epal, a ceratal, and a hypal
element, in the order named. The four elements or pieces in
the adult articulate one with the other, and the hypal element
articulates also at its distal end with a median, ventral line of
one or more basal elements.

The arch is always preformed in cartilage and it may persist
as such in the adult, or one or more of its elements may become
more or less ossified, the ends of the elements always remain-
ing cartilaginous even in fishes where the ossification is most
complete. The bones so preformed in cartilage may receive
membraneous additions, as the hyomandibular in Amia, and
the ossification of the cartilage may take place from two or
more different and apparently independent points or centers,
as in the ceratohyal of Amia and the hypohyal of Perca and of
Scomber, the element in such cases presenting, in the adult,
two or more osseous portions, usually, but perhaps not always,
separating by an interspace of cartilage.

The anterior branchial arches always present the most com-
plete and normal development. In the hyoid and mandibular,
or palato-mandibular, arches the development is irregular or
exaggerated, and in the posterior branchial arch it is always
abbreviated, one or more of the normal elements of the arch
not being found. In this last arch, and also in the other
branchial arches that may be incompletely or irregularly devel-
oped, the ceratobranchial is always found, and the outer, lateral
angle of the arch, in the adult, is always at the outer, upper
end of that element, between it and the next preceding element,
whatever that element may be. The pharyngobranchial is
probably also always found in all the arches, but it is subject
to much variation in different species, and also in the different
arches of the same individual. It may be found as a single
piece; as a small and unimportant piece detached from the
other pieces, as is probably the case in the fifth arch of Amia
and of Scomber ; as a piece with two processes or portions, as

662 ALETS. [Vou. XII.

in the second and third arches of Polypterus (van Wijhe); or
as two independent pieces, as in Acipenser and in the first arch
of Polypterus (van Wijhe), the two pieces, or two portions,
where they exist, being always the one anterior and inferior
and the other posterior and superior. Van Wijhe says
(No. 129, p. 225) that where but one pharyngobranchial is
found it always lies below the ‘“‘ Kiemenvene,”’ by which must
be meant the efferent artery of the arch, and that where two
are found the anterior of the two always lies below that artery
and the posterior always above it. He has, accordingly, called
them the infra- and supra-pharyngobranchials, respectively, and
has suggested that the latter should be considered as a fifth
element of a complete and normal arch.! He also states, as a
general rule, that the epibranchial of an arch always touches or
articulates with the infrapharyngobranchial of the next follow-
ing arch, a relation that would naturally arise if the arches are
primarily Y-shaped at their dorsal ends, and continuous one
with the other, as Stohr (No. 117, p. 9) gives them in early
stages of Triton.

At the point where the epibranchial touches the next follow-
ing infrapharyngobranchial, in Amia, the two elements are
always connected by an articular ligament. They are also
always connected, in Amia, by a second series of ligaments,
the interarcual ligaments, which connect the epibranchial with
the next following infrapharyngobranchial or epibranchial, one
or both. This important series of ligaments seems to have
escaped van Wijhe’s notice.

The first branchial arch may articulate with, or be closely
attached by ligament to the cranium by the anterior of its two
pharyngeal elements alone, as in Amia, by the posterior element
alone, as is apparently the case in Polypterus (van Wijhe), or
by both elements, as in Acipenser (van Wijhe). Where the
articulation or attachment is by the anterior element it is, in

1 Attention should here be called to a mistake in the lettering of the pharyngo-
branchials in van Wijhe’s Fig. 2 (No. 129) which shows the branchial arches of
Acipenser. If the pharyngobranchials on the first two arches in this figure be
compared with those in his Fig. No. 3, and with those in Parker’s Fig. No. 5

(No. 86, Pl. XVIII), it will be evident that the anterior processes in van Wijhe’s
figure should be the infra- and not the supra-pharyngobranchials, as they are marked.

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 663

the ganoids described by van Wijhe, with or to the parasphe-
noid and not with the cranium proper; where it is by the
posterior element, it is directly with or to the cartilage of the
cranium.

The second arch also may be attached to the cranium by
ligament or by tissue, but it is always less firmly and less closely
attached than the first arch. In ganoids this attachment of the
second arch is always to the parasphenoid and always by means
of the anterior pharyngeal element of the arch, the posterior
pharyngeal element, where it exists (and where it does not exist,
the corresponding process of the epibranchial), being as a rule
attached to the anterior pharyngeal element of the next
following arch.

The posterior arches do not reach the cranium. They are
simply suspended in the tissues beneath it or beneath the front
end of the spinal column.

In Amia the suprapharyngobranchial, on the arches where
it is found, lies behind the efferent artery of its arch rather
than above that artery, as van Wijhe states, although it would
certainly lie above the artery if prolonged. It lies also behind,
and hence, in the same limiting sense, above the posttrematic
branch of the nerve of its arch, and behind, or more properly
external to, the external levator muscle of its arch, which mus-
cle is innervated by a branch of the posttrematic nerve given
off as that nerve passes outward in front of it. The supra-
pharyngobranchial and levator muscle of each arch both lie in
front of the pretrematic branch of the nerve of the next follow-
ing arch.

If now, in Amia, the relations to the muscles, nerves, and
artery of the hyoid arch of that part of the hyomandibular
that lies posterior to the facial canal be considered, it is
seen: Ist, that it lies behind an artery which accompanies the
truncus hyoideo-mandibularis facialis, which truncus contains
the posttrematic elements of the nervus facialis; 2d, that it
lies external to the adductor hyomandibularis ; and 3d, that that
muscle lies behind the posttrematic branch of the facial nerve,
between that branch and the pretrematic nerve of the next
following arch, and that it is innervated by a branch of the post-

664 ALLIS. [Vor. XII.

trematic nerve sent to it as the nerve passes outward across it.
It thus fulfils all the conditions of a suprapharyngeal element,
and seems certainly to be that element of its arch. If it be
that element, that part of the hyomandibular that lies in front
of the facial canal must either be the infrapharyngeal element
of the arch completely fused with the suprapharyngeal element,
or an outgrowth of this latter element. Under the former
supposition the symplectic would be a special outgrowth of the
infrapharyngeal element ; under the latter it would be that
element itself. In either case the element would lie below
the nerve and artery of the arch, as it should, and be connected
by ligament, or even by fusion, with a part of the next preced-
ing arch. The adductor hyomandibularis then becomes the
levator of the hyoid arch, naturally transformed, by the great
deveiopment of the hyomandibular, into an adductor, and the
double ligament on the inner surface of the hyomandibular, con-
necting it with the epihyal and ceratohyal, can be accounted for
as two remnants of the articular and interarcual ligaments that
originally bound the epal element of the hyoid arch to the infra-
pharyngeal element of the next following, or first branchial arch.

The hyomandibular does not always, in the adult of fishes,
lie behind the posttrematic branch of the facialis, as it practi-
cally does in the adult Amia, and as it does in Heptanchus
(No. 124, Fig. 2) and in Silurus (No. 98, p. 19). In Acipenser
it lies entirely in front of the nerve (van Wijhe) ; in Spatularia
either entirely in front of it, according to van Wijhe, or behind
it, according to Collinge; in Ceratodus behind it (van Wijhe) ;
in Polypterus between its two branches (van Wijhe and Pollard) ;
while in Lepidosteus and many teleosts it is pierced by the
nerve as in Amia. If, then, the hyomandibular be the supra-
pharyngeal element of its arch, or that element and the infra-
pharyngeal element combined, it must, in moving upward to
acquire its articulation at a high level on the lateral aspect of
' the skull, have passed in some instances in front of the facialis,
in others behind it, and in others across it or between its two
branches.

In the mandibular arch, in Amia, the large median process of
the metapterygoid fulfils, in its arch, even better than the hyo-

No. 3.] MUSCLES AND NERVES [IN AMIA CALVA. 665

mandibular does in its arch, the conditions required of a supra-
pharyngeal element. The anterior process of the metapterygoid
also fulfils those required of an infrapharyngeal element, for the
truncus maxillaris trigemini, and the artery accompanying that
truncus, lie between the two processes, in front of the levator
of the arch, and that levator, the levator arcus palatini, is
inserted largely on the inner surface of the median process,
and lies entirely internal to that process. The anterior process
is connected by connective tissue with the skull at a low level,
and the median process by ligamentous tissue and by muscle
at a high level. If the metapterygoid be the pharyngeal element
of the arch, the quadrate becomes the epal element, and it fulfils
in a measure the conditions required of such an element. Its
distal end lies at the outer angle of the arch, as it should, and
it is firmly bound to the symplectic, which is the anterior and
inferior process of, or the infrapharyngeal element itself, of the
next following or hyoid arch. In the Characinidae (No. 106,
p. 65) the quadrate is not firmly and rigidly bound to the
hyomandibular as in Amia. It is loosely bound, so that con-
siderable movement is possible between the two pieces, as
between the corresponding pieces in the branchial arches.

The metapterygoid being the proximal element of the man-
dibular arch, the palatine would probably be part of a premandib-
ular arch, and the coronoid process of the mandible possibly
another. The four divisions of the levator maxillae superioris
would then be the levator, interbranchial, or interarcual mus-
cles of this arch. That there are two or more premandibular
arches has been often asserted ; that parts of one or more of
these arches are represented in the palato-quadrate, or so-called
palatine arch is indicated in many ways; by the innervation
of the levator maxillae superioris by a special branch of the
trigeminus ; by the general course of the trigeminus and its
branches, the ramus maxillaris inferior lying always internal to
the coronoid process where it would not naturally lie if the
process belonged to the outer, anterior edge of the mandibular
arch; by the fact that the palatine is found as a wholly
detached and separate piece in siluroids (Pollard) and possibly
also in Scymnus lichia (No. 99, p. 232), and that it is jointed

666 ALLIS. (Vou. XII.

to the rest of the so-called palatine arch in the Cyprinidae
(Sagemehl) ; that the coronoid cartilage is sometimes found as
a separate piece, as in Lepidosteus (No. 129, p. 268); and that
the palatine and pterygoid elements may (No. 118, p. I1), or
generally do (No. 97, p. 403), fuse late. Hubrecht’s figures
(No. 59, Figs. 1 and 2, Pl. XVII) showing the arrangement
of the labial cartilages in Chimaera and Callorhynchus also
strikingly suggest the probability of parts of two arches being
connected with the mandible.

Further evidence in support of this is found in the develop-
ment of the arches in Salmo, Pristiurus, and Raja, as given
by Parker. In Salmo (No. 84, Fig. 1, Pl. III) the hyomandib-
ular is shown as the proximal element of the hyoid arch, the
metapterygoid as the proximal or the two proximal elements of
the mandibular arch, and the palato-pterygoid as an entirely
separate piece lying in front of the upper end of the mandibular
arch. In Pristiurus and Raja (No. 85, pp. 214, 216, and 219)
the spiracular cartilage, or metapterygoid, and the hyomandibu-
lar arise as separate pieces at the distal ends of the mandibular
and hyoid arches respectively. The spiracle lies between
the two elements. In Raja it is an anterior process of the
upper end of the hyoid arch that becomes the hyomandibu-
lar, while a posterior process in the mandibular arch becomes
the metapterygoid. This seems to indicate that the hyoman-
dibular may be in Raja an infrapharyngeal instead of a supra-
pharyngeal element. It may possibly be such in Amia, also,
or it may be in one line of descent one of these elements, in
another the other, and in still another the two combined, thus
accounting for its varying and puzzling relations to the facial
nerve. That either one or the other of the two pharyngeal
elements may acquire an articulation with the skull is seen in
the branchial arches.

Pollard’s assertion (No. 98, p. 24) that “the supposed hyo-
mandibular of Elasmobranchs does not correspond to the
hyomandibular of Teleostei and sturgeons,” is based in part on
certain errors, for the levator maxillae superioris muscle of elas-
mobranchs is not, if I am right, the homologue of the levator
arcus palatini of teleosts, and there is some question as to its

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 667

being the homologue of the protractor hyomandibularis of
sturgeons. The spiracle is also probably not inconstant in its
relations to the hyomandibular if one does not first accept the
conclusion that “the hyomandibular of Teleostei must be sought
in the articular portion of the quadrate of Heptanchus.”

Regarding Pollard’s cirrhostomial theory, in general, I find
nothing in Amia, so far as my work has gone, that seems in any
way to support it. On the contrary, everything seems to indicate
the development of the mouth parts from visceral arches. As
to the tentacles of fishes, at least of all gnathostome fishes, I am
strongly inclined to think that they are simply special sensory
structures developed in connection with terminal buds, or per-
haps of such buds and also of other organs more nearly related
to the pit organs of Amia. That procartilaginous or even car-
tilaginous or bony pieces, such as Pollard describes, should
develop in connection with such accumulations of these organs
is not surprising, if dermal sense organs are, as Klatsch states
(No. 65), active centers of ‘“skleroblastic ’ formation.

It is perhaps worthy of note that in Amia there is a little
fold or flap of dermal tissue on the upper edge of the lower
jaw, just in front of the coronoid process, that is, about where
the coronoid tentacle of siluroids is found. A similar fold is
found in Esox and Scomber also, and doubtless in other teleosts
as well. If these folds be sensory structures, as they seem to
be, they may be the homologues of, or at least represent a
certain stage in the development or degeneration of, the ten-
tacles.

2. Levatores Arcuum Branchialium.

The levatores arc. branch. (Figs. 52-59, Pls. XXXIV and
XXXV) area group of straight, diverging muscles, arising close
together from the well-marked ridge that extends medianward
and forward along the side of the skull from the hind end of the
posterior process of the intercalar. The surface of origin lies
immediately median to and below that of the posterior portion
of the adductor hyomandibularis, and immediately lateral to and
above that of the anterior end of the trunk muscles, which, in
the adult, extend forward beyond the hind margin of the leva-

668 ALLIS. (Vo. XII.

tores and beyond the vagus foramen to the posterior surface of
the slight elevation representing, in Amia, the bulla acustica.

There are, in the adult, seven levatores on each side of the
head : two interni, an anterior and a posterior one; four externi,
one to each of the first four arches ; and a fifth muscle, which is
inserted on the clavicle instead of on the branchial arches, and
hence can be considered as one of this group of muscles simply
because of its origin with, or even as a part of, the fourth
externus.

The anterior internus (Laéz*) is, both at its origin and at its
insertion, the anterior muscle of the group. It runs median-
ward, downward, and forward across the posterior part of the
bulla acustica, external to the glossopharyngeus, the vagus, and
the efferent artery of the first arch, and is inserted partly on the
bone and partly on the cartilage of the proximal end of the
second infrapharyngobranchial. It is innervated by a branch
of the glossopharyngeus.

The posterior internus (Za@éz%) arises external to the poste-
rior portion of the anterior muscle. It runs downward, median-
ward, and backward external to the vagus, and is inserted on
the upper surface of the third infrapharyngobranchial, near the
middle of the piece, the insertion being partly on the hind edge
of the bony portion of the piece and partly on the cartilage
adjoining it.

The first externus, or levator arc. branch. ext. ad arcum pri-
mum (Laée./), the externus anterior of McMurrich, is a short
muscle arising immediately external to the anterior internus,
and running almost directly downward to its insertion on the
posterior edge of the first epibranchial, at the proximal end of
the bony portion of the piece. In one specimen the anterior
fibres of the muscle had separated as a small separate bundle.

The second, third, and fourth externi (Lade.//-/V) arise
close together, almost as a single muscle, the surface of origin
lying immediately behind that of the first externus and imme-
diately external to that of the posterior internus. They are
found in 40 mm. specimens as partly separated portions of a
single muscle, are considered by McMurrich as such in the
adult, and are called collectively by him the externus posterior.

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 669

They run downward, backward, and outward, the second muscle
lying external to the third, and the third external to the fourth,
and are inserted, the second and third muscles on the cartilag1-
nous processes on the posterior edges of the second and third
epibranchials, and the fourth muscle at the distal end of the
bony portion of the fourth epibranchial, the insertion extend-
ing sometimes onto the cartilaginous, distal end of the piece.
In one specimen there was an additional muscle, not found in
any other dissection. It arose with the fourth levator, apparently
as a part of that muscle, but wholly distinct from it, and was
inserted on the proximal end of the fourth epibranchial, instead
of at its distal end.

The fifth externus (Lade.V) arises from the extreme end of
the posterior process of the intercalar and runs downward and
backward, posterior to and internal to the posterior edge of the
fourth externus, to the posterior or outer end of the fourth
arch. It then turns downward over the outer end of that arch
and that of the fifth, and is inserted by a tendon, often long,
to the anterior edge of the clavicle at about one third the
length of that bone. It is found in 4omm. fishes as a part of
the fourth levator, and is possibly what McMurrich refers to
as a “few fibres” of that muscle ‘continued toward the degen-
erate fifth arch.” No muscle similar to it is described by
Vetter in any of the forms examined by him. It is innervated
by a separate branch, or by two delicate branches of, the vagus.

3. Interarcuales Dorsales.

The muscles of this group are not all entirely differentiated
one from the other; there can, however, be distinguished three
obliqui dorsales on each side of the head, two transversi dor-
sales, an anterior and a posterior one, and a median retractor
arcuum branchialium.

a. Obliqui Dorsales.

The most anterior of the obliqui (Od!, Figs. 52-59, Pls.
XXXIV and XXXV) arises from the upper surface of the
third infrapharyngobranchial internal to the insertion of the

670 ALETS. [Vox. XII.

levator arc. branch. int. post. and external to that of the retractor
arc. branch. It lies in part, at its origin, ventral to the anterior
end of the latter muscle, runs outward and backward across the
dorsal or internal surface of the levator arc. branch. int. post.,
and is inserted on the posterior edge of the median of the three
processes on the proximal end of the fourth epibranchial. The
second obliquus (Od?) arises from the dorsal surface of this
same process, runs almost directly backward to the anterior
end of the pharynx, then turns downward, and is inserted along
the posterior edge of the proximal end of the fifth ceratobran-
chial. The third obliquus (Od) arises from the upper surface
of the third infrapharyngobranchial median to the posterior
process of the proximal end of the fourth epibranchial, runs
outward and backward across the ligament connecting that
process with the infrapharyngobranchial, then turns downward
and becomes continuous with the constrictor of the pharynx.
The anterior obliquus is innervated by a branch of the vagus
that supplies also the retractor arc. branch. The other two
obliqui are innervated by branches of the vagus that innervate
the constrictor of the pharynx.

b. Transversi Dorsales.

The anterior transversus (7da) arises from the upper surface
of the anterior portion of the third infrapharyngobranchial, the
surface of origin extending to the extreme lateral edge of the
piece, and the muscle lying, in this part, under and external to
the levator arc. branch. int. post. The fibres of the muscle
have a curved course, running at first forward and medianward
under the levator, then medianward, and then medianward and
backward, to the middle line of the head. The fibres in the
middle portion only of the muscle have a fairly straight course
from one side of the head to the other. In its anterior portion,
in the middle line, it is somewhat fibrous, and is here closely
attached to the base of the skull. In 40mm. specimens the
fibres at this point interlace. The muscle is continuous in its
deeper portion with the retractor, and is innervated by the
nerve that innervates the anterior obliquus and also the retrac-
tor. The transversus posterior (7d) arises from the posterior

No.3.] MUSCLES AND NERVES IN AMIA CALVA. 671

portion of the dorsal surface of the third infrapharyngobran-
chial, runs directly across the middle line of the head, ventral
to the retractor, and is apparently simply an anterior portion
of the constrictor of the pharynx, with which it is continuous.

c. Retractor Arcuum Branchialium Dorsalis.

‘hiss muscle (Kaba, Figs: 62,53) and 62, Pls. XoXxOChy
and XXXVII) is a large and stout one arising from the
sides of the bodies of the third and fourth vertebrae. It runs
directly forward along the under surface of the spinal column
and of the posterior portion of the skull, and is inserted, on
each side of the head, on the median portion of the dorsal
surface of the anterior end of the third infrapharyngobranchial,
its lateral fibres sometimes turning outward and then backward
around the front edge of the levator arc. branch. int. post.,
ventral to the anterior transversus, to be inserted with the latter
muscle on the outer, anterior edge of the infrapharyngobranchial.
The two halves of the muscle are, in its anterior part, more or
less continuous, and the arrangement of the two differs in
different specimens. Some of the fibres of the muscle arising
from one side of the spinal column may cross the middle line
of the head and be inserted on the infrapharyngobranchial of
the opposite side. The muscle lies immediately ventral to the
dorsal aorta, but dorsal to the trunk of the united efferent
arteries of the third, fourth, and fifth arches on each side, which
trunks unite in the middle line to form a common trunk,
which runs upward through the muscle at about the middle
of its length to join the aorta.

4. Adductores Arcuum Branchialium.

There are two of these muscles, one the adductor of the
fourth arch, the other that of the fifth (Figs. 48 and 56,
Pls. XXXIII and XXXV). The former (Aad./V) arises from
the under surface of the distal cartilaginous cap of the fourth
epibranchial and from the under surface of the adjoining bony
portion of the piece, and is inserted on the entire dorsal sur-
face of the triangular cartilaginous cap of the fourth cerato-

672 ALLIS. (Vor. XII.

branchial, and somewhat, also, on the bone beyond it. The
latter (dab.V) arises from the ventral surface of the thin
projecting corner of the triangular cartilaginous cap of the
fourth ceratobranchial, and is inserted on the dorsal surface of
the cartilaginous cap of the fifth ceratobranchial, a portion of
its fibres passing over the end of the cartilage and there con- |
tracting into a tendon which is, in part, inserted on the ventral
surface of the cap, or of the bone adjoining it, and in part is
continuous with the outer, posterior, tendinous end of the
transversus ventralis posterior.

The adductores are each innervated by the nerve of its arch,
the branch that innervates it passing in each case over the
posterior edge of the arch.

5. Interarcuales Ventrales.

The muscles belonging to this group in Acipenser are called
by Vetter (No. 125, p. 478) the interarcuales ventrales, although
but one of them, the first one of the group, is in any way
interarcual. In teleosts Vetter retains the same name for the
group (No. 125, p.517), but subdivides it into two portions, the
obliqui ventrales and transversi ventrales. _McMurrich (No. 76,
p. 132) retains for the obliqui muscles in Amia (doubtless
because Amia is a ganoid) the name given by Vetter to the
corresponding muscles in Acipenser, and so separates the
muscles in Amia into two groups, the interarcuales ventrales
and transversi ventrales. The names given by Vetter to the
muscles in teleosts seeming also the more appropriate for those
in Amia, I have used them instead of the ones selected by
McMurrich.

a. Obliqui Ventrales.

The obliquus ventralis of the first arch (Ov./, Fig. 47,
Pl. XX XIII) arises in the groove on the ventral surface of the
first hypobranchial, the surface of origin occupying the full
extent of the groove from the bend near the anterior end of
the bone to its hind end. Most of the fibres of the muscle
converge proximally, that is, outward and backward, toward
the middle line of its under surface, and are there inserted on

No. 3.]} MUSCLES AND NERVES IN AMIA CALVA. 673

a longitudinal tendon which extends about one half the length
of the muscle, and is itself inserted on the anterior car-
tilaginous cap of the first ceratobranchial and on the under
surface of the bone of that element beyond the cap. A few of
the proximal fibres of the muscle have a direct insertion on
the cartilaginous cap itself on either side of the insertion of
the tendon. In front of this muscle, in the space between the
distal end of the hypobranchial and the posterior surface of
the hypohyal, McMurrich found a muscle, but I have been
unable to find one, or even a trace of one, in any specimen, old
or young. A strong interarcual ligament was, however, always
found running from the small process on the front edge of the
first hypobranchial to the posterior and ventral surface of the
hypohyal near its articulation with the ceratohyal, as already
described ; and internal to it, in the space where McMurrich
seems to have found the muscle referred to, I always found a
mass of fatty, connective tissue.

The obliquus ventralis of the second arch (Ov.//) is very
similar to that of the first, but it extends, at its origin, farther
forward and it is incompletely separated into a distal and a
proximal portion. The proximal portion arises, as does the
muscle of the first arch, from the grooved, ventral surface of
the shank or straight part of the hypobranchial. The anterior
portion arises from the ventral edge of the bent end of the bone,
and its fibres form on the ventral surface of the muscle a
rounded, nearly separate muscle-belly. In its deeper portion it is
continuous with the proximal muscle, and the fibres of both are
inserted on a median, longitudinal tendon lying on the ventral
surface of the proximal muscle, the tendon and a few fibres of
the proximal muscle being inserted, as in the first arch, on the
cartilaginous cap and on the anterior end of the bony portion
of the ceratobranchial of the arch.

The obliquus ventralis of the third arch (Ov.///) resembles
that of the second, but the proximal portion of the muscle,
owing to the smaller size of the hypobranchial of the arch, is
relatively much less important, and the fibres of the distal por-
tion contract more markedly into a tendinous end before their
insertion on the median tendon of the muscle. The distal

674 - PaaS. [Vor. XII.

portion arises from the anterior and ventral surfaces of the
small anterior and ventral articular process of the hypo-
branchial, and between its fibres, on this same process, is
inserted one of the two dorsal ends of the ligament from which
one of the two arms of the branchiomandibularis arises. The
other dorsal end of the ligament is inserted on the second
hypobranchial and on the interarcual ligament connecting that
hypobranchial with that of the third arch.

The obliquus ventralis of the fourth arch is found as three
distinct and separate muscles. One of them (Ov./V") corre-
sponds to the outer or proximal portion of the muscles of the
other arches. It arises from the anterior half of the ventral
surface of the fourth hypobranchial and from the base of the
small articular process of that piece, and is inserted on the
anterior end of the ceratobranchial of the arch. The second
muscle (Ov./V*) corresponds to the distal portion of the mus-
cles of the other arches. It is a large, spindle-shaped muscle,
the largest of all the obliqui, arises from the hind surface of
the small articular process of the third hypobranchial, and, run-
ning almost directly backward ventral to the proximal portion
of the muscle and ventral to the anterior transversus ventralis,
is inserted by a long and slender tendon on the ceratobranchial
of the fifth arch near its anterior end and along the anterior
side of the pharyngo-clavicularis externus. It may arise in part
from or be attached to the fourth hypobranchial where it passes
across it. The third muscle (Ov./V2) lies immediately dorsal
to the anterior end of the second muscle and is undoubtedly a
derivative of that muscle, for in 19 mm. specimens the two are
found as a single muscle with no trace whatever of separation.
It extends from the posterior surface of the small articular
process of the third arch to the anterior surface of the corre-
sponding process of the fourth arch. As all three muscles are
innervated by branches of the nerve of the fourth arch, the
large ventral one cannot be properly considered as the obliquus
of the fifth arch, as McMurrich from its insertion concluded

(No. 76, p. 133).

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 675

b. Transversi Ventrales.

The transversus ventralis anterior (Z7va) arises in a line along
the middle of the ventral surface of the fourth hypobranchial
on one side of the head, internal to and dorsal to the obliquus
of the arch but ventral to the artery of the arch, and extends
across the middle line, ventral to the second and third basi-
branchials and dorsal to the sinus arteriosus, to the fourth
hypobranchial of the other side. It is always composed of two
or more more or less distinct muscles or muscle bundles; the
muscle bundles, when they exist, —or when they do not the ven-
tral fibres of the muscle, —always crossing each other at a con-
siderable angle, the muscle fibres or muscle bundle of the right
side always lying ventral to that of the left side. The dorsal
muscle where the separation is complete, or the dorsal fibres of
the muscle where it is incomplete, always run directly across
the middle line of the head. The muscle is innervated in all
its parts by the nerve of the fourth arch and the muscle in its
posterior portion is often continuous with the obliquus of that
arch.

The transversus ventralis posterior (7zvp) is formed by two
muscles, each of which belongs distinctly to its own side of the
head and does not, as in the case of the anterior transversus,
cross the middle line of the head to be inserted on the opposite
side. The muscle, on each side, arises from the ventral edge of
the posterior portion of the third basibranchial and from a
narrow median aponeurosis extending ventralward from it. It
runs outward and backward posterior to the pharyngo-clavicu-
lares and is inserted on the ventral surface of the fifth cerato-
branchial, the surface of insertion extending nearly the full
length of the bone. At its outer, lateral end a portion of its
fibres give rise to a tendon which is closely attached to the
proximal cartilaginous cap of the fifth ceratobranchial, but con-
tinues across that cap and becomes continuous with a similar
tendon on the lower end of the adductor of the fifth arch, as
already described. In 40 mm. specimens the separation of the
muscle into two parts, one on each side of the head, is equally
or even more distinctly marked than in the adult, for in young

676 ALLIS. [Vou. XII.

fish the median aponeurosis from which the muscles in part
arise is not so well developed as in the adult, and there are not
the numerous intercrossing tendinous fibres found in the adult
on the ventral surface of the aponeurosis and of the muscles
themselves adjoining it. In its posterior portion the transversus,
both in the young and in the adult, is continuous with the ante-
rior portion of the constrictor of the pharynx. It is innervated
by the nerve of the fifth arch, closely resembles the fifth inter-
arcual of Acipenser (No. 125, Fig. 6), and is unquestionably the
homologue of the proximal portion of the obliqui muscles of
the anterior arches.

c. Pharyngo-Clavicularis Externus and Internus.

These two muscles (Pee and Pez) arise close together, usually
as a single muscle, from the anterior surface of the clavicle
immediately external to the dorsal portion of the sternohyoi-
deus. The two muscles separate at once and the externus,
which arises external to and below the internus, runs forward
and medianward across that muscle and is inserted on the pos-
terior or median edge of the ceratobranchial of the fifth arch
near the middle of the bone, lying, at its insertion, median or
posterior to the tendinous end of the second division of the
fourth obliquus, and in front of the transversus posterior.

The internus may be single or double, the two parts, when
it is double, lying closely together but wholly separate except
at their origins. The muscle runs forward and medianward
across the transversus posterior and is inserted between that
muscle and the transversus anterior on the anterior part of the
third basibranchial.

Both muscles are innervated by the nerve of the fifth arch
and they are, in all probability, the homologues of the distal por-
tions of the obliqui muscles of the anterior arches, that is, on
the fourth arch, to muscle Ov./V?._ They are called by Schnei-
der (No. III, p. 116), in ganoids and teleosts, the sterno-
pharyngei, one being a rectus and the other an obliquus; by
Albrecht (No. 2, pp. 29 and 34) they are called, in Acipenser,
the interbranchialis internus VI, or omozonio-branchialis, and,
in Gadus, the branchiretractor inferior and superior.

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 677

6. Review and Comparison.

In the primitive condition of all fishes there was, according
to Vetter (No. 124, pp. 443-446), a single, annular, general
constrictor muscle, which later separated into two layers, an
outer and an inner one. From the outer layer arose the inter-
branchial muscles of the visceral arches ; from the inner one
the adductor and interarcual muscles, and the constrictor of
the pharynx. Connected with the primitive general constrictor
there were a large number of visceral arches all lying “ ober-
flachlich in der diinnen Leibeswand,” their upper ends im-
bedded in and moved by the “Ringmusculatur des Schlundes.”
Whether the arches so described lay inside or outside the
muscle is not clear either in this statement of Vetter’s conclu-
sions, taken alone, or taken in connection with further state-
ments made in two footnotes. The arches would seem to lie
inside the muscle were it not that Vetter definitely states his
acceptance of Gegenbaur’s conclusion that the adductor man-
dibulae muscle, which is considered as the serial homologue of
the adductor muscles of the branchial arches, lay primarily on
the inner surface of its arch.

In Acipenser and all teleosts Vetter concludes (No. 125, pp.
485 and 534) that the adductor muscles of the visceral arches
become greatly reduced, and that they may, on certain or on
all of the arches, disappear entirely. The interarcuales also
disappear, or may possibly give origin to the third, fourth, and
fifth levatores arch. branch. in Acipenser and to the obliqui-
dorsales in teleosts (No. 125, pp. 485 and 534). The inter-
branchiales give origin to the levatores arc. branch. ext. and
int. of teleosts and to the interarcuales ventrales and levatores
arc. branch. (at least the first two) in Acipenser. The trans-
versi dorsales are said to be entirely new formations, or to be
derived, possibly, from the retractores; the retractores them-
selves are possibly the homologues of the subspinalis of Acan-
thias and Heptanchus.

With these general conclusions of Vetter I do not entirely
agree. The arrangement and innervation of the muscles in
Amia seem to me to strongly indicate that they all lay primarily

678 ALLIS. [Vou. XII.

external to the visceral arches. This seems confirmed by
Kaestner’s description of the embryological origin of the skele-
togenous elements of a segment, from the inner wall of the
protovertebra of the segment (No. 62, Pl. XII), and by Kupf-
fer’s positive statement (No. 71°, p. 110) that the cartilaginous
arches in selachians arise from the mesoderm of the arch.
That the muscles of the branchial segments lie primarily inter-
nal to the branchial basket in Petromyzon (No. I11, p. 69, and
No.) 713, Figs, ) has no bearing on the subject, that basket
being of ectodermal origin and not homologous with the
branchial arches of other fishes (No. 714, p. 118).

The general constrictor of the visceral arches lying primarily
external to those arches, the adductor muscle of each arch
must aquire, secondarily, its position on the inner surface of
that arch. This it does, in Amia, in two different ways, either
by passing over the posterior edge of its arch, as on the fourth
arch and probably the fifth also, or over its anterior edge, as
on the mandibular arch. The innervation shows this conclu-
sively, if it be not assumed that the nerve that innervates the
muscle of one or the other of the arches has cut through the
cartilage of its arch.

Before further considering the adductores it will be better
to seek the homologies of the interarcuales, those muscles
being derived by Vetter, in selachians, from the same deep
layer that gives origin to the adductores.

In Acanthias there are, according to Vetter, three interar-
cuales on each of the first three arches. They are called by
him interarcuales I, II, and III. The first of these muscles,
interarcuale I, arises on the upper, posterior process of the
pharyngobranchial of its arch, and is inserted on the corre-
sponding process of the pharyngobranchial of the next follow-
ing arch. The second arises on the epibranchial of its arch,
and is inserted at the base of the posterior process of the pharyn-
gobranchial of its arch. The third arises with the second, on
the epibranchial of its arch, and is inserted on the anterior
process of the pharyngobranchial of the next following arch,
crossing and touching the anterior end of the epibranchial of
that arch. Between the second and third muscles, or, more prop-

Noss) WOSCLES AND NERVES TN AMIA, CALVA: 679

erly, immediately median to the third muscle, a branch of the
ramus pharyngeus of the next following arch runs downward
to the pharynx.

In Heptanchus the interarcuales are found much as they are
in Acanthias ; in Scymnus interarcuale I is entirely wanting.
In all three fishes the muscles are said, by Vetter (No. 124, p.
443), to be innervated by pharyngeal branches of the vagus
nerve that belongs to the corresponding branchial cleft ; hence,
necessarily, the nerve of the next posterior cleft, for the glosso-
pharyngeus belongs to the cleft in front of the first arch. He,
however, only succeeded in definitely tracing the nerves to the
muscles in one fish, Scymnus. The corresponding muscles in
Mustelus and the rays are first said by Tiesing, in a somewhat
uncertain way (No. 123, pp. 111 and 115), to be innervated by
the nerve of the arch to which they belong. This statement
is then qualified as to the innervation of interarcuale I on the
first arch of Mustelus, but not as to the corresponding muscle
in the other fishes. Fiirbringer refers to this qualification of
the first positive statement regarding Mustelus, as confirmation
of his own statement (No. 36, p. 133) that the interarcuales I
are all innervated by branches of the spino-occipital or spinal
nerves, and the interarcuales II and III, on each arch, by the
nerve of the arch to which they belong.

There are thus three different innervations given for these
muscles. If Vetter were right, each vagus nerve would have
to innervate muscles found on the arch in front of its cleft as
well as others found on the arch behind that cleft. This is
directly opposed to what van Wijhe finds in larvae of Scyllium
and Pristiurus, the posttrematic branch alone of each branchial
nerve, in those fishes, innervating the muscles of the arch of
the segment to which the nerve belongs (No. 130, p. 1g). If
the branchial clefts were intrasegmental and the arches, there-
fore, intersegmental, as Hoffmann says they are in Acanthias
(No. 57, p. 647), the innervation given by Vetter would seem
possible or even natural. The clefts are, however, said to be
intersegmental, and the arches, therefore, intrasegmental in
Amphioxus (No. 54, p. 145), in Ammocoetes (No. 54, p. 159),
and in Necturus (No. 91, p. 955). The innervation given by

680 ALLTS, [Vou exIt;

Vetter is, therefore, probably wrong. Between Tiesing and
Fiirbringer there is no way to choose. The innervation given
by Tiesing seems more natural, and if he be right the muscles
in selachians can be readily homologized with those in Amia.
If Firbringer is right no comparison seems possible. I there-
fore assume Tiesing to be correct.

In Amia there are, belonging to each of the first three bran-
chial arches, two dorsal muscles and two dorsal ligaments.
One of the two muscles of each arch arises from the side of
the skull, and is inserted on the infrapharyngobranchial of the
arch next following the arch to which the muscle belongs, the
large so-called third infrapharyngobranchial being here con-
sidered as the fused third and fourth infrapharyngobranchials.
The other muscle, on each arch, arises with the first muscle
from the side of the skull, and is inserted on the epibranchial
of its ownarch. That the muscles belong, in each case, to the
arches to which they are assigned is shown conclusively by the
innervation. The two dorsal ligaments on each arch are,
one, the articular ligament connecting the epibranchial of its
arch with the infrapharyngobranchial of the next following
arch, and the other, the interarcual ligament connecting the
epibranchial of its arch with the infrapharyngobranchial, or
epibranchial, of the next following arch. Between these two
ligaments, that is, immediately median to the interarcual liga-
ment, the ramus pharyngeus of the next following arch always
passes downward to reach its destination.

If, now, muscles I and II in Acanthias should acquire at
their upper, anterior ends, an attachment to the side of the
skull, and muscle III should become a ligament, an arrange-
ment of muscles and nerves would arise so closely resembling
that found on the first three arches of Amia, that the muscles
in the two fishes can safely be considered homologous. Muscle
I, on each of the first three arches of Acanthias, is thus, in all
probability, the homologue of the internal levator of the corre-
sponding arch in Amia, the muscles of the second and third
arches, in Amia, being united to form a single muscle, the
levator arc. branch. int. posterior, the largest of all the levator
muscles. Muscle II, on each arch of Acanthias, is the homo-

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 681

logue of the external levator of the corresponding arch in
Amia ; and muscle III the homologue of the corresponding
dorsal interarcual ligament. It may here be stated that these
latter ligaments in Scomber are still partly muscular.

On the fourth arch, in Acanthias, there is no interarcuale I.
As it is found, in Heptanchus, on that arch and the next fol-
lowing ones, Vetter concludes that, in Acanthias, it has been
absorbed, before the disappearance of its arch, by the ever
advancing constrictor of the pharynx. It seems to me much
more probable that it, either with or without other more pos-
terior muscles of the series, has become the subspinalis of
Acanthias and the retractor arcuum branchialium of Amia and
teleosts.

In selachians no ventral interarcual muscles are given by
Vetter. He, however, describes them in Acipenser and in
teleosts, and derives them, in both, from the interbranchial
muscles of elasmobranchs, more particularly from the so-called
interbranchiales of Chimaera. These latter muscles, although
considered by Vetter the homologues of the muscles of the
same name in selachians and teleosts, differ greatly from
those muscles. They are found in the first three arches only
of Chimaera, and arise, on each arch, from a posterior process
of the pharyngobranchial of the arch, the surface of origin
extending backward beyond the hind end of the process, and,
on each of the first two arches, occupying part of the anterior
edge of the epibranchial of the next posterior arch. The
muscle runs the full length of its arch, lying on its outer sur-
face, and is inserted on the outer surface of the hypobranchial
of the arch. The muscle has, thus, almost exactly the origin of
the interarcuales II and III of selachians, and the insertion
of the obliqui ventrales of ganoids and teleosts. Furthermore,
interarcuale II in Scymnus is said to extend outward and
downward along the outer surface of its arch, to be inserted on
the upper end of the ceratobranchial, thus closely resembling
the interbranchiale of Chimaera. The muscle in Chimaera
may accordingly be considered as representing a primitive con-
dition in the differentiation of that deeper layer of the general
constrictor to which Vetter assigns the interarcuales and adduc-

682 ALLIS. [Vov. XII.

tores of all fishes, and not as a specially developed condition of
the interbranchiales of other fishes. From the dorsal end of
this primitive muscle the interarcuales of selachians and the
levatores arcuum branchialium of ganoids and teleosts arise ;
from its ventral end the obliqui ventrales of the latter fishes.

The adductores can now be considered. In Amia and
teleosts they seem to represent, on the branchial arches, the
greatly reduced and disappearing middle portion of the once
continuous interbranchiale of Chimaera. They cannot, how-
ever, always be derived from that muscle, for in Chimaera both
an adductor and an interbranchiale are found on each of the
first three arches. Whether they arise from that muscle, or
with it from the same deeper layer of the general constrictor,
they, in Amia, reach their position on the inner surface of the
arches by passing backward and inward over the hind edge of
the arch. This the innervation shows conclusively.

In the hyoid arch the hyohyoideus superior is unquestion-
ably the true interbranchial muscle of its arch; that is, a
muscle homologous with the interbranchiales of selachians and
not with those of Chimaera. From the dorsal portion of this
muscle the levator and adductor operculi of Amia arise. These
latter muscles are accordingly not the serial homologues either
of the external or the internal levator muscles of the branchial
arches. The homologue of one or both of those muscles is
certainly the so-called adductor of the arch, the adductor hyo-
mandibularis. The origin, insertion, and innervation of the
muscle all indicate this almost conclusively. The other inter-
arcual muscle or muscles of the arch are found, in Amia, in
one or both of the two ligaments on the inner surface of the
hyomandibular. The ligaments, in Amia, have entirely lost
their attachment to the first branchial arch. In certain selachi-
ans they still retain it (No. 38). Other ligaments, in selachians,
connect the hyomandibular with the skull and represent the
levator, or so-called adductor, hyomandibularis, which muscle is
otherwise entirely wanting. The hyohyoideus inferior may be
the obliquus ventralis of the arch.

In the mandibular arch the levator arcus palatini develops
from the superficial portion of the general constrictor, and is,

No: 3:] USCLES AND NERVES IN AMIA CALVA. 683

accordingly, homodynamous with the interbranchiales of sela-
chians, and not with the interbranchiales of Chimaera and the
levatores arcuum branchialium of other fishes. The levator
maxillae superioris, on the contrary, develops from the deeper
portion of the constrictor and is the serial homologue on its
arch, whatever that arch may be, of those muscles. The
adductor mandibulae lies along the outer surface and anterior
edge of the arch, excepting in its ventral portion, where it is
inserted in part on the inner surface of the arch. The nerve
and artery of the arch lie along the anterior surface of the
muscle, as they do along the anterior surface of the inter-
branchiales in Chimaera. The middle portion of that muscle
has accordingly, in the mandibular arch, slipped over the
anterior edge of the arch, instead of over its posterior edge, as
on the branchial arches. The adductor mandibulae is thus not
strictly homodynamous with the adductores arcuum branchia-
lium. The geniohyoideus and intermandibularis are probably
the obliqui ventrales of their arch or arches.

7. WNervus Glossopharyngeus.

The nervus glossopharyngeus ( g/, especially Figs. 25, 52 to
55,and 64, Pls. XXV, XXXIV, XXXV, and XXXVIII) arises,
in the adult, by a single root from the side of the medulla
oblongata, immediately ventral to and a little posterior to the
root of the nervus lineae lateralis vagi. It runs at first out-
ward and backward across the ventral surface of that root, and
then turns outward, backward, and downward immediately in
front of the middle membranous portion of the posterior
bounding wall of the labyrinth recess. It passes above the
ramulus papillae lagenae acustici, under the ramulus ampullae
posterioris acustici, between the sacculus and the sinus utriculi
posterior (Fig. 57, Pl. XX XV), and issues from the cranium
by its foramen (g/f7, Figs. 9, 10, 11, Pl. XXI), which lies
immediately behind the hind edge of the petrosal, in the angle
between that edge and the ventral edge of the posterior process
of the bone. The foramen lies entirely in the cartilage of the
cranium, but its front and upper edges are formed by the
petrosal.

684 ALLIS. [Vor. XII.

As the root of the glossopharyngeus issues from under the
root of the n. lineae lateralis vagi it receives, from that root,
the root of the so-called dorsal branch of the glossopharyn-
geus. This latter root runs outward with the main root of
the nerve, lying immediately behind that root, closely applied
to it, and in the adult issues with it, through the same fora-
men, and enters the same ganglion. As the two roots pass
outward between the two terminal branches of the ramus pos-
terior acustici, there is an apparent interchange of fibres, the
nature of which could not be satisfactorily determined. The
nerve in Amia thus seems to differ in this respect, though not
in position, from the nerve in Protopterus (No. 89, p. 316).
In larvae, as already stated, the dorsal root usually issues
from the cranium by a distinct and separate foramen lying
immediately behind the foramen of the main root, and forms a
ganglion separate and distinct from the ganglion of that root.
From this ganglion arises the ramus dorsalis of the nervus,
which runs upward, backward, and outward along the outer
surface of the intercalar, internal to the levator arc. branch.
int. ant. Immediately below the surface of origin of that
muscle, it pierces the intercalar and then the chondrocranium
by a special foramen, and issues on the upper surface of the
cranium on the bottom of the temporal groove. There it
separates into two portions, one of which turns outward,
enters the squamosal, and supplies organ 17 infraorbital of the
lateral sensory canals. The other portion turns medianward,
runs external to and superficial to, or through, the communica-
ting branches from the trigeminus to the vagus, superficial to
the anterior extension of the trunk muscles that fills the tem-
poral groove, and supplies the middle dorsal head line of pit-
organs (No. 3, p. 516). In Laemargus (No. 31) this dorsal
branch of the glossopharyngeus leaves the main root of the
nervus proximal to the ganglion of that root, and contains no
ganglion cells. It is accompanied by or contains fibres that
are distributed to fibrous tissues.

The main ganglion (gg/) of the glossopharyngeus, in the
adult, is a somewhat elongated structure, lying close against
the outer surface of the skull, and extending downward and

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 685

forward across the foramen of the nerve and then along the
outer surface of the petrosal. It lies immediately internal to
the levator arc. branch. int. anterior about midway of its length.
From it, in addition to the ramus dorsalis, two main branches
and several smaller ones arise. The main branches are the
ramus posterior, or posttrematicus (psg/), and the ramus anterior
(ag/), which latter soon separates into the ramus pretrematicus
(frg/) and the ramus pharyngeus, or palatinus (fg/7). Of the
smaller branches, one arises from the outer surface of the
ganglion near the base of the ramus posttrematicus and enters
at once the levator arc. branch. int. ant., which it supplies. The
others, two in the specimen used for illustration, arise from the
distal end of the ganglion between the two main branches.
One goes at once to the tissues along the anterior surface of
the distal end of the first arch. The other runs outward and
backward along the upper surface of that arch, parallel to, and
immediately anterior to the ramus posttrematicus, with which
it anastomoses near the middle of the epibranchial.

The ramus anterior runs downward and forward, dorsal to the
efferent artery of the first branchial arch, and to the infra-
pharyngobranchial of that arch, and there separates into its
two portions, or branches, the ramus pharyngeus and ramus
pretrematicus. The former runs forward, downward, and
slightly inward, ventral to the jugular vein and the adductor
hyomandibularis, lateral to and parallel to the common carotid
artery, dorsal to the anterior end of the first infrapharyngo-
branchial, ventral to and external to the hyo-opercularis and
external carotid arteries, and then external to and ventral to
the internal carotid. With this last artery it enters the palatine
canal by the internal carotid foramen at the hind edge of the
lateral wing of the parasphenoid, the foramen lying immediately
in front of the anterior end of the first infrapharyngobranchial.
The common carotid lies ventral to the anterior end of that
bone, the hyo-opercularis dorsal to it, the common carotid
separating into its internal and external branches immediately
beyond the bone. In the palatine canal the ramus pharyngeus
lies on the median side of the ramus palatinus facialis, and at
the point where that nerve separates into its anterior and

686 ALLIES: (VOL. XiL:

posterior branches, the pharyngeus, in one dissection, separated
also into two portions, one of which accompanied each of the
two branches of the palatinus. In other dissections this
separation was not evident, the pharyngeus glossopharyngei
accompanying the anterior branch of the facialis and issuing on
the ventral surface of the vomer to be distributed to dermal
tissues there. The nerve as it passes the external opening of
the pseudobranchial canal lies median to that opening, and
hence in all probability takes no part in the innervation of the
pseudobranch.

The ramus pretrematicus glossopharyngei, after separating
from the ramus pharyngeus, turns outward and forward along the
under surface of the adductor hyomandibularis, and when near
the anterior edge of that muscle, and close to the hind edge of the
pseudobranch, turns outward, downward, and backward along the
under surface of the muscle, and then along the inner surface
of the hyomandibular and symplectic. In this part of its course
it lies ventral to, or even in front of, the truncus hyomandibularis
facialis, and approximately parallel with it and with the ramus
hyoideus facialis. It sends several branches forward and back-
ward to the dermal tissues on the under surface of the hyoman-
dibular, passes beyond the hind edge of the symplectic, and
then turns downward and forward in the dermal fold between
that bone and the upper end of the ceratohyal. At this point,
in larvae, branches of the nerve were traced into the terminal
buds found in the lining membrane of the mouth cavity. The
nerve, in larvae, passes dorsal to the pseudobranchial canal,
dorsal to the hind end of the pseudobranch, lying close to that
organ. No branches of the nerve could be traced directly to
the organ, but a branch was found distributed to the wall of
the spiracular cleft.

The ramus posterior, or posttrematicus, glossopharyngei runs
downward, outward, and backward in front of the levator arc.
branch. int. anterior and the levator arc. branch. ext. primus, and
then along the anterior portion of the upper surface of the first
arch. From the posterior edge of the nerve at its base, or
from the outer surface of the main ganglion itself, one or two
small branches are sent backward to the inner surface of the

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 687

levator arc. branch. int. anterior, which they innervate. As
the nerve passes in front of the levator arc. branch. ext. primus
it sends a nerve to that muscle, and immediately beyond the
muscle is joined, on its anterior side, by the small branch
that arises directly from the main ganglion, with which it
anastomoses, as already stated. Immediately beyond this anas-
tomosis a branch is given off on the anterior side of the united
nerve, the branch being larger than the small anastomosing
branch, and hence apparently containing fibres from the main
nerve. This branch, shortly before reaching the outer end or
angle of the arch, turns downward along the anterior surface of
the epibranchial of the arch, reaches the dorsal surface of the
ceratobranchial, and then runs forward and inward along the
dorsal surface of that bone and of the hypobranchial as far
as, or farther than, the distal end of the latter element. Many
small branches are given off by it in its course.

Immediately beyond the insertion of the levator arc. branch.
ext. primus, the ramus posttrematicus glossopharyngei is joined,
on its posterior side, by three branches of the ramus pretremati-
cus vagi primi, and the four nerves continue side by side without
anastomosis to the outer end of the arch. They all lie external
to and superficial to the efferent artery of the arch, and internal
to, that is, deeper than, the afferent artery, the same relations
to the arteries being maintained on the ventral portion of the
arch as well. At the outer end of the arch the anterior of the
three branches of the first vagus crosses externally over the
glossopharyngeus, and on the ventral aspect of the arch lies
along the anterior side to that nerve. The three remaining
nerves pass to the ventral surface of the ceratobranchial, and
there anastomose completely, forming at first two nerves and
then a single one. From these nerves, before and after the
anastomosis, branches are given off on both sides, and, from
these branches, branches are sent forward and backward along
the entire length of the anterior and posterior edges of the
arch. They form numerous anastomoses with each other, as
shown in the figures, and anastomose also with that branch of
the vagus that crosses externally over the glossopharyngeus to
the anterior edge of the arch, The main portion or branch of

688 ALLIS. (Vou. XII.

the united nerves continues forward along the ventral surface of
the obliquus ventralis primus, sends a branch from its posterior
edge into that muscle, another inward and backward to the
tissues of the floor of the branchial chamber between the ventral
ends of the first and second arches, and then continues forward
and is distributed to the tissues between the ends of the hyoid
and first branchial arches. Terminal branches of it extend also
to the tissues dorsal to the ventral end of the hyoid.

8. Nervus Vagus.

The nervus vagus (v, especially Figs. 25, 52, 53, 59, and 64,
Pls. XXV, XXXIV, XXXV, and XXXVIII), excluding the
nerve of the lateral line which has already been described, arises
by several rootlets which are usually grouped in two main
bundles, the anterior of which arises by one or two rootlets
and the posterior always by three or more. The most posterior
of these rootlets always arises a little behind the others and
runs at first directly outward, or even outward and a little for-
ward, instead of outward and backward as do the others. The
long line of origin of the root lies about opposite to, or a little
in front of, the inner edge of the strongly projecting ridge that
forms the posterior wall of the labyrinth chamber. The root,
owing to its extended line of origin, is at first flattened horizon-
tally, but the posterior edge soon turns downward, thus twisting
the root through about go°, and so presenting its flattened sur-
face at first upward, backward, and inward, and then nearly back-
ward and inward. In this twisted position it is joined by the root
of the lateral nerve which lies along its upper, anterior edge, over-
lapping it somewhat. The united roots then run outward and
backward along the membranous and cartilaginous posterior
wall of the labyrinth chamber, and issue together by the large
vagus foramen. This foramen (vf/7, Figs. 9, 10, and 11, PI.
X XI), as Sagemehl has stated, lies between the intercalar and
the occipitale laterale ; but as the intercalar in Amia is a super-
ficial bone, and probably a membrane bone, the foramen is
properly a perforation of the anterior margin of the occipitale
laterale, nearly inclosed in that bone, but still lying between
it and the cartilage of the skull in front of it.

No.3.] MUSCLES AND NERVES IN AMIA CALVA. 689

As the two roots approach the foramen they pass through
what is apparently a ganglionic enlargement (gv), and there is
always here, or between the two roots proximal to the enlarge-
ment, an interchange, perhaps mechanical, of fibres, for one or
more small nerve bundles were always broken in attempting to
separate the roots. This apparent interchange of fibres, and a
mass of ganglion cells on the root of the vagus, were also found
at this point in larvae (No. 3, pp. 517 and 518), the ganglion
cells being always confined to the root of the vagus, none of
them appearing on the root of the lateral nerve.

Beyond this intracranial ganglion, which lies inside the
cranium, the nervus vagus separates immediately into its four
main portions, or into three portions, the posterior of which
soon separates into the two posterior main branches of the
nerve. On each of these branches of the root, after they issue
from the cranium, a well-marked ganglion is found, that on the
first branch lying close to the side of the skull, the second close
to the first, and the third and fourth at some little distance
from the other two. From the intracranial ganglion the first
branch of the nervus (stv) has its apparent origin, and runs
outward and backward through the vagus foramen above the
four main roots of the nerve. Immediately beyond the gang-
lion, and ventral to the first branch of the nerve, the nervus
lineae lateralis, now well rounded, runs backward across the
upper surface of the four main vagus roots, median to their
ganglia. On it, in the adult, there is no such ganglionic
enlargement as was found in larvae, but a slight difference
in the color of the nerve indicates undoubtedly its presence
immediately beyond the intracranial ganglion of the main
root. From the proximal portion of this discoloration, im-
mediately outside the vagus foramen, or even still inside the
cranium, the first branch of the lateral nerve (#//.s¢o) is given
off, and joins at once the first branch of the main vagus
root, with which it is intimately connected, but apparently
without any interchange of fibres whatever. The two nerves
issue from the vagus foramen together, turn upward and out-
ward along the side of the skull, pass across the posterior angle
of the intercalar median to the posterior process of that bone,

690 ALLIS. [Vot. XII.

and, lying lateral to, and superficial to, the trunk muscles, and
internal to the median process or leg of the suprascapular,
which in embryos is simply a ligament, reach the under surface
of the extrascapular near its lateral edge. Here the branch of
the nervus lineae lateralis separates into two portions, one of
which goes to supply organs No. 18, infraorbital in the extra-
scapular, and No. 19, infraorbital in the suprascapular, and the
other to the organs of the supratemporal cross-commissure and
those of the posterior dorsal pit line (No. 3, p. 517).

After giving off its first branch the nervus lineae lateralis
leaves the vagus, runs backward, outward, and slightly downward,
internal to the levatores arc. branch., between them and the
trunk muscles, and then slightly upward, internal to the gian-
dular formation which I have already described as probably
being the thyroid gland. At the front edge of the supracla-
vicular the nerve turns downward and backward, then, beyond
that bone, directly backward, and, lying slightly imbedded in
the fissure between the dorsal and ventral portions of the
trunk muscles, continues its slightly undulating course toward
the tail. Its branches are distributed to organs 20 and 21, in-
fraorbital, in the supraclavicular, and to the organs of the lateral
line and the body pit lines, as already described in my earlier
memoir (No. 3, p. 518).

The first branch of the vagus, the one arising from the root
of the nerve, gives off a small intracranial branch, and then,
after reaching the outer surface of the head with the first
branch of the nervus lineae lateralis, separates into two main
portions. One of these portions turns forward along the lateral
edge of the anterior extension of the trunk muscles and runs
directly into and anastomoses completely with one of the
branches of the first pair of dorsal branches of the ophthalmicus
trigemini, as already described in describing that nerve. This
branch of the vagus seems to be the nerve described by Pinkus,
in Protopterus, as a communicating branch from the lateralis
facialis to the lateralis vagi. Like the nerve in Protopterus, it
comes into intimate relations with a branch of the lateral, or
so-called dorsal, branch of the glossopharyngeus, as already
stated. From it branches are sent medianward and laterally to

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 691

the dermal and subdermal tissues in the temporal region of the
top of the head. The remaining and larger portion of the
nerve turns outward, sends, in all the specimens examined,
an important branch forward to unite with the other anterior
portion of the nerve, and then, turning downward, breaks up into
numerous branches which run downward, forward, and backward
in the pigment layer along the inner surface of the operculum,
lying immediately external to the terminal ramifications of the
ramus opercularis facialis on the outer surface of the levator
and adductor operculi. The anterior branches of this part of
the nerve run forward into the thick dermal tissue behind the
dorsal end of the preoperculum, and even slightly under the
free dorsal end of that bone. The entire nerve may, from its
distribution, be called the dorsal, or supratemporal branch of
the vagus. Its small intracranial branch (*, Figs. 25 and 59)
arises more from the main vagus root than from the dorsal
branch itself. It, however, arises from the fibres that go to
form that branch, and hence can be considered as a branch of
it. It runs upward and forward, internal to the root of the
lateral nerve, and issues on the top of the chondrocranium by
a small foramen (/¢fr, Fig. 8, Pl. X XI) lying near the middle
line of the head, near its hind end. Its further course could
not be traced. From its position, at its exit, it would seem to
be the ramus lateralis trigemini, or ramus recurrens facialis, of
teleosts. Its apparent origin is, however, from the vagus, and
not from the trigeminus or facialis. If it be the ramus lateralis
trigemini the first pair of dorsal branches of the ophthalmicus
superficialis trigemini can not be that nerve, or can only be a
part of it. Both nerves, and the main supratemporal nerve as
well, are distributed to regions where terminal buds are found,
this being especially true of the posterior branch of the main
nerve.

The supratemporal branch of the vagus arises almost entirely
from the anterior bundle of the root of the nerve. The fibres
of that bundle, where they traverse the intracranial ganglion,
have, apparently, an important interchange of fibres with the
posterior bundle of the root. Viewed superficially and from
below, they have, however, the appearance of a separate and

692 ALLIS, [Vou. XII.

distinct bundle lying on the ventral surface of the intracranial
ganglion, near its anterior edge. Immediately beyond the
vagus foramen, those fibres of the anterior root that have not
entered into the first, or supratemporal branch of the vagus,
turn directly outward and enter the first vagus ganglion, which
is a well-rounded ganglion lying close to the side of the skull
between the levatores arc. branch. int. posterior and anterior.

From this first vagus ganglion (gv!) the truncus branchialis
primus n. vagi arises, and close to it, from the posterior surface
of the truncus at its base, two small, short nerves which run
outward and backward to the inner surface of the levator arc.
branch. int. post., which muscle they entirely or in large part
supply. The truncus then runs outward and downward be-
tween the two levatores arc. branch. int., and in all the speci-
mens examined, excepting one, there separated into a ramus
anterior and a ramus posterior. The former separates at once
into two portions, the ramus pharyngeus n. vagi primi (fv"), and
the ramus pretrematicus n. vagi primi (frzv"), the latter being
the ramus branchialis posterior arcus primi of Stannius. The
ramus pharyngeus turns forward, and then forward and inward,
above the efferent artery of the second arch and between the
levator arc. branch. int. ant. and the levator arc. branch. ext.
primus, passes through the opening between the ligaments
that connect the posterior edge of the first epibranchial with
the anterior edge of the second infrapharyngobranchial, and,
turning inward and forward between the first and second
infrapharyngobranchials, is distributed to the tissues on the
roof of the branchial cavity.

The ramus pretrematicus (frv') runs at first forward, out-
ward, and downward, and then separates into four branches,
one of which turns inward and forward and joins the ramus
pharyngeus, with which it passes through the opening between
the articular and interarcual ligaments, connecting the first
and second arches. The other three branches turn outward,
downward, and backward, and lying close against the posterior
surface of the levator arc. branch. ext. primus, reach the upper
surface of the first arch beyond that muscle, and have the
course and distribution already described.

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 693

The ramus posterior, or ramus posttrematicus n. vagi primi
(fsv'), runs outward and downward immediately in front of the
levator arc. branch. ext. secundus, sends a branch to that muscle,
and beyond it reaches the upper surface of the second arch,
where, like the corresponding nerve on the first arch, it is the
anterior and larger of four nerves that run outward and back-
ward on that surface. From it, as on the first arch, a branch
is sent downward across the anterior surface of the epibran-
chial of the arch to the dorsal surface of the ceratobranchial,
and, as also on the first arch, the second nerve of the four
crosses over the first nerve at the outer end of the arch, and
on the ventral surface of the ceratobranchial lies along the
anterior edge of that bone, anastomosing with a branch of the
larger nerve. The other nerves anastomose and branch irregu-
larly, the main portion of the nerve so formed continuing for-
ward and inward onto the ventral surface of the interarcualis
ventralis of the arch. There it sends branches from its hind
edge into the deeper and proximal portion of that muscle, and
then, turning inward through the superficial and distal bundle
of the muscle, or passing around the anterior and then the
dorsal surface of that bundle, sends a branch to that bundle,
and is then distributed to tissues near the basal line between
the second and third arches. One of the branches of the
united nerves is sent along the anterior edge of the arch to
tissues between its arch and the next preceding arch. On
each of the four complete arches a similar branch is found.

In one specimen (Fig. 59, Pl. XXXV) what was apparently
a second ganglion was found on the truncus primus n. vagi, be-
yond the first ganglion, at the point where the truncus separated
into its several portions. In this specimen four large nerves
arose from this second ganglionic swelling, —a posttrematicus,
a pretrematicus, and a pharyngeus, anda fourth nerve which ran
directly to the ganglion of the glossopharyngeus, and had on its
distal end, where it joined that ganglion, a marked swelling,
ganglionic in appearance. No such nerve or ganglion was
found in any other specimen.

The truncus branchialis secundus n. vagi arises from the
second vagus ganglion (gv?). It runs backward, downward, and

694 ALLIS. [Vou. XII.

outward, internal to, and then outward and downward, behind,
the levator arc. branch. int. post. (to which muscle, in one dis-
section (Fig. 52, Pl. XXXIV), it sent a delicate nerve, appar-
ently innervating it in part), and then separates into an anterior
and a posterior portion. The former, or ramus anterior n.
vagi secundi, turns forward along the outer surface of the
levator, and gives off three branches which together constitute
the ramus pretrematicus of the nerve (pvv?). The remainder
of the nerve continues forward, as the ramus pharyngeus n.
vagi secundi (fv?), passes through the large oval opening
between the two ligaments connecting the epibranchial of the
second arch with the infrapharyngobranchial and epibranchial
of the third arch, and then turns inward and forward between
the second and third infrapharyngobranchials, and is distrib-
uted to tissues there. The three branches which together
constitute the ramus pretrematicus of the nerve, the ramus
branchialis posterior arci secundi of Stannius, run outward and
downward close against the posterior surface of the levator arc.
branch. ext. secundus, reach the upper surface of the second
arch beyond the insertion of that muscle, and have the course
and distribution already given.

The posterior portion of the nerve, the ramus posttrematicus
n. vagi secundi (fsv2), gives off from its base two nerves, one
of which runs directly to the levator arc. branch. ext. tertius,
which it innervates, and the other to the levator arc. branch.
ext. quartus, which it innervates in part, but in part only, for
the muscle receives also an important branch from the posttre-
matic branch of the third vagus, which runs downward and
outward to the upper and posterior surface of the muscle.
From this latter nerve a terminal branch is sent into the
adductor muscle that lies between the epibranchial and cerato-
branchial of the fourth arch.

After giving off the two branches described above, the ramus
posttrematicus secundus runs outward and downward in front
of the levator of the third arch, reaches the dorsal surface of
the arch, sends an anterior branch to the dorsal surface of the
ceratobranchial of the arch, and is joined by two or three
nerves, which together constitute the ramus pretrematicus n.

NO. sa), JZOSCLES AND NERVES IN AMIA CALVA: 695

vagi tertii (~7v3), one of which, as on the other arches, crosses
over the posttrematicus to reach its anterior side, and there
anastomoses with other anterior branches distributed to the tis-
sues along the anterior edge of the ventral surface of the arch.
The main nerve separates, near the distal end of the cerato-
branchial of its arch, into two portions, the posterior of which
enters and supplies the deeper portion of the obliquus ventralis
of the arch, while the anterior one enters and supplies the
superficial bundle of that muscle, as on the second arch.

The truncus branchialis tertius n. vagi separates, soon after
leaving its ganglion, into two parts, one of which is a ramus
pharyngeus, and the other what may be called the united pre-
and posttrematic branches of the nerve. The pretrematic
branch, however, contains both pharyngeal and pretrematic
elements and corresponds, in its relation to the muscles and
ligaments of the arches, to the entire ramus anterior of the two
preceding vagus nerves. The ramus pharyngeus (v3) runs
outward and downward, dorsal to the efferent artery of the third
arch, turns backward and inward behind that artery, perforates
the cartilage of the third infrapharyngobranchial immediately
behind the hind edge of the ossified portion of that element, in
front of, or under, the proximal end of the fourth epibranchial,
and is distributed to tissues in the region of the pharyngeal
bones. In one specimen it sent a branch to the posterior sur-
face of the third arch, and in all the specimens examined it
varied somewhat in its arrangement and distribution. The
pre- and posttrematic branches of the nerve soon separate.
The former sends an important branch, which corresponds to
the pharyngeal branches of the two preceding vagus nerves,
forward and inward through the opening between the ligaments
uniting the third and fourth epibranchials, the rest of the
nerve, the true ramus pretrematicus (v3), running outward
and backward behind the levator of the third arch to the upper
surface of that arch, where it is distributed as already described.

The ramus posttrematicus n. vagi tertii (fsv3) runs outward,
downward, and backward, sends a branch to the posterior sur-
face of the levator of the fourth arch and, on the right side
of the specimen used for illustration, another to the second

696 ALLIS. [Vou. XII.

obliquus dorsalis. From the main nerve, near its base, in one
specimen (Fig. 59), a long and delicate branch was sent to the
fifth levator. In another specimen, the one used for illustra-
tion, this nerve arose from the main truncus of the nerve, or
from the united trunks of the third and fourth vagus. The
branch that goes to the levator muscle innervates that muscle,
and then the adductor of the fourth arch, passing to the latter
muscle over the outer and posterior end of the arch. The
main nerve, after giving off these branches, passes under and
anterior to the levator of the fourth arch, and reaches the upper
surface of that arch beyond the muscle. There it is not joined,
as the corresponding branch on the other arches is, by pretre-
matic branches of the next following vagus nerve. It, however
itself, separates, on the ventral aspect of the arch, into two,
three, or four branches, the smaller ones of which lie along the
anterior and posterior edges of the arch, the main central
portion running forward and inward onto the ventral surface of
the obliquus ventralis of the arch. There it breaks up into
several branches which enter and supply the obliquus of the
arch, the two divisions of the so-called obliquus of the fifth
arch, and the several divisions of the transversus ventralis
anterior. Other branches are doubtless distributed to the
cutaneous tissues between the fourth and fifth arches, though
this was not definitely traced.

The fourth vagus ganglion (gv4) is a large one, either well
rounded or somewhat elongated, and from it arise two nerves.
The outer or anterior of these nerves is apparently the truncus
pharyngeus inferior of Stannius, the other the united truncus
pharyngeus superior and truncus intestinalis. The truncus
pharyngeus inferior separates at once into two branches, the
outer or anterior of which is, from its distribution to the
muscles of the fifth arch, the ramus branchialis anterior of
that arch, or the ramus posttrematicus n. vagi quarti (rv).
The other branch is the ramus pharyngeus inferior of Stan-
nius, but as it is in part, at least, a motor nerve, it cannot
be compared with the pharyngeal branches of the other vagus
nerves. The two nerves run backward and downward, close
together, to the hind end of the branchial arches, lying

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 697

internal to the fourth and fifth levators, and dorsal to, and
then posterior to, the posterior of the two upper ends of the
efferent artery of the fourth arch. Behind that artery and
behind the hind end of the fifth ceratobranchial, the ramus
posttrematicus turns downward, sends a branch to the adductor
muscle of the fifth arch, and then runs forward and median-
ward along the ventral surface of the fifth ceratobranchial. It
crosses the lateral end of the transversus ventralis posterior,
sends branches to that muscle, and then to the pharyngo-
claviculares externus and internus, and ends in several delicate
branches, which enter the tissues near the basal line. McMur-
rich says (No. 76, p. 138) that the claviculares in Amia are
innervated by branches from the first spinal nerve. This is
an error due, doubtless, to the fact that the fourth occipital
nerve, mistaken by him for the first spinal, sometimes enters,
but simply transverses, or gives off a branch which transverses,
the inner of the two muscles. The nerve does not innervate
the muscle; it simply perforates it.

The ramus pharyngeus inferior (fzv+) turns downward at
the hind end of the branchial apparatus, then medianward, and
separating usually into two main portions, sends branches
forward and backward along the outer surface of, and into,
the constrictor oesophagei.

The truncus pharyngeus superior is the median and posterior
of the two divisions of the median and posterior of the two
main nerves that arise from the fourth vagus ganglion. It runs
medianward, backward, and downward toward the front end of
the oesophagus, where it turns directly medianward ventral to
the efferent artery of the third and fourth arches, and passes
between the transversus dorsalis posterior and the retractor arc.
branch. dorsalis. It then turns forward and medianward toward,
and then along, the median edge of the third infrapharyngo-
branchial, sending in its course some branches through that
element, the branches perforating first the transversus posterior,
or the anterior end of the constrictor oesophagei. It is distrib-
uted mainly, if not entirely, to dermal tissues near the hind and
median edge of the pharyngeal bone. On the left-hand side
of the specimen used for Fig. 52, a branch of this nerve

698 ALLIS. [VoL. XII,

formed a loop with a branch of the truncus intestinalis, and
another branch entered, and apparently innervated, the second
obliquus dorsalis. The second obliquus on the right side was
apparently innervated by a branch of the ramus posttrematicus
of the third vagus, as already stated.

From the base of the truncus pharyngeus superior, or even
from the main fourth ganglion itself, one or two branches are
sent medianward and backward to the dorsal surface of the
retractor arc. branch. dorsalis, and then forward and backward
upon that muscle and the transversus dorsalis anterior, inner-
vating those two muscles and also the anterior obliquus dorsalis.

The truncus intestinalis runs downward, medianward, and
backward to the side of the oesophagus, sends several branches
to the dorsal surface of the constrictor oesophagei, and then
continues backward, its further course and distribution not
being traced.

IV. MUSCLES INNERVATED BY THE POSTVAGAL OR OCCIP-
ITAL NERVES, AND THE NERVI OCCIPITALES.

1. Branchiomandibularis.

The branchiomandibularis (Gm, Figs. 44-47, and 58, Pls.
XXXI-XXXV) is a long muscle, double at its origin and at its
insertion, but single in its middle portion, where it lies in the
median plane of the body, immediately ventral to the tough, semi-
cartilaginous mass that lies between, and in front of, the hypohy-
als, and forms the tongue of the fish. Posterior to this mass, and
posterior to and between the hypohyals, the muscle, when at
rest, turns directly downward in front of the anterior tendinous
edges of the hyohyoidei, and passes through a large vascular
or lymphatic space, which, in the adult, forms an annular vein-
like vessel, tightly holding the muscle, about midway of its
length, and compressing its fibres into a small, round neck.
This neck lies nearly perpendicular to the anterior ventral por-
tion of the muscle, and somewhat in front of the posterior
extremity of that portion, the ventral fibres of the muscle
turning upward and then forward to enter it, as if, in the com-
pression of the fibres here, the dorsal ones had been more

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 699

closely and more tightly gathered in than the ventral ones,
these latter being thus left to project loop-like beyond the neck.

The anterior, ventral portion of the muscle lies nearly hori-
zontal, running forward from the neck of the muscle, immediately
dorsal to those portions of the geniohyoidei that have their in-
sertion in median aponeurosis of that muscle, and immediately
dorsal to the intermandibularis, but ventral to those portions of
the superior geniohyoidei that have their insertions as a tendin-
ous sheet on the inner surface of the mandible. It is double
throughout the greater part of its length, the two parts lying
close together at the middle line of the head, except near their
front ends, where they diverge slightly, and are inserted, in
part, on the dorsal surface of the hind margin of the interman-
dibularis, and in part beyond that margin, on the inner surface
of the mandible, near the symphysis. That portion of the
muscle that has its insertion on the mandible is mainly or
entirely tendinous, and forms a thin, tendinous sheet, which,
in the adult, is closely applied to, and not easily separated from,
the ventral surface of the anterior, tendinous ends of the genio-
hyoidei.

The posterior portion of the muscle runs upward and back-
ward from the neck of the muscle, in front of the anterior
tendinous margins of the hyohyoidei, and between the two
tendons of the sternohoideus, being compressed somewhat by
these latter tendons as it passes between them. It then sepa-
rates at once into two portions, which diverge to the right and
left, and are inserted on either side to the arms of a Y-shaped
tendon (Caa4, Figs. 46 and 47, Pls. XX XII and XXXIII), which
arises, by its median portion, from the symphysis of the clavicles,
or from the median aponeurosis of the sternhyoideus, and is
inserted, by its two arms, to the second and third hypobran-
chials on each side of the head. This ligament, therefore, corre-
sponds to one or both of those portions of the coraco-arcuales
anteriores of Acipenser (No. 125, p. 480) that have their in-
sertions on the second and third arches. As the insertion on
the third arch, in Amia, is apparently the most important, and
as the branchiomandibularis, in Acipenser, is in intimate relation
with the tendon to the third arch, I have called the tendon or

700 ALLIS. [VoL. XII.

ligament in Amia, as Vetter does the muscle in Acipenser, the
coraco-arcualis anterior¢.

The branchiomandibularis, in Amia, varies greatly in different
individuals. The posterior portion of the muscle, both in old
and young specimens, is as often single as double, the insertion,
when single, being to one arm only of the Y-shaped tendon, no
trace of the other arm of the muscle being found. In some
instances the second arm of the muscle is found reduced in
length, and ending without insertion of any kind, and in one
specimen both arms were wholly wanting, the muscle ending,
without insertion of any kind, a little beyond its neck. The
anterior portion of the muscle, which usually, in the adult, has, in
part, a tendinous insertion on the inner surface of the mandible,
sometimes extends only as far as the intermandibularis, and in
larvae it can never be traced definitely beyond that muscle.
The anterior, tendinous portion of the superior division of the
geniohyoideus, that portion that has its insertion on the mandi-
ble, can also not be traced in larvae. In 10 mm. specimens, the
branchiomandibularis extends but little beyond the large vas-
cular space through which the muscle passes, and is there lost
in the general tissues. In one young specimen it passed above
the fold in the lining membrane of the mouth cavity at the base
of the tongue, instead of below it, as it is usually found, and
was lost in the connective tissues toward the tip of the hyoid
and at the base of the tongue. It seems therefore prob-
able, that the tendinous insertions of the geniohyoideus and
branchiomandibularis on the inner surface of the mandible are
acquired in postlarval, or even in adult life, Acipenser (No. 125,
p. 480), where the branchiomandibularis is firmly attached to
the integument of the floor of the mouth, representing an
intermediate stage. The general relations of the branchioman-
dibularis to the tongue, in larvae, and the actual termination
of the muscle in one young Amia in that organ, seem also
to corroborate McMurrich’s statement (No. 76, p. 152) that
that muscle, in lower forms, corresponds to the muscle from
which the tongue muscles of Amphibia are developed. The
tongue of the adult Amia may, therefore, represent a condition
of that organ in which its partial muscularization has been lost,

No. 3.} MUSCLES AND NERVES IN AMIA CALVA. 7O!I

rather than not yet acquired, as Gegenbaur (No. 45) states to
be the case for fishes in general. That such a condition should
be found in Amia is only what might reasonably be expected
if amphibians and ganoids both descend from some proto-
urodelian type, as the muscles and nerves of the eyeball indi-
cate (Fig. 12, Pl. XXII). It is perhaps worthy of note, in this
connection, that the branchiomandibularis is found in Acipenser
(Vetter) and Polypterus (Pollard), that it is not found in teleosts,
and that it is found, undifferentiated, as Muscle Cm1, or Cmz,
in Chimaera (Vetter).

The branchiomandibularis is innervated, in Amia, by a deli-
cate terminal branch (7.6m) of the fused ventral branches of
the first, second, and third occipital nerves.

2. Sternohyoideus.

The sternohyoideus (S%, Figs. 33-35, 46, and 57, Pls.
XXVITI-XXXV) arises from the dorsal and anterior surface of
the lower half of the clavicle, the surface of origin extending
from the symphysis of the bones up to the place of origin of
the pharyngo-claviculares. It is a strong, conical muscle lying
immediately below the bulbus arteriosus, and tapering grad-
ually into a stout, rounded tendon which is inserted on the
ventral surface of the ossified portion of the hypohyal. It is
closely united with its fellow of the opposite side of the head
from its origin forward, through about one half its length, the
two muscles being separated by a median aponeurosis, from
which the fibres of both in part arise. This aponeurosis is
strongest toward the dorsal surface of the muscle, and from its
anterior end arise the two arms of the tendon Caa+, to which
the branchiomandibularis is inserted.

Through each sternohyoideus there are two transverse inter-
muscular septa, resembling the intermuscular septa of the
trunk muscles, and dividing each muscle into three parts.
The middle point of each septum (Fig. 35) is pulled forward
into a pocket, and on the anterior face of the septum, around
the pocket, there is a conical tuft of connective tissue. A
similar tuft is found on the corresponding point of the ante-

702 ALLIS. Vion. sxaily

rior face of the clavicle. In the skin on the outer surface of
the muscle, opposite the posterior septum, the anterior end of
the posterior flagellum (//%) has its attachment. No special
attachment to the clavicle, such as Sagemehl describes (No. 106,
p. 63), could be found, nor were there what could be called
special muscles connected with it. Certain fibres of the sterno-
hyoideus were, however, inserted on the inner surface of its
anterior end, or base, as McMutrrich has described. The ante-
rior flagellum (//a) is attached to the skin alone, a little in
front of the anterior septum.

The trunk muscles, although projecting, internal to the
clavicle, slightly beyond the inner, curved edge of that bone
immediately above the sternohyoideus, are in no place con-
tinuous with that muscle. No indication was found of a sepa-
ration of the sternohyoideus into a hyodorsalis and hyoventralis,
such as Schneider describes in Esox (No. III, p. 116).

The sternohyoideus is innervated by the united ventral
branches of the first three occipital nerves, and by the ventral
branch of the fourth occipital nerve, or by a branch of that
branch. The branch of the fourth occipital nerve innervates
the posterior segment of the muscle, the first three nerves in-
nervate the two anterior segments, a terminal branch of these
united nerves being sent to the branchiomandibularis as already
stated.

The muscles of the pectoral fin are innervated by branches of
the ventral branches of the first six spinal nerves, and, in the
two dissections made, by a corresponding branch also of the
ventral branch of the last, or fourth, occipital nerve.

3. Occipital Region of the Skull.

The occipital region of the skull, in the adult of the lower
vertebrates, and of the head, in their embryos, is not defined
and limited in the same way by different authors. Gegenbaur
(quoted No. 104, p. 189) defines it, definitely, in adult sela-
chians, as that part of the skull that lies posterior to the vagus
foramen; and Froriep (No. 34, p. 226), in chick embryos,
as that part of the vertebral column that is included between

No: 30) MWUSCEES, AND NERVES [IN AMIA CALVA. 703

the first cervical nerve and the vagus. By these definitions
that part of the skull that lies between the glossopharyngeal
and vagus foramen, in the adult, and the glossopharyngeal
somite or segment of the head, in embryos, are definitely
excluded from the region. In selachian embryos the glosso-
pharyngeal somite is similarly excluded from the occipital
region of the head by van Wijhe (No. 130), Kaestner (No. 62),
Killian (No. 63), and others, Hoffmann (No. 57) even excluding
with it the first vagus somite. Sagemehl (No. 104, p. 189), in
opposition to these views, defines the occipital region of the
skull, in Amia and in all bony ganoids and teleosts, as that
part of the skull that lies posterior to the glossopharyngeal
foramen and to the hind edge of the petrosal, the reason
given for this definition being that, in some rare instances in
teleosts, the glossopharyngeus and vagus issue through the
same foramen. Mollier (No. 82), in embryos of reptiles, gives
practically the same definition as Sagemehl, for he takes the
hind edge of the auditory vesicle, which lies immediately in
front of the giossopharyngeus, as the anterior limit of the
occipital region of the skull. Froriep, in a recent work on
embryos of mammals, birds, and reptiles (No. 35, p. 575)
apparently assigns the same anterior limit to the region, though
he does not definitely so state. This latter definition differs,
therefore, from his earlier one. According to it, and to the
definitions of Mollier and Sagemehl, the glossopharyngeal
somite must either be absorbed by the auditory vesicle or be
included in the occipital somites of the head. In either case,
occipital and postauditory would be, in the adult, synonymous
terms, and as such I consider them in Amia in all stages
included in this investigation.

In the interior of the skull of the adult Amia, there is a
large postauditory portion, or chamber, occupying almost one
third the length of the cavum cranii. This chamber is sepa-
rated from the labyrinth recess by a ridge of cartilage which
projects, in a curved line, forward and inward from the inner
surface of the lateral wall of the skull. This ridge of cartilage
is, in young larvae, the bulging hind wall of the auditory cap-
sule and of the skull, for, as will be later shown, that part of

704 ALLIS, (Vou. XII.

the skull that lies posterior to the capsule develops, in its dor-
sal portion, later than, and independently of, that capsule, and
then fuses with it to form the occipital or postauditory region.
Whether the skull of these young larvae, that is, that part of
the skull of the adult that lies anterior to the hind wall of the
auditory capsule, is the protometameric skull of Sagemehl (No.
105, p. 527), the equivalent of the skull of adult amphibians
and selachians, or is equivalent to that skull less the occipital
vertebra of Stohr, I am unable to judge, not having been able
to see the particular publication referred to by Sagemehl. My
work, however, strongly inclines me to accept the assertion
attributed to Stohr, that a vertebra, lying originally between
the vagus and first spinal nerves, has fused with the hind end
of the amphibian skull, and also to consider the skull, as first
formed in young Amia, as the true protometameric skull of Sage-
mehl, but as something less than that skull as defined by him.
The anterior limit of the ventral portion of the postauditory,
or occipital chamber of the adult Amia is marked by the trans-
verse ridge of cartilage that forms the hind wall of the eye-
muscle canal, and gives support to, and doubtless also origin
to, the median processes of the petrosals. The antero-ventral
end of the lateral bounding ridge of the chamber, the hind
wall of the labyrinth recess, lies opposite the ventral ridge,
and opposite, or immediately in front of this point, the glosso-
pharyngeus and vagus have their origins from the brain. The
glossopharyngeus perforates at once the median, membranous
wall of the auditory cavity, and transverses that cavity to reach
its foramen. The vagus, on the contrary, lies always median
to that membranous wall, running outward and backward along
it, and then along the posterior surface of the hind wall of the
labyrinth recess, to its foramen, which lies immediately behind
the point where the wall or ridge arises from the inner surface
of the skull. The nerve of the lateral line has a similar course.
Both nerves, therefore, lie morphologically behind the hind
wall of the auditory chamber, or, in larvae, behind the hind
wall of the skull. In the occipital chamber the occipital
nerves of Sagemehl arise, all of them running outward and
backward, in the chamber, to their points of exit.

No. 3.]} MUSCLES AND NERVES IN AMIA CALVA. 705

The course and position of all these nerves indicate that the
skull in its posterior portion, as in its anterior portion, develops
more rapidly than the inclosed brain, and that posterior to
some central point, approximately the nervus acusticus, the
issuing nerves are, by that development, pulled or pushed
backward, just as in front of that point they are pulled forward.
The glossopharyngeus is an exception to this rule, for it runs
almost directly outward through the auditory cavity instead of
behind it. The auditory vesicle, which ‘develops between the
facial and glossopharyngeal nerve roots’’ (No. 4, p. 232), seems,
in fact, as it developed, to have invaded and occupied the
glossopharyngeal region of the head, and, in so doing, to have
pushed the vagus and lateral nerves before it, but to have left
the glossopharyngeus in its natural position, the sinus utriculi
posterior pushing backward above that nerve, and the sacculus
below it. In the tadpole, the glossopharyngeus is apparently
not treated with similar consideration, for Strong shows it
lying, with the vagus, posterior to the auditory capsule.

The brain, in the adult Amia, does not lie on the floor of the
occipital chamber; on the contrary, it traverses the chamber
at such a height that about one fourth the chamber lies below
its ventral surface. The extreme posterior end of this ventral
portion of the chamber ends (Fig. 11, Pl. X XI) in a blunt point
in the line of the center of the vertebral column, and hence, of
the center of the notochord. This part of the chamber lies
on a level with the main portion of the eye-muscle canal, and,
like that canal, is developed in late larval or postlarval stages.
It is filled, in the adult, with fatty tissue.

In the walls of the occipital chamber there are three large
ossifications, which extend from the outer to the inner surface
of the skull. They are the two occipitale laterale, the exoccip-
itals of Bridge, one on each side of the head, and the median
basioccipital.

The basioccipital (BO, Figs. 8-11, Pl. XXI) is described by
Sagemehl as having the shape of a muscle-shell. It is a stout
bone, has approximately the length of the first six vertebrae,
and its posterior third is solid and shaped like a vertebra
(Sagemehl). Both ends of this solid portion are concave, and

706 APLTS. VO Ie

its posterior portion is marked off from the rest of the bone by
a groove, sometimes faint and sometimes well defined, which
runs across the ventral surface of the bone and then upward
and forward along its sides, but not across its dorsal surface.
In front of this groove, on the ventro-lateral surfaces of the
bone, there are two large, oblong, flattened surfaces, one on
either side of the middle line. They are inclined at an angle
to each other, raised somewhat above the level of the rest of
the bone, and extend to its anterior end. On them the
hind ends of the parasphenoid rest, and the development of
this part of the skull shows that these flattened elevations, in
their posterior portion at least, must be, like portions of the
vertebrae, of membranous, and not of cartilaginous, origin.
Above the dorso-lateral edges of the two surfaces, and in front
of the posterior, solid end of the bone, the basioccipital becomes
thin, and its thin upper edges diverge from behind forward.
Between the two surfaces, on the ventral surface of the bone,
there is a median, depressed portion, or groove, and at the
anterior end of this groove, in the middle line of the head and
extending directly forward, there is a long, flat, thin process,
which, viewed from beneath, has the appearance of a thin
wedge or blade of bone driven into the basioccipital from
its ventral surface, and slightly separating the bone into two
halves. The anterior end of this process almcst reaches the
posterior end of the hypophysial fenestra.

Between the hind ends of the two flattened and slightly
elevated ventro-lateral surfaces of the basioccipital, and hence
between the hind ends of the parasphenoid, are the two little
openings (caz) described by Sagemehl (No. 104, p. 195), one
on each side of the middle line of the head. From each of
them arise one or two canals which run outward, upward, and
forward, and have their dorsal opening, or openings, at about
the middle of the lateral surface of the basioccipital, dorsal to
the flattened, ventro-lateral surface of that bone, and hence
dorsal to the lateral edge of the parasphenoid. The anterior
of the two canals, where there are two, is the smaller, and
through it a small sympathetic nerve passes; through the
posterior and larger canal a spinal artery passes, This artery

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 7O7

arises from the dorsal aorta, traverses its canal, runs upward,
outward, and forward along the lateral surface of the basioc-
cipital and then of the occipitale laterale, and then inward
along the dorsal surface of the latter bone. It apparently
receives branches from all the prespinal muscle segments, but
this was not definitely established. In 20 mm. larvae it lies
wholly outside the cartilage of the base of the skull, passing
through a region where membrane bone is forming, internal to
the parasphenoid.

Immediately posterior to the ventral openings of the two
canals, above described, there are, on the ventral surface of
the basioccipital, on each side of the middle line, one or two
slightly projecting processes of cartilage (vp°). The posterior
of these two processes is always found, and has been described
by Sagemehl (No. 104, p. 195). The anterior one, which is
sometimes wanting, and is sometimes fused with the posterior
one, to form a single elongated piece or process is not
described by him. The posterior process lies on the posterior,
vertebra-like, terminal piece or portion of the basioccipital ; the
anterior one lies immediately in front of the groove or line sep-
arating that posterior portion from the anterior portion of the
solid end of the bone. Sagemehl states that these little cartil-
aginous processes correspond to similar processes found on the
ventral surfaces of all the vertebrae. This statement I can
confirm for the six anterior trunk vertebrae. On the vertebrae
of the tail, which unfortunately I have been unable to examine, I
am inclined to think the processes become the ventral arches,
which is evidently not what Sagemehl intended to state. It
seems to me also extremely probable that these processes, con-
trary to Sagemehl’s statement (No. 102, p. 516), are the homo-
logues of the pharyngeal process of Cyprinoids, and that these
latter processes are thus also the homologues of the ventral
arches of the tail. One reason given by Sagemehl for not so
considering them is, that the pharyngeal process in Chondros-
toma nasus is not performed in cartilage. This may be, in part,
an error, as are certainly his statements that the intercalar in
Amia is so performed, and that the prefrontal in Amia is in part
a dermal bone. The other important reason given by him for

708 ALLIS. [Vou. XII.

not considering the pharyngeal process as formed by the fusion
of ventral arches, is, that the ribs in Amia are ventral arches,
and that they inclose the body cavity instead of simply the
aorta, as the process does. As will be seen later, the statement
that the ribs in Amia are ventral arches is not wholly above
question. If the pharyngeal process of Cyprinoids is not
homologous with the cartilaginous processes of Amia, it must
have its homologue in Amia in the bony pads that give sup-
port to the hind ends of the parasphenoid.

On the dorsal surface of the solid end of the basioccipital,
extending longitudinally the full length of that end, there is,
on each side of the middle line, a projecting strip of cartilage
(dp’, Fig. 18, Pl. XXII), which has its origin slightly below
the outer surface of the bone. This strip is thickest along
its median edge, and has, extending transversely across it, two
wedge-shaped ridges. The anterior ridge is much the larger
of the two, its summit corresponding with the hind edge of
the occipitale laterale. The posterior ridge lies on the dorsal
surface of the posterior, vertebra-like, terminal portion of the
basioccipital. Between the two processes lies the anterior
occipital dorsal arch (DA?) of Sagemehl; between the poste-
rior process and the first vertebra, the second or posterior
occipital arch. This second occipital arch of Sagemehl,
although fused with the hind end of the skull, is possibly,
though not probably, the first spinal arch, as will be later
shown.

From the lateral surface of the basioccipital, immediately
dorsal to the parasphenoid, two ligaments arise (oc* and Joc*),
one from the posterior, vertebra-like, terminal portion of the
bone, and the other just in front of that terminal portion. The
posterior ligament is much the stronger of the two. It runs
outward, downward, and backward in the first prespinal or
last occipital intermuscular septum (Figs. 33-35, Pls. XXVIII
and X XIX), and is inserted in dense tissue lying between the
lower, posterior end of the supraclavicular, and the upper end
of the clavicle, the attachment being rather to the former, than
to the latter, bone. The anterior ligament runs outward, down-
ward, and backward in the second prespinal intermuscular

No. 3.] WUSCLES AND NERVES IN AMIA CALVA. 709

septum, and is inserted on the inner surface of the supracla-
vicular. In young larvae the large ligament is well developed
and easily traced in sections ; the small ligament, on the con-
trary, could not be distinguished from the septal tissues in
which it lies. In one larva the large ligament lay in the
second prespinal septum instead of in the first.

The occipitale laterale (LO) has two exposed surfaces, a
lateral and a dorso-posterior. The two surfaces lie almost at
right angles to each other, and form, where they unite, the
dorso-lateral edge of the bone and of the hind end of the
skull. The posterior and lower end of this dorso-lateral edge is
thickened, and rests upon, or abuts against, the anterior face of
the anterior wedge-shaped process of the cartilaginous strip
on the dorsal surface of the basioccipital. The occipitale
laterale is thus separated entirely by cartilage from the solid
vertebral portion of the basioccipital, excepting, possibly, along
its extreme outer edge. The antero-dorsal end of the dorso-
lateral edge of the bone is also thickened, and that portion
of the lateral surface of the bone that lies between this thick-
ened upper corner and the upper edge of the vagus foramen is
raised slightly above the level of the rest of the bone, and has
its outer surface flat and roughened. On this roughened sur-
face, which resembles, in general appearance, the surfaces
found elsewhere, in Amia, between the dermal and cartilagi-
nous bones, the lower, posterior corner of the intercalar rests.
The intercalar thus, in this particular, presents the character-
istics of a membrane rather than of a cartilage bone. It is
also easily detached from the skull, as the dermal and mem-
brane bones are, and it leaves beneath it, when removed, a
clean, unbroken surface. In larvae its first appearance is
decidedly that of a membrane bone. Sagemehl’s positive
statement (No. 107, p. 556) that it is, in Amia, a primary
ossification is, therefore, greatly to be doubted. It is much
more probable that it is of exactly the same character as the
corresponding bone in the Cyprinidae, that is, simply a “ Deck-
knochen.”’

In front of the thickened postero-ventral end, or base of the
occipitale laterale, the ventral edge of the bone, is thin, and

710 ALLTS. [VoL. XII.

rests directly upon the thin, dorsal edge of the anterior portion
of the basioccipital. Through this thin portion, and hence in
front of the thickened base of the bone, there is, as Sagemehl
states, a foramen (ofr?) giving passage to the ventral root of a
spinal-like occipital nerve; the foramen is, however, strictly on
the lateral face of the bone instead of on its dorsal surface, as
his figure shows. In front of this foramen, slightly above it,
and about halfway between it and the vagus foramen, there was,
in all the specimens I examined, another very small foramen
(ofr') giving passage to another similar but very delicate ventral
root. Between these two foramen there is no thickening of the
bone, but opposite the space between them, on the median edge
of the bone, directly above the foramen magnum, there is a con-
siderable thickening. Along the lateral surface of the basi-
occipital, starting from the exit foramen of the spinal artery
already described, there is a slight depression or groove which
continues upward and forward onto the occipitale laterale, across
the posterior occipital foramen, and then onto the dorsal sur-
face of the bone. It apparently marks the course of the spinal
artery, and is apparently the line shown by Sagemehl in his
Fig. 3, and considered by him as of segmental value.

The first vertebra (V‘) is, as Sagemehl states, slightly
convex in front and strongly concave behind. It is thinner
longitudinally than the second vertebra, the second is thinner
than the third, and the third thinner than the fourth. The
progression is regular, and there is no reason whatever, from
their dimensions alone, to consider the two first vertebrae as
half vertebrae (Schmidt, No.110). The first six vertebrae, the
only ones that were examined, had each, on each side, a dorsal
and a ventral cartilaginous process (ap and vf), and all except-
ing the first had also a lateral, bony process (/p) directed down-
ward, outward, and backward, tipped with cartilage, and giving
articulation to a rib. Schmidt (No. 110, Fig. 2) shows the first
lateral process and the first rib on the third vertebrae. I
have always found them on the second. Arising from the
lateral surface of the first vertebra, exactly in line with the
lateral processes and ribs of the other vertebrae, there is
always a small ligament (/vZ), similar to and in line with those

No. 3:] “USCLES AND NERVES [N AMIA CALVA, 7it1

arising from the posterior end of the basioccipital, and, like
them and like the ribs, lying in an intermuscular septum.

The ventral cartilaginous process on each vertebra is small,
is round or oval, arises slightly beneath the surface of the
vertebra, and projects but slightly beyond that surface. The
dorsal process is much larger and much more important than
the ventral one. It arises, like the ventral one, slightly beneath
the surface of the vertebra, but it projects, at its median edge,
considerably above that surface. It extends longitudinally
across the dorsal surface of the vertebra, and is wedge-shaped,
the edge of the wedge lying transversely, as do the edges of
the wedge-shaped projections on the dorsal surface of the basi-
occipital. The posterior face of each process is intimately
united with a dorsal spinal arch (DA), so intimately, in fact,
that the two form a single piece and cannot be separated with-
out fracture. A surface line, however, indicates a plane of
separation between them.

The first six spinal arches, the only ones examined, are wedge-
shaped at their lower ends, and fit in between the wedge-shaped
dorsal processes of two adjoining vertebrae. They are there-
fore, in position, strictly intervertebral, as Franque described
them (quoted and contradicted by Schmidt, No. 110, p. 751).
They are capped above and below with cartilage, and on the
median side of the ventral end of each there is an indentation
which extends entirely through the cartilaginous cap of the
piece, and, on the sixth arch, slightly into the bone itself. On
the sixth arch the indentation or groove extends from the
median face of the arch almost to its lateral surface. The
cartilaginous cap on this arch thus had the appearance of two
oval pieces placed transversely, and connected, at their lateral
edges, by a narrow strip of cartilage. Whether this is an
indication that the arches are double in origin, as Goette (No.
49) finds them in reptiles, or not, I am unable to state.
Nothing else in their development, in Amia, from 12 mm.
larvae upwards, indicates such a double origin. The inden-
tation on all the arches is filled with ligamentous tissue con-
nected with, and doubtless derived from, the ligamentum longi-
tudinale dorsale inferior. The anterior face of the cartilaginous

7i2 ALLIS. [VoL. XII.

cap of each arch sits directly upon the posterior face of the
dorsal process of the next preceding vertebra, and is so inti-
mately connected with that face that, as already stated, it can-
not be separated from it, without fracture, even in specimens
that have been treated with acids. The posterior face of the
cap rests upon the anterior face of the dorsal process of the
next following vertebra, but is always separated from that face
by a thin, flat pad, which, in the one specimen examined, was
of tough, semi-cartilaginous tissue on the first two vertebrae,
but, to all appearance, of true cartilage on the posterior ones.
The pad or plate is, at its lower anterior end, continuous with,
or connected with, the ligamentous tissue filling the median
indentation in the base of the arch. It is also connected,
at its lateral edge, by tendinous tissue, extending downward
and backward along the lateral surface of the vertebra next
posterior to the arch, with the intermuscular septum of that
vertebra.

The intermuscular septum on each vertebra runs from the rib
of that vertebra directly upward toward the top of the vertebral
body, where it turns sharply backward onto and across the
lower, or lower and posterior, end of the dorsal arch that lies
next behind that vertebra. It then runs in a nearly horizontal
direction across the lower ends of the two dorsal arches next
posterior to that arch, onto the third posterior arch. There it
turns upward and backward along that arch, and onto the pro-
cessus spinosus of the arch. At the distal end of the spine it
turns sharply forward and upward, and reaches the dorsal sur-
face of the body about opposite the vertebra to which it belongs.
The angle formed where the septum first turns backward from
its own vertebra, onto and across the dorsal arch next posterior
to that vertebra, is much sharper on the anterior of the six verte-
brae than on the posterior ones. It is at this angle that each
septum receives the tendinous formation from the semi-carti-
laginous pad that lies between its vertebra and the dorsal arch
that lies next anterior to that vertebra.

In larvae from 20 mm. to 50 mm. in length, the six vertebral
processes of the adult, above described, are found as relatively
large blocks of cartilage, three on each side of each vertebra,

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 713

and they form, with the notochord, the principal part of each
vertebral body. Membrane bone has, however, even in 20 mm.
larvae, begun to form between the processes, and between
those of opposite sides. There is not the slightest indication
of a separation of the vertebra into two lateral halves, or into a
dorsal and a ventral half.

The dorsal process, on each side, is much larger than the
other processes, and is wholly independent of them. Of the
other two the lateral one is the larger, but it is not independ-
ent of the ventral one, the two being connected by a thin strip
of cartilage, which extends downward from the anterior end of
the lateral process to the posterior end of the ventral one. The
ventral processes lie on either side of the dorsal aorta, and, in
old larvae, are sometimes continuous at their hind ends, dorsal
to the aorta.

The dorsal arches lie, as in the adult, with their bases
wedged in between two succeeding dorsal processes, but they
are not continuous with either process. The fusion of the
arches, at their bases, with the dorsal process next in front
of them, is thus a secondary arrangement, arising in post-larval
stages. That this late and secondary fusion of the arch proper
with the vertebral body is not peculiar to Amia, and that it is
probably the regular method of development in all bony ganoids
and teleosts, if not in all animals, is shown by Sagemehl’s
statement (No. 107, p. 587) that even in the adult of the
Cyprinidae and Characinidae, the dorsal arches are not ‘fest
verbunden”’ with the vertebral bodies. The cartilaginous pad
found in the adult Amia, between each arch and the next
posterior dorsal process, is not found even in 50 mm. speci-
mens. It, also, must therefore develop in later stages.

The vertebral processes all rest, at their bases, upon a thin
layer of strongly refractive material, apparently bone, formed
around the notochord. This layer is considered by Schmidt,
in the adult Amia, as the “selbstandig verknécherte dussere
zellige Chordascheide”’ (No. 110, p. 754), and by Hasse (No.
52, Fig. 14), as the “‘ Faserscheide der Chorda,” a part of the
cuticula chordae. What is apparently the same layer in teleosts
and selachians is called by Géppert (No. 48) the ‘“Chorda-

714 ALBIS: [Vou. XII.

scheide,” while in Amphiuma it is considered by Field (No.
33, p. 346) as the cuticula sceleti. This layer in Amia, what-
ever its origin and character, is, with the notochord, in 30 mm.
specimens, strongly constricted. There are, however, no
breaks in it, as there are in Amphiuma, and no indication
whatever of intervertebral or intercuticular cartilage, as in
that animal. The vertebral constriction and subsequent seg-
mentation of the notochord in Amia are, therefore, due to the
compressive action of the cartilaginous processes, which always
lie directly opposite and in the constrictions, and not to an
intervertebral expansion of the notochord, as in Amphiuma
(Field). The vertebral processes in Amia are, accordingly, in
origin as well as in position, strictly vertebral; the dorsal
arches, which lie between the dorsal processes and directly
opposite the crests of the notochordal constrictions are, on
the contrary, so far at least as their position is concerned, as
strictly intervertebral.

These relations do not change in 14 mm. or 12 mm. larvae.
The dorsal processes on each side in 12 mm. larvae are, how-
ever, found forming a continuous line, the dorsal “ Langleiste”’
of Goppert (No. 48, p. 168). In this line the dorsal arches
are imbedded at their bases. The outlines of their bases
are, however, distinctly indicated, and they lie opposite the
places where later the crests of the notochordal constrictions
appear.

Whether there is also a continuous ventral line, such as
Goppert describes, formed by the fusion of the lateral processes,
or two such lines, one representing the line of the lateral proc-
esses, and the other that of the ventral processes, was not
investigated.

From the literature at my disposal I am unable to satisfac-
torily homologize the cartilaginous portions of the vertebrae in
Amia with those described by others in other animals. The
lateral processes of all, or at least most, fishes — Amia being
especially mentioned —are called by Baur (No. 7, p. 119) para-
pophyses, and are defined by him as the proximal, independent
ends of the ribs or haemapophyses. What, then, are the ventral
processes in Amia which Baur does not mention? They have

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 715

exactly the position of ventral arches, and structures resembling
them in every respect in larvae of Salamandra maculosa are
considered by Goppert (No. 47, p. 444) as probable remnants
of such arches. They are also, almost unquestionably, the
homologues of the two little bony processes described by
Goppert on the ventral surface of the body vertebrae of Cala-
moichthys (No. 48, p. 150). These processes in Calamoichthys
become, in the tail, the basal processes (Basalstiimpfe) of the
haemal arches, the remaining, distal portion of those arches
being formed by the “ Pleuralbogen,” or lower ribs, the haema-
pophyses, therefore, of Baur, which shift from the under surface
of the lateral processes onto the distal ends of the ventral
ones. In the trout, in marked distinction to Calamoichthys,
no such shifting process takes place (No. 48, p. 187). The
lower ribs, or Pleuralbégen, in that fish, simply disappear, and
the haemal arches are formed by direct and independent growth
from the ventral surface of the base of the Basalstiimpfe, the
parapophyses of Baur. Judging simply from Schmidt’s figures,
Amia must agree with the trout in this respect. The haemal
arches in the tail of Amia would then be homodynamous with
the ventral processes found on the trunk vertebrae, and with
those processes only. The ribs would then take no part in the
formation of the haemal arches in Amia, contrary to what
Goppert states to be the case in all ganoids excepting Calamo-
ichthys (No. 48, p. 153), and would agree with those of teleosts
(No. 48, p. 161). The ribs in Amia may, therefore, be the
homologues of the upper ribs of Calamoichthys, and not of the
lower ones. If this be so, the ventral processes in Amia are
the parapophyses of Baur. The lateral processes would then
of necessity be, as in Polypterus, diapophyses (Baur), and the
ribs pleurapophyses or true ribs.

Goppert states (No. 48, p. 153), based on the observations
of Balfour and Parker, that the ribs in Lepidosteus present some
of the characteristics of upper ribs ; and Scheel states (No. 109,
Abstr.) that the ribs in teleosts are such ribs, and not lower
ribs. The ribs in teleosts lie in the lining wall of the peritoneal
cavity. If they be upper ribs they must therefore have lost
the position assigned by origin to them, that is, in the horizontal

7110 ALLIS. [Vou. XII.

muscular septum where that septum is crossed by the transverse
intermuscular septa (Rabl, Goppert). In Lepidosteus, on the
contrary, they still lie in the horizontal septum. A condition
intermediate between these two is indicated in the anterior ribs
of larvae of Amia, which lie buried, at their proximal ends, in
the trunk muscles, approximately at the level of the horizontal
septum. AQ still less advanced condition is shown in the rib-
like ligaments which lie in front of the anterior ribs in Amia,
for although lying below the horizontal septum they lie, even in
the adult, entirely in the trunk muscles.

The horizontal muscle septum is said to be a relatively late
acquisition (Rabl, quoted No. 48, p. 166). Being such, one
would naturally suppose that the “ Anlagen”’ of the upper ribs
would lie in the transverse septa, immediately ventral to the
horizontal septum, rather than in the latter septum itself. Why,
then, should a relatively slight difference of position, in those
septa, be considered of such great morphological importance
that ribs found in the horizontal septum should be called upper
ribs, and ribs found out of it, in totally different classes or
orders of animals, lower ribs? Goppert says (No. 48, p. 196)
that “Gleichzeitig mit dem Auftreten der Pleuralbogen hebt
sich das Niveau des horizontalen Septums.”’ Why not say that
as the horizontal septum began to rise in level, the ribs, devel-
oped primarily in connection with it, began to descend, relatively
to it, and finally acquired a position in the wall of the peritoneal
cavity? That in sucha process, in certain fishes, small remnants
of the original cartilaginous “‘ Anlage” should remain attached
to the horizontal septum, and there appear as cartilaginous parts
of the “ Seitengraten,” or as the cartilagines intermusculares,
would not seem improbable ; nor that, in other fishes, ribs
should be occasionally found on the first tail vertebrae, as in
Salmo (No. 48, p. 160). As the true upper ribs thus acquired
a new position, the lower ribs, which Géppert considers as the
more primitive structures, would naturally disappear. They are
still retained in Polypterus and Calamoichthys, which thus, in
this as in many other things, show a more primitive organi-
zation than that of Amia or the teleosts. In selachians also,
remnants of them seem sometimes to be still retained, as in the

No. 3.] MUSCLES AND NERVES I[N AMIA CALVA. Fees

little pieces of cartilage described by Gegenbaur in Cestracion
(quoted No. 48, p. 154).

For the dorsal processes of Amia Baur gives no name, and
I find no description, that I can satisfactorily identify, of such
processes in any fish or other animal. They may be centra or
pleurocentra, or the interarcualia of Klaatsch in Chimaera
(No. 64, Aédstr.), or possibly the basi-dorsals of Gadow and
Abbott (No. 39). They are certainly not rudimentary dorsal
arches (Schmidt), nor can they be considered as the proximal
ends of dorsal arches, for they are segmented, after the forma-
tion of those arches, from a continuous cartilaginous layer, or
« Anlage,” and they lie, in larvae, between the dorsal arches.
The arches themselves are certainly neuropophyses, but they
are not, in their origin, processes of the vertebrae. The whole
subject is to me much too obscure to venture upon names.

The dorsal portions of the anterior intermuscular septa, in
larvae of Amia, differ considerably, in their relations to the
vertebral processes, dorsal arches, and dorsal spines, from what
has been described in the adult. In 12 mm. larvae they run
downward and forward along the antero-lateral edges of the dor-
sal arches, then turn sharply forward, practically at a right angle,
as is well shown in vertical sections, onto the dorsal process
in front of the arch to which they belong, and then downward,
at a right angle again, onto the corresponding lateral process
and rib. Ventral to the rib they turn backward, and seem to
have their attachment, or origin, in the region where, later, the
crest between two constrictions of the notochord will appear.
They are thus, at this age, restricted to a single dorsal arch,
and are, in their central portions, vertebral in position. In
their dorsal portions they are intervertebral in position, as the
dorsal arches themselves are.

Between the age represented by 12 mm. larvae, and the adult,
there must, therefore, have been a gradual backward displace-
ment of the middle portion of that part of each of the anterior
intermuscular septa that lies above the level of the vertebral
column, the displacement becoming continually more pronounced
up to its complete development. The amount of this displace-
ment certainly varies, not only at different ages, but also in

718 ALLTS. [Vor cat

different parts of the vertebral column and of the body. Where
it first begins, at what age, and the amount of it at different
ages, I have not attempted to determine, as it belongs more
properly to a later study. It certainly begins at very early
ages, for even in 12 mm. larvae it is already plainly evident.
The course alone of the septa, in part vertebral and in part
intervertebral, sufficiently indicates this, for it it generally ac-
cepted that the septa are, primarily, through their entire length,
strictly vertebral in position. As the dorsal arches arise in the
septa, even these arches must, in 12 mm. larvae, have shifted
somewhat from their original position. In this shifting, the
arch and septum would naturally carry with them the nerve
that innervates the muscle segment next posterior to them.
Such at least is the case in Amia ; for in larvae, as in the adult,
the nerve that issues behind an arch innervates the muscle
segment that lies behind the septum of the vertebra next
anterior to that arch. My work thus tends to confirm Koll-
mann’s conclusion (quoted No. 24, p. 20) that the dorsal arches
in human embryos shift backward relatively to their vertebrae.

Schmidt, from an anatomical and histological investigation
of the vertebral column of the adult Amia, concludes (No. 110,
pp. 751 and 755) that there were, originally, in each body seg-
ment, two vertebral “Anlagen.” These “Anlagen”’ in the
trunk fuse to form a single complete vertebra, the dorsal arch of
the posterior half-vertebra, or intercentrum, remaining as the
dorsal arch of the complete one, and that of the anterior half-
vertebra, the centrum, persisting as a cartilaginous process,
on the dorsal surface of the vertebra, immediately in front
of that arch. This process, or rudiment of a dorsal arch,
as he considers it, can be nothing more nor less than the
anterior half of what I have described as the dorsal cartila-
ginous process of the vertebra, a part of that vertebra and not
of an arch. In larvae, as in the adult, there is not the slightest
indication of an arch that has disappeared, or of two vertebral
bodies that have fused. Schmidt’s conclusion, which simply
confirms Ihering’s earlier statement (No. 61), thus finds no
confirmation in my work, so far as it has gone. Regarding
Gadow and Abbott’s conclusions (No. 39, p. 298), based on

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 719

the study of a 57 mm. Amia, I can form no opinion as yet.
It is to be noted, however, that they consider the precentrum,
or arch-bearing element of a complete vertebra, the intercen-
trum, therefore, of Schmidt, as the anterior of the two elements
of a complete vertebra, instead of the posterior. This anterior
element of the complete tail vertebra is described by them
as bi-protovertebral or bi-myomeric in origin, being formed by
the fusion of “the basidorsal of one sklerotome and the basi-
ventral of the next previous sklerotome.’”’ The posterior, arch-
less disk of a complete tail vertebra, the postcentrum, is formed
by the interdorsalia and interventralia of the same sklerotome.
“The intermuscular septum runs obliquely across the precent-
rum.” The complete trunk vertebrae are thus, according to
them, partly bi-myomeric, instead of being myomeric, as
Schmidt finds them ; and the dorsal arch that belongs to any
particular vertebra must be, in 57 mm. specimens, and hence
also in the adult, the one that lies in front of that vertebra
instead of the one that lies behind it. Gadow and Abbott,
therefore, assign the arches to the vertebrae next posterior to
the ones to which Schmidt assigns them. Schmidt’s con-
clusions in this respect agree with the earlier statements of
Sagemehl (No. 104) and Ihering (No. 61).

Although my own work as yet warrants no opinion on this
subject, I cannot avoid the impression that each “ Anlage”’ of
a dorsal arch, arising, as it is said to, from parts of two adjoin-
ing somites, must, because of this double origin, have inherently
in itself the possibility of forming two separate pieces ; and
that these two pieces may remain separate, giving origin to
two arches, or that they may fuse, or that one of them may
disappear, thus giving origin in either case to a single arch.
The single arches so formed would doubtless always remain
septal at their bases ; they would, however, tend, because of
their different origin in different species, to occupy different
positions on, or relative to, their vertebrae. Those vertebrae,
also, even in nearly related forms, may not be homologous
structures, for Gadow and Abbott (No. 39, p. 298) state that
the vertebrae of Lepidosteus are truly bi-myomeric, while those
of Amia are not so.

720 AIMELS. [Vou. XII.

In addition to these possibilities of some difference in the
developmental composition of the parts concerned, there is the
strong probability, already explained, that the dorsal arches
may, in early stages of development, shift somewhat, with their
septa, relatively to their vertebrae. Can a dorsal arch, then,
be definitely assigned in every case, without proper investi-
gation, to the vertebra to which it has, in the adult, its apparent
attachment? It seems to me not, and yet it is evidently of the
first importance to know, definitely, to which of two adjacent
vertebrae a dorsal arch should be assigned, when the vertebral
composition of the occipital part of the skull is under consider-
ation. It is also of first importance in the naming or number-
ing of the spinal and occipital nerves.

The spinal nerves are usually assigned to the vertebrae that
bear the dorsal arches next in front of them, and named in
consequence. This does not, however, apply to the com-
plete serial numbering of the nerves, for the nerve in front of
the first vertebra, between that vertebra and the hind end of
the skull, is usually counted as the first spinal nerve. Contrary
to this generally accepted method Willis considered it in man
as the last cranial nerve, and Froriep (No. 34, p. 230) considers
the corresponding nerve in the chick as the last occipital nerve,
or nervus postoccipitalis. That the nerve, in Amia, is in reality
the first spinal nerve, and not the last occipital one, seems
certain if Ihering is correct in his statement (No. 61) that
in the tail of Amia, the nerve of each segment issues in
front of the arch of that segment. It is evident that the
general distribution of the nerve itself can give no definite
solution of the problem, for the nerve belongs, in that distri-
bution, to the muscle segment between two vertebral septa,
and not to the vertebrae to which those septa were primarily
attached.

Unable from my own work to arrive at any satisfactory
conclusion on this subject, I follow, for the present, Schmidt
in the serial numbering of the arches, and Ihering — that is, the
generally accepted method — in the numbering of the nerves.
The first spinal arch in all my figures and descriptions is, there-
fore, the arch that lies between the first and second vertebrae,

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. Jan

the first spinal nerve the nerve that lies between that arch and
the hind end of the skull.

In front of the first spinal dorsal arch, defined as above,
between it and the occipitale laterale, there are still, in Amia,
two dorsal arches, the occipital arches (DA?) of Sagemehl.
The anterior of these two arches rests in the hollow between
the two transverse wedge-shaped ridges on the cartilaginous
strip on the dorsal surface of the basioccipital, the posterior one
in the hollow between the posterior ridge and the dorsal cartila-
ginous process on the first vertebra. These transverse occipital
ridges are, therefore, dorsal vertebral processes, morphologically
similar to the dorsal processes of the vertebrae. The posterior
occipital arch is fused with the posterior face of the posterior
process, while the anterior arch is fused, not with the posterior
face of the anterior process, but with the hollow between the
two processes. This hollow lies, as already stated, in the
plane of the slight groove or line that separates the posterior,
vertebra-like end of the basioccipital from the rest of that
bone.

Between the dorsal end of the two pieces forming each occipi-
tal arch, there was, in all young specimens, a processus spinosus
(PSP), performed, in part at least, in cartilage, as are the cor-
responding processes on the spinal arches. In some of the
adult specimens there were, also, two such occipital processes,
as Sagemehl has already stated (No. 104, p. 190). In others
there was but one, the process, or plate (Sagemehl), in such cases,
having two upper ends or spines, one directed backward and
the other forward. No separate piece was found, in any speci-
men, between the first spine and the occipitale laterale, as
shown in Sagemehl’s Fig. 3 (No. 104), and described by him in
a footnote to page 190. Sagemehl suggests (No. 107, p. 524),
that one of these two occipital spines may fuse with, or become
incorporated in, the skull of teleosts and Amniota, and there
give origin to the supraoccipital. He also makes the general
statement (No. 107, p. 526), that the latter bone is found in all
vertebrates in which one or more nerves issue from the cranium
posterior to the vagus. This is certainly not entirely true, for two
such nerves issue from the cranium of Amia, and Amia does

722 ALLIS. [Vou. XII.

not possess the bone. Sagemehl himself was well aware that
one such nerve is found in Amia, for he himself first described
it. Doubting the origin of the bone assigned to it by Sagemehl,
I am inclined to consider it as primarily of membranous origin,
developed in connection with the tendinous attachment of
muscles, as the intercalar evidently is. That it should not be
developed in ganoids, dipnoids, selachians, or amphibians is not
singular, if the schema given in Fig. 12, Pl. XXII, is correct.
The bone is, in that case, simply a late acquisition, and hence
found naturally, and developed independently, in teleosts and
Amniota.

The intermuscular septum that has its attachment on the
posterior occipital arch runs downward onto the vertebral-like
end of the basioccipital, and then onto the large ligament that
arises, rib-like, from that end. The intermuscular septum that
has its attachment on the anterior occipital arch runs down-
ward in front of the vertebra-like end of the basioccipital, onto
the small anterior occipital ligament.

Starting from the hind end of the skull, partly from the
cartilage of the skull, and partly from the adjoining median
edges of the two occipitalia lateralia, and running directly
backward between the dorsal arches, there is a strong liga-
mentum longitudinale vertebrale dorsale superior (/vd’, Fig. 9,
Pl. XXI). In the 50 mm. larva examined in transverse sections,
there were, immediately below this ligament, between the two
halves of the last dorsal spinal arches included in the sections,
two independent pieces of cartilage, one on each side, forming
a bridge between the arches, and between the ligament and the
spinal cord. These little pieces are described by Scheel (No.
109) in Trutta and other teleosts, and are considered by him
as the “urspriingliche Fortsetzung der Neuralbogen.”’

The ligamentum longitudinale dorsale inferior (4vd’, Fig. 18,
Pl. XXII) starts from the front edge of the dorsal surface of
the solid end of the basioccipital, and runs backward between
the dorsal occipital processes, and then between the dorsal
spinal processes. From it arise, as already described, the little
lateral projections that fill the indentations in the bases of the
dorsal arches.

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 723

We thus have in the solid end of the basioccipital of the adult
Amia all the parts of two vertebrae, or vertebral segments ;
the longitudinal ligaments, the intermuscular septa, the lateral
processes and ribs represented by the two occipital ligaments,
the dorsal arches, the dorsal processes, and the ventral proc-
esses. That still other vertebrae also have fused with the skull
anterior to these two vertebrae is indicated by the nerves and
muscle segments, and by the development of the occipitale
laterale.

4. Occipital Nerves and Corresponding Muscle Segments.

From each of the spinal ganglia (gsf) in the adult Amia four
well-developed nerves or branches arise (Fig. 35, Pl. XXIX):
a dorsal branch (spd), which runs upward and backward along
the posterior face of the septum in front of the segment to
which the nerve belongs ; a horizontal branch (sp), which runs
outward to, and then outward and backward along, the same
face of the same septum, lying always above the horizontal
septum between the dorsal and ventral muscles; a small
branch (spa), which runs upward and backward, but mor-
phologically forward, through the septum in front of the
segment to join the dorsal branch of the second anterior
spinal nerve; and a large ventral branch (spv), which runs
downward and backward in front of the septum posterior to its
segment, along the anterior edge of the rib that lies in that
septum.

In larvae the spinal ganglia lie wholly on the dorsal root of
their nerve. The ventral root issues close to the ventral end
of the ganglion, but does not touch it. Here it gives off the
dorsal and horizontal branches above described, which are,
respectively, the ramus dorsalis and ramus ventralis lateralis of
von Lenhossék (No. 73, p. 173). It then turns downward to
form the principal part of the large ventral branch of the nerve,
the ramus ventralis medialis of von Lenhossék. The remaining
portion of this ramus medialis is formed by a large nerve which
arises from the ventral end of the ganglion of the dorsal root,
and runs downward and slightly outward in front of the dorsal
and horizontal branches of the ventral root. No fibres of this

724 Teds. [VoL. XII.

nerve were seen going to either of the other two branches of
the ventral root, but they may nevertheless exist, as in the
chick (von Lenhossék), for my sections not being especially
prepared for such work, I made no special search for them. I
also did not attempt to find whether the small communicating,
or anterior, branch of the nerve arose from the dorsal or from
the ventral root. This communicating branch in young larvae
runs directly forward instead of upward and backward as in the
adult. The sympathetic ganglia did not lie, as in the chick,
upon the ventral branches of the ventral roots. From the dorsal
end of the spinal ganglion, in some sections, there was a dorsal
prolongation and possibly also a delicate dorsal branch. I could
not, however, trace the latter either in larvae or in the adult.

The spinal nerves in Amia thus seem to differ considerably
from the schema given by Hatschek of the nerves in Amphi-
oxus (No. 54, p. 140).

The nervus postoccipitalis, or first spinal nerve, agrees in
every respect with the other spinal nerves, and in front of it
there are still two nerves (oc4 and o0c3, Figs. 33-35, 63, and
64, Pls. XXVIII, XXIX, XXXVII, and XXXVIII) also agree-
ing in every respect with those nerves. They issue and lie,
one in front of each of the two occipital arches, and are
described by Sagemehl (No. 104, p. 193) as occipital nerves.
Sagemehl, however, did not find, or at least does not mention,
the ganglion found on the dorsal root of each of them. In
front of these two, complete, spinal-like, occipital nerves there
are two ventral roots (oc? and oc!), one of which, the posterior
one, is described by Sagemehl as the first occipital nerve. The
other he did not find, nor did he find the foramen by which it
issues from the skull. No other still more anterior nerve
being found, I have called these nerves the first and second
occipital nerves. As they belong, as will be shown later, to
the second and third occipital muscle segments, they might,
perhaps with better reason, be called the second and third
occipital nerves. From each of them arise the dorsal, hori-
zontal, and ventral branches found on the ventral roots of all
the spinal nerves. The small, anterior, communicating branch
found on all the spinal nerves was not, however, found on them.

No. 3.] MUSCLES AND NERVES [IN AMIA CALVA. 725

In young larvae the two anterior occipital nerves arise
directly ventral to the two posterior roots of the vagus, often
lying in the same transverse section with those roots. These
two roots of the vagus arise, even in large larvae, considerably
behind the other roots of the nerve, and run forward a rela-
tively long distance internal to the occipitale laterale to join
those roots. In one 12 mm. larva, cut in vertical sections, all
the roots of the vagus were cut in one section. In that speci-
men there were six distinct roots, all of fairly equal size, and
all converging together to form the common root of the
nerve. The posterior of these six roots was formed by the
fusion of the two roots above described, and on these roots
was found a considerable part of the intracranial ganglion of
the nerve. If these two roots are, as they seem to be, the
dorsal roots corresponding to the two independent ventral
ones, we have, in Amia, four complete spinal-like occipital
nerves, and a postoccipital, or first spinal nerve which must
also be assigned, as its dorsal arch is, to the occipital region
of the skull. The nerves thus indicate five vertebral segments
fused with the hind end of the skull. If the last occipital arch
be assigned to the first vertebra, there would be but four.

In 12 mm. larvae the occipitale laterale is a piece of carti-
lage running upward, outward, and forward, touching, or almost
touching, at its upper end, the hind wall of the auditory cap-
sule, but not fused with that wall. It is, however, connected
with the wall by a line of tissue, in which cartilage later
develops. At its lower end, which is much enlarged, it is
imbedded, like the occipital and spinal dorsal arches, in the as
yet unsegmented cartilaginous line of the dorsal processes.
On the dorsal surface of its lower end there is a slight ele-
vation, so that, in vertical sections passing through this por-
tion, it has exactly the shape and appearance of the following
occipital and spinal arches. It is, however, inclined forward,
instead of backward as these arches are. The cartilaginous
line of the dorsal processes here drops gradually from near the
dorsal surface of the notochord to its lateral aspect, and in
front of the occipitale laterale it is continuous with, or be-
comes, the base of the skull. The notochord everywhere lies

726 ALLIS. [Vor. XII.

between and separates the two sides of the hind end of the
skull, the cartilage in no place passing either above it or below
it. Farther forward, where the notochord becomes smaller, the
sides of the skull recede from it, and a wide cleft is left which
continues beyond the notochord, and there becomes the hypo-
physial fenestra of the adult. In later stages the median proc-
esses of the petrosals develop, approximately at the anterior
end of the notochord, and the hypophysial fenestra, or anterior
portion of the cleft, is thus separated from its posterior, occipi-
tal portion. The ventral portion of the original cleft persists,
however, in this intermediate portion, as a groove, and is found
as such even in the adult. The occipital portion of the cleft
is filled, in the adult, by the median, anterior process of the
basioccipital. That structure may, therefore, be the ossified
notochord, if that structure ossifies, or membrane bone, if it
simply disappears.

The hypophysial fenestra of the adult Amia seems to be the
opening, the posterior end of which, in the chick, is said, by
Froriep (No. 34), to mark the anterior limit of the occipital
region of the skull. This limit, in the Characinidae, is marked,
according to Sagemehl, by the anterior end of the basioccipital,
which extends forward to the hind edge of the petrosal (No.
106, pp. 42 and 61). In Amia it is marked by the bases of the
transverse processes of the petrosals. The occipital cleft of
Amia is, therefore, not found in adult teleosts or in embryos of
the chick. In the latter the notochord and parachordalia seem,
from Froriep’s description, to extend as far forward as the
sphenoid bone, the basisphenoid apparently of Amia. The
parachordalia in Amia, whatever their anterior limit may be,
are, posteriorly, a direct continuation of the cartilaginous line
of the dorsal processes of the vertebrae.

Antero-ventral to the occipitale laterale in 12 mm. larvae, in
the triangular space inclosed between it, the base of the skull,
and the bulging hind end of the auditory capsule, the nervus
vagus and the two anterior occipital ventral roots have their exit.
The vagus issues at the upper, anterior angle of the triangle,
the second occipital nerve near its lower, posterior angle, and
the first occipital nerve about midway between them. The

No. 3.] MUSCLES AND NERVES [IN AMIA CALVA. F274

two ventral occipital roots and the nervus vagus thus issue, at
this age, through the same foramen. The occipitale laterale,
at this age, may therefore be considered as a dorsal occipital
arch that has inclined forward against the hind end of the
auditory vesicle, or as parts of three such arches that have
fused with each other and are about to fuse with the auditory
vesicle, as later stages show. In the former case, the arch,
inclining forward, has pushed before it, against the vagus, the
dorsal roots of two occipital nerves; in the latter, the two
dorsal occipital roots, if they do not abort, have, in joining and
fusing with the vagus, cut through the ventral ends of the
«“Anlage”’ of two dorsal arches, so retarding or deranging their
development that no trace of their ventral ends is found in
12 mm. larvae. The vagus, under either supposition, would lie
immediately behind the hind wall of the auditory vesicle, and
would, if the dorsal occipital roots have not entirely disappeared,
be formed by the fusion of the dorsal roots of three or more
postauditory nerves.

That the occipitale laterale represents three, rather than one,
dorsal arch is shown by its relations to the muscle segments
and intermuscular septa. The first septum, the one that lies
between the first and second segments, has its attachment, in
I2 mm. larvae, near the upper, outer end of the occipitale
laterale ; the second has its attachment near the middle of
that piece, and the third, to the elevated portion of its lower
end. The fourth septum is attached to the first occipital arch,
the fifth to the second occipital arch, the sixth to the first
spinal arch, etc.

The dorsal portion of the first muscle segment, the seg-
ment lying in front of the first septum as above defined,
arises, at its anterior end, from the hind wall of the auditory
capsule, and extends across the upper, unformed end of the
occipitale laterale. The ventral portion of this same segment
arises from that thickened portion of the base of the skull
that forms the hind boundary of the eye-muscle canal, and lies
along the ventral portion of the lateral surface of the hind
end of the skull, appearing, in vertical sections, below the
notochord. It extends nearly, if not quite, to the anterior end

728 ALLIS. [Vor. XII.

of the notochord, and it lies, in sections, approximately opposite
the united vagus roots. Opposite the second segment, and
distributed to that segment, lies the anterior occipital
nerve ; opposite the third, the second nerve ; and opposite the
fourth and fifth, the first and second occipital ganglia. No
ventral root was found opposite to, or belonging to, the first
muscle segment, nor could the innervation of that segment by
branches of other roots be established. It seems, however,
probable that it is innervated by branches of the first occipital
nerve.

In 20 mm. larvae the occipitale laterale has fused, at its
upper end, with the hind end of the auditory capsule, and
at its lower end, with the cartilaginous base of the skull.
In certain vertical sections through it, it is seen to have two
thin portions, with a third thin portion between it and the
skull. It therefore presents three transverse enlargements.
From the hind end of the first or anterior of these enlarge-
ments the first intermuscular septum arises; from the hind
end of the second, the second septum; and from the third, the
third septum. The third enlargement is fused with, and forms
part of, the enlarged hind end of the base of the skull. Pos-
terior to that base, between it and the first separate cartilagi-
nous, dorsal vertebral process, lies the first dorsal occipital
arch, and other arches and processes follow regularly. The
dorsal process between the two occipital arches is thus, at this
age, not fused with the hind end of the cartilaginous base of
the skull, and we have a vertebra or part of a vertebra in
process of absorption.

The vagus foramen in 20 mm. larvae is beginning to be en-
closed by a cartilaginous process which forms between it and
the first occipital nerve. In 25 mm. larvae it is entirely in-
closed, and at 30 mm. the two first occipital nerves are also
separated by a line of cartilage. The other arrangements and
relations of the parts here under consideration are the same as
at 20 mm., except that the trunk muscles already push forward
somewhat into the temporal groove, and that a thick, tough
line of tissue begins to be evident extending from the upper,
hind edge of the skull backward into the first muscle segment.

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 729

This ligament is found strongly developed in the adult, arising
from the hind edge of the parietal under the extrascapular.

The suprascapular lies opposite the second muscle segment,
and the nerve innervating the lateral sense organ contained in
it, organ No. 1g infraorbital, lies opposite the outer edge of the
first intermuscular septum. The pedicle of the suprascapular
is, at this age, still a ligament, lies opposite the first septum,
and has its attachment to the skull immediately in front of that
septum. It may, therefore, be homologous with the ligaments
that arise from the hind end of the basioccipital and lie in the
fourth and fifth septa. The upper end of the supraclavicular
lies opposite the third and fourth segments, and the nerves to
organs 20 and 21 infraorbital lie opposite the second and third
septa.

In the adult the muscle segments (J/s) have not the simple
arrangement found in young larvae. The attachment of the
intermuscular septa (27s) to the vertebral column, and the line
followed by their median, inner edges has already been given.
Their outer edges do not have such a zigzag course. Starting
from the mid-dorsal line they run first strongly backward, and
then curve downward and forward to a point approximately
in the line of the lateral edge of the skull. There they turn
backward and downward in a gentle curve, cross the line of
the lateral-line canal, turn downward and slightly forward to
the line of the outer ends of the ribs, and then curve gently
forward and downward toward the mid-ventral line. As they
cross the lateral line and the line of the outer ends of the ribs
there is a slight notch or interruption in the line of the curve.

The septa and segments in front of the first rib are always
short, and while they cannot strictly be said to be irregular,
they vary somewhat in different specimens and present a much
more complicated arrangement than the posterior segments.

Ventral to a radial, longitudinal plane passing, approximately,
through the line of the lateral edge of the skull, the septa
and muscle segments incline regularly outward and backward,
their outer ends lying posterior to the inner ones. Above that
plane they are not so regularly arranged. The septa in the
middle point of the muscle mass so defined have been pulled

730 ALLTS: [VoL. XII.

backward, so as to form a pocket, the hind end of which lies
considerably posterior both to the outer and to the inner edges
of the septa. The septa and segments in this part run inward
and backward from the outer surface of the body to the hind
end of the pocket, and then inward and forward to the mid-
vertical plane of the body. The cause, whatever it may be,
that has led to the formation of these pockets is certainly the
one that has also caused the shifting of the inner edges of the
septa. A good and sufficient reason would seem to be found
in the fact that the tail is capable of, and subject to, violent
movements in early larval stages when the egg-yolk has not yet
been absorbed. The trunk, or at least its anterior portion, is
held by the yolk practically immovable, and as something must
yield when the muscles contract, it is apparently the trunk
septa, which are thus pulled backward by the functionally more
active muscles of the tail. If the tail movements began before
the dorsal arches had become fixed in position, those arches
would doubtless yield and shift with their septa, the amount of
the shifting depending upon the state of development of the
vertebral column when the tail movements first began. The
varying and perplexing relations of the dorsal arches to the
vertebral bodies might thus be simply and naturally accounted
for.

At its anterior end the limited muscle mass here considered
extends into the temporal groove and above and below the
vagus foramen, the vagus and nervus lineae lateralis, as they
issue, lying in a notch in the front edge of the mass. By this
notch the first muscle segment is always cut into two portions,
and the second segment almost but not entirely so. The first
intermuscular septum, the septum between the first and second
segments, is not, however, usually cut through; it is simply
interrupted and pressed back by the issuing nerves toward or
against the second septum, so that the second segment is
pinched off at that point. Amia thus shows, in this, practically
the same arrangement as that described by Hoffmann (No. 57)
in Acanthias. It seems to me, however, that the vagus cuts
into the muscle segments, rather than that the latter push for-
ward above and below the nerve.

No. 3.] MUSCLES AND NERVES [IN AMIA CALVA. 73!

Above the vagus, the first intermuscular septum, follows the
postero-lateral edge of the skull to its upper surface, and then
runs medianward along the hind edge of the squamosal and
parietal under the extrascapular. There it is somewhat broken,
so that two or three parts of it appear as short ligaments, vary-
ing slightly in every specimen. Below the line of its attachment
to the dorsal edge of the skull it is pushed forward into the
temporal groove, crowding out the first muscle segment, so
that it is the anterior surface of this part of the first septum
that forms the upper surface of the muscle mass that fills that
groove. In the first septum there are thus two pockets, a median
one directed backward, as in the other septa, and a lateral one
directed forward into the temporal groove. Similar but much
less marked lateral pockets are found in the second, third, and
following septa. The second septum, and even the third, may
acquire, with the corresponding muscle segments, an insertion
in the temporal groove.

Into the posterior pocket of the first septum, from the hind
edge of the parietal, a strong, flat, ligamentous structure ex-
tends, and ends free without insertion of any kind. From its
dorsal and ventral surfaces muscle fibres arise, and have their
insertions on the adjoining surfaces of the pocket. They
represent all that can be found of that part of the first muscle
segment that lies above the vagus. Below the vagus the first
muscle segment is represented by a small pocket of muscle fibres
found at the anterior end of the trunk muscles. Ventral and
median to this point the edge of the trunk muscle runs _ back-
ward and inward, and opposite the hind end of the basioccipital
almost reaches the ventral cartilaginous processes of that
bone (Fig. 63, Pl. XX XVII). At this point the trunk muscles
cover ventrally the lateral occipital ligaments. The edge of
the muscle mass then turns outward and backward, but the
mass still covers ventrally the lateral ligament on the first ver-
tebra and the lateral process and proximal end of the rib on
the second. The lateral process and rib on the third vertebra
are left fully exposed. One rib and three rib-like ligaments
thus lie imbedded at their proximal ends in the upper end of
the ventral portion of the trunk muscles.

732 ALES: [VoL. XII.

The course of the outer edge of the first intermuscular
septum has been given. The second runs from the mid-dorsal
line, outward and backward, approximately along the posterior
edge of the extrascapular. It then turns abruptly downward
and forward, and runs between the posterior end of the supra-
scapular and the anterior and upper end of the supraclavicular.
The third septum reaches the under surface of the supracla-
vicular near its upper end. It then turns downward and back-
ward, and is attached to the corium, approximately opposite
the anterior end of the first scale of the lateral line. The
same part of the fourth septum corresponds, approximately, to
the anterior end of the second scale of the lateral line, etc.
The seventh septum lies close against the inner, posterior sur-
face of the muscles of the pectoral fin, and is the first septum
that extends unbroken from the mid-dorsal to the mid-ventral
line. Between the ventral end of this septum and the ventral
end of the clavicle, below the fin, there were, in the several
specimens dissected, but one muscle segment and no muscle
septum. The sixth septum always ended as a ligament attached
to the clavicle at the lower edge of the fin, and the fifth as a
similar ligament attached to its upper edge. The fin always
lies in a deep impression in the trunk muscles, the muscle mass
being pushed inward so that its anterior edge, at this place, lies
internal to the inner edge of the clavicle. By this arrange-
ment the sixth segment is apparently cut off. As there are,
however, muscle fibres between the sixth septum and the clavi-
cle, above the fin, the clavicle, at its ventral end, probably corre-
sponds to the fifth intermuscular septum, that is, to the septum
that has its attachment to the second occipital arch. The
sternohyoideus, which lies in front of the clavicle, thus belongs,
in its relation to the muscle segments, as it will be shown that
it does in its innervation, entirely to the region of the head,
and not in part to the trunk.

In one specimen shown in Fig. 34, Pl. XXVIII, the septa
were still further complicated, for the seventh septum had been
pushed back by the fin against the eighth septum, and was
found partly fused with that septum. The eighth septum had
a surface septal connection with the ninth septum. In front

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 733

of the ventral end of the seventh septum there was a small
piece of septum and a small muscle pocket which doubtless
represented the missing sixth segment.

The ventral branches (ocv and spv, Figs. 33, 34, and 57,
Pls. XXVIII and XXXV) of the four occipital and first spinal
nerves come to the outer surface of the trunk muscles poste-
rior to the segments to which they belong, and vary greatly in
their points of exit. The ventral branch of the first occipital
nerve is a very delicate nerve which usually issues at, or imme-
diately behind, the third septum. It always joins and fuses
with the ventral branch of the second occipital nerve, which
issues from the fourth segment or at the fourth septum. The two
nerves united join the ventral branch of the third occipital nerve,
which issues between the fourth and fifth septa; and the nerve
or trunk so formed sends a communicating branch to the ven-
tral branch of the fourth occipital nerve. It then turns down-
ward and forward, behind the hind end of the visceral arches,
along the anterior surface of the ventral portion of the trunk
muscles, and near the inner, anterior edge of the clavicle.
It passes internal to both the pharyngo-claviculares, and
reaches the dorsal surface of the sternohyoideus near the
median line of the body. It there gives off several cutaneous
or general branches, and, continuing forward, sends branches
first to the middle, and then to the anterior, divisions of the
sternohyoideus, and ends in a delicate branch which innervates
the branchiomandibularis.

In one specimen, and on one side only of that specimen, a
large branch was found which arose from the third occipital
nerve, and passed dorsal to, that is, internal to, the second
occipital nerve. There it received an anastomosing branch
from that nerve, and then ran outward and upward internal to
the nervus lineae lateralis. As the specimen had been cut, in
dissection, above that point, the ultimate destination of the
nerve could not be determined. Nothing anything like it was
found in any other fish, and I can only think that it was some
marked irregularity due to accident of some kind.

The ventral branch of the fourth occipital nerve issues,
as it should, in its own segment, the fifth, along the anterior

734 ALLIS. [Vou. XII.

surface of the septum posterior to that segment or imme-
diately behind that septum. It always has a double dis-
tribution, going in part to the posterior segment of the
sternohyoideus and in part to the pectoral fin. In some
specimens it separates into two parts, one going directly to the
posterior segment of the sternohyoideus, and the other joining
and fusing with the ventral branch of the nervus postoccipi-
talis, or first spinal nerve. In others it touches and anasto-
moses with, or receives an anastomosing branch from, the ven-
tral branch of the third occipital nerve, and then joins the first
spinal nerve, or the contrary, anastomosing with the first
spinal nerve, and then sending a branch to the sternohyoideus.
The branch that goes tothe sternohyoideus accompanies closely
the united ventral branches of the first, second, and third
occipital nerves, running sometimes, with those nerves, internal
to the pharyngo-claviculares, but sometimes perforating the
internal one of those two muscles. It always gives off one or
more general cutaneous branches, one of which either perfo-
rates the pharyngo-clavicularis internus, or passes around the
anterior edge of that muscle to its outer surface. The nerve
then enters the posterior division of the sternohyoideus along
its median face.

The ventral branches of the first spinal and last two occipi-
tal nerves often separate, in dissection, into two strands, and
the two strands of one or more of the nerves may be found
issuing as separate nerves from the trunk muscles. This is
often confusing, and to be certain of the number of nerves
under consideration, the dissection must always be carried to
the ganglia themselves.

There are thus in Amia, in the adult, five muscle segments
and five intermuscular septa in the postauditory region of the
skull. In larvae there are but four. The sixth intermuscular
septum, in the adult, is attached to the first vertebra, and the
sixth muscle segment extends from it to the fifth septum,
which is attached to the hind end of the skull. In larvae it is
the fifth septum and segment that have that position. If the
intermuscular septa are vertebral, as is generally conceded, the
five postauditory septa of the adult must indicate five vertebrae

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 735

fused with the hind end of the skull to form the postauditory
or occipital region. In larvae but four such vertebrae are in-
dicated. -Whether the dorsal arches may have shifted from the
anterior end to the posterior end of their respective vertebrae,
or not, makes no difference in the count. If, however, the
last occipital arch of the adult is to be assigned to the first
vertebra, the count would be diminished by one. The septa
seem to preclude such a supposition.

5. Review and Comparison.

In comparing the occipital region in Amia with that in other
animals, some fixed point must first be found. The anterior
end of the first muscle segment seems to be such a point ; but
there may have been still other, more anterior, muscle segments
that have disappeared, as there certainly is one nerve that has
disappeared, that of the first existing muscle segment. Another
point that would seem to be fixed, and wholly independent of
the possible disappearance of such segments, is the fifth inter-
muscular septum. This septum seems to mark, in both fishes
and reptiles, the posterior limit of the muscle segments, or
myotomes, from which the muscles of the tongue primarily
develop; and the anterior limit, in fishes, of the segments that
enter into the formation of the muscles of the pectoral fin.

In Scyllium and Pristiurus the coraco-hyoideus, or sterno-
hyoideus, develops, according to van Wijhe (No. 130, p. 16),
from ventral prolongations of the last three head myotomes
and the first trunk myotome. Corning (No. 21), in Scyllium,
assigns the same posterior limit to the myotomes so concerned,
but adds the first head myotome to the number. In teleosts,
also, Corning finds the “‘ Hypoglossusmusculatur”’ arising from
the first five myotomes, confirming in this Harrison’s earlier ob-
servation (quoted by Corning). In reptiles both he and Mollier
(No. 82) find the muscles of the tongue arising from the first
five myotomes, but Corning states that in late stages these
muscles receive additions from the sixth and seventh myotomes.
In Acanthias the coraco-hyoideus develops, according to Hoff-
mann (No. 57, p. 639), from the fourth occipital myotome, that

6 ALLIS. VoL. XII.
1d

is, from the tenth somite, or first trunk myotome of van Wijhe.
The tenth somite of van Wijhe thus corresponds, in all these
forms, in so far as its relations to the sternohyoideus are con-
cerned, with the fifth muscle segment in Amia. In all these
forms, also, the muscles of the pectoral fin in fishes, and
of the neck and anterior extremity in reptiles, develop from
myotomes posterior to van Wijhe’s tenth somite; that is,
in Scyllium (Corning), from the sixth to the fourteenth
myotomes ; in teleosts (Corning) and reptiles (Corning and
Mollier), from the sixth to the thirteenth. In Scyllium the
«‘Vornierenblaschen,” and in reptiles the ‘ Urniere,” begin at
the sixth segment. The fifth intersegmental space in all these
animals seems, therefore, to mark a definite limit, and this limit
corresponds to the fifth intermuscular septum in Amia. The
first muscle segment in Amia is therefore, in all probability, the
sixth somite of van Wijhe, and there are, in Amia, the same
number of head segments as in Acanthias (Hoffmann), that is,
one more than in the head of Scyllium, of Pristiurus, and of
reptiles. The one additional segment in Amia fuses with the
hind end of the skull in post-larval stages.

Opposite the sixth somite in Amia, as in Scyllium and
Pristiurus (van Wijhe) and in teleosts (Harrison, quoted by
Corning), lie the united vagus roots with no corresponding
ventral root ; opposite each of the next two somites lies a ven-
tral root, as in Scyllium and Pristiurus (van Wijhe) and Acan-
thias (Hoffmann). Directly dorsal to these two ventral roots
in Amia, there arise two distinct and separate roots of the
vagus, in all probability the dorsal roots of the somites, and on
these roots thereis part of an intracranial ganglion. Oppo-
site these two somites in Scyllium and Pristiurus, van Wijhe,
in his earlier work (No. 130), also found separate roots of the
vagus. In later work he found a root opposite the seventh
somite only (quoted by Rabl, No. 1o1, p. 118). In Acanthias,
opposite the seventh somite, Hoffmann finds the posterior root
of the vagus with a part of the vagus ganglion ; opposite the
eighth somite he finds a rudimentary ganglion which soon dis-
appears. Opposite the ninth somite, or fourth muscle segment,
in Amia, there is a spinal nerve with dorsal and ventral roots,

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 737

and a dorsal ganglion. Similar roots and ganglia are found in
Scyllium (Corning), Pristiurus (Kaestner, No. 62, p. 188), and
reptiles (Corning), while in Acanthias (Hoffmann) there is
simply a rudimentary ganglion which soon disappears. Van
Wijhe did not find in Scyllium the dorsal root or ganglion
described by Corning in this somite, and Mollier also (No. 82,
p- 483), in reptiles, seems to have found the first ganglion
opposite the fifth segment and not opposite the fourth.

Amia, therefore, in its occipital region presents nothing new
or different from what has already been described in selachians,
teleosts, and reptiles. With Protopterus, as described by
Pinkus, it is impossible to make any comparison, for in Protop-
terus (No. 89, p. 325) there are, in the occipital region of the
head, three ventral roots, and three complete spinal nerves,
two of which are called by Pinkus the first and second hypo-
glossi. The two posterior ventral roots are distributed to the
“umliegende Visceralmuskulatur.” The anterior one is simply
said to run backward along the under side of the basis cranii,
but as it joins, in the figures, the two posterior roots, it prob-
ably has the same distribution that they have. The two hypo-
glossi are distributed in part to the first and second trunk
muscle segments, in part to the tongue, and in part they join
branches of the first three spinal nerves to form the brachial
plexus and take part in the innervation of the fin.

GENERAL SUMMARY.

1. The eye-muscle canal of Amia and its tall, orbital open-
ing present, in certain respects, striking resemblances to the
pituitary fossa, or sella turcica, and the sphenoidal fissure of
the human skull. The pituitary fossa, in man, is bounded
behind by the dorsum sellae of the sphenoid bone, and in
front by the olivary eminence of the same bone. In Amia the
eye-muscle canal is bounded behind by the median, horizontal
wings of the petrosals, and in front by a transverse cartilagi-
nous bar, on the sloping, dorsal surface of which the optic
chiasma rests. On this sloping surface, in Amia, as on the
olivary eminence, in man, the optic nerves pass outward to
their foramina, accompanied by the ophthalmic arteries. In the

738 ALLIS. [Vor. XII.

fossa, in man, on each side of the pituitary body, are the
cavernous sinuses. These sinuses receive, anteriorly, from the
orbits, through the sphenoidal fissures, the ophthalmic veins ;
and posteriorly, behind the pituitary body, they communicate
with each other by the intercavernous sinuses. In Amia, in
the same position, the eye-muscle canal lodges, on each side, the
ophthalmic veins, which communicate with each other, in the
adult, under the hypophysis, but in embryos behind it. In
the cavity of each sinus, in man, lie the internal carotid artery
and the sixth nerve ; in the canal, in Amia, lies the latter one
only of these two structures, — the internal carotid, however,
lying near the edge of the hypophysial fenestra, in such a posi-
tion that it would of necessity enter the canal if the fenestra
were slightly enlarged. In the outer wall of each cavernous
sinus, in man, lie the third and fourth nerves and the ophthalmic
and superior maxillary divisions of the fifth nerve ; in the upper,
lateral chamber of the canal, in Amia, are found the same
nerves. In man, after traversing the fossa, the internal carotid
artery turns upward, between the optic and third nerves, and
internal to the anterior clinoid process; in Amia it has the
same relation to the same nerves and to the hind end of
the basisphenoid. In man, the sphenoidal fissure transmits to
the orbit the third, fourth, and sixth nerves, the ophthalmic
division of the fifth nerve, and the ophthalmic veins ; in Amia,
the orbital opening of the canal similarly transmits the third,
fourth, and sixth nerves, the ophthalmic veins, and the radix
profundi, which latter nerve is, in all probability, the homologue
of the ophthalmic division of the fifth nerve in man. The
superior maxillary nerve, both in man and in Amia, does not
issue from the fossa with the other nerves, through the sphe-
noidal fissure, but through a separate and more posterior fora-
men. This foramen, in man, is said by Thane (No. 100) to be
cut off from the sphenoidal fissure ; in Amia it is certainly not
so cut off. The recti muscles arise, in man, from the sphenoid
bone, near the lower end of the sphenoidal fissure ; in Amia
they arise from the basisphenoid, at the lower end of the
orbital opening of the eye-muscle canal, or from the floor of
that canal, inside the opening.

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 739

2. The canalis transversus of selachians is, in Amia, a groove
extending transversely from orbit to orbit, immediately in front
of the transverse, cartilaginous bar that marks the anterior
limit of the eye-muscle canal, and immediately in front of the
two ossifications of that bar, which are said to represent, in
Amia, the basisphenoid of teleosts.

The eye-muscle canal of Amia and teleosts is not, there-
fore, the canalis transversus of selachians, as Sagemehl con-
cluded it to be. In Amia it is a cavity or space formed in late
larval stages around the hypophysis cerebri and saccus vascu-
losus ; and entrance to it, for the muscles of the eye, has been
acquired through the foramen of the nervus abducens, or along
a canal which serves for the passage of an important ophthal-
mic vein. This vein arises from the under surface of the
hypophysis, in connection with the corresponding vein of the
opposite side of the head, and is almost directly continuous,
anteriorly, with another vein, which arises from the choroid
gland. Both veins are branches of that part of the orbital
venous sinus that lies in the upper, lateral chamber of the eye-
muscle canal.

3. The internal carotid artery enters the cranial cavity
through a special canal which traverses the basis cranii along
the median edge of the basisphenoid of its side of the head.
The basisphenoid of Amia thus lies lateral to the artery, while
in Scomber and Perca it lies median to it. While traversing
its special canal, the artery, in Amia, sends a communicating
branch to the efferent pseudobranchial artery, and in the
cranial cavity, after giving off optic and olfactory branches, it
forms a circulus cephalicus opposite the auditory region and
behind the lobi inferiores.

4. The external carotid artery enters the upper, lateral
chamber of the eye-muscle canal by a special foramen through
the petrosal, but does not, so far as could be determined by
sections, enter or send any branches into the cranial cavity
proper. Branches of it issue from the eye-muscle canal with
branches of the trigeminus and facialis, which they accompany.

5. The efferent pseudobranchial artery enters the orbital
opening of the eye-muscle canal by a special foramen, the

740 AELES. [VoL. XII.

foramen lying immediately in front of the lateral wing of the
parasphenoid. It thus, like the external carotid, pierces a
part of the chondrocranium, but does not enter the cranial
cavity proper. In the orbital opening of the eye-muscle canal
it receives a small branch from the internal carotid, and
issues into the orbit with the recti muscles.

6. The hypophysis cerebri and saccus vasculosus, in the
adult, are both glandular structures. Both receive a consider-
able nervous supply directly from the base of the infundibulum,
and the glandular cavities of both communicate directly with
the infundibular cavity.

7. The olfactory nerve, contrary to Sagemehl’s statement,
lies exposed to the orbit through a limited part of its course.
The opening through which it is so exposed lies at the extreme
front and upper end of the orbit, and gives passage to a vein
coming from the nasal pit. The olfactory canal in front of
the opening is, therefore, formed by the fusion of two canals,
the anterior part of the olfactory canal proper, and what is
probably the homologue of the orbito-nasal canal of selachians.
The opening can, therefore, be called the orbito-nasal opening
or fenestra.

8. The “hitherto undescribed cranial nerve”’ of Pinkus, in
Protopterus, is found in Amia, part of its fibres arising with
the olfactorius and part of them having the intercranial course
described by Pinkus, though their definite origin from the
brain, as in Protopterus, was not satisfactorily determined. In
the nerve the large cells described by Pinkus are found, scat-
tered along the nerve in old larvae, but in 12 mm. larvae
gathered into a knob-like protuberance on the under surface of
the nervus olfactorius at about the middle of its length. In
the latter larvae the collection of cells on the olfactorius
resembles, in general histological appearance, the ciliary gan-
glion found on the nervus oculomotorius in much the same
relation to that nerve. As the ciliary ganglion is unquestion-
ably, in part at least, a sympathetic ganglion, the ganglion on
the olfactorius is, perhaps, of a similar character, and is pos-
sibly the sphenopalatinum of higher vertebrates.

g. The muscles rotating the eyeball in the several orders of

5]

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 741

the Ichthyopsida are not homologous structures, if existing
descriptions of their innervation can be relied upon. On the
contrary, the group Ichthyopsida (the Pharyngobranchii ex-
cluded) can be divided by the innervation of the muscles of
the eye into two great groups and other sub-groups, which, if
reversions have not taken place, indicate distinct and definite
lines of descent.

In one of the two great groups, represented by the Cyclosto-
mata alone, the nervus abducens innervates two recti muscles,
the inferior and the externus. In the other group that nerve
innervates but one rectus muscle, the externus, but it inner-
vates also a retractor bulbi when that muscle is found.

The second group is subdivided into two sub-groups, in one
of which the superior branch of the oculomotorius, lying
dorsal to the ophthalmicus profundus, innervates two recti
muscles, the superior and the internus, while in the other it
innervates but one, the superior. In the first sub-group are
found Elasmobranchii, Dipnoi, and Urodela; in the second,
Ganoidei, Teleostei, and Anura. The Amphibia are thus
separated into two sub-groups, corresponding to the two sub-
groups of Pisces. Inthe prototype of this second group there
must have been an arrangement of muscles and nerves most
nearly represented by that found in the Holocephala, where
the rectus internus arises near the front end of the orbit, as in
Petromyzon, and the obliquus superior from the edge of the
orbit, as in Petromyzon. From this prototype two lines lead
to the two sub-groups of Pisces and two to the two sub-groups
of Amphibia.

Between the two lines leading to Amphibia lies Ichthyophis,
in which the muscles rotating the eyeball are innervated as in
Anura, but in which a retractor tentaculi has been formed
from one of the muscles of the eye (Sarasins), probably from
the rectus internus of Urodela, and not from the retractor
bulbi, as the Sarasins suggest. Ichthyophis seems therefore,
in the arrangement of the muscles of the eyeball, to represent
the beginning of the line leading to higher vertebrates, as
Burckhardt states that it does in the arrangement of the parts
of the brain.

742 ALETS. [Vor. XII.

10. The profundus ganglion is found both in larvae and in
the adult as a separate and distinct ganglion, connected with
the ciliary ganglion by a radix longa, and with the brain by a
profundus root, which, in larvae, is wholly separate and distinct
from the root of the trigeminus, but, in the adult, is somewhat
fused with that root.

From the ganglion arise, in addition to the radix longa, two
ciliares longi, a large and important portio ophthalmici pro-
fundi, and often, but not always, a small, delicate, and appar-
ently degenerating nerve.

11. The portio ophthalmici profundi may be single, double,
or even triple. It runs forward and upward, dorsal to all the
muscles of the eyeball and the nerves innervating them, and
joins and completely fuses with the ramus ophthalmicus super-
ficialis trigemini, while that nerve is still in the orbit. Its
position thus shows that it cannot be the homologue of the
ramus ophthalmicus profundus trigemini of selachians and
other animals, that nerve always lying below the superior
branch of the oculomotorius, and below the nervus trochlearis.
It must, therefore, correspond to certain of those frontal
branches of the ophthalmicus profundus of selachians that arise
from that nerve before it passes under the rectus superior.

There is, therefore, no true ramus ophthalmicus profundus in
Amia, that nerve being apparently represented by the small and
delicate nerve often found arising from the profundus ganglion.

12. The ciliary ganglion is a separate and well-developed
ganglion, connected with the profundus ganglion by a long
radix longa, and with the nervus oculomotorius by short fibres,
which represent the radix brevis. From the ganglion a single
nerve arises, the ciliaris brevis.

13. The trigemino-facial ganglionic complex lies in the
upper, lateral chamber of the eye-muscle canal, and consists
of three parts between which no communicating fibres could be
traced. The three parts are the profundus ganglion, which is
entirely separate and distinct, the main trigemino-facial ganglion,
and the ganglion of the buccalis and ophthalmicus facialis.
The latter ganglion lies on, and partly imbedded in, the upper
surface of the anterior part of the main ganglion.

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 743

The main ganglion is in young larvae imperfectly separated
into three parts or regions, an anterior and upper, a median,
and a posterior one.

The anterior and upper of these three portions of the gan-
glion is connected with the brain by the anterior, or trigeminal,
root of the ganglion. This root arises from the anterior and
lateral edge of the medulla, and is there distinctly double, its
dorsal and posterior portion arising superficially in the brain,
its ventral and anterior portion deeper in the brain, in part
from the posterior longitudinal fasciculus. These latter fibres,
which are the motor fibres of the root, either alone or with
other of the deeper fibres, traverse the anterior part of the
main ganglion as its anterior commissure, and enter the
truncus maxillaris trigemini.

The median and posterior portions of the main ganglion are
connected with the brain by the posterior, or facial, root of the
ganglion, which arises from the lateral surface of the medulla
in front of, and close to, the root of the nervus acusticus.
It arises as two bundles, which immediately receive a third
bundle coming from the root of the ganglion of the buccalis
and ophthalmicus facialis. The anterior of the two bundles of
the root proper arises at a high level in the brain, apparently
from the fasciculus communis tract, and enters mainly or
entirely into the median portion of the main ganglion. The
posterior of the two bundles arises deep in the brain, and con-
tains the motor fibres of the root. It enters the posterior
portion of the main ganglion, and, with the third bundle,
traverses it as its posterior commissure. The third bundle
contains the lateral, or sense-organ, elements of the root.

The root of the ganglion of the buccalis and ophthalmicus
facialis is the lateral or sense-organ root of the ganglionic
complex. It arises from a slight swelling on the lateral surface
of the medulla, close to, and almost as a part of, the root of the
nervus acusticus. The branch of this root that is sent to the
posterior root of the main ganglion, may arise as a separate
root or bundle.

14. The rami ophthalmicus superficialis and buccalis facialis
arise from their own special ganglion; the ramus oticus facialis

744 TTS: [Vou. XII.

from that same ganglion, but as a part of the ramus buccalis.
The ophthalmicus superficialis innervates the sense organs
of the supraorbital sensory canal, and the anterior head line
of pit organs; the buccalis innervates the sense organs of
the suborbital and antorbital parts of the main infraorbital
canal; and the oticus, the otic part of that canal, and the
sense organ of the spiracular canal.

15. The ramus ophthalmicus superficialis trigemini arises,
in young larvae, from the upper, anterior end of the main
trigemino-facial ganglion, but many of its fibres can be traced
backward to the superficial portions of the median portion of
the ganglion. They accordingly belong to the fasciculus com-
munis portion of the posterior root of the ganglion, and not
to the anterior root. The nerve is concerned largely in the
innervation of the terminal buds found on the top of the head
and snout, and its branches are sent to the outer surface of the
head on both sides of the ophthalmicus facialis, which nerve it
closely accompanies, lying immediately beneath it.

16. The truncus maxillaris trigemini arises directly from the
anterior commissure of the main trigemino-facial ganglion, but
receives important additions to its fibres from the median,
fasciculus communis portion of that ganglion. From it arise
the rami maxillaris superior and inferior, and from it, also, or
from the main ganglion itself, several accessory trigeminal
nerves which seem destined to supply the terminal buds on
the side of the cheek.

17. The ramus maxillaris superior trigemini closely accom-
panies the buccalis facialis in its course forward, behind, and
below the eye. It lies immediately internal to and below the
buccalis, its branches are given off on both sides of that nerve,
and from its distribution it seems destined in part to the inner-
vation of terminal buds. Its branches form anastomoses both
with the posterior and the anterior portions of the ramus
palatinus facialis.

18. The ramus maxillaris inferior trigemini innervates all
the levator and adductor muscles of the mandibular arch, the
dilatator operculi, and probably also the intermandibularis, and
all, or a part of, the inferior division of the geniohyoideus.

No. 3.] WUSCLES AND NERVES IN AMIA CALVA. 745

The branch that innervates the four divisions of the levator
maxillae superioris arises from the base of the nerve, or even
from the main truncus maxillaris, and may be double through-
out its entire length.

Certain branches of the ramus have a peculiar and circuitous
course, and are distributed to regions where terminal buds
abound. One of these branches becomes a mandibularis inter-
nus trigemini, distributed to the inner surface of the mandible ;
another goes to the tissues of the mouth cavity near the upper
end of the ceratohyal; and two others pass under the ramus
mandibularis externus facialis, and are distributed to the outer,
ventral surface of the head, where they form anastomoses with
branches of the ramus hyoideus facialis. These branches of
the trigeminus are all prespiracular nerves, and are all dis-
tributed to regions where terminal buds are found. The
mandibularis internus trigemini, with or without other of these
special branches, seems to correspond to the inferior branch of
the palatinus facialis in Protopterus and Polyodon, and to be
the homologue of the chorda tympani of higher animals.

19. The truncus hyoideo-mandibularis facialis arises largely
from the posterior commissure of the main trigemino-facial
ganglion. Its rami opercularis and hyoideus innervate the
adductor and levator muscles of the hyoid arch, the hyohyoi-
deus, and probably also the superior division of the genio-
hyoideus, and a part, at least, of the inferior division of that
muscle. Its ramus mandibularis externus contains all the
lateral fibres of the nerve, and innervates the sense organs of
the operculo-mandibular portion of the lateral canal system.
Its ramus mandibularis internus is distributed to the inner
surface of the hyoid and mandibular arches. As this last nerve
is a postspiracular nerve, it seems impossible that it should be
the homologue of the chorda tympani, which is apparently a
prespiracular nerve.

No branches of any of these nerves could be traced definitely
to the pseudobranch.

20. The ramus palatinus facialis arises, in larvae, from the
median, or fasciculus communis, portion of the main trigemino-
facial ganglion. It separates into an anterior and a posterior

746 ALETS. iVor. Xi.

portion, both of which are distributed to the upper and lateral
surfaces of the mouth cavity. Both portions form anastomoses
with branches of the ramus maxillaris superior trigemini, and
the anterior portion, at its anterior end, pierces the basis cranii
to reach the upper surface of the anterior end of the chon-
drocranium. One of the branches of this nerve was traced
into one of the large premaxillary teeth.

21. The nervus acusticus arises, in larvae, from the summit
of the tuberculum acusticum by three distinct roots, on each
of which a somewhat distinct part of the large acoustic gan-
glion is formed. The more dorsal of the three roots forms the
posterior root of the nervus, and gives origin to the ramus
cochlearis. The other two roots together form the anterior
root of the nervus, and give origin to the ramus vestibularis.
The two nerves innervate the cristae ampullae as in other
fishes.

No ductus or saccus endolymphaticus could be recognized
in any of the specimens examined, old or young.

22. The nervus lineae lateralis vagi arises, in larvae, imme-
diately posterior to the tuberculum acusticum. It either
traverses in part, or is traversed by, the root of the nervus
glossopharyngeus, an important interchange of fibres there
taking place, the fibres apparently running from the nervus_
lateralis into the glossopharyngeus, and there giving origin to
the so-called dorsal root of that nerve.

The nervus issues from the cranial cavity with the nervus
vagus through the vagus foramen, enters its own ganglion, and
is then distributed to the lateral sense organs of the supra-
temporal cross-commissure, and to those of that part of the
lateral sensory system that lies posterior to that commissure.

23. The nervus glossopharyngeus, immediately after its exit
from the brain, pierces the median, membranous wall of the
auditory cavity and traverses that cavity, lying, in its passage,
behind the membranous ear, between the sacculus and the
sinus utriculi posterior, and between the two terminal branches
of the ramus cochlearis acustici. It apparently receives fibres
both from the root of the nervus lineae lateralis and from the
ramulus ampullae posterioris, the fibres so received going to

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 747

form the so-called dorsal root of the nervus. On this dorsal
root, in larvae, a separate ganglion is formed, and from it
arises the dorsal or suprabranchial branch of the nervus, which
branch is destined entirely to supply certain of the sense organs
of the lateral system. From the ganglion of the main root
arise pharyngeal, pretrematic, and posttrematic branches, which
have the distribution indicated by their names. The ramus
pharyngeus and ramus pretrematicus traverse regions where
terminal buds are found. No branch of the nervus could be
traced definitely to the pseudobranch.

24. The vagus arises by several roots, all of which traverse
an intracranial ganglion, and then separate into four portions,
on each of which an extracranial ganglion is formed. From
each of the first three of the latter ganglia arise pharyngeal,
pretrematic, and posttrematic branches ; from the fourth arise
the superior and inferior pharyngeal nerves, and the nervus
intestinalis. The pharyngeal branches of the first three nerves
pass downward, to their destination, between the articular and
interarcual ligaments that connect the infrapharyngobranchial,
or epibranchial, of the arch of the nerve, with the epibranchial
of the next preceding arch.

From the anterior fibres of the root of the nervus, or in part
from its intracranial ganglion, a supratemporal branch arises,
which accompanies closely the supratemporal branch of the
nervus lineae lateralis, and is distributed to regions where
terminal buds abound. A part of this nerve fuses completely
with a branch arising from one or both of the first pair of
dorsal branches of the ramus ophthalmicus superficialis trige-
mini. Another branch, arising from the base of the nerve, or
perhaps from the main vagus root itself, has an intracranial
course, and issues on the top of the cranium, about where the
ramus lateralis trigemini, or ramus recurrens facialis, of other
authors issues. It may represent that nerve or a part of it,
the rest of the nerve being represented in the main supra-
temporal branch of the vagus, or in that nerve and the two
branches of the ophthalmicus trigemini together.

25. Terminal buds are said to represent a condition through
which the canal organs of the lateral lines have passed in

748 ALLIS, [Vou. XII.

their development. They and the nerves innervating them
should, therefore, arise from, or in connection with, sensory,
ectodermal thickenings, as do the canal organs and their nerves.
Certain pharyngeal or pretrematic nerves are known to inner-
vate certain terminal buds, and the pharyngeal and pretre-
matic branches of all the cranial nerves are said to arise from,
or in connection with, epibranchial or pretrematic ectodermal
thickenings. Terminal buds should, therefore, arise from, or
in connection with, those same thickenings.

Those trigeminal and facial nerves, in Amia, that are known
to innervate terminal buds, or that innervate regions where
those buds abound, all arise either from the median, fascicu-
lus communis portion of the main trigemino-facial ganglion,
or receive important additions to their fibres from that portion
of that ganglion. It seems, therefore, probable, as Strong has
suggested, that the fasciculus communis tract of the brain is
largely or entirely concerned in the innervation of terminal
buds. The arrangements presented by different fishes indi-
cate that the fibres of this tract may issue from the brain as
a somewhat separate root, and have a somewhat separate gan-
glion, or that they may issue as parts of the roots of certain of
the cranial nerves, their ganglia, in that case, being completely
fused with the ganglia of the nerves with which they issue.
The nerves, their ganglia, and the organs they innervate thus
form an epibranchial or pretrematic series comparable to the
nerves, ganglia, and organs of the lateral or dorso-lateral series.
The organs and nerves of the epibranchial series are formed
before those of the lateral series, for the nerves of the former
lie always below those of the latter, and may hold them be-
tween branches issuing on both sides of them.

26. The eye is said to belong to the line of epibranchial
thickenings. The lens of the eye may, therefore, be a modified
terminal bud or buds, and one or more of the ciliary nerves
the nerve innervating it.

27. The chorda tympani is distributed to the tongue, and
contains taste fibres. As taste bulbs are supposed to be
developed from terminal buds, the chorda probably belongs, in
part, at least, to the fasciculus communis nerves. AS itis

No. 3.] MUSCLES AND NERVES IN AMIA CALVA.
749

apparently a prespiracular nerve, it cannot be the ramus man-
dibularis internus facialis of Amia and other fishes, as that
nerve is a postspiracular nerve. It, therefore, seems probable
that it is represented, in Amia, by the mandibularis internus
trigemini, and in Protopterus and Polyodon by the inferior
branch of the palatinus facialis.

28. The ramus ophthalmicus superficialis trigemini, in
Amia, arises largely from the fasciculus communis portion
of the trigemino-facial ganglion, and is largely or entirely con-
cerned in the innervation of the terminal buds found on the
top of the head and snout. Where those buds are few or
entirely wanting in that part of the head, as in selachians and
amphibia, the nerve is, naturally, small or wanting.

29. The suprapharyngeal elements of the branchial arches,
in Amia, are usually found as special processes on the posterior
edges of the epibranchials of their arches, and not as separate
elements. They lie at the distal edge of the surface of inser-
tion of the external levator muscle of the arch to which they
belong; behind, rather than above, the efferent artery of the
arch; behind the posttrematic branch of the nerve of the arch;
and in front of pretrematic branches of the nerve of the next
posterior arch.

The infrapharyngeal elements lie proximal to, or in front
of, the external levator muscles of their respective arches, and
give insertion to the internal levator muscle of the next pre-
ceding arch, when that muscle exists. They lie, on each arch,
below the efferent artery, and below the posttrematic and pha-
ryngeal branches of the nerve of the arch.

The epibranchial of each arch is connected, normally, with
the infrapharyngobranchial of the next posterior arch by
articular and interarcual ligaments. Where the infrapharyngo-
branchial of the next posterior arch is not normally developed,
the connection is wholly or in part with the epibranchial of
that arch.

30. In the hyoid arch the hyomandibular and symplectic
have practically the same relations to the muscles, nerves,
artery, and ligaments of their arch, that the suprapharyngo-
branchials and infrapharyngobranchials, respectively, have to

750 ALLIS. [VoL. XII.

those structures in their arches. They seem, therefore, to be
those elements of the hyoid arch, and the interhyal, or stylo-
hyal, to be the epal element.

The opercular bones and the branchiostegal rays, from
12 mm. larvae upward, contain no trace of cartilage, and give
no indication whatever of having been preformed in cartilage.
They all lie external to the nerve and artery of the arch. They
are, accordingly, probably entirely of dermal origin, and not
homologous with the gill-rakers of selachians.

31. In the mandibular arch the metapterygoid process of
the metapterygoid and the anterior process of the same bone
fulfil even better the same conditions. They seem, therefore,
to be the supra- and infrapharyngeal elements, respectively, of
the mandibular arch, and the quadrate to be the epal element.
The palatine is then, probably, part of a premandibular arch,
the coronoid process of Meckel’s cartilage being possibly an-
other part of the same arch.

32. The muscles of the visceral arches probably all lay
primarily external to those arches, as a simple constrictor
muscle. From that muscle the arcual and interbranchial
muscles arose, by a separation of the simple constrictor into an
inner and an outer layer. From the arcuals arose the adduc-
tors and the interarcuals. The adductors, however, on the
different arches, probably arose in different manners, and thus
are not all strictly homodynamous structures. The same is
true of the levators, which probably arose in part from the
interarcual muscles, and in part from the interbranchials.

33. The adductor mandibulae probably arose from a muscle
resembling the so-called interbranchiale of Chimaera. That
muscle in Chimaera is the simplest form of arcual muscle
known, and, in giving origin to the adductor mandibulae, it had
simply to slip over the anterior edge of the mandibular arch
into its actual position on that arch.

In Amia the adductor mandibulae is in large part a con-
tinuous muscle extending from the upper ends of the hyoid
and mandibular arches downward and forward, internal to the
coronoid process of Meckel’s cartilage, into the hollow of the
mandible. The muscle has become entirely tendinous in its

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 751

middle portion, where it passes into the mandible, and is
almost completely separated into two portions, a deeper and a
superficial one. The latter, in its superior portion, is again
incompletely separated into two or three portions. The most
superficial of these portions is probably the homologue of the
entirely independent superficial muscle of some teleosts.

34. The adductor hyomandibularis, adductor operculi, and
levator operculi are derived from the dorsal half of the general
constrictor of the hyeid arch, the adductor hyomandibularis from
the deeper layers of that muscle, the other two, or at least the
levator operculi, from its superficial portions. The adductor
hyomandibularis is probably developed from a muscle com-
parable to one or more of the interarcual muscles of the
branchial arches of selachians, and is thus homodynamous
with the levators of the branchial arches of teleostomes, and
not with the adductor mandibulae. The adductor operculi and
levator operculi, at least the latter, are derived from the inter-
branchial muscles of their arch, and are thus homodynamous
with the levator arcus palatini, and not with the levator mus-
cles of the branchial arches.

35. The adductors of the branchial arches arise from, or in
connection with, the arcuals of their respective arches, muscles
comparable to the so-called interbranchiale of Chimaera. They
represent the middle portion only of that muscle, and have
acquired their actual position on their respective arches by
passing over the posterior edges of those arches. They are
thus not homodynamous either with the adductor of the man-
dibular arch or with that of the hyoid arch.

36. The levator arcus palatini and the dilatator operculi are,
in Amia, as in teleosts, separate and distinct muscles. They are
developed from a single muscle derived from the dorsal half of
the general constrictor of the mandibular arch. Their innervation
and position indicate that they are, together, the homologue of
the muscle called by Vetter, in selachians, Addy, and that they,
and that muscle, arise from the dorsal portion of the inter-
branchial muscle of their arch. They are thus homodynamous
with the levator operculi, and not with the levators of the
branchial arches.

52 ALETELS: VOL. Delt:
75

37. The levators of the branchial arches probably arise from
the first and second interarcuals of selachians. There are, nor-
mally, two of them to each arch, —an external muscle, inserted
on the epibranchial of its own arch, in close connection with the
suprapharyngobranchial of that arch, and an internal muscle,
inserted on the infrapharyngobranchial of the next following
arch. As the posterior arches, or the pharyngobranchials of
those arches, disappeared, the internal levators, or interarcuales
I, of those arches, seem to have acquired attachment to the
under surface of the spinal column, and thus to have given
origin to the retractor arcuum branchialium.

Interarcuales III of selachians are found, in Amia, as a
series of dorsal interarcual ligaments.

38. The four muscles called by McMurrich the second, third,
fourth, and fifth divisions of the levator arcus palatini are
probably not parts of that muscle. The second and third of
these muscles are probably derived from the levator maxillae
superioris of selachians, or from that muscle and one of the
spiracle muscles. The fourth and fifth muscles are undoubt-
edly derived from the muscle called by Vetter, in selachians,
Addgf. As all four muscles are, in Amia, innervated by branches
of a single nerve, I have called them the second, first, third,
and fourth divisions, respectively, of the levator maxillae supe-
rioris. This name is, however, probably not a proper one for
the third and fourth muscles.

The nerve that innervates the muscles was found double in
one Amia, and it is frequently, if not always, double in sela-
chians. It arises, both in Amia and in selachians, from the
truncus maxillaris trigemini, and is sometimes assigned to the
inferior maxillary nerve and sometimes to the superior maxil-
lary. The muscles may, therefore, belong in part to one or
more preoral arches. They apparently belong to the inter-
arcual, and not to the interbranchial muscles.

In teleosts these muscles either disappear entirely, are
entirely absorbed by the adductor mandibulae, or persist as
what are called special insertions or special divisions of the
latter muscle.

39. No interarcuales ventrales are described by Vetter in

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 753

selachians. In Acipenser and in teleosts he describes them,
calling them, in the latter, the obliqui ventrales. In Amia
these muscles resemble, in a general way, the dorsal inter-
arcual muscles ; that is, there is on each arch an interarcual
ligament, a muscle that has its insertion on its own arch, and
another that tends to have its insertion on the next posterior
arch. They therefore probably arise from the ventral portion
of the interbranchiale of Chimaera, as Vetter concludes, but
they belong, with that muscle, to the deeper layer of the
primitive constrictor, and not to the superficial layer.

The pharyngo-claviculares externus and internus are probably
the obliqui muscles of the fifth arch.

40. The geniohyoideus and intermandibularis arise together.
The geniohyoideus has an inferior and a superior portion.
From, or in connection with, the anterior end of the inferior
portion, the intermandibularis arises. The latter muscle is
double, both parts extending from mandible to mandible. The
superior division of the geniohyoideus extends from mandible
to hyoid, and seems to be one of the obliqui ventrales of the
mandibular arch. The inferior division occupies a position
intermediate between that of the superior division and that of
the intermandibularis, and may also be an obliquus ventralis,
or more probably the interbranchial muscle of the arch, in
which case the intermandibularis probably belongs to the same
muscles.

41. The hyohyoideus, in its superior portion, is the inter-
branchial muscle of its arch. In its inferior portion it may be
partly the obliquus ventralis.

42. The branchiomandibularis is found in Acipenser and
Polypterus as well as in Amia. It is represented in Elasmo-
branchii by the coraco-mandibularis, or by part of that muscle.
In teleosts it is not found. In Amia it varies greatly in dif-
ferent individuals, and undoubtedly represents a muscle in
process of deterioration and disappearance. It is probably, as
McMurrich has suggested, a muscle from which the muscles of
the tongue, in higher vertebrates, have developed. It is inner-
vated by a terminal branch of the united first, second, and
third occipital nerves.

754 ALLIS. [Vou. XII.

43. The sternohyoideus is separated by two transverse inter-
muscular septa into three portions. The posterior of these
portions is innervated by a branch of the fourth occipital
nerve, and thus belongs to the fifth and last occipital muscle
segment, which is innervated by the same nerve. The two
anterior portions of the muscle are innervated by branches of
the united first, second, and third occipital nerves, and thus
belongs to the fourth, and one or more anterior, occipital
muscle segments.

44. The development of the occipitale laterale indicates that
that bone is formed of parts of three occipital dorsal arches.
It is pierced by two ventral occipital nerves, and the nervus
vagus issues along its anterior margin between it and the hind
wall of the auditory capsule. It gives attachment to three inter-
muscular septa. Posterior to it, on the dorsal surface of the
basioccipital, there are two complete dorsal arches, each giving
attachment to an intermuscular septum. In front of each of
these two arches there is a complete spinal nerve with dorsal
and ventral roots and dorsal ganglion. There are thus five
vertebral segments indicated in the occipital region of the
skull of Amia, the first, or most anterior, segment not being
represented by a nerve, unless the vagus be, in part, that
nerve. The posterior of these five segments fuses with the
hind end of the skull in post-larval stages.

45. A complete trunk vertebra, in larvae of Amia, contains
six blocks of cartilage, three on each side. The dorsal block,
on each side, is wholly independent of the other two, and is
fused, in the adult, with the ventral end of the dorsal arch that
lies next posterior to it in larval stages. The lateral block is,
in larvae, connected at its anterior end with the posterior end
of the ventral block. It bears at its outer end the rib. The
ventral blocks lie on either side of the aorta, and resemble
vertebral processes found in Calamoichthys and Salamandra,
which, in those animals, are said to become the “ Basal-
stiimpfe”’ of the haemal arches. If they be the homologues
of those structures, as seems most probable, they are haema-
pophyses, and the ribs of Amia may be true ribs, or pleuro-
pophyses.

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 755

Io.

If.

13.

14.

15.

16.

BIBLIOGRAPHY.

AHLBORN, FREDERICK. Untersuchungen iiber das Gehirn der Petro-
myzonten. /nzaug. Dissert. Leipzig, 1883. Wilhelm Engelmann.

ALBRECHT, PAUL. Beitrag zur Morphologie des M. omo-hyoides und
der ventralen, inneren Interbranchialmusculatur in der Reihe der
Wirbelthiere. Jnzaug. Dissert. Kiel, 1876.

ALLIS, EDWARD PHELPS, JR. The Anatomy and Development of
the Lateral Line System in Amia Calva. Journ. of Morph. Vol. ii,
No. 3. April, 1889.

ALLIS, EDWARD PHELPS, JR. The Cranial Muscles and Cranial and
First Spinal Nerves in Amia Calva. Journ. of Morph. Vol. xi,
No. 2. October, 1895.

AYERS, HOWARD. Vertebrate Cephalogenesis. II. A Contribution
to the Morphology of the Vertebrate Ear, with a Reconsideration of
its Functions. Journ. of Morph. Vol. vi, No.1. May, 1892.

BALFOUR AND PARKER. On the Structure and Development of
Lepidosteus. PAzl. Trans. Roy. Soc. 1882.

Baur, G. Ueber Rippen und dhnliche Gebilde und deren Nomen-
clatur. Anat. Anz. Bd. ix, No.4. Dec. 11, 1893.

BEARD, J. The Ciliary or Motoroculi Ganglion and the Ganglion of
the Ophthalmicus profundus in Sharks. <Axat. Anz. Jahrg. ii,
Nos. 18 andi1g. Aug. 15, 1887.

BEARD, J. A Contribution to the Morphology and Development of
the Nervous System of Vertebrates. Anat. Anz. Jahrg. iii, Nos.
29 and 30. 1888.

BEARD, J. Morphological Studies. II. The Development of the
Peripheral Nervous System of Vertebrates. Part i, Elasmobranchii
and Aves. Quart. Journ. Micro. Sct. October, 1888.

BICKFORD, ELIZABETH E. The Hypophysis of the Calamoichthys
calabaricus (Smith). Amat. Anz. Bd.x, No. 15. Marchrs, 1895.

BRANDIS, F. Untersuchungen iiber das Gehirn der Vogel. II. Theil:
Ursprung der Nerven der Medulla Oblongata. Arch. f. mikr.
Anat. Bd.xli, Heft 4. July 25, 1893.

BRANDIS, F. Untersuchungen tber das Gehirn der Vogel. II. Theil:
Ursprung der Nerven der Medulla Oblongata. Arch. fi mikr.
Anat. Bad. xliii, Heft 1. Feb. 19, 1894.

BRANDIS, F. Untersuchungen tiber das Gehirn der Vogel. III. Theil:
Der Ursprung des N. Trigeminus und der Augenmuskelnerven.
Arch. f. mikr. Anat. Bd. xliv, Heft 4. March 30, 1895.

BRIDGE, T. W. The Cranial Osteology of Amia Calva. Journ. of
Anat. and Phys. Vol. xi, Part 4. July, 1877.

BURCKHARDT, RUDOLF. Das Centralnervensystem von Protopterus
annectens. Berlin, 1892.

756

Ly:

20.

2I.

22.

23;

24.

253

26.
27.

28.

29.

30.

2,

32.

ALLIS. [VoL. XII.

BURCKHARDT, RUDOLF. Die Homologieen des Zwischenhirndaches
und ihre Bedeutung fiir die Morphologie des Hirns bei neideren
Vertebraten. Anat. Anz. Bd. ix, Nos. 5 and 6. Dec. 23,
1893.

CHIARUGI, GIULIO. Contribuzioni allo studio dello sviluppo dei nervi
encefalici nei Mammiferi in confronto con altri vertebrati. G, Car-
nesecchi e figli. Firenze, 1894.

COLLINGE, WALTER EDWARD. The Sensory Canal System of
Fishes, Part i, Ganoidei. Quart. Journ. Micro. Sct. No. 144.
August, 1894.

COLLINGE, WALTER EDWARD. On the Sensory Canal System of
Fishes. Teleostei— Suborder A. Physostomi. Proc. Zool. Soc.
London, April 2, 1895.

CorninG, H. K. Ueber die Entwickelung der Zungenmusculatur bei
Reptilien. Verh. anat. Ges. 9th Vers. Basel, April 17-20,
1895.

DOHRN, ANTON. Bemerkungen uber den neuesten Versuch einer
Lésung des Wirbeltierkopf-Problems. Axat. Anz. Jahrg. v, Nos.
2and3. Jan. 18 and Feb. 1, 1890.

DoLLo, Louis. Sur la Phylogénie des Dipneustes. Bruxelles,
Juillet, 1895.

EBNER, V. VON. Ueber die Beziehungen der Wirbel zu den Urwirbeln.
Sitzungsbh. k. Akad. ad. Wiss. Math. Natur. Cl. Bd. ci, Abt. iii.
Wien, 1892.

EMERY, CARLO. Le Specie del Genere Fierasfer nel Golfo di Napoli
e Regioni limitrofe. Fauna und Flora des Golfes von Neapel. ii.
Monographie. 1880.

Ewar7, J. C. On the Cranial Nerves of Elasmobranch Fishes.
Prelim. Com. Proc. Roy. Soc. Vol. xlv. March 7, 1889.

Ewart, J. C. On the Development of the Ciliary or Motor Oculi
Ganglion. Proc. Roy. Soc. Vol. xlvii. 1890.

Ewart, J. C. The Cranial Nerves of the Torpedo (preliminary
note). Proc. Roy. Soc. Vol. xlvii. 1890.

Ewart, J. C. The Lateral Sense Organs of Elasmobranchs. I.
The Sensory Canals of Laemargus. Zyvans. Roy. Soc. Edinburgh.
Vol. xxxvii, Part i, Nos. 5 and 6. July 6, 1891.

Ewart, J. C. Supplementary Note on the Cranial Nerves of Elas-
mobranchs. University of Edinburgh. November, 1892.

Ewart, J. C. On the Dorsal Branches of the Cranial and Spinal
Nerves of Elasmobranchs. Proc. Roy. Soc. Edinburgh. March 18,
1895.

Ewart, J. C., AND MITCHELL, J. C. The Lateral Sense Organs of
Elasmobranchs. II. The Sensory Canals of the Common Skate
(Raia batis). Zrans. Roy. Soc. Edinburgh. Vol. xxxvii, Part i,
Nos:.5 and.6, "Dec: 21,1861.

;
ne)
ps

No. 3.]} MUSCLES AND NERVES IN AMIA CALVA. 757

33:

34.

35.

36.

37:

38.

39.

40.

4l.

42.

43.

44.

45.
46.
47.
48.

49.

FiELD, H. H. Bemerkungen tiber die Entwicklung der Wirbelsdule
bei den Amphibien ; nebst Schilderung eines abnormen Wirbelseg-
mentes. Morph. Jahrb. Bad. xxii, Heft 3. April 23, 1895.

FRORIEP, AUGUST. Zur Entwickelungsgeschichte der Wirbelsdule
insbesondere des Atlas und Epistropheus und der Occipitalregion.
I. Beobachtungen an Hiihnerembryonen. Arch. f. Anat. u. Phys.
Anat. Abt. 1883.

FRORIEP, AUGUST. Entwickelungsgeschichte des Kopfes. Evrged. d.
Anat. u. Entwickelungsgesch. Bd. i. 1891.

FURBRINGER, MAx. Ueber die mit dem Visceralskelet verbundenen
spinalen Muskeln bei Selachiern. Jen. Zeitschr. f. Naturw. Bad.
xxx, Heft 1. October 18, 1895.

FURBRINGER, PAUL. Untersuchungen zur vergleichenden Anatomie
der Muskulatur des Kopfskelets der Cyclostomen. /ex. Zeitschr. f.
Naturw. Bd. ix, Heft 1. Jan. 30, 1875.

Gabow, Hans. On the Modifications of the First and Second
Visceral Arches, with especial Reference to the Homologies of the
Auditory Ossicles. PAzl. Trans. Roy. Soc. London. Vol. 179. 1888.

GAbDow, H., AND ABpoTT, Miss E. C. On the Evolution of the
Vertebral Column of Fishes (abstract). Proc. Roy. Soc. Vol. lvi,
No. 337. Sept. 21, 1894.

GAGE, SUSANNA PHELPS. The Brain of Diemyctylus viridescens,
from Larval to Adult Life, and Comparison with the Brain of Amia
and Petromyzon. The Wilder Quarter-Century Book. Ithaca, 1893.

Gaupp, E. Ueberdie Anlage der Hypophyse bei Sauriern. Arch. f.
mik. Anat. Bd. xlii, Heft 3. Nov. 30, 1893.

GEGENBAUR, CARL. Untersuchungen zur vergleichenden Anatomie
der Wirbelthiere. 2 Hefte. (1) Schultergirtel der Wirbelthiere.
(2) Brustflosse der Fische. Leipzig, 1865.

GEGENBAUR, CARL. Ueber die Kopfnerven von Hexanchus. en.
Zettschr. f. Naturw. Vol. vi. 1871.

GEGENBAUR, CARL. Untersuchungen zur vergleichenden Anatomie
der Wirbelthiere. 3 Hefte. Das Kopfskelet der Selachier, ein
Beitrag zur Erkenntniss der Genese des Kopfskelets der Wirbel-
thiere. ) Leipzig, 1372.

GEGENBAUR, CARL. Zur Phylogenese der Zunge. Morph. Jahrb.
Bd. xxi, Heft 1. Feb. 27, 1894.

GOPPERT, ERNST. Der Musculus Obliquus superior oculi der Mono-
tremen. Morph. Jahrb. Bd. xxi, Heft 2. April 6, 1894.

GOpPERT, Ernst. Zur Kenntniss der Amphibienrippen. Vorlaufige
Mittheilung. Morph. Jahrb. Bd, xxii, Heft 3. April 23, 1895.

GOpPERT, ERNST. Untersuchungen zur Morphologie der Fischrippen.
Morph. Jahrb. Bd. xxiii, Heft 2. Nov. 12, 1895.

GorETTE, A. Ueber die Zusammensetzung der Wirbel bei den
Reptilien. Zool. Anz. Jahrg. xvii, p. 359. Oct. 8, 1894.

758

50.

ie

52.

53:

54.

55.

56.

57-

58.

59:
60.
61.
62.

63.

64.

65.

ALLIS. (VoL. IT.

GoronowItTscu, N. Das Gehirn und die Cranialnerven von Acipenser
ruthenus. Ein Beitrag zur Morphologie des Wirbeltheirkopfes.
Morph. Jahrb. Bad. xiii.

GROSSER, O., AND BREZINA, E. Ueber die Entwicklung der Venen
des Kopfes und Halses bei Reptilien. Morph. Jahrb. Bd. xxiii,
Heft 2. Nov. 12, 1895.

HasseE, C. Die Entwicklung und der Bau der Wirbelsdule der
Ganoiden. Fiinfte Abhandlung tiber die Entwicklung der Wirbel-
saule. Zettschr. f. wiss. Zool. Bd. vii, Heft 1. Dec. 12, 1893.

HASSE, C. Die Entwicklung der Wirbelsaule der Cyclostomen. Sechste
Abhandlung iiber die Entwicklung der Wirbelsdule. Zeztschr. f.
wiss. Zool. Bd. lvii, Heft 2.. Dec. 31, 1893.

HATSCHEK. Die Metamerie des Amphioxus und des Ammocoetes.
Verh. der anat. Gesellsch. 6th Vers. Wien, June 7-9, 1892. Azxat.
Anz. Erganzungsheft, Bd. vii.

HERRICK, C.JupDsoN. The Crania Nerves of Amblystoma punctatum.
Journ. Comp. Neur. Vol.iv. December, 1894.

HoFFMAN, C. K. Weitere Untersuchungen zur Entwickelungsge-
schichte der Reptilien. Morph. Jahrb. Ba. xi.

HoFFMANN, C.K. Zur Entwickelungsgeschichte des Selachierkopfes.
Anat. Anz. Bd. ix, No. 21. July 18, 1894.

Hotm, JoHN F. Some Notes on the Early Development of the
Olfactory Organ of Torpedo. <Axat.Anz. Bd. x, No.6. Oct. 27,
1894.

Huprecut, A. A. W. Beitrag zur Kenntniss des Kopfskeletes der
Holocephalen. Ved. Arch. f. Zool. Bad. iii, Heft 3. May, 1877.

Hux.ey, T. H. A Manual of the Anatomy of Vertebrated Animals.
D. Appleton & Co., New York. 1872.

IHERING, H. von. Ueber Wirbelverdoppelung bei Fischen. Zool.
Anz. Jahrg.i, No.4. Aug. 12, 1878.

KAESTNER, SANDOR. Ueber die allgemeine Entwickelung der Rumpf-
und Schwanzmusculatur bei Wirbelthieren. Mit besonderer Beriick-
sichtigung der Selachier. Arch. f. Anat. u. Phys. Jahrg. 1892.
Anat. Abth. Heft 3-4. Nov. 3, 1892.

KILLIAN. Zur Metamerie des Selachierkopfes. Azat. Anz. Ergan-
zungsheft, Jahrg. vi. Verh. anat. Gesellsch. 5th Vers. Miinchen,
May 18-20, 1891. Amat. Anz. Erganzungsheft, Bd. vi.

KLAATSCH, H. Beitriige zur vergleichenden Anatomie der Wirbel-
saule. II. Ueber die Bildung knorpeliger Wirbelk6rper bei
Fischen. Morph. Jahrb. Bd. xx, Heft 2. August 4, 1893.
Abstr. in Zool. Jahrb. f. 1893. p. 117.

KiLaatscu, H. Ueber die Bedeutung der Hautsinnesorgane fiir die
Ausschaltung der Skleroblasten aus dem Extoderm. Verh. anat.
Gesellsch. goth Vers. Basel, April 17-20, 1895. Anat. Anz.
Erganzungsheft, Bd. x.

No. 3.] MUSCLES AND NERVES I[N AMIA CALVA-: 759

66.

67.

68.

69.

70.

71a.

72)

73:

74.
75:

76.

77:

78.

79:

80.

KupFFER, C. von. Die Entwicklung von Petromyzon Planeri.
Arch. f. mikr. Anat. Bd. xxxv, Heft 4. July 22, 1890.

KUPFFER, C. von. Die Entwickelung der Kopfnerven der Verte-
braten. Verh. anat. Gesellsch. 5th Vers. Muinchen, May 18-20,
1891. Anat. Anz. Erganzungsheft, Jahrg. vi.

KupFFER, C. von. Entwickelungsgeschichte des Kopfes. £yrged. d.
Anat. und Entwickelungsgeschichte. Bd.ii. 1892.

KuPFFER, C. VON. Studien zur vergleichenden Entwicklungsge-
chichte des Kopfes der Kranioten. Heft 1. Die Entwicklung des
Kopfes von Acipenser sturio an Medianschnitten untersucht.
Miinchen und Leipzig, 1893.

KUPFFER, C. VON. Studien zur vergleichenden Entwicklungsges-
schichte des Kopfes der Kranioten. Heft 2. Die Entwicklung des
Kopfes von Ammocoetes Planeri. Miinchen und Leipzig, 1894.

KUPFFER, C. VON. Die Deutung des Hirnanhanges. Sztzungsd. d.

Gesellsch. f. Morph. u. Phys. in Miinchen. 1894.
KUPFFER, C. von. Ueber die Entwickelung des Kiemenskelets von
Ammocoetes und die organogene Bestimmung des Exoderms. Verh.
anat. Gesellsch. gth Vers. Basel, April 17-20, 1895. Anat. Anz.
Erganzungsheft, Bd. x.

LEE, STuART. Zur Kenntniss des Olfactorius. Bericht. d. Naturf.
Ges. zu Freiburg t. B. November, 1893.

LENHOSSEK, M. von. Beitrage zur Histologie des Nervensystems
und der Sinnesorgane. IX. Ueber das Ganglion sphenopalatinum
und den Bau der sympathischen Ganglien. Wiesbaden, 1894.

LeypiGc, F. Integument und Hautsinnesorgane der Knockenfische.
Weitere Beitrage. Zool. Jahrb. Bad. viii.

McMurricu, J. PLAYFAIR. The Myology of Amiurus catus L. Gill.
Proc. Canad. Inst. Vol. ii, Fasc. 3. Toronto, October, 1884.

McMurricu, J. PLAYFAIR. The Cranial Muscles of Amia Calva
L., with a consideration of the Relations of the Post-Occipital and
Hypoglossal Nerves in the Various Vertebrate Groups. Studies
Jrom the Biol. Lab. of Johns Hopkins Univ. Baltimore, June,
1885.

MARSHALL, A. MILNES, AND SPENCER, W. BALDWIN. Observations
on the Cranial Nerves of Scyllium. Studies from the Biol. Lab.
of Owen’s College. Vol. i, p. 87. 1886.

Maurer, F. Die Elemente der Rumpfmuskulatur bei Cyclostomen
und héheren Wirbelthieren. Ein Beitrag zur Philogenie der quer-
gestreiften Muskelfaser. J/orph. Jahrb. Bd. xxi, Heft 4. Sept.
18, 1894.

MERKEL, Fr. Ueber die Endigungen die sensiblen Nerven in der
Haut der Wirbelthiere. Rostock, 1880.

MiTRopHANOW, Paut. Etude embryogénique sur les Sélaciens.
Arch. de Zool. Exp. et Gén. Tome i, série 3.

760
St.
82.

83.

84.

85.

86.

87.

88.

89.

go.

92.

93.
94.

95.

96.
97.

98.

ABIELS: [Vor. XII.

MIVART, ST. GEORGE. The Cat. An Introduction to the Study of
Backboned Animals, especially Mammals. London, 1881.

MOLLIER, S. Die paarigen Extremitaten der Wirbelthiere. II. Das
Cheiropterygium. xvi Hefte (Bd. v, Heft 3). 1895.

PARKER, J. JEFFERY. XVII. On the Blood-vessels of Mustelus
Antarcticus: a Contribution to the Morphology of the Vascular
System in the Vertebrata. Pz. Trans. June to, 1886.

PARKER, W. H. II. On the Structure and Development of the
Skull in Salmon (Salmo salar L.). Bakertan Lectures. Read May
30,1872. Ai. 17ans. 11573:

PARKER, W. H. V. On the Structure and Development of the
Skull in Sharks and Skates. Read Nov. 7, 1876. Zvrans. Zool. Soc.
Vol. Io.

PARKER, W. H. III. On the Structure and Development of the
Skull in Sturgeons (Acipenser ruthenus and A. sturio). Read May
5, Pool. PAW. Trans. 1862;

PARKER, W. H. VIII. On the Development of the Skull in Lepi-
dosteus osseus. Read Dec. 8, 1881. PAzl. Trans. 1882.

Pinkus, FELIx. Ueber einen noch nicht beschriebenen Hirnnerven
des Protopterus annectens. Azat. Anz. Bd. ix, No. 18. June 23,
1894.

PINKUS, FELIx. Die Hirnnerven des Protopterus annectens. Morph.
Arbeit. Bd. iv, Heft 2. 1894.

PLATT, JuL1A B. A Contribution to the Morphology of the Verte-
brate Head, based on a Study of Acanthias vulgaris. /ourn.
of Morph. Vol.v, No. 1. June, 1891.

PLATT, JULIA B. Ontogenetische Differenzirung des Ectoderms in
Necturus. I. Studie Archiv f. mikr. Anat. Bd. xliii, Heft 4.
June 30, 1894.

PLESSEN, Baron Jos. v. and RABINOVICZ, JOHN. Die Kopfnerven
von Salamandra maculata in vorgeriickten Embryonalstadium.
Munchen, J. F. Lehmann. 1891.

POLLARD, H. B. On the Anatomy and Phylogenetic Position of
Polypterus. Zool. Jahrb. Bd. v, Heft 3 and 4. Oct. 20, 1892.

POLLARD, H. B. The Lateral Line System in Siluroids. Zool.
Jahrb. Bd. v, Heft 3 and 4. Oct. 20, 1892.

POLLARD, H. B. Observations on the Development of the Head in
Gobius capito. Quart. Journ. Micro. Sci. No. 139. January,
1894.

PoLLarpD, H. B. The “ Cirrhostomial” Origin of the Head in Ver-
tebrates. Anat. Anz. Bd.ix, No. 11. March 30, 1894.

PoLLARD, H. B. The Oral Cirri of Siluroids and the Origin of the
Head in Vertebrates. Zool. Jahrb. Bad. viii, Heft 3. May 11, 1895.

POLLARD, H. B. The Suspension of the Jawsin Fishes. Avzat. Anz.
Bd. x, No.1. «Sept. to; 1s04.

»

100.

Iol.

102.

103.

104.

105.

106.

107.

108.

109.

Ilo,

Ill.

112.

0X3:

114.

.3:.] MUSCLES AND NERVES [N AMIA CALVA. 761

. POLLARD, H.B. Ueber Labialknorpel. Verh. anat. Gesellsch. 9th

Vers. Basel, April 17-20, 1895. Amat. Anz. Ergiénzungsheft,
Bd. x.

Quan. Elements of Anatomy, edited by E. A. Schafer and G. D.
Thane. Longmans, Green & Co., London. 1892-1895.

RABL, C. Ueber die Metamerie des Wirbelthierkopfes. Verh.
anat. Gesellsch. 6th Vers. Wien, June 7-9, 1892. Anat. Anz.
Erganzungsheft. Bad. vii.

RABL, C. Theorie des Mesoderms (Fortsetzung). Morph. Jahrb.
Bd. xix. Abstr. in Zool. Jahrb. f. 1892. p. 132.

Retzius, GustaF. Ueber das Ganglion ciliare. Avat. Anz. Bd.
ix, No. 21. July 18, 1894.

SAGEMEHL, M. Beitrage zur vergleichenden Anatomie der Fische. I.
Das Cranium von Amia calva, L. Morph. Jahrb. Bad. ix, Heft 2.
1883.

SAGEMEHL, M. Beitrage zur vergleichenden Anatomie der Fische.
II. Einige Bemerkungen tiber die Gehirnhaute der Knochenfische.
Morph. Jahrb. Bd. ix, Heft 4. 1884.

SAGEMEHL, M. Beitrage zur vergleichenden Anatomie der Fische.
III. Das Cranium der Characiniden, nebst allgemeinen Bemerk-
ungen uber die mit einem Weber’schen Apparat versehenen Phy-
sostomenfamilien. Morph. Jahrb. Bd. x, Heft 1. 1884.

SAGEMEHL, M. Beitrage zur vergleichenden Anatomie der Fische.
IV. Das Cranium der Cyprinoiden. Morph. Jahro. Bad. xvii, Heft
Ane OCEs23, WSOl.

SARASIN, PAUL AND Fritz. Ergebnisse naturwissenschaftlicher
Forschungen auf Ceylon. Bd. ii, Heft 4. Wiesbaden. C. W.
Kreidel’s Verlag. 1890.

SCHEEL, C. Beitrage zur Entwickelungsgeschichte der Teleostierwir-
belsdule. Morph. Jahrb. Bd. xx. Abstr. in Zool. Jahrb. f. 1893.
Pa.t20}

SCHMIDT, LuGwic. Untersuchungen zur Kenntnis des Wirbelbaues
von Amia calva. Zeztschr. f. wiss. Zool. Bd. liv, Heft 4. Oct.
18, 1892.

SCHNEIDER, ANTON. Beitrage zur vergleichenden Anatomie und
Entwicklungsgeschichte der Wirbelthiere. Berlin, 1879.

SCHNEIDER, H. Ueber die Augenmuskelnerven der Ganoiden.
Inaug. Dissert. Jena, 1881.

SCHWALBE, G. Das Ganglion oculomotorii. Ein Beitrag zur ver-
gleichenden Anatomie der Kopfnerven. Jen. Zettschr. f. Naturw.
Bd. xiii, Heft 2. July 16, 1879.

SEDGWICK, ADAM. On the Inadequacy of the Cellular Theory of
Development, and on the Early Development of Nerves, particularly
of the Third Nerve and of the Sympathetic in Elasmobranchii.
Quar. Journ. Mic. Sct. No. 145. November, 1894.

762

11S.

116.

17:

118.

119.

120.

12!I.

122.

123.

124.

125;

126.

127:

128.

129.

ALLIS. (VoL. XII.

SHUFELDT, R. W. The Osteology of Amia Calva, including certain
special References to the Skeleton of Teleosteans. Gov. Printing
Office, Washington. 1885.

STANNIUS, H. Das peripherische Nervensystem der Fische. Ros-
tock, 1849. .
STOHR, PHILIPP. Zur Entwicklungsgeschichte des Urodelenschadels.

Leipzig, W. Englemann. 1879.

STOHR, PHiLipp. Zur Entwickelungsgeschichte des Kopfskeletes der
Teleostier. Festschrift 2. Feter ad. 300 jahr. Best. da. Julius-Max.-
Univ. 2. Wirzburg. Leipzig, C. W. Vogel. 1882.

STRONG, OLIVER S. The Structure and Homologies of the Cranial
Nerves of the Amphibia as determined by their Peripheral Distribu-
tion and Internal Origin. Zool. Anz. No. 348. 1890.

STRONG, OLIVER S. The Structure and Homologies of the Cranial
Nerves of the Amphibia as determined by their Peripheral Distribu-
tion and Internal Origin. Part 2. Anat. Anz. Jahrg. vii, No. 15.
1892.

STRONG, OLIVER S. The Cranial Nerves of Amphibia. A Contri-
bution to the Morphology of the Vertebrate Nervous System.
Journ. of Morph. Vol. x, No.1. January, 1895.

STUDNICKA, F. K. Zur Lo6sung einiger Fragen aus der Morphologie
des Vorderhirnes der Cranioten. Anat. Anz. Bd. ix, No. to.
Feb. 28, 1894.

TIESING, BERTHOLD. Ein Beitrag zur Kenntnis der Augen-, Kiefer-
und Kiemenmuskulatur der Haie und Rochen. Jen. Zettschr. f.
Naturw. Bd. xxx, Heft 1. Oct. 18, 1895.

VETTER, BENJAMIN. Untersuchungen zur vergleichenden Anatomie
der Kiemen -und Kiefermusculatur der Fische. Theil I. Selachier.
Jen. Zeitschr. f. Naturw. Bd. viii, Heft 3. Sept. 12, 1874.

VETTER, BENJAMIN. Untersuchungen zur vergleichenden Anatomie
der Kiemen- und Kiefermuskulatur der Fische. Theil II. A. Chimaera
monstrosa. B. Acipenser sturio. C. Knochenfische. Jen. Zeztschr.
fj. Naturw. Bd. xii, Heft 3. Aug. 1, 1878.

WALDSCHMIDT, JULIUS. Beitrag zur Anatomie des Zentralnerven-
systems und des Geruchsorgans von Polypterus bichir. Amat. Anz.
Jahrg. ii, No. 11. May 15, 1887.

WHuiteE, PuiLip J. The Existence of Skeletal Elements between the
Mandibular and Hyoid Arches in Hexanchus and Laemargus.
Anat. Anz. Bd. xi, No.2. Aug. 22, 1895.

WIEDERSHEIM, ROBERT. Grundriss der vergleichenden Anatomie
der Wirbelthiere. Jena, 1893.

VAN WIJHE, J. W. Ueber das Visceralskelett und die Nerven des
Kopfes der Ganoiden und von Ceratodus. Méederl. Archiv f. Zool.
Bd. v, Heft 3. July, 1882.

No. 3.] MUSCLES AND NERVES IN AMIA CALVA. 763

130. VAN WIJHE, J. W. Ueber die Mesodermsegmente und die Ent-
wickelung der Nerven des Selachierkopfes. Amsterdam, 1882.

131. WiLson, Henry V. The Embryology of the Sea Bass (Serranus
alvarius). Gov. Printing Office, Washington. 1891.

132: WRIGHT, R. RAMSEY. On the Nervous System and Sense-Organs
of Amiurus. Proc. of the Canad. Inst. Vol. ii, tasc. No. 3.
Toronto, 1884.

133. WRIGHT, R. RAMSEY. On the Hyomandibular Clefts and Pseudo-
branchs of Lepidosteus and Amia. /ourn. of Anat. and Phys.
Vol. xix. July, 1885.

EXPLANATION OF PLATES.

INDEX LETTERS.

a. ossicle a of Bridge.
aa. ampulla anterior.
A,. superficial portion of adductor mandibulae; A.’, A2”, A2’”, its
three divisions.
A,;. deeper portion of adductor mandibulae.
Aw. mandibular portion of adductor mandibulae; Aw’, Aw”, its two
divisions.
A,Aw’. tendon connecting A, and Aw’.
4;Aw”’. tendon connecting A; and Aw”.
Aab. lV, Aab.V. adductores arcuum branchialium of fourth and fifth arches.
ab. nervus abducens.
abfr. foramen for nervus abducens.
ac. nervus acusticus.
acufr. foramen for anterior cerebral vein.
ae. ampulla externa.
agi. ramus anterior glossopharyngei.
Ah. adductor hyomandibularis.
allg. accessory lateral-line groups of pit organs.
ana. anterior nasal aperture.
ANT. antorbital (Sag.), preorbital (Bridge).
Ao. adductor operculi.
ap. ampulla posterior.
aps. afferent pseudobranchial artery.
ART. articular.
AS. alisphenoid.
AUP. autopalatine.
6. ossicle b of Bridge.
BB.*-3 first, second, and third basibranchials.
éf. ramus buccalis facialis.
éfi0.'-** branches of buccalis facialis to infraorbital sense organs, Nos.
I to I4.

704

ALLIS.

branchiomandibularis.

basioccipital (Bridge), occipitale basale (Sag.).

branchiostegal rays.

basisphenoid.

ossicle c of Bridge.

carotid artery.

coraco-arcualis anterior‘.

canal for intervertebral artery and sympathetic nerve.

ciliaris brevis.

first to fifth ceratobranchials.

common carotid artery.

ceratohyal.

clavicle.

ciliares longi.

coronoid process.

constrictor pharyngis.

anterior semicircular canal.

external semicircular canal.

posterior semicircular canal.

canalis transversus.

canalis utriculo-saccularis.

dentary.

ossicle d of Bridge.

dorsal arches of first six vertebrae.

dorsal occipital arches.

dorsal aorta.

dilatator operculi.

dermopalatine.

dorsal cartilaginous occipital and vertebral processes.

efferent arteries of first to fourth branchial arches.

first to fifth epibranchials.

external carotid artery.

foramen for external carotid artery.

mandibular branch of external carotid artery.

branch of external carotid accompanying ophthalmic branches
of trigeminus and facialis.

ectopterygoid.

epihyal.

eye-muscle canal.

upper lateral portion of eye-muscle canal.

entopterygoid.

exoccipital, occipitale externum (Sag.), epiotic (Bridge).

epiphysis.

efferent pseudobranchial artery (arteria ophthalmica, Sag.).

foramen for efferent pseudobranchial artery.

extrascapular (Sag.), supratemporal (Bridge).

ethmoid.

fore-brain.

facial canal through the hyomandibular.

(Vor. XII.

No. 3.) MUSCLES AND NERVES IN AMIA CALVA.

ifr.
Jia.
f-

IOP.
ip.

IPB. I-III.
115570,

TG:

LA.

765

facial foramen.

anterior flagellum.

posterior flagellum.

foramen occipitale magnum.

frontal.

gular.

ciliary ganglion.

Gasserian ganglion.

glossopharyngeal ganglion.

inferior portion of geniohyoideus.

superior portion of geniohyoideus.

nervus glossopharyngeus.

glossopharyngeal foramen.

ganglion of rami ophthalmicus superficialis and buccalis facialis.

ganglia on third and fourth occipital nerves.

profundus ganglion.

spinal ganglia.

trigemino-facial ganglion.

intracranial vagus ganglion.

ganglia of first to fourth vagus.

hind-brain.

first to fourth hypobranchials.

ramus hyoideus facialis.

hypophysial fenestra.

subcutaneous branches of ramus hyoideus facialis.

hypohyal.

inferior portion of hyohyoideus.

superior portion of hyohyoideus.

horizontal pit line of cheek.

hyomandibular.

truncus hyoideo-mandibularis facialis.

hyo-opercularis artery.

hypophysis.

intercalar (Sag.), opisthotic (Bridge).

internal carotid artery.

internal carotid canal.

internal carotid foramen.

ethmoid incisur.

infraorbital groups of pores Nos. I to 21, or trunk canals lead-
ing to those pores.

intermandibularis.

intermuscular septa.

infraorbital lateral canal.

interoperculum.

preorbital incisur.

first to third infrapharyngobranchials.

double trunk at union of infraorbital and supraorbital canals.

jugal.

lachrymal.

766
Labe. l-V.

Labia,

LabiP.

Lap.

Zi.

lid.h. to lid. ILI.

liv.h. to liv. LV.

lig.
“mob.
lmh.
lmi.
Lms.3-4
EO:
loca,
loce.
lol.
op2-6
ltfr.

ludi,
luds.
ul.

M.
MM.
mob.
mde.
mef.
mefe.
mef.hl.

mef.madl.

mef.omo.*-*©

mef.vl.

mf.

mif.
mit.
mitfr.
MP.
ls. I-III
mst.
MSLMX.
MX.

Mx.

ALLIS, Vor, X11,

levator arcus branchialis externus of first to fifth branchial
arches.

levator arcus branchialis internus, anterior muscle.

levator arcus branchialis internus, posterior muscle.

levator arcus palatini.

lobus inferior.

ligamenta interarcualia dorsalia of hyoid and first three bran-
chial arches.

ligamenta interarcualia ventralia of hyoid and first four bran-
chial arches.

lateral-line groups of pores.

ligament from mandible to first branchiostegal ray.

ligamentum mandibulo-hyoideum.

ligamentum mandibulo-interoperculum.

first to fourth divisions of the levator maxillae superioris.

levator operculi.

anterior lateral occipital ligament.

posterior lateral occipital ligament.

lobus olfactorius.

lateral cartilaginous vertebral processes.

foramen of ramus lateralis trigemini, found in Amia as a branch
of the vagus.

ligamentum longitudinale vertebrale dorsale inferior.

ligamentum longitudinale vertebrale dorsale superior.

lateral vertebral ligament.

Meckel’s cartilage.

mentomeckelian ossification.

mid-brain.

mandibular lateral canal.

ramus mandibularis externus facialis.

canal in articular for ramus mandibularis externus facialis.

branch of r. mandibularis externus facialis to horizontal cheek
line of pit organs.

branch of r. mandibularis externus facialis to mandibular line
of pit organs.

branches of r. mandibularis externus facialis to operculo-man-
dibular sense organs, Nos. I-16.

branch of r. mandibularis externus facialis to vertical cheek
line of pit organs.

truncus mandibularis facialis.

ramus mandibularis internus facialis.

ramus maxillaris inferior trigemini.

foramen for ramus maxillaris inferior trigemini.

metapterygoid.

first eleven muscle segments of trunk.

ramus maxillaris superior trigemini.

maxillary branch of ramus maxillaris superior trigemini.

maxillary.

maxilla.

Rey

No. 3.] MUSCLES AND NERVES IN AMIA CALVA.

n.
NA.
net.

ne.

nil.
nil.all.

nll.dl.
ndl.io.38-23

nil.llo.
nli.pl.

nll.sto.3-3

707

nerve “n” of Pinkus.

nasal.

ramus nasalis or nasociliaris trigemini.

nasal epithelium.

nervus lineae lateralis.

branches of n. lineae lateralis to pit organs of accessory lateral
line.

branches of n. lineae lateralis to pit organs of dorsal body line.

branches of n. lineae lateralis to infraorbital sense organs, Nos.
18-21.

branches of n. lineae lateralis to sense organs of lateral line.

branches of n. lineae lateralis to pit organs of posterior line of
head.

branches of n. lineae lateralis to sense organs of supratemporal
commissure.

nasal tube.

nervus opticus.

occipitale basale (Sag.), or basioccipital (Bridge).

occipital nerves.

anterior or communicating branches of occipital nerves.

dorsal branches of occipital nerves.

horizontal branches of occipital nerves.

ventral branches of occipital nerves.

nervus oculomotorius.

inferior branch of nervus oculomotorius.

superior branch of nervus oculomotorius.

obliqui dorsales.

ramus oticus facialis.

canal for ramus oticus facialis.

branches of ramus oticus facialis to sense organs, Nos. 15 and
16, of infraorbital canal.

optic fenestra.

foramen of first and second occipital nerves.

branch of oticus facialis to sense organ of spiracular canal.

obliquus inferior.

occipitale laterale (Sag.), or exoccipital (Bridge).

nervus olfactorius.

olfactory canal.

olfactory foramen.

olfactory pit.

operculo-mandibular lateral canal.

operculo-mandibular groups of pores, Nos. 1-16, or trunk
canals leading to those pores.

orbito-nasal canal.

orbito-nasal fenestra.

operculum.

optic perforation of sclerotic.

ramus ophthalmicus superficialis facialis.

branch of ramus ophthalmicus superficialis facialis to pit organs
of anterior line of head.

768

PA Ss I ESS:

(VoL. XII.

opfr. ophthalmic foramen.
opf.so.-7 branches of ramus ophthalmicus superficialis facialis to sense
organs of supraorbital canal.
opp. ramus ophthalmicus profundus trigemini.
oprf. ramus opercularis facialis.
opt. ramus ophthalmicus superficialis trigemini.
opt.sc. subcutaneous branches of ramus ophthalmicus trigemini.
optstv.sc. fused subcutaneous branches of ramus ophthalmicus trigemini
and supratemporal branch of vagus root.
OS. orbitosphenoid.
os. obliquus superior.
Ov. /-/77. obliqui ventrales of first three branchial arches.
Ov. /V.*-3 three divisions of obliquus ventralis of fourth arch.
ov and vo. vein coming from eyeball.
ov’and vo’. vein coming from eyeball.
p. pectoral fin.
PA. parietal.
paf. ramus palatinus anterior facialis.
pafir. foramen for ramus palatinus anterior facialis.
pe. palatine canal.
pee. pharyngo-clavicularis externus.
pet. pharyngo-clavicularis internus.
PE. petrosal.
pf ramus palatinus facialis.
pfr. palatine foramen.
pel. ramus pharyngeus glossopharyngei.
ph. pit for hypophysis cerebri and saccus vasculosus.
piv4 ramus pharyngeus inferior n. vagi quarti.
PMX. premaxillary.
pn. prenasal process (Bridge).
pra. posterior nasal aperture.
POP. preoperculum.
popp. portio ophthalmici profundi.
POR. first postorbital.
POR. second postorbital.
ppc. canal for ramus palatinus posterior facialis between the auto-
and dermo-palatine bones.
ppf ramus palatinus posterior facialis.
ppfr. foramen for ramus palatinus posterior facialis.
fq. cartilaginous portion of palato-quadrate.
pr. preorbital process.
PRE. prefrontal, or antorbital ossification.
prel. ramus pretrematicus glossopharyngei.
prv.-4 pretrematic branches of first to fourth vagus nerves.
PS. parasphenoid.
psb. pseudobranch.
PSF. postfrontal (Sag., Allis), dermo-postfrontal (Bridge).
psgl. ramus posttrematicus glossopharyngei.
PSP. processus spinosus.

No. 3.] MUSCLES AND NERVES IN AMIA CALVA,

PST.

psu.

psu4

Duo

Q.

7.Ca.

7.Ua.

r.aa.

r.ae.

r.ap.

r.mfu.

rms.

Y.mu.

r.pl.

Vllry PAC, TO,
Vibsy TCAbry FACE;
Ay Abs, We.
r.am.

v.om.

r. ght.

r.ehs.

rim.

vr lap.do.
rims.

r.mait.
r.mit.SC.
r.MSt.SC.
Rabd.

RB.

rb.

rbb.

rdoc.3-4

re.

rel.

rit.

rl.

robf.

rp.

r. phegl.

VS.

rtfa.
refe.
rvoc 3+,
SANG.
SIGs

SC.

SCL.

769

postorbital ossification (Allis), postfrontal (Sag.), sphenotic
(Bridge).

posttrematic branches of first to third vagus.

ramus pharyngeus superior n. vagi quarti.

pharyngeal branches of first to third vagus.

quadrate.

ramus cochlearis acustici.

ramus vestibularis acustici.

ramulus ampullae anterioris.

ramulus ampullae externae.

ramulus ampullae posterioris.

ramulus maculae acusticae fundi utriculli.

ramulus maculae acusticae sacculi.

ramulus maculae acusticae recessus utriculi.

ramulus papillae lagenae.

accessory branches of nervus trigeminus.

branch of r. max. inf. trig. to A, and A3.
branch of occipital nerves to Bm.

branch of r. max. inf. trig. to Ghi.

branch of r. max. inf. trig. to Ghs.

branch of r. max. inf. trig. to Im.

branch of r. max. inf. trig. to Lap. and Do.
branch of r. max. inf. trig. to Lms.

ramus mandibularis internus trigemini.
subcutaneous branches of r. max. inf. trig.
subcutaneous branches of r. max. sup. trig.
retractor arcuum branchialum dorsalis.
ribs.

radix brevis.

retractor bulbi.

dorsal roots of third and fourth occipital nerves.
rectus externus.

root of nervus glossopharyngeus.

rectus inferior.

rectus internus.

radix longa.

root of rami ophthalmicus superficialis and buccalis facialis.
radix ophthalmici profundi.

ramus pharyngeus glossopharyngei.

rectus superior.

anterior root of trigemino-facial ganglion.
posterior root of trigemino-facial ganglion.
ventral roots of third and fourth occipital nerves.
supra-angular.

suprascapular.

sacculus.

supraclavicular.

779
Sourae

Sh.
SMX.
SOP.
SOR.*
SOR?
SP.
sp.
spa.*-®
spa.r-®
sph.r®
SD Uae
SPB. Lf-L11.
Spe.

SQ.
stgl.
stgle.
stgl.io.™7

stgl.ml.

stv.
StU.SC.
sup.
SUS.
SU.

ALLIS.

supraorbital groups of pores, Nos. 1-8, or trunk canals leading
to those pores.

sternohyoideus.

septomaxillary.

suboperculum.

first suborbital.

second suborbital.

splenial.

first six spinal nerves.

anterior or communicating branches of first six spinal nerves.

dorsal branches of first six spinal nerves.

horizontal branches of first six spinal nerves.

ventral branches of first six spinal nerves.

second and third suprapharyngobranchials.

spiracular canal.

squamosal.

supratemporal branch of nervus glossopharyngeus.

canal for supratemporal branch of nervus glossopharyngeus.

branch of supratemporal branch of nervus glossopharyngeus to
sense organ, No. 17, of infraorbital canal.

branches of supratemporal branch of nervus glossopharyngeus
to pit organs of middle head line.

supratemporal branch of vagus root.

subcutaneous branches of supratemporal branch of vagus root.

sinus utriculi posterior.

sinus utriculi superior.

saccus vasculosus.

symplectic.

sympathetic nerves.

transversus dorsalis anterior.

transversus dorsalis posterior.

trigeminal foramen.

temporal groove.

truncus maxillaris trigemini.

nervus trochlearis.

subcutaneous branches of truncus trigeminus, or of accessory
trigeminal nerves.

truncus intestinalis n. vagi.

transversus ventralis anterior.

transversus ventralis posterior.

first spinal vertebra.

nervus vagus.

vagus foramen.

vomer.

vein coming from eyeball.

vein coming from eyeball.

ventral cartilaginous occipital and vertebral processes.

transverse ‘“‘ Wulst” or bar of cartilage.

branch of vagus apparently corresponding to the ramus lateralis
trigemini.

9 -

» 1) ALY Binks

Un

Fic.
Fic.
xX 1%.
FIG.
Fic.
’ Fic.
x 1%.
Fic.
Fic.

oS

ALLIS.

EXPLANATION OF PLATE XxX.

Side view of complete skull of adult Amia. xX 1%.
Side view of pterygo-palatine arch, hyoid arch, and Meckel’s cartilage.

Side view of quadrate. X 1¥4.
Side view of symplectic. X 1%.
Inside view of upper end of hyoid arch with opercular bones attached.

Inside view of right mandible. x 1%.
The same with splenial removed. X 14.

ART

IG it

CGE : ; inh limb
SANG mdc

lini

PUN

mitfr

b

pasa! aty or Toe
Beate ws vu
ey eae ay a ’
if

nary r

nt - HG gor Ae Wil Qe :
iieh fe Gita 4 ildpig
. NUN,

PT a aie
ay |

Te ¥ | : ei

Nana an shih

ty CURD t yoy Si a y

774 ALLIS.

EXPLANATION OF PLATE XxXI.

Fic. 8. Top view of primordial cranium and first three vertebrae of adult
Amia. The spinal and occipital dorsal arches have been removed and also the
intercalar bone on the left side of the skull. X 1%.

Fic. 9. Side view of same, the spinal and occipital dorsal arches in place.
xX 1%.

Fic. 10. Bottom view of same. X 1%.

Fic. 11. Inside view of primordial cranium, vertically bisected. X 1%.

acv fr
ofn
ct.

w
eps{r

UCC +--«--»

pir
ecfr

Wie
PE

glf

stgle

st gle

|
4. |

OS

os
acuyTr

ct. AS
Ww
tfr
enic of ~
ecfr
em ti
ff
PE

er - rng
. ,
= perry
ariemv vee’ on m4

776 ALES.

EXPLANATION OF PLATE XXII.

Fic. 12. Genealogical tree based on the innervation of the muscles of the eye-
ball.

Fic. 13. Side view of basisphenoid of adult Amia. X 1¥.

Fic. 14. Top view of same. X 1}.

Fic. 15. Bottom view of same. X 1%.

Fic. 16. Rear view of primordial cranium, looking slightly from above. Same
specimen as Figs. 8-11. The intercalar bone removed on left side. X 1%.

Fic. 17. Top view of parasphenoid of a different and larger specimen. X 1%.

Fic. 18. Top view of himd end of cranium and first six vertebrae of a third
specimen. On the right side the occipitale laterale and the dorsal, spinal, and
occipital arches have been entirely removed. On the left side the ventral half of
the occipitale laterale and the ventral, cartilaginous ends of the dorsal arches are
left in place. xX 2.

cals
Journal of Morphology Vol. XM.

Sauropsida
pee eg ce hao

16.

Teleostet Ganotdet

= re
Plagiostomata Holocephala

-

br.

F SS
Cyclostomata © Proto -urodele
eS AE

18.

Prototype We OL BO loc® toc? Wl pt Ve Wed

Nomura del

eth Wr

778 ALLIS.

EXPLANATION OF PLATE XXIII.

Fic. 19. Top view of head of adult Amia. On the left side the dermal bones
have been entirely removed ; on the right side they have been left in place, but
scraped so as to expose the lateral canals and the ends of the nerves that innervate
the canal organs. On the left side the anterior extension of the trunk muscles
that fills the temporal groove has been removed. X 1%.

Fic. 20. The same, a deeper dissection. The eyeball has been removed on
the left side, the surrounding tissue being left in place and represented as trans-
parent in its ventral portion so that the nerves below it can be seen. X I.

aa

opf --

msh-—-

opt... Fs P-.- : ' eae .- eee 5 OPESC.
opfso” : : :
mit

ifio! PP EE Y shee 2 . US ee FOE bf io

Lap i

spe : & { , ee ae a «fl .- RAS SY apfso”

Fo

¢ opfal el ‘anf
Por so
of to” Ay. eae iy i. \ a fee, De fis ofio®
PA Cor 422..6.b.2 te. ATA 9 Mi ae A / \ ee th SC.
SQ mefhl » Lf ‘ , 5

voptst:sc MINE LN . z i : stgl. ml
stgl.mi.. Lt.) eee) ae ' x r stgl.io”.
| siglo”. — a i ; >. ‘so ies ofvo"®

optstysc

-stgl.io”

ome t ‘ Sa ali ee - - nll. pl
nll ste 3 : ™, ‘
opr.

nil pl. yay
iste ia nllisto?
SWSC. = BEE % : Tera es ~ stisc
nlLio® >. fe ee “he + : -nll.to" a j ; oS : 2 > rll. io”

- stgl. lo

rl. sto*

| nll.io”- nllia®®

i nlhio” ih. &
| nllto

| nilio’
| .

| nll. : f ; : i) ig rat Uw . “4 ve rs 2 wine ee a ig*

nll. dy ; ; ;
nlllle’. j A — } re ee ee - ullg

\allall. lg?
g~

J Womuara det

we ti i

ast iats Mi Gin " AV A mi
i) oy nous i)

dit aw

Lead WY ’ Wy?)
res Lt i

ie ae ae an

ay

a i. ae

% i

A ne
sie | | |
yy ‘nia

EXPLANATION OF PLATE XXIV.

i”,
Fic. 21. Same as Figs. 19 and 20, but a still deeper dissection. On the left
_ side only a thin outer shell of the chondrocranium has been left, and this shell

thas been cut through in one place to show the full course of the supratemporal

Fig. 20, is shown cut in this figure. x 114.

_ Fic. 22. Top view of right eyeball of another, unusually large specimen, with

ng associated nerves and muscles; the rectus superior cut and turned back.

xX 2,

_ Fic. 23. Bottom view of same, the rectus inferior cut and turned back. 2.
Fic. 24. Top view of cartilaginous capsule (sclerotic) of same, the dorsal part

of capsule removed. x 2.

PLAXIV:

CACC hee Ae 2 :
LP

rit’ or

£5.

ocms

ocme

stgl.io”

nlilio? +
nilall

|
|

“

) ria

“ah S

EXPLANATION OF PLATE XXV.

"1G 25. Same as Figs. 19, 20, and 21, but a still deeper dissection. Both

alls have been entirely removed. On the left side the adductor mandibulae,

levator arcus palatini, the dilatator operculi, and the upper half of the hyo-
ular have also been removed. X 13%.

“Fic. 27. Same as Fig. 22, but showing the nerves only. X 2.
Fic. 28. Reconstruction from a series of transverse sections of a 35 mm.
Amia to show the dorsal aorta and its branches.

bee ©

Journal of Morphology Vol AM _— a =

|
|
|
|

-----INSEMX

---- mst

Of

oemi----

: -opl
~opf

rlms---

PPF
ims**—-4

St ---- J ~ptlerremme PS = 3 Z : 2, a Opp
J \ | B STA Soc

yy, \ .
wo Paw) yf )
\ 6 ff eps Lipset

/
j y

TE BA 7 vo z
| i 7
gl ; = f 200 9
och Tr) } =
-stgl
* y f cb A
; j cl
Ss ne sty j
j
ee 00
nil.vol? ee -
J opt .
nil Popp =
x
# call )\- =
/ cal
.| da
opp
OCmMs Fy 10|\0
sot)?
¥ call |
f \ rae
y \ > 25
ee * -
eal | \
Pe \ | -
OCm NY -
XN 350} ©
SS (

J

784 ALES,

EXPLANATION OF PLATE XXVI.

Fic. 29. Side view of specimen used for Figs. 19, 20, 21 and 25. The skin
and dermal bones have been removed, the dissection corresponding nearly with
that shown in Fig. 19. X 1%.

Fic. 30. Same as above, a deeper dissection, corresponding nearly with that
shown in Fig. 21. X 1%.

.
fo.
Tits, ¥y :
a oui uAe

al

Journal of Morphology. Vol. Xl — a | | ui rice

|

|

ee"
° bf to*
a

.) i >. seve -LOC

Ie

yer OPESC
opt ‘so’

s----- bf to®

1
| ']

INSEL NIX

Py

ea:

- opfso*

"2 ———
Stow
8

5By.

4 “pope.se.
bf to? AF coophsce. ay, TE ER AAA |
op so”
woes TL

\---opf so” ache t os ioe t : ae
bf io™ 5

ra

rede

7 |, aay

[fe ~-rlap.do
mefirndl fs

/ . 4 ram
af (sy : } ia aoe te Lap

| mtef Vb} gs

| f \. a \ \ a ‘ 3 rac. hf A\\ \\ ; > ; ---Do
mef hl... pee ht P : Sots .-- eat : A... optstirsc

|....stgl. to”
stgl.md

OTN

nl. sto" ae
; “sly
| - SU sy .
| ae ]
nll.io” bw i
do Le |
| ‘ Lo % 4 ’ : |
; ay |
4 AA. 222d. 20°" .
' ond |
aa 5 NN nil.dl. - | ACN \
\ 4 \ 1A
1} Caer

\ a

as.

eta ia
Ge a

0 fee v4 al ik

OF iat) ) hoe P

hb vee Re
‘

ul

at ae
:

re vise at a j
ne as)

7
'

786 ALLIS.

EXPLANATION OF PLATE XXVII.

FIG. 31. Same as Figs. 29 and 30, but a still deeper dissection. The coronoid
process of Meckel’s cartilage has been removed, the lateral canals in the preoper-
culum exposed, and the gill filaments removed from the branchial arches. X 1%.

Fic. 32. Same as above, but still deeper. The hyoid and pterygo-palatine
arches have been almost entirely removed so as to expose the branchial arches
and the levatores arcuum branchialium. X 1%.

ny

ay

ij ,

oe

PAaemal of Morphology Vol. xi

’ Nomura del.

Pp
=== SL TX

m----Lims*

Se On

vesee rlms

---tlapdo

.--Do

-.--hinf-
---- mife

mef
hf

i mef oma®
-omg”
sty

q

| nilid® ’

[ip =

of at hi
isp an
Sth aa
. aia ‘

788 AELTS.

EXPLANATION OF PLATE XXVIII.

Fic. 33. Side view of occipital and anterior trunk regions of adult Amia. The
branchial arches, excepting only their ventral ends, the corresponding portion of
the clavicle with the pectoral fin, and the dermal bones between the clavicle and
the skull and those on the top of the latter have been entirely removed. The
vagus and glossopharyngeal nerves are cut near their exits from the skull, the
spinal nerves near the points where they enter the pectoral fin. The occipital
nerves are left entire. The pharyngo-claviculares are shown, the externus entirely
cut, the internus partially so. X 1%.

Fic. 34. Same as Fig. 33, but another specimen. The occipital and spinal
nerves have been pulled down out of their natural positions so as to show their
anastomoses. The position of the pectoral fin is indicated by dotted lines, as is
also the position of the deeper portion of the eighth intermuscular septum. The
muscle fibres between the dorsal portions of the intermuscular septa have been
removed so as to show the pockets in the latter. * 1%.

=
=

Len ———_
T Nomura. ae

a

Ps

790 AE ETS:

EXPLANATION OF PLATE XXIX.

Fic. 35. Side view of occipital and anterior trunk regions of a third specimen.
The muscle fibres of the muscle segments have been entirely removed, leaving
the nerves and intermuscular septa in position, but considerably cut. The inter-
muscular septa of the sternohyoideus are shown in a similar manner. The hori-
zontal intermuscular septum, in the twelfth muscle segment, has been cut so as
to show the ventral branch of the spinal nerve of the segment in position. X 1%.

Fic. 36. Side view of the deeper muscles on the side of the head, the two
divisions of the adductor mandibulae turned downward and forward. X 1%.

Fic. 37. The same, the upper end of the hyomandibular, with the levator
arcus palatini and dilatator operculi, removed. X 1%.

a =
vi

F Plana
mal of Morphology Vol. | | .

imh

we
mefomo

ra EH mi f HMD

i |
JS Nomura del = — < P

792 PAU ETEUESs

EXPLANATION OF PLATE XXX.

Fic. 38. Enlarged side view of trigemino-facial ganglionic complex, with nerves
arising from it. The hyomandibular and palatine branches of the facialis are not
shown. Same dissection as shown in Figs. 20 and 30.

Fic. 39. The same, still more enlarged. The ophthalmic and buccal branches of
the facialis have been turned back to expose the Gasserian ganglion, and the acces-
sory trigeminal nerves have been turned forward and pulled apart to show their
origins and interconnections.

Fic. 40. Side view of Meckel’s cartilage, and the adductor mandibulae and
levator maxillae superioris muscles. X 1%.

Fic. 41. Inside view of same. X 13%.

Fic. 42. The same with muscle 4;, and the second and third divisions of the
levator turned down. X 1%.

:
\
«
7
7 me
a

Wie” bio!” “s
rde

opt poppe

rl gp ocms rp

y fi tard rh rere gg
iA X: rd a

\

gobf Of opf Uf

re re mit rlapdo

A, Aw’

rmdtt

vlins*

aa

Wi
Lat At
i”

i ‘7
iPods litnteh Paa
ee aN

'
Rly) warn.
PR ett kt

i ieee used
;

794 AETTS:

EXPLANATION OF PLATE XXXI.

Fic. 43. Bottom view of head of adult Amia, the skin, gular plate, and part
of right mandible removed. X 1%.

Fic. 44. The same, a deeper dissection. The hind corner of the right mandi-
ble and ramus ».¢/2 turned outward so as to show the internal mandibular branches
of the trigeminus and facialis. x 1%.

-

a

it

ao

4
=
-
:
*
oo?

'
34
.
,
n
~
1
op ae
of

.

on

i
»
4
*

—~_
——— =
ea

=

PEE

Journal of Morphology. Vol.XU- : ~ — 2a? = | oe :

Im.

Ghs » +4
7 2

i i 7
Bm fj

ME fr

oe |
mef = if UN) Wa He \ Y :
ii B\ a) VIZ | \
er Wf U, Hhi Sim Lm 3

PONALE -no-n-95

rmdit-- Pres

rghs---4
FEQHS --v-n fere-ca ose

Mef OMA Mon: faenat-nnnen

* fo :

oe mefomo :

lmh --

CH}

_—

rt oe

7

796 ALLELES:

EXPLANATION OF PLATE XXXII.

Fic. 45. The same as Figs. 43 and 44, with the geniohyoidei, the anterior half
of the branchio-mandibularis, and part of the tendinous ends of the hyohyoidei

removed. X 1%.
Fic. 46. The same, the hyohyoidei and one half of the sternohyoideus

removed. X 1%.

i
ex.

r “s ’
+ ae ; ; Heed: 9ey) ? '
Ce ad BaP S i SF 2 , : : ike h : Ms = +
ae ' = in aml ’ » i
~irt, | 'y . i > ie ; ;
14 i) Sabu ii: 17> B ;
: fi : fins ~ 1) 7 : ao} / 4 f ~* 4

Siete

a oe rea
gh ii) aaa Sh Phoe at . .
hese, :
Cae ” 7

py hey ‘ ae 3
9 fi : :.
an ’ 7 + ‘
> a .
hie . é
ory , 4 3 _
dl aie we - »
aa An fi »
i ia) = .
- ‘
ete mje
fi us 4
ST ot :
‘a A 7
ner
4
. i
a.
a)
re Ly :
M ; ie
sige _ eh
aq! ,
eae dig *G

wp ytd

Jie iva ci

798 ALLIS.

EXPLANATION OF PLATE XXXIII.

Fic. 47. The same as Fig. 46, but still deeper. The hyoid arches and sterno-
hyoidei have been cut so as to leave only their anterior ends, and the clavicles
and the heart and blood-vessels have been entirely removed, so as to expose the
under surfaces of the branchial arches and basal line, with the nerves and muscles
that lie upon them. X 1%.

Fic. 48. The same as Fig. 47, showing the ventral surfaces of the fourth and
fifth ceratobranchials of the right side of the head, and the muscles associated
with them. X 2.

Fic. 49. Dorsal view of the branchial and hyoid arches of the left side of the
head. X 1%.

Fic. 50. Ventral view of the branchial and hyoid arches of the right side of
the head. X 1%.

Fic. 51. Side view of the basal line. X 1%.

= Sh

CH
HH

TER Teen

Sele HRT
IPB

er Bm lin liv liv I

~- Ov

OV NE
Ov?
--Tva
- OV?
----Tva

-Pce FRY

“Od* H
AabV BB? BR BR
Aab lv

I Moms =
J Nomura deh

reo)
ma)
id ia
fa s2

800 ALETS

EXPLANATION OF PLATE XXXIV.

Fic. 52. Dorsal view of the branchial arches and the floor of the mouth cavity,
with the muscles and nerves in place. The cranium has been entirely removed
excepting a small piece where the levator muscles have their origin. On the
left side the transversus dorsalis and the retractor arc. branch. dors. have been
removed. X 1.

FIG. 53. Same as above, with the muscles entirely removed on the left side so
as to show their insertions on the branchialarches. On the right side the levator
muscles are cut about the middle of their length, and the vagus nerve is turned
medianward and downward so as to expose its branches. The ventro-lateral sur-
face of the nerve is thus shown in the drawing. X 134.

=

3
f
i
1 i
;
7 ,
- sell}
a ‘
-
* PA ¥
i
a
‘
"e
;
ve
}
s
3 ‘
p
;
‘
i
,
: ;
6
;
i} .
H 7
1
Lj
;
;
i
cc
Fi
Ne
oe ;
= 7 -
i. n
u :
:
7
«
;
=
thie
v

_
=
>
Ce
A pad
-
ae
=
a4
=r -
——— —
7 ae
a
ro
_

i

a * : >

= =
—— -
= =
~—_
7 =

ye
a.

= 4

~ ~
. :
.

=

“Journal of Morphology VolXM

—

Od’
Label.
Odé....ta

Ta? need
Og AN

Label:

hid

habe N

lid

}
) f \
i > habeV

Rabd

es gs deadly Saal

802 AUEELES:

EXPLANATION OF PLATE XXXV.

Fic. 54. Same dissection as Figs. 52 and 53. Shows the dorsal surfaces of the
branchial arches on the right side of the head, the muscles and arteries entirely
removed, and the nervus vagus turned medianward and downward as in Fig. 53.
x 1%.

Fic. 55. Same as above, but showing the arches and efferent arteries only,
with the sympathetic nerves that surround and follow the arteries. X 1%.

Fic. 56. Top view of the outer ends of the fourth and fifth arches, with the
adductor muscles and the second obliquus dorsalis. X 2.

Fic. 57. Top view of the left ear with the associated nerves. X 2.

Fic. 58. Top view of the sternohyoideus, showing the distribution of the
occipital nerves. X 1%.

Fic. 59. Top view of the vagus and glossopharyngeal nerves in a specimen in
which the first vagus nerve and the glossopharyngeus were united by a communi-
cating branch. The glossopharyngeus has been cut away between its ganglion
and its point of exit from the brain, and pulled slightly forward. X 3.

Aabv

Labi”.

Od’

Od*

Rabd

habe W

4 oO Od a eee)
at a aha aye ieee
Ree eer ee |
) Pye Pe) Lay

804 AELIS:

EXPLANATION OF PLATE XXXVI.

Fic. 60. Dorsal surface of mouth cavity and branchial chamber seen from
below. The lower jaw, the lower half of the branchial arches, and the lower half
of the oesophagus have been cut away. On the right side of the head the first
branchial arch is pressed back slightly so as to show the ventral opening of the
spiracular canal. X 1%.

Fic. 61. Same as above, the skin and underlying tissues removed so as to
expose the ventral surface, of the cranium and the ventral surfaces of the dorsal
halves of the branchial arches. The efferent arteries and the pharyngeal and pre-
trematic branches of the facialis, glossopharyngeus, and vagus nerves are seen.
The pseudobranch on the right side of the head has been removed. X 1.

+ ay ~ ) 7 ee
P 5 =
ue +e
“a
‘ eae ‘4
- ae - oy i na 7 - = Bg
Ls ’, ¥ ' Le
.
7
et : ni 2 .

. a «|
i S
- = 7s
? _ =
; a
5 A of
— 7 i _ \y
= rn
: » 3 M
: Yo
‘ J, —
>

-

,
i]
e
©
° _
..
{
= ~~

Iwrnal of Morphology Vol. Xi PLAANYL.
pene Ro

pp

PPT
P sb
icf?
IPBI

ce

\ gt

\
{

IPBa

* da

Tila
PRM

806 ALLTS.

EXPLANATION OF PLATE XXXVII.

Fic. 62. Same as Figs. 60 and 61, but a deeper dissection. The pharyngeal
bones on the right side of the head and the constrictor pharyngei have been entirely
removed. On the left side of the head the first two infrapharyngobranchials and
a part of the third are left in place. The parasphenoid has been partly removed,
exposing a part of the base of the skull. A part of the eye-muscle canal is seen
filled with membranous and fatty tissues. The pterygo-palatine arch on the right
side of the head is partly cut away so as to show the muscles and nerves of the
orbit. X 1%.

Fic. 63. Same as above, but still deeper. The median portions of all the vis-
ceral arches with their associated muscles, the retractor arcuum branchialium in-
cluded, have been entirely removed. The larger part of the base of the chondro-
cranium has also been removed. The ventral surfaces of the vertebral column
and the trunk muscles, of the hypophysis cerebri, the saccus vasculosus, and the
eyeballs and their associated nerves and muscles, are thus exposed. X 1%. -

—
7 >
——
4 _

Journal of Morphology pe : — PLXXXYIL.

62. 635
--Lms*
ims*
; : mst

Ppr Wst mx

Of -

ocme-

rims
ocme
paf- ims'*
pf
99 -
re
mphgl-
hinf---
hop SI
Ah
agl
ag! .
C0] ~~
Tila aa yt
J
cv
py
Rabd * x

Sy f }

oie

808 AGES:

EXPLANATION OF PLATE XXXVIII.

Fic. 64. Same as Fig. 63, but still deeper. The base of the skull has been
entirely cut away so as to show the base of the brain and the roots of the nerves
as far back as the fourth occipital nerve. The sympathetic nerves have all been
removed. X 4.

Journal of Morphology Vol. Xm. PUXXXVI

OU

Oc -----

Pa

pr

Pf
OCcIN -
gg

rphgl
Life

THE ORIGIN OF THE CLEAVAGE CENTROSOMES.

KATHARINE FOOT, Evanston, Illinois.

DurinG the past two years the weight of authority has
greatly increased in favor of the conclusion that the cleavage
centrosomes are of spermatic origin.

These centrosomes are asserted to be the daughter centro-
somes of the so-called male centrosome, and this is assumed to
originate from the middlepiece of the spermatozo6n, —to be
of the very substance of which the middlepiece was formed.

The phenomena of fertilization in the egg of AUolobophora
foetida do not sustain this view; they suggest that this centro-
some has the same origin as the so-called egg centrosome,
and that both are cytoplasmic elements, of like origin and
constitution.

The sperm attraction sphere and the fertilization cone have
several points in common; both structures appear to be
dependent, not alone upon the entrance of the spermatozoon,
but also upon a definite stage of maturation reached by the
egg ; for I have never found a cone after the first polar body is
constricted off, nor a sperm attraction sphere before the ana-
phase of the first spindle, no matter how far the spermatozoon
may have penetrated into the egg ; both appear to be formed
of substances not confined to the structures themselves, but
belonging to the entire cytoplasm.?

These observations suggest that both the cone and the
attraction sphere are expressions of a definite effect produced
upon the cytoplasm by the entrance of the sperm, the phenom-

1 In the summer of 1895 I differentiated from the cytoplasmic network a sub-
stance (archoplasm) which is present in the attraction spheres, in the fertilization
cone, in the spindle, and throughout the cytoplasm (see Journ. of Morph., vol.
XII, Plate I). Recently I have differentiated the microsomes and centrosomes
from the cytoplasm and archoplasm (Figs. 2, 3, 4, 7,9, 10). These microsomes
do not appear to be merely thickenings of the cytoplasmic threads ; they appear
to be morphological elements, many of them scattered throughout the cytoplasm

in a relatively independent manner, others imbedded in the cytoplasmic network.
809

810 FOOT. (Vor. XII.

ena differing in that the anterior end of the head of the sper-
matozo6n produces a cone, whereas it is the middlepiece that
produces the attraction sphere.

Before the first polar body is formed we find a cone the
moment any part of the head of the spermatozoon penetrates
the egg, and the extent of the cone appears to depend upon the
length of head which has entered: if a relatively short piece
of the head has entered, we find a relatively small cone. The
sperm attraction sphere, on the contrary, does not appear until
the middlepiece is within the egg, and in other forms this fact
has been cited as evidence for the assumption that the centro-
some of the sperm attraction sphere is of the substance of
which the middlepiece of the spermatozo6n was formed.

The more important observations adduced to prove that the
male centrosome is formed from the middlepiece are the fol-
lowing: First, the attraction sphere does not appear until
after the spermatozoon has entered the egg. Secondly, it
appears at or near the point formerly occupied by the middle-
piece of the spermatozoon. Thirdly, the middlepiece and the
centrosome have been shown to select the same stains. Might
not the first and second observations be cited with equal perti-
nence to prove that the spermatozoén produces merely a physi-
ological effect upon the cytoplasm? The third observation has
a more definite bearing upon the point in question. In the egg
of Allolobophora foetida, however, I have differentiated by two
methods the middlepiece of the spermatozoén from the male
centrosome. This differential staining might be explained by
the assumption that the middlepiece undergoes a physiological
change on entering the egg; but as the head gives no such
evidence of physiological change, are we justified in assuming
without further evidence that this takes place in the middle-
piece?

The centrosomes of the egg of Allolobophora foetida \end
support to the view that the centrosome is a mechanical center
— the expression, rather than the cause, of cell activity. The
egg attraction sphere is present during the two maturation
divisions, but after the second polar body is formed and the
female pronucleus begins to develop, it totally disappears (Fig.

\

No. 3.] ORIGIN OF THE CLEAVAGE CENTROSOMES. 811

5). The sperm attraction sphere is present until the head of
the spermatozoon begins to develop into the male pronucleus,
when it also totally disappears (Fig. 6). Both spheres are
absent during a relatively long period (z.e., while the young
pronuclei are developing) ; and when the pronuclei have at-
tained their maximum size and are in contact, two attraction
spheres appear again in the cytoplasm, and the cleavage spindle
is formed.

Method.— The middlepiece of the spermatozoén was first
differentiated from the centrosomes by iron haematoxylin fol-
lowed by erythrosin. The same results were obtained by
staining with a mixture of orange and methyl green. (Satu-
rated aqueous solution of orange two parts, in eight parts
water. Saturated aqueous solution of methyl green one part,
in four parts water. Wash in absolute alcohol.) This stain
was the outcome of a series of experiments with the anilins,
undertaken with the aim of supporting the method of double
staining (Figs. 8-10) by one containing fewer possibilities
of error. This method demands relatively no technical manip-
ulation, and practically the same results were obtained whether
the sections remained in the stain fifteen minutes or twenty-
four hours. The stain was used after several fixatives, with the
same result, but all the figures for this paper were drawn from
chromo-acetic preparations.

An examination of the figures will show that a variety of
structures select the methyl green, — vzz., the nucleoli, micro-
somes, centrosomes, chromosomes, and sperm granules, —and
that the orange is selected by the chromatin of the vesicles and
resting nucleus, the cytoplasmic network, and the archoplasm
(which is present in the attraction spheres, fertilization cone,
spindle, and throughout the cytoplasm).

This method of staining sharply differentiates the very young
nucleoli of the vesicles found at the telophase of the second
maturation spindle, and destined to become the female pro-
nucleus. These vesicles when first formed are eleven in num-
ber, thus corresponding with the chromosomes. As _ they
progress towards the formation of the female pronucleus, their
walls break and fuse with each other, thus forming the chro-

812 FOOT. [VoL. XII.

matic network of the pronucleus, while their nucleoli grow and
fuse with one another. Fig. 5 shows several small vesicles
and a few relatively large vesicles, the result of the fusing of
two or more small ones. Just before the vesicles appear, the
chromosomes select the methyl green stain, but as soon as
the vesicles are formed we have a tiny, green nucleolus sur-
rounded by ye//ow chromatin, this clearly suggesting that the
nucleolus is a substance which a moment before was contained
in the chromosomes. The chromatin which at first surrounds
each vesicle continues to select the orange stain until it has
again assumed the form of chromosomes. Fig. 6 shows the
male pronucleus forming in exactly the same manner as the
female pronucleus, the head of the spermatozoon breaking up
into vesicles similar to those of the young female pronucleus.

Vesicles entirely similar to those found at the telophase of
the second maturation spindle are found also at the telophase
of the first maturation spindle, though I have never seen any
evidence of their development into a resting nucleus.

Sperm Granules, Fig. 2.— The sperm granules are not con-
stant structures. In eggs found during the height of the
breeding season (when they are less likely to present ab-
normalities), the sperm granules are either not present at all
or are relatively insignificant in both size and number, and
in such cases they are often found near the posterior end
of the head of the spermatozoon. Near the close of the
breeding season, however, when few normal eggs are found
(sometimes only one in a cocoon that contains fitty eggs, show-
ing various degrees of degeneration),

at this season nearly
all the eggs having one or more fertilization cones contain
relatively many and large sperm granules. These granules
appear to be formed at the expense of some of the surrounding
substance, and such preparations as are represented by Fig. 2
suggest that the substance sacrificed to the formation of these
bodies is archoplasm, for there is relatively very little archo-
plasm at the side of the cone occupied by the sperm granules.

The suggestion that these granules are metamorphosed
archoplasm raises several questions that I am entirely incom-
petent to answer. Wherein do they differ from apparently

No. 3.] ORIGIN OF THE CLEAVAGE CENTROSOMES. 813

similar; and also inconstant, large and small granules (or bodies)
found outside and sometimes within the first and second matu-
ration spindles (Fig. 4), outside and sometimes within the
attraction spheres (Figs. 3, 4, 9, 10), and, again, often scattered
throughout the cytoplasm? Furthermore, as no exact distinc-
tion can be made as to size, how do these bodies differ from
the microsomes? (Figs. 2, 3, 4, 7,9, 10.) Can these sperm
granules and large granules throughout the cytoplasm —
apparently indicative of incipient degeneration — be at first
merely an expression of abnormal activity in the cell? These
questions must be left to the more experienced student to
answer.

I am greatly indebted to Dr. Whitman for criticism of my
work, and for other courtesies in connection with this paper.

814 FOOT.

EXPLANATION OF PLATE XXXIX.

Zeiss hom. immer. 2mm. 140 ap. Abbe Camera.

Frcs. 1-7. Orange, methyl green. See Method, p. 811.
Fics. 8-10. Iron haematoxylin followed by erythrosin.
Fics. 1, 2, 3, 4, 8, X about 687.

FIGs. 5, 6, 7, 9, 10, X about 500.

Fic. 1. Spermatozo6n. (Middlepiece yellow.)

Fic. 2. Longitudinal section through fertilization cone. (Only such micro.
somes represented as appear to be independent of the cytoplasmic threads.)

Fic. 3. Sperm rod and sperm attraction sphere. (Microsomes and centrosome
green.)

Fic. 4. Longitudinal section through second maturation spindle.

Fic. 5. Section of egg. (Telophase of the second maturation spindle. Vesicles
forming female pronucleus. Egg attraction sphere has nearly disappeared. Micro.
somes not represented.)

Fic. 6. Vesicles forming male pronucleus. (Male attraction sphere has dis.
appeared. Microsomes not represented.)

Fic. 7. Pronuclei in contact. (Attraction spheres present.)

Fic. 8. Spermatozobn. (Middlepiece red.)

Fic. 9. Male attraction sphere. (Microsomes and centrosome black.)

Fic. 10. Egg attraction sphere. (Microsomes and centrosome black.)

4

- Peete

era Mater, FrankfurtoM.

0b K Hayacki

enue

4
Sof AN,

agg
re hy,

i
pty?

iD
Wate!
RP is

AJA
LPN a ea

Kt.

SMITHSONIAN INSTITUTION LIBRARIES

WENA

3 9088 0123