CONTRIBUTIONS TO EMBRYOLOGY

Volume VIII, Nos. 24, 25, 26

Published by the Carnegie Institution of Washington
Washington, 1918

CARNEGIE INSTITUTION OF WASHINGTON

Publication No. 271

Ifftr

PRESS OF GIBSON BROTHERS
WASHINGTON

CONTENTS.

No. 24. The developmental alterations in the vascular system of the brain of the human

embryo. By George L. Streeter (5 plates, 12 text-figures) 5

25. The mitochondrial constituents of protoplasm. By E. V. Cowdry (1 plate, 9

text-figures) 39

26. The development and reduction of the tail and of the caudal end of the spinal cord.

By K.\NAE KuNiTOMO (4 plates, 2 text-figures) 101

CONTRIBUTIONS TO EMBRYOLOGY, No. 24.

THE DEVELOPMENTAL ALTERATIONS IN THE VASCULAR SYSTEM
OF THE BRAIN OF THE HUMAN EMBRYO.

By George L. Streeter.

With five plates and twelve text-figuies.

CONTENTS.

PAOE.

Introduction , ........,....;.... i .>J ,., . ........ .: .^...- 7

Establishment of primordial vascular system of head 9

Differentiation of primordial system into arteries, veins, and rapilUuies 11

Cleavape of blood-vessels of head into separate systems 10

Adjustments of blood-channels in adaptation to growth and change in form of brain and ear 21

Development of superior sagittal sinus 30

Summary 34

References cited 38

Explanation of plates 38

6

THE DEVELOPMENTAL ALTERATIONS IN THE VASCULAR SYSTEM
OF THE BRAIN OF THE HUMAN EMBRYO.

By George L. Streeteb.

INTRODUCTION.

One of the most striking features in the development of the blood-vessels of
the head is the clear way in which they demonstrate the embryological principle
of what may be termed integrative development. It is quite evident that the
vascular apparatus does not independently and by itself "unfold" into the adult
pattern. On the contrary, it reacts continuously in a most sensitive way to the
factors of its environment, the pattern in the adult being the result of the sum
of the environmental influences that have played upon it throughout the embry-
onic period. We thus find that this apparatus is continuously adequate and
complete for the structures as they exist at any particular stage; as the environ-
mental structures progressively change, the vascular apparatus also changes and
thereby is always adapted to the newer conditions. Furthermore, there are no
apparent ulterior i^reparations at any time for the supply and drainage of other
structures which have not yet made their appearance. For each stage it is an
efficient and complete going-mechanism, apparently uninfluenced by the nature of
its subsequent morphology.

With these factors in mind, one can better understand the architectural
arrangements of the vascular system of the head that appear in different periods of
development. In the primordial or precirculatory period the vessels that then
exist are engaged principally in growth and in the elaboration of a plexus, and
their form then is little influenced by conditions that would favor the circulation
of the contained fluid. As the cu-culatory flow of the blood becomes estabUshed
we find that the vascular plexus responds by conforming to the hydrodynamic
requirements, becoming adequately adapted to the form of the neural tube as then
existing, with favorably situated aortic feeders and simple and direct drainage
channels. As the brain becomes more complicated, and as the skull-membranes
form, there occur, step by step, the necessary adaptations on the part of the blood-
vessels. Finally, when the permanent form is attained, the vessels lose their transi-
tory character and develop permanent and more highly difi"erentiated walls, properly
suited to the adult functional requirements.

It is possible to subdivide the development of the blood-vessels of the brain
into five successive periods, each showing special adaptations to their changing
environmental conditions. To facilitate the description of this process an arbitrary
order of that kind will be adopted in this paper. During the first of these five
periods there are established the primordial endothelial blood-containing channels,
which are neither arteries nor veins, but constitute the source from which all the
arteries, veins, and capillaries of the brain are derived. These primordial blood-

7

8 THE DEVELOPMENTAL ALTERATIONS IN THE VASCULAR SYSTEM

vessels lie bilaterally close along the brain-A\all, at first either in the form of a
single, slender, longitudinal tube, as seen along the hindlwain, or a plexiform space
as seen near the forebrain and midbrain. Soon after the primordial vessels are
estabUshed, endothehal buds sprout from their walls and in conjunction with them
form an irregular endothelial vascular meshwork which tends to spread over the
surface of the brain-wall, especially in the region of the forebrain and midbrain.
There is thus formed a plexiform system which constitutes a germinal bed of endo-
thelium rather than a circulatory apparatus.

During the second developmental period the primordial blood-vessel plexus
of the head slowly resolves itself into veins, arteries, and capillaries, and becomes
architecturally suited to the circulatory flow of the blood. That portion of the
plexus which lies against the brain spreads out as a flattened capillary sheet, con-
forming everywhere to the shape of the brain-wall and its attached ganglia and
sense-organs. The more superficial part of the plexus develops a coarser mesh and
forms larger channels, which tend to unite into continuous trunks and gradually,
by virtue of their communications, can be recognized as definite arteries and veins.
The intermediate loops of the plexus maintain the anastomosis between the deep
capillary sheet and the more superficial trunks, forming tributaries of the veins
and branches of the arteries. The second period thus establishes the primary
type of the circulation of the head, in which there is a capillary bed, fed by arterial
branches from the aortic system and drained bilaterally by a continuous venous
trunk which extends back to the venous end of the heart.

The third period is inaugurated by the differentiation of the membranous
skull, the dura mater, and the arachnoid-pial membranes. As a result of this stratifi-
cation of the tissues of the head, the more ventral of the anastomosing channels
that connect the deep capillary plexus with the superficial vessels become closed
off and there is a general separation or cleavage of the vessels immediately sur-
rounding the brain-wall from those belonging to the membranous skull and its
coverings. This process begins at the base of the skull and extends bilaterally
upward toward the middle fine of the vault, in which region the communications
between the deeper and more superficial systems are to some extent maintained.
In this way the cerebral vessels are gradually separated from the dural vessels and
in a similar manner the superficial vessels of the head become isolated b.y the laying
down of the primordium of the membranous skidl, after which their course of
development is quite independent of that of the dural and cerebral systems.

By the time the third period is well under way it is overlapped by the fourth
period, under which we include the remarkable series of adjustments in the arrange-
ment of the blood-vessels, in adaptation to the developmental alterations in the
form, size, and condition of the structures of the head region. The brain is one of
the chief factors in this process. The marked change in its form, and especially
the prolonged relative growth of the cerebral hemispheres, render necessary a
continuous series of alterations in the blood-channels that extend far into the late
fetal stages. In the earlier stages a fundamental change results from the growth
of the labyrinth and its cartilaginous capsule, whereby a mechanical obstruction is

OF THE BRAIN OF THE HUMAN EMBRYO. 9

introduced that results in the obUteration of a considerable part of the largest vein
of the head. This is compensated for by a new channel which takes a more dorsal
course, and which eventually forms the sigmoid portion of the lateral sinus.

Finally, under the fifth period we would include the late histological changes in
the walls of the vessels that convert them into the adult arteries, veins, and the
various types of sinuses. These histological factors, however, are not considered
in the present paper and are merely mentioned to complete the sequence, as they
have no determining influence on the phenomena of the preceding four periods.

The observations that are reported are chiefly concerned with the third and
fourth developmental jjeriods; that is, after the primary circulation of the head is
established and during the cleavage and adjustmental stages. A review, however,
will be made of the history of the vascular system of the head previous to that
time, which will be based in large part on the important observations of Evans
(1909, 1912) and Sabin (1915, 1917f/, 19176). The study of the development of
the vascular system of the head of the human embryo was initiated in this laboratory
by Professor Mall (1905). Later I continued the same investigation and reported
(Streeter, 1915) some of the features of the adaptive metamorphosis of the dural
veins. In the present paper I shall include much of the same subject-matter and
incorporate with it further observations made on additional material, thus making
it possible to treat the general subject more completely and to demonstrate the
relations of the dural veins to the principal arteries of the head.

ESTABLISHMENT OF PRIMORDIAL VASCULAR SYSTEM OF HEAD.

A distinct advance in our knowledge of the origin of the vascular system has
recently been made by Professor t^abin through the study of living preparations
of the growing chick. By that means it was possible to observe directly the prin-
cipal steps in blood-vessel formation. Certain steps in the process that were still
under dispute have thus become established, as well as others which are new
observations. According to her description (Sabin, 19176), the vaso-formative cells
or angioblasts are differentiated from the mesoderm, and on proUferation they form
small, dense, syncytial masses which join one another by means of tiny processes
of cytoplasm. In this way there are formed plexuses of solid angioblastic cords,
the growth of which is maintained partly by proliferation of their constituent cells
and partly by the further addition of new angioblasts which differentiate from the
adjacent mesoderm.

During the formation of these angioblastic plexuses a liciuefaction of their
cytoplasm occurs in such a manner as to convert the sohd angioblastic cords into
vessels filled with clear fluid. This is brought about by a process of vacuolization
and the formation of an "intracellular" lumen. Vacuoles appear near the nuclei
and rapidly enlarge, so that in from one to two hours there occurs a complete de-
struction of the cytoplasm and nuclei of the central i^art of the angioblastic cords,
resulting in the formation of a clear plasma, the periphery of the angioblastic mass
being jireserved as an endothelial boundary. The liquefaction of cytoplasm occurs
both in the loops of the angioblastic plexuses and in masses of angioblasts

10 THE DEVELOPMENTAL ALTERATIONS IN THE VASCULAR SYSTEM

that are still isolated, in the latter case resulting in small vesicles that subsequently
join the main plexus. These observations were made for the most part in the area
pellucida of the yolk-sac, but the same phenomena were seen also within the body
of the embryo. The dorsal aorta in the trunk region and a portion of it within
the head was seen to differentiate in situ in the above manner.

The angioblasts are not all converted into endotheUum and blood-plasma;
some of them take part in the formation of red blood-corpuscles. During the
lumen formation in the angioblastic strands, small clumps of the original angio-
blasts become partially separated by the liquefaction of the cytoplasm around
them. Such masses soon show the presence of hemoglobin and constitute blood-
islands which eventually break apart as free red blood-cells and float away in the
blood-plasma. Blood-islands can also be seen to be derived secondarily from the
endothelium.

Concerning the distribution of the earUest blood-vessels and concerning the
form and development of the primordial vascular system of the head, I have
followed the descriptions published by Evans (1912) and Sabin (1917a), both of
whom studied injected specimens of the chick and pig. On account of the
difficulty of making observations of this region in living preparations, the details
in the growth of this endothehal meshwork have not been actually seen. The
plexiform transformation of the aortic arches has been demonstrated only partly
and it is not known how much of the first endothelial system of the head is derived
from angioblasts that are differentiated locally and how much to the proliferation
of the cells of the aortic arches. It is possible, even after the laying-out of the main
parts of the primordial system, that angioblasts continue to differentiate in the
new territory around its margins and become incorporated with it. The essential
features, however, in the development of this s^ystem are admirably shown by
injected material, as can be seen in figures 398 (duck, 13 somites) and 393 (chick,
15 somites) of Evans (1912), and plate 1, figure 3 (chick, 9 somites), and plate 2,
figure 1 (chick, 14 somites) and figure 2 (chick, 16 somites) of Sabin (1917a).

At about the time that the dorsal aortse become established in the head region,
it is seen that the first pair of aortic arches connecting them with the heart are
plexiform in character. Sprouting from this plexiform area an endothelial mesh-
work arises and extends dorsalward and caudalward toward the forebrain and
midbrain, to become the primary head-plexus. Very soon afterwards, a slender
longitudinal channel is formed bilaterally along the ventro-lateral margins of the
hindbrain which communicates with the primary head-plexus in front and cau-
dally with the anterior cardinal vein by way of the transverse veins of the first
two interspaces. A few slender communications are established very early between
it and the dorsal aorta of its respective side. This channel has been designated by
Sabin (1917a) as the "vasa primitiva rhombencephali." It seems probable from
her observations that this channel is not a derivative from sprouts from the dorsal
aorta or from the primary head-plexus, but is differentiated i7i situ, and its com-
munications with them are secondary. This slender channel, together with the
primary head-plexus, constitutes the primordial system from which all the blood-

OF THE BRAIN OF THE HUMAN EMBRYO. 11

vessels of the brain and its membranes are derived. Sabin has suggested a greater
restriction in the use of the terms "artery" and "vein" and has warned against
making a too early identification of the adult vessels in the embryo. The
primordial vascular system of the brain as seen in the chick of 15 somites consists
morphologically of a sprouting meshwork rather than a set of definite supply and
drainage channels. It is to be considered as a bed of proliferating endothelium
rather than as a circulatory apparatus. Even after the blood begins to move
through its meshes it is some httle time before the cnculatory function becomes the
dominating influence in the determination of its architectural features, and thus,
in these early stages of the vascular system, we meet arrangements which, as
regards their form, size, and communications, are distinctly inefficient as to the
circulatory flow of the contained blood, but are quite characteristic of the germinal
period of an endothehal meshwork. Any attempt to distinguish arteries from
veins in this primordial system results only in hopeless confusion.

DIFFERENTIATION OF PRIMORDIAL SYSTEM INTO ARTERIES. VEINS, AND CAPILLARIES.

Our information regarding the angioblastic period in which the primordial
blood-vessels are laid down has been derived mainly from study of chick embryos.
The transition into the second developmental period in which the primordial
system gradually undergoes resolution into arteries, veins, and capillaries has
been demonstrated in mammalian material (pig embryos), and toward the end of
the period, as the primary circulation becomes established, the chief features have
been described in human embryos. As illustrating these transitions, the reader
is referred to figures 394, 395, and 400 of Evans (1912), and figure 1, plate 4, of
Sabin (1917a). From the time when the primordial blood-vessel system of the
head is first laid down, its endothehal walls undergo active prohferation and a
sprouting meshwork extends from it, invading the interval between the ectoderm
and the brain-wall. This spreading of the endothelial plexus is more active in
some directions than in others. In general it tends to spread toward and over
the surface of the neural tube and the nerve structures connected with it. It
extends early over the midbram and forebrain and soon encircles the optic stalk.
Along the hindbrain the plexus formation is somewhat slower. In fact, in the chick
the primordial vessel in this region persists as a simple channel until the embryo
has acquired about 29 somites, by which time the growth in the more cephalic
region is quite extensive. Persisting so long unchanged, as it does in the chick,
it was named the "vena capitis media," in contradistinction to the more laterally
situated channel that develops later, the so-called "vena capitis lateraUs." It
was shown, however, by Sabin (1917a) that what exists here in the chick at this
time is the persisting primordial channel and not a true vein.

In its earUer stages the prohferating meshwork shows considerable irregularity
in the form and size of its constituent channels. Gradually it can be seen that the
more superficial loops of the mesh are taking the form of larger and more directly
continuous channels. The deeper loops of the mesh flatten out in a more uniform
capillary sheet that lies in direct apposition to the brain-wall and its attached

12 THE DEVELOPMENTAL ALTERATIONS IN THE VASCULAR SYSTEM

structures. The intermediate portions of the mesh maintain the communications
between the deep capillary sheet and the superficial main channels. The few
communications of the aortic system with the intermediate and deeper jxart of
the plexus are maintained and unite with selected loops of the plexus to form slen-
der continuous channels which are eventually lost in the plexus. In this way the
irregular plexus of proUferating endothelial tubes is resolved into a definite circula-
tory system, consisting of a capillary mesh that is fed by deep branches from the
aortic system and drained by superficial tributaries into a main channel that
extends backward to empty into the common vitello-umbilical vein. By the time
that this is accomplished the contained blood is already slowly circulating through

this primitive system.

The main drainge-channel which thus becomes established on each side of the
head was originally called the anterior cardinal vein until it was recognized that
only the caudal portion of it properly belonged to the cardinal system. Evans
(1912, p. 676) suggested, as more appropriate for it, the name "primitive head-
vein,'' and the same term was utilized by Sabin (1917a). This primary head- vein
develops essentially in the same manner both in the chick and the pig— that is,
through the elaboration of a simpler and more direct channel through the superficial
part of the proliferating plexus.

The difference, however, in the anatomy of chick and pig embryos is associated
with some difference in the details of the development of this primary head-vein.
The pre-trigeminal portion of it is identical in both forms and consists of a super-
ficial main channel, receiving tributaries from the deeper part of the plexus. It
forms in the interval between the thalamus and the optic bulb and leads backward,
median and ventral, to the trigeminal ganglion, where it temporarily connects with
the primordial system. The middle portion of the i)rimary head-vein, the portion
in the interval between the trigeminal and vagus nerves, is formed later than
the pre-trigeminal portion, and is formed relatively later in the chick than in the
pig. In the chick, it has been shown by Sabin (1917o) that the primordial channel
running along the ventro-lateral margin of the hindbrain persists, in embryos of
29 somites, as a single large channel communicating in front with the elaborate
plexus of the midbrain region and the already formed pre-trigeminal portion of
the main drainage channel. Caudally it communicates through the transverse
vein of the first intersixice with the anterior cardinal vein, the latter forming the
caudal portion of the primary head-vein. Thus, in the chick the hindbrain portion
of the primordial blood-channel (the vasa primitiva rhombencephah of Sabin)
serves temporarily as a part of the circulatory apparatus before its proliferative
function has been completed, but in the stage to which we are referring proliferating
loops have hegun to spread from the main channel over the wall of the hindbram
and its attached ganglia. Derived parti}- from these and partly from the plexuses
of the gill-arches, there is formed a series of superficial loops which link them-
selves together into a slender longitudinal channel communicating in front with
the pre-trigeminal portion of the main drainage-channel and extending backward
lateral to the otic vesicle, to empty into the anterior cardinal vehi. This consti-

OF THE BRAIN OF THE HUMAN EMBRYO. 13

tutes the middle portion and completes the formation of the main drainage-channel
of the head, the primary head-vein. In the pig, the primordial blood-channel
along the margin of the hindbrain, instead of enlarging as a simple temporary
channel as seen in the chick, becomes resolved almost at once into a prolifer-
ating plexus, and from its superficial loops, and perhaps also from some of the
loops of the adjacent branchial plexuses, there is evolved the middle portion of the
primary head-vein, which is completed nearly as soon as the pre-trigeminal portion
and before there is any considerable circulation of the blood. The caudal part of
the primary head-vein is made up of the anterior cardinal vein, which, as we have
seen, is originally continuous with the primordial channel, joining it (in the pig)
in front of the occipital myotomes instead of through the transverse vein of the
first interspace, as seen in the chick. As the more superficial loops, sprouting from
the primordial system, become established lateral to the otocyst in the formation
of the middle portion of the primary head-vein, their communication with the
anterior cardinal becomes larger, whereas the original communication of the ante-
rior cardinal with the primordial channel becomes more restricted and breaks up
into a plexus; thus, by the linking-ui) of these three parts, a continuous superficial
channel is formed which extends the whole length of the head and provides for the
adequate and efficient drainage of all its structures.

It has been noted that, from the beginning, communications exist between the
aortic system and the more dorsally placed primordial vascular system of the head.
The largest and most constant are the paired trunks that connect the first aortic
arches with the deeper loops of the ventral part of the forebrain plexus. What
was originally a plexiform communication is later resolved into a single trunk
that eventually forms part of the internal carotid artery of its respective side.
Along the hindbrain region are other irregularly placed, slender communications,
and on reaching the myotome region there are the segmental dorsal branches
of the dorsal aortse, which anastomose with the proUferating plexus of the neural
tube. The primordial blood-channel along each side of the hindbrain proliferates
in the form of a plexus, and as the plexuses of the two sides spread on the surface
of the brain-wall they gradually establish an anastomosis along the mid-ventral
line. At the same time a series of the more superficial loops of the plexus, in
association with the communications from the dorsal aortse, become elaborated
into a slender bilateral longitudinal channel which is continued caudally into the
spinal region, and orally it connects with similar loops which are associated with
the embryonic internal carotid artery. There is thus established, on each side
along the ventral surface of the neural tube, a continuous arterial channel which is
connected dorsally by many loops with the neural capillary plexus and ventrally
by a few branches with the aortic system. From these simple channels are derived
later the main arteries of the brain and spinal cord.

With the estabUshment of the primary head-vein, we may regard the first
type of circulation of the head as completed. It consists essentially of a series of
arterial feeders from the aortic system, which lose themselves in the sheet of capil-
laries that mvests the neural tube, which capillaries in turn are drained by many

14 THE DEVELOPMENTAL ALTERATIONS IN THE VASCULAR SYSTEM

anastomosing loops into the more superficially placed primary head-vein, and thereby
are connected with the duct of Cuvier and the venous end of the heart. It is this
primary arrangement that exists in human embryos 4 mm. long, and this is the
earliest stage that was examined in connection with the present study. In figures
22 and 23, plate 2, are shown the left and right profiles of a model made by the Born
construction method of such an embryo (Carnegie Collection, No. 588, 4 mm.
long). This is shghtly younger than the stage shown by Mall (1905) in his figure 3,
although the conditions in the two are very similar. It is distinctly younger
than both the Ingalls (1907) and the Elze (1907) specimens, although they also
show the same primary type of circulation. In figures 22 and 23 the neural tube,
eye-stalk, trigeminal ganglion, and ear-vesicle are shown in gray and the recon-
struction was planned so as to show the relation to these structures of the main
arteries and veins, the latter being colored red and blue respectively. On ex-
amining these figures it is to be remembered that only the definite supply and
drainage channels are shown. In order to complete the system one must imagine
the neural tube as almost completely ensheathed by a capillary sheet into which
the arterial feeders open and from which the small venous tributaries arise. In
figure 23 fragments of this capillary sheet showing this relation to the drainage-
vessels are indicated in the hindbrain region.

The morphological details of the models will perhaps be more readily under-
stood by the study of these two figures, in which the attempt has been made to
indicate the form and relations as clearly as possible, than by the following of a
descriptive text. Attention, however, may be directed to a few of the general
features. In the first place, it seems to be the rule that, throughout the whole
body of the embryo, the source of blood-supply has a central position and its flow
to the tissue is in a peripheral direction, whereas the return drainage system lies
more superficially and the flow is consequently in a central direction. As for the
drainage of the head, there is provided a simple and adequate channel, the pri-
mary head-vein. Its only deflections from a perfectly direct course are those
rendered necessary by the structure of the parts. Its position is affected by the
trigeminal, facial, glossopharyngeal, and vagus nerves, due to their respective
placodal relations to the ectoderm. It could not pass lateral to the trigeminal
mass, and so is deflected inward. The facial and glossopharyngeal gangha Ue in
positions sufficiently ventral so as not to interfere with its superficial position; the
former, however, deflects it upward slightly. The ganglion nodosum of the vagus
lies dhectly in its course and, as in the case of the trigeminal, the channel is thereby
forced to take a median course, since the preferred superficial course is ruled out
by the placodal attachment. We thus meet with a median deflection at that
point, so that the primary head-vein curves caudally around the vagus trunk.

Just as the main trunk of the primary head-vein follows the most simple and
direct course possible, so its plexiform tributaries are favorably situated for draining
the various areas which are present at that time. Those from the capillary sheet
of the brain-tube join it for the most part on its dorsal and medial border. In
accordance with the topography of the region, these tributaries are arranged in

OF THE BRAIN OF THE HUMAN EMBRYO. 15

three groups: (1) an anterior group from the forebrain and midbrain, leading into
the pre-trigeminal portion of the primary head- vein; (2) a middle group from the
cerebellar region, emptying into its otic or middle segment, that portion between
the trigeminal and glossopharyngeal nerves; and (3) a posterior or occipital group
which accompanies the vagus rootlets and empties near the junction of its middle
and cardinal portions. This posterior group usually empties by a common trunk.
This trunk corresponds to the original communication between the anterior cardinal
vein and the primordial blood-channel of the hindbrain, which now has proliferated
into the meshwork forming the capillary sheet that invests the brain-wall. Before
the completion of the middle or otic portion of the primary head-vein, this trunk
from the occipital group of tributaries was the only communication between the
anterior cardinal and the vessels of the more oral region of the head. Especial
attention is directed to these three tributary groups, as their arrangement is sig-
nificant for the later stages, as will presently be seen. In addition to these tribu-
taries from the brain, the primary head-vein receives ventral tributaries from
the eye region, from the nerve-ganglion masses, and from the region of the first and
second gill-arches, which communicate with the plexuses derived from the pro-
liferating elements of the first two aortic arches.

The arterial supply to the head region at this time is primarily through the
internal carotid arteries, which form relatively slender though direct channels.
On each side they are made up of the trunk that connected the primary head-
plexus with the first aortic arch and the portion of the dorsal aorta corresponding
to the first two arches. It will be noted that the first and second vascular arches
are more or less incomplete, having broken up into irregular plexuses ramifying in
the tissues of their respective gill-bars. A part of the plexus of the second arch
apparently becomes incorporated in the external carotid artery, although the trunk
of this artery is probably represented by the short stem seen in figures 22 and 23,
projecting from the oral border of the third vascular arch. The capillaries and
venous drainage plexuses of the first and second arches are shown only at their
point of entrance into the primary head-vein above. The internal carotid artery
extends forward to the root of the optic stalk, where it bifurcates into its terminal
branches, which soon become lost by anastomosis with the capillary sheet of the
brain-wall. The continuation of this channel backward along the ventral wall
of the brain could not be satisfactorilj^ modeled, though this channel and its anas-
tomosis with the basilar and vertebral arteries doubtless aheady exist. It is
shown in the reconstruction by Ingalls (1907) in a 4.9 mm. embryo.

The cut ends of communications such as originally connected the dorsal aorta
with the primordial vascular system of the brain are indicated. As long as they
persist, these must be regarded as arterial feeders through the basilar arteries to
the capillary sheet of the brain. One of these may doubtless be looked upon as
the stapedial artery (for example, the one between the first and second arches).
The vertebral arteries apparently were established in this specimen, but they could
not be satisfactorily outhned and so were omitted. They must, however, be
included as part of the source of blood-supply. With this manner of blood-supply,

16 THE DEVELOPMENTAL ALTERATIONS IN THE VASCULAR SYSTEM

and with this manner of drainage, as illustrated in figures 22 and 23, one must feel
that the provision for the blood-circulation is well designed and is perfectly adequate
for the structures as then existing, and were there no further alterations in the
structures themselves no further change in their blood-vessels would be necessary.
On the other hand, there are no superfluous channels present and no evident vascular
provision for structures not yet developed, that is, disregarding the potential pro-
liferative power which exists in such an endothelial system. The establishment of
this primary type of the circulation of the head completes our second period in
the growth of the cranial vascular system.

CLEAVAGE OF BLOOD-VESSELS OF HEAD INTO SEPARATE SYSTEMS.

The differentiation of the dura mater and the formation of the arachnoid mesh
begins in the region of the base of the skull, and from there the process spreads
slowly upward toward the vertex of the head. As these structures form it can be
seen that the anastomosing channels connecting the capillary sheet of the brain
with the more suj^erficial drainage channels are gradually closed off, the more
ventral ones first and the more dorsal ones later. In this way the dura forms a
partition that results in a general separation or cleavage of the superficial vessels
(consisting mainly of the primary head-vein and its tributaries) from the deeper
vessels in intimate contact with the brain-wall, including the capillary sheet and
the vessels supplying and draining it. The latter or fleeper system, however,
continues to drain into the former at certain restricted places. As this cleavage
occurs, we can distinguish between dural vessels which are chiefly veins and cerebral
or pial vessels which include arteries as well as veins. Soon after, coincident with
the formation of the membranous skull, the dural system becomes more or less
completely separated from the vessels of the integument and its subjacent soft
parts. We then have three different main strata of blood-vessels — the external,
the dural, and the cerebral. The most conspicuous of the early external vessels
are those belonging to the integument. They make their appearance around the
base of the skull in embryos between 12 mm. and 20 mm. long and spread ujiward
toward the vault. In spreading upward they exhibit a characteristic growing
edge consisting of anastomosing loops of the mesh which can be seen with the
naked eye as an advancing narrow line marking off the non-vascularized area
above from the vascularized part below, as has been clearly pointed out by Hoch-
stetter (1916) and as indicated in our figure 10.

The first steps in this cleavage are well under way in embryos 14 mm. long.
An embryo of this stage (Carnegie Collection, No. 940) is shown in figure 1.
The veins of the head were distended with a natural blood injection and at the
same time the surrounding tissues were quite transparent; as a result, it was pos-
sible to determine their arrangement from a surface examination with consid-
erable detail. A photograph was made of the specimen, and in it were added
the details that could be seen with the aid of a binocular microscope. It was in
this way that figure 1 was obtained. What is seen there is the primary head-vein
and its tributaries. To see the deeper structures it was found necessary to make a

OF THE BRAIN OF THE HUMAN EMBRYO.

17

plastic model of the region. Such a model is shown in figures 24 and 25, plate 3,
being a wax-plate reconstruction made from an embryo 11.5 mm. long (Carnegie
Collection, No. 544). Mall (1905) has pictured about the same stage in his figure
9, and this stage is also pictured by Markowski (1911) in his figure 1.

In their main points all of these embryos correspond rather closely, and ap-
parently the vascular system at this time does not show any great variation. A
large venous channel is formed in the region lateral to the diencephalon and passes

PLEXUS MEDIALIS

V CAPITIS PRIMA

VESICULA AUDITIVA

PLEXUS POST.

Figure 1.
Drawing of the primary head-vein and its main tributaries in a human embryo 13.8 mm.
long (Carnegie Collection, No. 940). The tributaries are arranged in the form of three plex-
uses— the anterior, middle, and posterior dural plexuses. The trunk of the posterior plexus
passes through the foramen jugulare in order to empty into the primary head-vein, and
this point marks the junction of the intrinsic head portion of the primary head-vein wnth the
neck portion or anterior cardinal vein. The latter eventually becomes the internal jugular
vein. Enlarged about 10 diameters.

backward median to the trigeminal nerve and lateral to the otic capsule through
the region of the future middle ear, where it bends sharply downward in the neck
region to finally empty into the duct of Cuvier. All the veins of the cranial region
drain into this main channel. This constitutes the primary head-vein with which
we are already familiar. This vein at 14 mm. differs from its condition at 4 mm.
only in being more deflected in its course by the structures through which it threads
its way. It still forms a fairly direct and efficient drainage-channel. It was this

18 THE DEVELOPMENTAL ALTERATIONS IN THE VASCULAR SYSTEM

primary head-vein that was described by different writers as the anterior cardinal
vein until Grosser (1907) showed that only the caudal portion of it— the part that
is found in the region of the somites and later forms the internal jugular vein —
could be properly spoken of as the anterior cardinal. Salzer (1895) designated
the portion in the presegmental region in the guinea-pig as vena capitis medialis
and vena capitis lateralis, depending on whether it was found median or lateral to
the cranial nerve-trunks. The more cephalic portion, in the trigeminal region, is
alwaj's found median to the nerve and hence is always vena capitis mediaUs. Caudal
to the trigeminal nerve, Salzer describes it as at first coursing medial to the facial,
glossopharyngeal, and vagus nerves, and subsequently, by a process of "island
formation," migrating lateral to these same nerves — that is, changing from vena
capitis medialis to vena capitis lateralis. These terms were advocated on the basis
of an homology with similar veins in the lower vertebrates, and were used in the
recent paper by Shindo (1915). We now know, however, from the work of Sabin
(1917a) that what Salzer called the vena capitis medialis is the primordial channel
of the hindbrain, whose purpose is primarily the proliferation of endothelium, and
hence is not to be regarded as a pure drainage-channel. The primary head-vein
is the first true drainage-channel in this region. Its composite origin has already
been pointed out. It has been shown that it belongs in part to the trunk (the
anterior cardinal vein) and in part is intrinsic to the head. As we shall presently
see, it is the trunk portion or anterior cardinal that forms the internal jugular vein,
whereas the intrinsic head portion in its more anterior segment becomes the cavern-
ous sinus, the otic or more posterior portion (the so-called vena capitis lateralis)
disappearing entirely and being replaced by a more dorsally situated channel.

The tributaries draining into the primary head- vein join it mainly along its
dorsal margin, though there are also ventral tributaries which are especially large
and numerous in the neighborhood of the optic stalk and the trigeminal ganglion,
as seen in figures 24 and 25. A large plexiform sheet Ues median to the maxillary
trunk of the trigeminal nerve, draining the structm-es of the maxillary arch. It is
continuous with the plexus that envelops and penetrates into the substance of the
trigemmal ganglion. It is a modification of this plexus that forms the infraorbital
vein and the venous plexus in the region of the pterygoid fossa. The ophthalmic
vein corresponds to the ventral tributaries just in front of this and median to the
first division of the trigeminal nerve. This is contrary to the view of Markowski
(1911, p. 600), who thought that it was the more caudal and larger of these tribu-
taries, the one draining the maxillary process, that becomes the ophthalmic vein;
whereas the more anterior tributaries, arising from the orbital fossa, he described
as undergoing retrogression and disappearing. That it does not disappear, but
forms the main portion of the ophthahnic vein, we shall be able to see in older stages.

The tributaries drainmg dorsally mto the primary head-vein are arranged
in three plexiform groups, as was pointed out by Mall (1905), the first group empty-
ing into the main channel in front of the semilunar ganghon, the second group
between the semilunar and the acustico-facial gangUa, and the third group caudal to
the otic capsule. .These were designated respectively the anterior, middle, and pos-

OF THE BRAIN OF THE HUMAN EMBRYO. 19

terior cerebral veins. The last one empties into the main channel through a single
trunk, but the other two groups tend to maintain the character of the original
plexus and usually have multiple openings into the primary head-vein. Further-
more, due to the cleavage effect of the dura, which tends to separate them from the
deeper vessels, the veins forming these three groups belong chiefly to the dura
mater and the tissues forming the membranous cranium. There is, therefore, an
advantage in adopting for the temporary description of this period of develop-
ment a terminology something like that of Markowski (1911). In doing so, a
distinction between the lateral and mesial portions will not be made, but the three
groups as given by Mall will be retained. We shall thus speak of the anterior, middle,
and posterior dural plexuses, or (more formally) plexus durce. matris anterior, plexus
durce matris medialis, and plexus dura' fnatris posterior, as they are indicated in
figures 1, 24, and 25. In these figures only the larger channels of the plexus are
shown and it is to be understood that an intervening smaller venous mesh con-
nects them more or less completely.

These three plexuses are not exactly uniform as regards their pattern, but
from the very first they exhibit an individuality that seems to correspond to the
difference in structure of the areas which they drain. The anterior plexus is
modified along its oral margin in adaptation to the form of the bulging hemisphere.
The middle plexus has two larger trunks, one of which drains the capillaries on the
oral surface of the cerebellar plate; the other drains the anterior part of the roof
of the fourth ventricle. It is this latter one that was pictured by His (1904) and
incorrectly described as the sinus transversus (p. 121). In the posterior plexus
there can be recognized usually a single channel that is larger than the others and
that tends to cross the dorso-median line to anastomose with the plexus of the opposite
side, the main channel of the other side being correspondingly smaller, thus giving
rise to a bilateral asymmetry. Another place at which a larger channel is frequently
seen crossing dorsally over the median line in an asymmetrical manner is at the
junction of the midbrain and hindbrain. A third favorable place of this kind is
over the diencephalon along the caudal margin of the cerebral hemisphere. Here
is formed the beginning of the transverse sinus, which consequently shows an
asymmetrical relation to the superior sagittal sinus, as will be pointed out later.

In the ventral portions the dural plexuses are more or less completely separated
from the deeper-lying plexus or capillary sheet that closely invests the wall of the
neural tube, in which are developed the cerebral veins and main arterial supply.
In tracing the plexus dorsalward toward the median line, we find an increasing
frequency of communication between the two, and near the median line they are so
intimately connected that it is impossible to distinguish between them; in other
words, in this region the cleavage between these two layers is not yet estabUshed.
The manner in which the tips of the dural plexuses communicate with the capillary
sheet of the brain-wall is indicated in figures 24 and 25, where a portion of the sheet
is drawn in over the occipital pole of the cerebral hemisphere. It is to be remem-
bered that a capillary meshwork of this kind invests the central nervous system
everywhere, the pattern of the mesh varying somewhat according to the region.

20

THE DEVELOPMENTAL ALTERATIONS IN THE VASCULAR SYSTEM

The main arterial trunks are well established at this stage and afford a more
abundant supplj' of blood to the brain than existed in the 4 mm. embryo. Whereas
the aortic system in the latter conformed to the branchial arches, it now presents a
definite aortic arch derived from the truncus arteriosus and the fourth branchial
arch of the left side. The innominate artery is formed by the fourth arch of the
right side. The third arch has been taken up on each side in the formation of the
common carotid and its bifurcating portion, including the plexiform external
carotid. The internal carotid, basilar and vertebral arteries are present in prac-
tically the adult form. It was noticed in studying the left vertebral artery in this
specimen that its communication with the aorta was more caudal than one would
expect, judging by its position in the adult. It is probable, however, that the
arrangement is a temporary one, and that one of the communications above this,
too slender to model, is destined to become the final trunk of the vertebral artery,
thus affording an instance of spontaneous migration of a blood-channel.

PLEXUS .MEDIALIS

SIN TRANSVERSUS

PLEXUS POST.

SIN.
PETROS. SUP

FORAMEN
JUGULARE

FiGUBE 2.

Profile reconstruction of the dural venous system in a human embryo 18 mm. long fC'arnegie Collection, No.
144). The primary head-vein is still intact; a more dorsal channel, however, is forming through the
meshes of the middle dtiral plexus, coursing backward into the posterior plexus. This new channel
becomes the permanent sinus transversus by which the greater part of the brain is finally drained through
the foramen jugulare. A course completely intercranial (sinus transversus) thus replaces one that was in
part extracranial (primary head-vein). Median to the second division of the trigeminal neri-e can Ije seen
the plexiform maxillary vein, which drains the structures of the maxillarv- process. Enlarged 13 diameters.

Similarly, as was seen in the 4 mm. embryo, the distribution of the arterial
supply and the arrangement of the venous drainage are here (in the 14 mm. embryo)
efficiently laid out from the standpoint of the structures of the head as then exist-
ing. The primary head-vein is already somewhat bent out of its course by the
ear- vesicle and nerve-trunks, and would necessarily become much more so by the

OF THE BRAIN OF THE HUMAN EMBRYO.

21

growth that rapidly follows were it not that a new provision is established to take
care of this, as we shall now see.

ADJUSTMENTS OF BLOOD-CHANNELS DUE TO GROWTH AND CHANGE IN
FORM OF THE BRAIN AND EAR

Before the cleavage between the dural and cerebral vascular systems is com-
pleted, certain alterations in their pattern, associated with the changes in their
environrnent, rapidly follow one another. If one examines a number of series of
about the same age as that just described, and a little older, it is seen that the
primary head-vein maintains the same general course and relations, but the pat-
tern of the dural plexuses is constantly changing, which in the end results in
a change in the primary head-vein itself. In embryos about 18 mm. long an

PLEXUS TENTORII

SIN. TRANSVERSUS

PLEXUS
SAGITTALIS

V OPTHAL

N. TRIGEMINUS

SIN. CAVERNOSUS

Figure 3.

rofile reconstruction of the dural venous system in a human embryo 21 mm. long (Carnegie Collec-
tion, No. 460). The sinus transversus is now clearly established and there is left of the primary
head-vein only that portion which persists as the sinus cavernosus. There is a remnant of it-s otic
portion (marked x) that originallj- connected the cavernous region with the internal jugular vein.
A more complete reconstruction of the blood-vessels of the right side of this same embryo is shown
on plate 4. Enlarged about 11 diameters.

important change occurs by which the blood from the middle dural plexus, which
heretofore had drained into the primary head-vein, in the interval between the
trigeminal and the acustico-facial ganglia, now drains caudalward into the posterior
dural plexus through anastomosing loops that exist between these two plexuses,
passing dorsalward to the otic capsule and just lateral to the endolymphatic sac.

22 THE DEVELOPMENTAL ALTERATIONS IN THE VASCULAR SYSTEM

This can be seen in figure 2, which shows a graphic reconstruction of a human
embryo 18 mm. long (No. 144, Carnegie Collection, CR length 18 mm. in formalin,
14 mm. on slide). This is the same embryo shown in Mall's figure 11 and is about
the same age as the embryo pictured in figure 2 of Markowski. In some respects
the reconstruction referred to differs from both of these. From Mall it differs in
that the greater part of the midbrain and forebrain is still drained by the primary
head-vein. From Markowski it differs in that in our specimen the single large
channel passing backward from the anterior and middle dural plexuses is not yet
established, but instead the region occupied by the anterior and middle plexuses
still shows an extensive anastomosing network not differing much from the pattern
we have already seen in figures 1, 24, and 25. The basis for this new channel
from the middle to the posterior plexus, dorsal to the otic capsule, already existed
in slightly younger stages (fig. 25) in the form of a venous plexus extending across
this region. We are not to conclude, however, that this plexus was placed there
for this particular purpose; it is only such a plexus as tends to form everywhere
throughout the dural .system.

An interesting feature in connection with the establishment of the new channel
just described is the fact that the trunks that originally drained the middle dural
plexus into the primary head-vein nearly disappear, owing to the fact that the
blood that they heretofore carried — i. e., from the cerebellar region and the pos-
terior part of the midbrain — adopts the new channel that is formed dorsal to the
otic capsule and is thus drained into the posterior dural plexus. As a result of this
the original trunks that connected the middle plexus with the primary head-vein
become relatively small and partially break up into a small plexus. We shall see
later, however, that with the next change in the head-vein a trunk will open up here
again as an important channel.

In taking up the question of terminology for figure 2, it is found that most of
the terms used in figures 1, 24, and 25 are still applicable. There are the three dural
plexuses draining into the primary head-vein and also the ophthalmic and maxillary
veins. The anterior dural plexus, however, can be seen to be reshaping itself so
as to come into a more free anastomosis with the middle dural plexus. The middle
dural plexus, by draining as it does over the otic capsule, presents the first stage in
the formation of the transverse sinus — that is, the sigmoid portion of it. The
posterior dural plexus shows less change in its form and connections than any other
group of the head-veins, and this is true also in the later stages. There are some
minor alterations in its pattern, but otherwise it simply extends to become the
occipital sinus of the adult. The primary head-vein can be subdivided into the
trigeminal portion that is to form the cavernous sinus and the otic portion which
passes lateral to the otic capsule accompanying the seventh nerve, and lastly, the
cervical portion or internal jugular vein, the boundary of which is indicated in
figure 2 by the label for. jug. The otic portion already shows a diminution in
volume as a result of the establishment of the new drainage-channel dorsal to
the otic capsule. Dorsal to the otic capsule there is sufficient free space for the
development of a vascular channel, whereas the region ventro-lateral to it becomes

OF THE BRAIN OF THE HUMAN EMBRYO. 23

crowded by the develojjment of the cochlea and the structures of the middle ear
This constitutes a mechanical factor that doubtless has a determining influence
upon the change in the course of this blood-channel.

In embryos about 20 mm. long the veins of the head have an arrangement that
is intermediate between the embryonic type and the adult type. The veins in the
basal portion of the skull closely resemble those of the adult, while the dorsal veins
still have many embryonic features. In figures 3 and 26 are shown reconstructions
of the head of such an embryo (Carnegie Collection, No. 460, 21 mm. long). Figure
3, showing left side of head, is a profile reconstruction, and figxire 26, showing right
side of head, is from a wax-plate reconstruction. The reconstruction of the blood-
vessels in this case was greatly facilitated by the work already done on the head of
this embryo by Professor Lewis, who kindly put all of his tracings and photographs
at my disposal. The study was further facilitated through the fact that the blood-
vessels had been injected through the umbilical vein with India ink by Professor
t^-^abin while the heart was still beating, so that there is a beautiful injection of the
entire vascular system. Before the embryo was cut, sketches and photographs
of the vessels that could be seen from the surface were made by Professor Evans.
For the sake of comparison another embryo sHghtly older (Carnegie Collection,
No. 632, 24 mm.) was studied, and a profile reconstruction of it is shown in figure 4.

On examining figure 3, showing left side of the head, it is seen that the primary
head-vein is now separated into its adult parts. In the trigeminal-nerve region we
can speak of it as the cavernous sinus, receiving as tributaries the oi^hthalmic and
maxillary veins and a large cerebral vein draining the lateral wall of the dienceph-
alon. This latter vein belongs to the cerebral- vein system, eventually becoming
the middle cerebral vein, and runs for the greater part of its course through the
pia-arachnoid membranes. It penetrates the dura and runs a short dural course
before joining the cavernous sinus. It may be regarded as one of the diminishing
number of channels that drain the cerebral venous system into the dural system.
There are also smaller tributaries from a network in the region of the semilunar
ganglion. No tributaries were detected flowing into the cavernous sinus from the
caudal pole of the cerebral hemisphere, such as were found up to this time; all of
this blood now flows in the opijosite direction, caudalward through the middle
dural plexus and the developing transverse sinus. On the right side of this same
embryo (see fig. 26) a communication still exists between the cavernous sinus and
the anterior dural plexus, though it is thinning out. In this respect, then, figure 26
is just before figure 3, and a comimrison of the two shows just how this interesting
reversal of the blood-current takes place.

Tracing the cavernous sinus backward, it can be seen that the interruption
between it and the internal jugular vein is complete, though there is still a remnant
of that connection which extends as a blind channel a short way along the facial
nerve. It is interesting to note that there is occasionally found in the adult skull
a persistent foramen, the foramen jugidare spurium of Luschka, which corresponds
to the exit of this decadent channel. The vein itself, however, has never been
described as persisting, although it exists normally in lower forms as a drainage for

24 THE DEVELOPMENTAL ALTERATIONS IN THE VASCULAR SYSTEM

the anterior part of the brain, passing through this extracranial course to empty
into the internal jugular vein. In the stage we are studying the drainage of the
cavernous sinus is upward over the semilunar ganglion into what may now be
recognized as the transverse sinus. This communication is through a short channel
that approximately represents the original trunk of the middle dural plexus and
constitutes the superior petrosal sinus. This channel is designated by Markowski
(1911) as the vena prontica, and he gives a different origin for the superior petrosal
sinus. According to him (p. 599), it takes its origin from a small cerebral vein
derived from the basal surface of the hindbrain which empties into the vena prootica.
In the further development, the oj^ening of the vein migrates by anastomosis along
the vena prootica toward the sinus transversus and empties either into that sinus
or into the vena prootica near it. According to Markowski, the superior petrosal
sinus has little connection with the cavernous sinus and morphologically represents
a metencephalic vein. Regarding the eventual fate of the vena prootica, he has
apparently made no observations, though he ]iictures it as a large channel in an
embryo 46.5 mm. long. From the si^ecimens I have examined I can not confirm
Markowski's description of the superior petrosal sinus, and I feel convinced that his
vena prootica and the superior petrosal sinus are one and the same thing, and that
which he regards as the superior petrosal sinus is, instead, one of its tributaries.
Mall (1905, p. 17) also described the superior petrosal sinus as the adult form of the
"vena cerebralis media," which, it will be remembered, is the same as the trunk
of our middle dural plexus.

With the alterations in the primary head-vein the anterior, middle, and pos-
terior dural plexuses are drained by means of the new dorsal channel which empties
through the jugular foramen into the internal jugular vein. This channel can be
at once recognized as the transverse sinus, and the sigmoid portion of it presents
relations that are much the same as are found in the adult. The three dural plexuses
are still of the embryonic type. The posterior or occipital plexus is practically
the same as was seen in 18 mm. embryos. Only its coarser meshes are shown in
figure 3. It is more completely shown in figure 26. It will be noted that it is
rather of a different character from the now combined anterior and middle dural
plexuses and (as we shall later see) it is to take less part in the further meta-
morphosis of these vessels.

The whole dural area lying lietween the cerebral hemispheres and the margin
of the cerebellum constitutes the tentorium cerebelli. It is very broad dorsally
and is more constricted ventrally; thus in profile it is wedge-shaped. In the loose
tissue composing it are found the meshes of the dural plexus, the combined anterior
and middle plexuses. As this region becomes more compressed, consequent upon
the growth of the cerebrum and cerebellum, there is a continual adjustment of the
contained venous channels with repeated alterations in the pattern of the meshes.
In general we find the larger channels radiating upward toward the midbrain region,
and as we approach the median line the plexus becomes finer and there is an inti-
mate anastomosis with the subjacent plexus belonging to the brain-wall,

OF THE BRAIN OF THE HUMAN EMBRYO.

25

On comparing embryos 21 mm. long with those 18 mm. long two character-
istic changes are observed in the pattern of the anterior dural plexus at this time.
(Comjjare figs. 3 and 26 with fig. 2). In the first place, the anterior dural plexus
annexes itself to the middle dural plexus and drains backward through this into the
newly estabhshed channel dorsal to the otic capsule. We will therefore, from now
on, refer to the combined anterior and middle dural plexuses as the tentorial plexus,
on the basis of its distribution, which is probably a more satisfactory terminology
than to group both of them under the designation anterior dural plexus, as was
done in the former paper (Ptreeter 1915). In the second place, there is differ-
entiated along the margin of the cerebrum and between the hemispheres a sub-
division of this ))lexus that is eventually to constitute the superior sagittal sinus
(marked plexus sagittalis in figs. 3, 4, and 26).

PLEXUS TENTORII

PLEXUS

SAGITTALIS

SIN. TRANSVERSUS

V. OPTHAL.

SIN. CAVERNOSUS

Figure 4.

Profile reconstruction of the main dural veins in a human embryo 24 mm. long (Carnegie Col-
lection, No. 6.32). The sigmoid portion of the transverse sinus can be identified as that part
between the point of entry of the superior petrosal sinus and the jugular foramen (marked X).
Enlarged about 5 diameters.

Examination of jihotographs and sketches of embryos of about this age shows
that there is a tendency to the formation of a larger channel along the anterior
margin of the tentorial plexus — that is, along the caudal margin of the cerebrum.
This was designated by Markowski (1911) as the anterior marginal vein (vordere

26 THE DEVELOPMENTAL ALTERATIONS IN THE VASCULAR SYSTEM

Grenzvene) ; the large tributary, draining the lateral surface of the cerebrum, that
empties into it, he calls the lateral telencephalic vein, of which there may be several.
Markowski describes the anterior marginal veins of the two sides as extending
forward and toward the median line and uniting in the formation of a plexus out
of which is to be derived eventually the superior sagittal sinus. From examination
of figures 2, 3, and 4 it can be seen that there is no sharp line between the sagittal
plexus as described by us and the more ventral loops of what was the anterior
dural i:)lexus of which it is a part. The anterior marginal vein of Markowski is a
part of both of them, as can be plainly seen in figure 26. The discussion regarding
the formation of the superior sagittal sinus will be reserved for a subsequent part
of this paper. We may, however, point out at this time that the anterior marginal
vein of Markowski is apparently not a definite vein, but rather a constantly changing
channel. What we find is that the more anterior loops of the tentorial plexus are
constantly dropping out and are replaced by the development of the more caudal
channels. By comparing figures 3 and 4 we can see this change occurring. Our
interpretation of the condition found in figure 4 is that what had been a larger
channel along the cerebral margin of the tentorial plexus is now dwindling.into a
small mesh, whereas the main blood-stream forms for itself a new course in a more
caudal loop of the plexus.

In this connection it may be pointed out that "migration of veins" may occur
in at least two ways. There may be a passive change in position or direction of
the endothelial tube itself, due to mechanical causes arising from alterations in its
environment; this is illustrated by the sigmoid portion of the transverse sinus
and its change in form in the later stages (embryos more than 20 mm. long). On

Direclionof ^p.V?-- --

F'«uRE 5. migrcrtion.-

Camera-Iucida sketch of a portion of the right transverse sinus in

a human embrj'O 27..5 mm. long (Carnegie Collection, No. 1458). ■'/ ^ JL

It illustrates the manner of channel-migration in the loops of , ,'//y/ -Dwindling channel,

the tentorial plexus, whereby the main channel follows a course
in adaptation to the caudal growth of the cerebral hemisphere.
While this process is under way there are found along the cere-
bral margin the slender remnants of discarded channels.

the other hand, a vein may change its position by forming or adopting a new
endothelial channel and at the same time relinquishing its original endothelial chan-
nel. The embryonic plexiform character of the veins in the region of the tentorium
is especially favorable for this procedure, and we find this type of alteration in
the blood-channels repeatedly illustrated in this region. In other words, under
migration of veins we are to distinguish between passive migration (where there is a
change in position due to some flexion or traction on the vein-wall itself) and spon-
taneous migration (where there is a change in position of the blood-stream only), and
where, by a process of what might be called circumfluent anastomosis or anasto-
motic progression, the blood-stream develops a new channel in the adjacent loops
of the plexus, with a corresponding dwindling of the previously used loop, as is
illustrated in figiu-e 5. From the observations of Evans (1912) on the ventral

OF THE BRAIN OF THE HUMAN EMBRYO. 27

branches of the aorta, it is apparently possible to obtain a spontaneous migration
without the aid of collateral loops. Here the result is obtained by unequal growth
of the endothehal walls. As a subhead under spontaneous migration we might
include replacement channels. In this process there is the formation of a new channel
and the obliteration of an old one. A replacement channel differs from other
spontaneously migrating channels in that it is not a gradual and progressive change
in position, but an abrupt and immediately complete one. Furthermore, the new
channel lacks the morphological characteristics of the old one. An illustration of a
replacement channel is the channel dorsal to the otic capsule (transverse sinus),
which supplants the otic portion of the primary head- vein.

The lateral telencephahc veins of Markowski apparently correspond to the
inferior cerebral veins of the adult, so we will label them in that way. Though
emptying into the dural system, they develop their course through the intradural
membranes and become typical cerebral veins. It is interesting to note that in the
21 mm. embryo certain definite topographical points in the transverse sinus are
already determined, namely, the jugular foramen, the location of the endolymphatic
sac, the points of entry of the superior petrosal sinus and of the inferior cerebral
veins. Thus we see that more than half of the sinus is already established and
that it is the terminal or jugular portion that is established first. The remainder
of the sinus is relatively late in assuming a permanent form, which is doubtless
the result of the prolonged period of growth of the cerebrum, making a continued
adjustment of the tentorial plexus necessary. Even in embryos 50 mm. long,
which we are about to examine, the proximal end of this sinus is still in the forma-
tive stage. Before leaving this stage, and once more comparing figures 3, 4, and
26, it should be pointed out that the great drainage-channels of the head are effi-
ciently adapted to the drainage of its parts as then existing. We are not to think
of them as busily engaged in building the transverse and sagittal sinuses, but as
carrying on their functional activity in the best manner possible for the moment
and with regard to the available space and the amount of given work. The com-
pleted transverse and sagittal sinuses will come in good time, as determined by
later conditions.

To cover the period of embryos about 50 mm. long, the writer examined four
series belonging to the Carnegie Collection: No. 886, 42 mm. coronal; No. 84, 50
mm. transverse; No. 96, 50 mm. sagittal; and No. 448, 52 mm. sagittal, injected.
There was also an embryo of about the same age (No. 458, 54 mm.) that had been
injected with India ink, the head of which was removed and partly dissected, and then
cleared after the Spalteholz method. This gave excellent total views of the blood-
vessels. The profile reconstruction shown in figure 6 is based on series No. 96 and
was made by preparing tracings on transparent paper which were then superimposed
and a composite tracing made of the whole series. This is about the same stage
that is shown by Markowski in his figure 4. The reconstruction shown in figure 27
is of a younger embryo and was made after the Born-Lewis method.

At this period the arterial supply and the venous drainage of the head are
established along channels that correspond fairly well to those found in the adult.

28 THE DEVELOPMENTAL ALTERATIONS IN THE VASCULAR SYSTEM

It is clearly subdivided into three separate systems: (1) the suijerficial system
belonging to the integument and soft parts, (2) the dural sj'stem lying between the
dura and bone, and (3) the cerebral system. All three are originally outgrowths
of the same capillary j^lexus. The separation of the dural veins and the cerebral
veins we have traced through ste]) by step. The sujierficial vessels in embryos
20 mm. long are already separated off from the dural system by the membranous
and cartilaginous cranium. They appear first in the lower parts of the head,
where, in consequence of the earlier maturation of this region, they are originally
separated off from the deeji system and are in the form of a plexus that gradually
spreads upward over the vault. They maintain a few anastomoses with the dural
system, which constitute the so-called emissary veins. One of these is shown in
figures 6 and 27. Aside from the channel maintained through the orbit, the chief
drainage from the superficial system is through the external jugular vein, which is
jiictured by J^alzer (1895) as already ]iresent in guinea-pig embryos 20 mm. long.

On examining the dura in embryos 50 mm. long it will be seen that for the
greater part it closely invests the interior of the developing cranium and is relatively
poor in blood-vessels. Tliis is true especially in those portions where the carti-
laginous and bony cranium is more advanced in its differentiation, as in the base of
the skull and in the frontal, temporal, and lower occipital regions. In other regions
the dura projects within the cranial cavity, being separated from the future bony
skull bj'' a layer of areolar tissue, in the meshes of which are found the large blood-
channels and their tributaries. The largest area of this kind is situated over the
midbrain, extending from the caudal margin of the cerebral hemispheres to the
cerebellum. This area extends laterally down to the base of the skull, narrowing
as it does so. It constitutes what is known later as the tentorium cerebelli, and in
it is included the greater jiart of the dural venous sj^stem. A basal extension of
the tentorium widens out in the region of the semilunar ganglion and in its meshes
is formed the cavernous sinus. A thinner area of the same tissue extends caudal-
ward from the cavernous sinus, median to the otic capsule, to join the jugular region.
The slender plexus of veins extending through this constitutes the inferior petrosal
sinus. Along all the sinuses we find this same areolar meshwork. It is not to be
confused with the developing arachnoid tissue, from which it is everywhere separated
by the dura. Blood-vessels supjilying and draining the brain are also found in
the arachnoid at this time and in some regions they are quite numerous, such as
the region of the Sylvian fissure and along the more ventral parts of the midbrain
and hindbrain. These cerebral vessels are everywhere separated and distinct
from the dural blood-channels, with the exception of the few points where they
empty into the big dural channels, as occurs in the adult. The connection between
the dural system and the cerebral system is no longer by a multiple anastomosis of
small vessels, but instead by isolated larger veins.

Examination of figures 6 and 27 shows that we have in fetuses at this time an
arrangement of the dural venous system that in most respects follows the adult
arrangement. The cavernous sinus still has a simpler character than is found in
the adult. It is situated median and ventral to the semilunar ganglion and has

Of the brain of the human embryo.

29

the large ophthalmic and maxillary tributaries in front. In figure 27 it receives a
large terminal trunk lateral to the infundibulum made up of tributaries coming
from the region of the Sylvian fissure; this corresponds to the middle cerebral
vein of the adult. Caudally the cavernous sinus communicates with the main
blood-stream by means of the superior and inferior petrosal sinuses. The superior
petrosal sinus passes over the cochlear part of the otic capsule and empties above

SIN. RECTUS

SIN. SAGITTALIS SUP.

PLEXUS TENTORII

SIN. TRANS.

EMISSAR.
MAST.

V. OPTHAL

FORAMEN
JUGULARE

Figure 6.

Profile reconstruction of the dural veins in a human embryo 50 mm. long (Carnegie Collection, No. 96). The
oiitlines of the veins of the falx cerebri can be seen through the cerel)ral hemisphere. That part of the
transverse sinus situated caudal to the superior petrosal sinus (sigmoid) shows practically the adult rela-
tions. The more anterior part is still in the form of a plexus embedded in the embryonic tissue of the ten-
torium. Enlarged 4 diameters.

into the transverse sinus. The inferior petrosal sinus consists of a ple.xus of veins
that passes median to the otic capsule to empty at the point of origin of the internal
jugular vein. It is shown in figure 6 but not in figure 27.

As regards the transverse sinus, it has been pointed out that the terminal or
jugular portion of it is established first. In both figures 5 and 27 it can be seen

30 THE DEVELOPMENTAL ALTERATIONS IN THE VASCULAR SYSTEM

that it consists of a single large channel from the point of entry of the superior
petrosal sinus to the jugular fossa (in other words, the sigmoid portion) and has the
same tributaries and the same general relations that are found in the adult. The
remainder or proximal portion of the transverse sinus is less well established, and
the large capillary meshwork found along its dorsal margin shows that the blood-
channels here are still in the formative stage and must still be spoken of as the
temporary tentorial plexus. The main channel is forming along the anterior margin
of this plexus, into which the inferior cerebral vein empties. It can be seen how
this portion of the transverse sinus undergoes spontaneous migration backward
in adjustment to the growth of the hemisphere and thus comes to assume a more
and more horizontal course. This change in direction, together with an increase
in length and diameter of the main channel at the expense of the formative mesh-
work, remains to be completed before the adult condition can be considered as
estabhshed. The variations found in the adult in the region of the confluens sinuum
can be readily understood as variations in channel selection through this tentorial
meshwork.

In the region of the forebrain a fold of dura is interposed between the two
hemispheres and is compressed into a flattened sheet which is to constitute the fak
cerebri. This and the vascular meshwork belonging to it are directly continuous
with the tentorium. Like the tentorium, it passes through a prolonged adjust-
mental period. In embryos 50 mm. long two of its permanent channels, which are
to belong to the dural sinus system, can be readily recognized; these are the superior
sagittal sinus and the straight sinus. In figure 6 the superior sagittal sinus is
quite irregular in outline, which is a result of shrinkage of the sjiecimen. In the
normal state, as seen in other embryos, it passes evenly along the margin of the
cerebrum. Certain details regarding the vessels belonging to the falx cerebri and
the drainage of the chorioidal masses will now be taken u]) in connection with the
formation of the superior sagittal sinus.

DEVELOPMENT OF SINUS SAGITTALIS SUPERIOR.

Under the description of embryos 21 mm. long mention was made of the
formation of a plexus sagittalis as a subdivision of the anterior dural plexus. At
that stage the plexus is clearly differentiated from the remainder of the anterior
dural plexus, as can be seen in the dorsal view of an embryo of about that age shown
in figure 8 and in the reconstruction shown in figure 26. Earlier than this, in em-
bryos about 14 mm. long (fig. 7), the plexus can be recognized, though here it is
not so clearly separated from the general plexus. In such embryos it can be seen
that the larger tributaries of the anterior and middle dural plexuses stop short of
the median line, with the exception of anteriorly, where they merge into a longi-
tudinal plexus that dips in between the developing hemispheres. It is in the meshes
of this plexus that we find the beginning of the superior sagittal sinus; and the
principal steps in its transformation can be seen by comparing figures 7, 8, and 9.
Sketches like these necessarily have to be simplified, and on examining them it
should be remembered that only the larger channels are shown and that in between

OF THE BRAIN OF THE HUMAN EMBRYO.

31

there is everywhere a fine anastomosing network. Also, the channels do not lie
all in the same plane. Furthermore, it is to be noted that there exists in embryos
of the same age a considerable variation in the pattern formed by these channels.
The three specimens selected, however, may be regarded as illustrating fairly definite
stages in this transformation.

In figure 7 is shown a dorsal view of the head of the same embryo previously
shown in figure 1 (Carnegie Collection, No. 940, 13.8 mm. long). It is about the
same age as the embryos shown in figures 24 and 25. Here we find the sagittal
plexus represented in its simplest form. It will be noted that it possesses two
characteristic features: In the first place, there is a tendency to an enlargement of

TELENCEPHALON

SIN.
SAGITTALIS

DIENCEPH.

MESENCEPHALON

MESENCEPHALON

CEREBELLUM'

8

9

Figs. 7, 8, and 9.— Three stages in the formation of the sagittal plexus, showing its asymmetrical character and its conversion
into the superior sagittal sinus. Draining into it from below is the drainage-channel from the <;horioidal bodies that ia
later known as the straight sinus. The channels marked x are interpreted as undergoing retrogression, being replaced
by more caudal channels. Large dural channels do not persist in the areas covering the cerebral hemispheres; they are
found only in the loose embryonic tissue filling the spaces and fissures that lie between the different subdivisions of the
brain. Figure 7 is a vertex view of a human embryo 13.8 mm. long (Carnegie Collection. No. 940). Figure 8 is a ver-
tex view of an embryo 20 mm. long (Carnegie Collection, No. 349). Figure 9 is a drawing of an injected and cleared
specimen 54 mm. long (Carnegie Collection, No. 458) reversed.

certain portions of the plexus, irrespective of a continuous channel. We thus have
a series of small lakelets connected by narrow channels. A definite single superior
sagittal sinus can not yet be said to exist. In the second place, the plexus is dis-
tinctly asymmetrical and shows a tendency to drain more freely to one side than
the other, in this case to the right.

A more definite and simpler channel system is found in 20 mm. embryos, an
example of which is shown in figure 8 (Carnegie Collection, No. 349). Here
one might possibly speak of a superior sagittal sinus. The channels, however,
are still in the form of a plexus, and hence the term plexus sagittalis is retained.
This view regarding the early identity of the superior sagittal sinus differs from
that given by Evans, who pictures the primitive capillary plexus creeping up on
each side of the forebrain in 8 mm. pig embryos. A portion of the dorsal margin of
this plexus he labels as the primitive superior sagittal sinus (Evans, 1909, fig. 156;

SIN.
TRANS.

32

DEVELOPMENTAL ALTERATIONS IN THE VASCULAR SYSTEM

Evans, 1912, figs. 399 and 400). According to him it is thus originally paired
and bilaterally symmetrical. According to the present writer, it is not until
later that we can speak of a superior sagittal sinus. It is not until the plexuses,
described by Evans, have anastomosed across the median line and have formed a
longitudinal network in the meshes of which an asymmetrical channel is finally
established, that we can speak of a superior sagittal sinus. In some cases two or
more larger longitudinal channels are formed in the mesh, as is shown in figure 10;
but in such cases they are not strictly bilaterally symmetrical.

Owing to the growth of the cerebral hemispheres in 20 mm. embryos, there is
formed a well-marked cerebral longitudinal fissure which is occupied by embryonic
tissue. This rapidly takes the form of the adult falx cerebri. It is in the dorsal
part of this loose dural tissue that the meshes of the sagittal i^lexus are found.

TELENCEPHALON

PLEXUS
CUTIS

Figure 10.

Caniera-lucida sketch of the vertex of the
head of a human embryo 27.5 mm. long
(Carnegie Collection, No. 1458), in
which the lilood-vessels were conspic-
uous through a natural injection. It
is the exception where two channels are
found so nearlj' the same size. From
a single specimen, like this, one might
incorrectly gain the impression that the
superior sagittal sinus is at first bilat-
erally symmetrical. Even here, how-
ever, it can be seen that Ijoth of the
larger channels drain caudally to the
right side. The minute anastomosing
loops that connect the two main chan-
nels with each other and with the chan-
nels in the more ventral ijaits of the
falx are not shown. On each side can
be seen the serrated upper margin of
the advancing capillary plexus of the
skin. Above this margin the skin is as
yet nonvascular. On account of the
dilatation of these vessels along this
.advancing edge the margin is conspic-
uous.

MESENCEPHALON

SIN. TRANS.
RECT.

At this time it can be seen that one or more larger channels are opening along the
dorsal mid-line, which will form the superior sagittal sinus, and connected with
them by anastomosing loops is a more ventrally situated large channel that con-
stitutes the sinus rectus. This latter extends forward and drains the lower part
of the falx. It has two converging limbs in front that drain the choroidal masses
of the hemispheres. In figure 26 a portion of the right hemisphere is removed to
expose the region of the falx cerebri. Here the straight sinus can be seen as dif-
ferentiated out from the sagittal plexus, with which it anastomoses freely at its
caudal end in a plexiform manner. Anteriorly the straight sinus bifurcates, enters
the choroidal fissure on each side, and terminates in the sinus-like choroidal bodies.
These, on the other hand, are fed from the caudal end by the choroidal arteries.

On coming to embryos 50 mm. long, figure 9 (Carnegie Collection, No. 458,
54 mm.), we find that here the superior sagittal sinus is established, at least in part.
In its cephalic portion there is a large characteristic channel, lacking only the dural

OF THK BRAIN OF THE HUMAN EMBRYO.

33

connective-tissue investment to make it an adult type. In its more caudal portion
it still exhibits a plexiform character that indicates its transitional state. Upon
comparison of a number of series, the writer is led to interpret the formation of a
single channel as the outcome of more than one process; in some segments there
seems to be the selection of a favorable loop of the plexus which enlarges and be-
comes the main channel, and in other segments there is apparently an enlargement
of two or more collateral loops which subsequently fuse into a more or less common
channel. Both processes are apparently represented in figure 9. It is to be
expected that we will find a considerable variation in this respect in different
brains. In figure 10 is shown a specimen which is about the same age as that
shown in figure 8. In this case two collateral channels of about equal size have
formed, both draining, however, to the same side.

11 12

Fii:. 11. — Scption showins the sagittal plexus in a human embryo It iiiiii. loii!^ (Carnegie Collection. No. 940. slide 1.5, row
3, section 1). The section shows the falciform area, the hemispheres lieing retracted from its lateral margins. It will
be noted that there are two main plexiform vascular sheets — a superficial one near the skin and a deeper one directly
against the brain-wall, the latter draining into the former by anastomosing loops. The superior sagittal sinus develop.s
in the meshes of the superficial plexus, and the straight .sinus develops in the meshes of the deep plexus over the area
corresponding to the third ventricle.

Fig. 12. — Section showing the sagittal plexus in a human embryo 21 mm. long (Carnegie Collection, No. 460, slide 11, row 3,
section 4), injected with India ink. The arrangements are similar, but more advanced than those shown in figure 11.
The straight sinus can be recognized. The superior sagittal sinus still has the form of a bilaterally asymmetrical plexus
from which a finer mesh work extends into the loose falciform tis.sue intervening between the two hemispheres and anasto-
moses with the straight sinus. On the left side a characteristic communication can be seen connecting the deep cerebral
plexus with the coarser sagittal plexus above.

No attempt was made to study the histological changes that occur in the
completion of the superior sagittal sinus, or of the cavernous sinus. These involve
details with which the present paper is not concerned. The caudal ward growth,
however, of the superior sagittal sinus in adjustment to the corresponding growth
of the hemispheres is of interest in our general problem. By comparing figures 7,
8, and 9 it can at once be seen that this caudal development is accomplished at the
expense of the meshes of the tentorial plexus, in which process the transverse and
straight sinuses also take part. These channels gradually obtain- a more caudal
course by what we have already described as spontaneous migration. The channel
repeatedly shifts into a more caudal loop of the plexus, the new loop enlarging and
the old loop dwindling. The veins marked X in figures 8 and 9 may thus be inter-
preted as discarded channels. The eventual confluens sinuum (torcular Herophili)

34

DEVELOPMENTAL ALTERATIONS IN THE VASCULAR SYSTEM

represents the point at which this caudal development reaches its completion — or,
in other words, is a remnant of the embryonic tentorial plexus and usually retains
a trace of the plexiform character that is found throughout the embryonic stages.
It is interesting to note that the asymmetry of the superior sagittal sinus
expresses itself in the embryo as well as in the adult by a tendency to drain more to
one side of the head than to the other. This becomes established by the time the
embryo is 20 mm. long. The drainage is preponderantly toward the right side.
It happens that in figure 9 the main drainage was in reaUty toward the left side.
In reproducing the sketch the figure was reversed right for left, in order to facilitate
its comparison with figures 7 and 8. In the accompanying table is given a list
of embryos which were examined as to this point, and it will be seen that of 18
specimens all but 2 drained predominantly toward the right side, that is, about 89
per cent. In order that account should be taken of the artificial element introduced
in those specimens where the vascular system had been injected with coloring
matter, such specimens are indicated in the table by an asterisk. No explanation
has thus far been reached to explain this interesting asymmetry. The drainage
of the straight sinus could not be determined as well in the younger stages, and
there were not enough of the older stages upon which to base an average. A
similar asymmetry might be expected here.

Superior Sagittal Sinus.

No.

CR length.

Drainage outlet.

No. ■

CR length.

Drainage outlet.

mm.

mm.

296

17

Right transverse sinus.

199

35

Right transverse sinus.

350

19

Do.

*449

36

Do.

128

20

Do.

224

40

Do.

*349

20

Do.

886

43

Do.

*460

21

Do.

95

46

Left transverse sinus.

966

21

Do.

84

50

Right transverse sinus.

*382

23

Do.

96

50

Do.

*632

24

Do.

*458

54

Left transverse sinus.

145

33

Do.

613

80

Right transverse sinus.

* Injected specimens.

SUMMARY.

In describing the development of the blood-vessels of the brain the process
has been subdivided into the following five arbitrary periods: (1) angioblastic
period, in which the primordial vascular system becomes established; (2) resolution
of the primordial system into arteries, veins, and capillaries, and the estabhshment
of the primary type of circulation; (3) cleavage of the vascular system of the head
into the external, dural, and cerebral layers; (4) adjustments of vascular channels,
due chiefly to growth and change in form of the otic capsule and the brain; and
(5) completion of histological differentiation of the walls of the vessels. Aside from
adding perhaps more emphasis to the morphological aspects of the precirculatory
type of the vascular system, the contributions of this paper are concerned only
with the third and fourth developmental periods, and even there they are more or
less restricted to the main drainage-channels and their gradual metamorphosis into
the adult dural veins and sinuses. It has been possible to present a rather com-
plete series of stages from which the essential factors in this process can be clearly

OF THE BRAIN OF THE HUMAN EMBRYO. 35

deduced. In order to facilitate a review of the successive steps in this interesting
metamorphosis, there have been assembled on plate 1, figures 13 to 21, a series of
simplified sketches, and through the aid of these it is hoped that the steps that
are outhned in the following summary can be readily identified.

In the primary type of circulation the arrangement for the drainage of the
capillaries of the head (figs. 13, 14) consists bilaterally of one main channel, the
"primary head-vein," that starts in the region of the midbrain, runs caudalward
alongside of the brain-tube, and terminates at the duct of Cuvier. The primary
head-vein is composite in origin. That portion of it rostral to the vagus nerve is
an intrinsic vein of the head; the remaining caudal portion is in reahty a neck-vein
and constitutes the anterior cardinal vein — eventually the internal jugular vein.
Together these portions form a continuous channel, the primary head-vein, into
which the blood from the capillary sheet immediately investing the brain-tube is
drained by means of anastomosing venous loops. These loops are arranged more
or less in the form of three plexuses — the anterior dural plexus, the middle
dural plexus, and the posterior dural plexus. Other small tributaries which are
not all shown in the figures empty into the primary head- vein, thereby draining
the structures ventral and lateral to the brain-tube, such as the nerve-ganglion
masses and the maxillary and mandibular gill-bars. A large one comes from the
eye region and eventually is modified into the ophthalmic vein.

From this simple group of drainage channels are eventually derived all the
adult venous sinuses. The metamorphosis which they undergo is based on a series
of circulatory adjustments that are made necessary by certain changes in their
environment, the two most conspicuous being the changes in the region of the
cartilaginous capsule of the labyrinth and the still greater changes involved in the
growth and marked alteration in the form of the brain. Among the factors involved
in these circulatory adjustments may be mentioned the reduction of plexuses into
simple channels, the conversion of channels into plexuses, the total obliteration
of established channels, and the change in position of channels. Under this latter
phenomenon there is to be recognized a "passive migration," where there is a change
in the position of the vein-wall itself, due to the movement of its environment,
which exerts a flexion or traction force upon it. We also recognize a "spontaneous
migration," where there is a change in position of the blood-stream only, where in a
circumfluent manner the blood-stream develops a new channel in the adjacent loops
of the plexus, with a corresponding dwindling of the previously used channel.
The "replacement channel" might be mentioned as another type of spontaneous
migration, in which the venous channels are changed in position and direction in
this process of adjustment. In the replacement channel there is the formation of
a new channel and the obliteration of an old one, as in other types of spontane-
ous migration. It, however, differs from them in that it is not a gradual and
progressive change in position, but an abrupt and immediately complete one.
Furthermore, the new channel lacks the morjihological characteristics of the
old one. With these various factors in mind one can readily follow the .stei:)s by
which the primary head-vein and its tributaries gradually merge into adult dural
sinuses.

36 DEVELOPMENTAL ALTERATIONS IN THE VASCULAR SYSTEM

While the three head-plexuses are spreading ui)ward (figs. 14 and 15), the out-
lines of the dura mater and the arachnoid spaces make their apjiearance, and first of
all in the ventral parts. This results in a general separation or cleavage of the
more superficial i:)rimary head-vein and its three tributary ple.xuses from the sub-
jacent vessels that arise from and drain the capillary sheet directly investing the
brain-tube. This deeper system, however, continues to drain into the former at
certain places, notably in the more dorsal parts. The primary head- vein and its
three tributary plexuses thus become established as a true dural system as dis-
tinguished from the deeper "cerebral veins" belonging to the arachnoid-pial
membrane. The di])loic veins are a later sul^idivision of the dural system. The
superficial vessels of the head belonging to the integument and soft parts are sep-
arated ofT in the more ventral regions and from there sjjread ujiward over the head
independently of the dural system. We then have for the head three separate
systems: (1) the superficial layer belonging to the integument and soft parts;
(2) the middle layer belonging to the dura and dijiloe; (3) the deeji layer of cerebral
ve.ssels belonging to the brain. It is the middle layer, or dviral system, that is
exclusivelj' concerned in the formation of the dural sinuses and whose changes in
form and position we are now following.

In the region of the cartilaginous capsule of the labyrinth adaptive changes
in the dural channels occur early (figs. 14, 15, and 16). Owing to the marked
elaboration of these structures in this region, the course of the primary head-vein,
ventro-lateral to the otic capsule, becomes an vmfavorable one. If it persisted it
would be tortuous and remote from the area drained; instead, this part of it becomes
obliterated, and during this obliterating process an adjustment is made in two ways
(figs. 14, 15, and 16): first, a channel is established in the venous plexus above
the otic capsule, and through this the middle dural plexus thereafter drains caudally
into the loops of the posterior dural plexus; second, the anterior dural plexus,
which originally drained into the primary head-vein, completely reverses its direc-
tion of flow and drains through anastomosing loops into the middle dural plexus and
through the newly established channel dorsal to the otic capsule.

In this way a complete trunk for the drainage of the head becomes established
which is everywhere dorsal to the primary head-vein as far as the jugular foramen,
where it is continuous with the internal jugular vein. Of the i)rimary head-vein
there is left, in addition to the cardinal portion of it or internal jugular vein, only
that part in the region of the trigeminal nerve. This may now be spoken of as the
"cavernous sinus." Into it drain a vein from the base of the brain and the veins
from the orbital and maxillary regions; whereas it, in turn, drains ui)ward through
the original trunk of the middle plexus, which is now the superior petrosal sinus,
into the newly established dorsal channel. By comparing with later stages (figs.
17 to 21) it will be seen at once that this dorsal channel is the transverse sinus, of
which that part between the superior petrosal sinus and the jugular foramen forms
its sigmoid portion. Thus in the 21 mm. embryo the dural chaimels in the region
of the temjioral V)one have acciuired essentially all tlieir permanent connections, with
the exception of the inferior petrosal sinus, which appears a little later (fig. 19).

OF THE BRAIN OF THE HUMAN EMBRYO. 37

Otherwise there remains to complete the aduU condition only a certain amount of
passive migration in accommodation to the changes in the adjacent parts.

The adjustment in the dural channels rendered necessary by the protracted
growth of the hemispheres extend much later in fetal life. A large part of this
adjustment is accomplished by spontaneous migration of the principal channels,
and for this reason a venous i)lexus is essential. We thus find in the neighborhood
of the advancing occipital pole of the hemispheres a continuous persistence of the
transitory or embryonic dural plexus from which are evolved all the veins of the
falx cerebri and of the tentorium cerebelli.

An anterior subdivision of the plexus extends forward in the median hne as the
plexus sagittalis, being interposed as a vertical curtain between the hemispheres.
Among its dorsal meshes is developed an asymmetrical longitudinal channel which
we know as the "superior sagittal sinus." In its early stages this channel is made
up of several collateral anastomosing veins. The eventual single channel is formed
in the anterior portions by the selection and enlargement of the most favorable
vein with a corresponding disappearance of the others. In the posterior portions
there is apparently some coalescence of adjacent veins. The anterior part of the
sinus is completed first. As the hemispheres extend backward the sinus corre-
spondingly elongates itself by incorporating the more caudal loops of the plexus.
Transverse sections through this part of the sinus in older fetuses thus usually
reveal incomplete coalescence of the separate loops. The sagittal plexus very early
exhibits a tendency to drain more to one side of the head than to the other and
usually toward the right side. As the superior sagittal sinus becomes established
we thus find that caudalward it is usually continuous with the ventral main channel of
the right anterior plexus (or tentorial plexus as it is better called in the late stages),
which eventually forms part of the right transverse sinus. The straight sinus is
formed in the ventral part of the sagittal plexus and its caudal adjustment is
essentially like that of the superior sagittal sinus. It may drain chiefly toward the
right or left plexus or equally toward both.

In embryos between 35 and 50 mm. long (figs. 18 and 19) we can recognize a
main channel of the tentorial plexus that, is to become the transverse sinus.
If we disregard the sigmoid portion of it, it forms a fairly straight line with the
internal jugular vein. In the interval between the 50 mm. embryo and the adult
the transverse sinus bends backward until it comes to lie at an angle of 90° with
the internal jugular. This marked change in ijosition is accomplished in large
part by spontaneous migration, by the repeated shifting l)ack of the main blood-
current into more caudal loops of the plexus, with subsequent dwindling of the
discarded anterior loops. As the sinus becomes more definitely established the
tentorial plexus becomes relatively smaller (fig. 20) and the final change in position
is completed by passive migration, that is, actual traction on the vein-wall by its
environment. In this change in position of the transverse sinus the superior
sagittal sinus and the straight sinus participate and we find in the adult, at the point
where they meet, an anastomosis, the conflucns sinuiini which is usually ]ilexi-
form in character and represents the last trace of the embryonic tentorial plexus.

38

DEVELOPMENTAL ALTERATIONS IN THE VASCULAR SYSTEM.

REFERENCES CITED.

Elze, C. 1907. Beschreibung eines menschlichen Em-
bryo von za. 7 mm. grosster Lange. Anat.
Hefte, No. 106, vol. 35, Ahth. 1.

Evans, H. M. 1909. On the development of the aortse,
cardinal and umbilical veins, and other blood-
vessels of vertebrate embryos from capillaries.
.\nat. Rec, vol. 3.

. 1912. The development of the vascular system.

Keibel-Mall Manual of Human Embrj'ology, vol. 2.

Grosser, O. 1907. Die Elemente des Kopfvenensystem
der Wirbcltiere. Verh. .\nat. Gesell., Anat. Am.,
Bd. .30, Erg.inz. Hft.

und E. Brezix.\. 189.5. Uel)er die Entwicklung

der Venen des Kopfes und Halses bei Reptilien.
Morph. Jahrbuch, Bd. 23.

His, \V. 1904. Die Entwickelung des menschlichen
Gehirns. Leipzig.

HocHSTETTER, F. 1916. Ucbor die V.askularisation der
Haut des .Schadeldaches menschiicher Embr>'onen.
K. Akad. d. Wiss. Wien, Math.-Naturwis.s. Kl.,
Bd. 93.

Ingalls. N. W. 1907. Beschreibung eines menschlichen
Embryo von 4.9 mm. Archiv. f. mikr. Anat.,
vol. 70.

Mall, F. P. 1905. On the development of the blood-
vessels of the brain in the human embryo. Am.
Jour. Anat., vol. 4.

Markowski, J. 1911. Ueber die Entwicklung der Sinus
durae matris und der Hirnvenen bei menschlichen
Embryonen. Bull. d. I'Acad. d. Sci. de Cracovie.
CI. d. Sci. math. -Nat., Serie B.

Sabin, F. R. 1915. On the fate of the post-cardinal veins and
their relation to the development of the vena cava
and azygos in the embryo pig. Contributions to
Embryology, Carnegie Inst. Wash. Pub. No. 223.

. 1917a. Origin and development of the primitive

vessels of the chick and of the pig. Contributions
to Embrj'ology. Carnegie Inst. Wash. Pub. No. 226.

. 19176. Preliminary note on the differentiation of

angioblasts and the method by which they produce
blood-vessels, blood-plasma, and red blood-cells
as seen in the living chick. Anat. Rec, vol. 13.

Salzer, H. 1895. Ueber die Entwicklung der Kopfvenen
des Meerschweinchens. Morph. Jahrbuch, Bd. 23.

Shindo, T. 1915. Ueber die Bedeutung des Sinus caver-
nosus der Siiuger mit vergleichend anatomischer
Beriicksichtigung anderer Kopfvenen. Anat.
Hefte, Abth. 1, Heft 157, Bd. 52.

EXPLANATION OF PLATES.

Plate 1.

Simplified profile drawings of the dural veins, showing the manner in which they adapt themselves to the growth and change
in form of the brain in human embrjos from 4 mm. to birth.

Fig. 13, embryo No. 588, 4 mm.; fig. 14, embryo No. 940, 14 mm.; fig. 15. embrj-o No. 144. 18 mm.; fig. 16,
embryo No. 460, 21 mm.; fig. 17. cmbrj'o No. 632, 24 mm.; fig. 18, embrjo No. 199. 35 mm.; fig. 19,
embryo No. 96, 50 mm. CR length: fig. 20, embryo No. 234a, 80 mm. CR length; fig. 21, adult.

Plate 2.

Right and left profile views of a wax-plate reconstruction of the main arteries and veins in a human embryo 4 mm. long
(Carnegie Collection, No. 588). Enlarged about 40 diameters.

This stage illustrates the character of the first type of the circulation of the head and its relation to the other
blood-vessels of the body. The primary head-vein and its tributaries which form the main drainage-channels of the
head aie shown in blue. These communicate by anastomosing loops with the capillarj' plexus everywhere investing the
brain-wall, only patches of which are shown in the model. The capillary plexus of the brain-wall is fed by aiterial
feeders, the stumps of which, as shown in the model, arise from the aortic system. The trunk that persists as the
internal carotid artery is already quite definite.

Plate 3.

Right and left profile views of a wax-plate reconstruction of the blood-vessels of the brain in a human embryo 11.5 mm. long
(Carnegie Collection, No. 544). Enlarged about 14 diameters. The primary head-vein stiil constitutes the main
drainage-channel of the head. The manner in which its tributaries tap the deep capillary sheet investing the brain
is indicated over a small area of the cerebral hemisphere. A capillary mesh of that kind invests the entire central
nervous system, but is not shown in the model.

Plate 4.

Left lateral view of a wax-plate reconstruction of the larger blood-vessels of the brain in a human embryo 21 mm. long (Car-
negie Collection, No. 460). Enlarged 16.4 diameters. Instead of the head being drained by the primary head-vein,
this is now accomplished by a more dorsally situated channel that has formed through the meshes of the middle and
posterior dural plexuses to become the transverse sinus. (Compare with text-figure 3, which shows a left profile of
the same specimen.) All that is left of the primary head-vein is that portion which is to become the cavernous sinus.
In this model the right cerebral hemisphere has been dissected so as to expose the chorioidal body with its arterial
feeder and the straight sinus draining it. The plexiform character of the superior sagittal sinus and of the caudal end
of the straight sinus is indicative of their transitory condition.

Plate 5.

Left profile view of a wax-plate reconstruction of the blood-vessels of the brain in a human embryo 43 mm, long (Carnegie
Collection. No. 886). Enlarged 8 diameters. The vascular architecture is at this stage beginning to approximate
the adult condition, although the whole tentorial region retains its embiyonic character, d\ie to which the marked
subsequent migr.ation of the transverse sinus is possible. It will be noted that the siguioid portion of the sinus is
fairly well established.

STREETER

PLATE 1

.■tt/otn &Co fAfh

STREETER

e

Q_

PLATE 2

CD

eg

_0)
D-

v^r »rt>i

^l^/« '^

*^

a.

£

\o

>

A Hoen S Co Ufh

J, F, Didusch fecrt

STREETER

PLATE 3

I/)
CM

O)

d

D
X

a-

in

a

X)

<

X

CD

a.

e

0)

Z

0

X

0

"ta

*»-<fl^

05

E

I

T**

■K

!^*

in

X

0)

J. F". Didusch fecit

/

AHoen&Co Uth

STREETER

PLATE 4

U)

LL

U)

5

0

o

L

0

%

<

CL

>

<

J.F, Didusch fecit

STREETER

PLATE 5

Plexus Sagittalis

Plexus tentorli

A et V.

Emissan mast.

Bulbus V.jug. sup
A, car int.

" car com sin

Asubclav.

A pulmon.

Fig.27

J F. Oidusch fecit

CONTRIBUTIONS TO EIMBRYOLOGY, No. 25.

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM

By E. V. CowDRY.

Ouc plate and uinc text-figures.

39

CONTENTS.

PAGE.

I. Historical review 41

II. Nomenclature 47

III. Literature:

Varieties of cells 51

Different types of organisms 5.3

IV. Technique 58

Examination of living cells unstained 58

Vital staining 58

Supravital staining 59

Fixation and staining 59

V. Morphology:

In organisms 66

In tissues 66

Filamentous mitochondria 66

Filaments with hleb-like swellings, rods,
granules and dumb-bells, networks and

branching forms 68

Rings and vesicles 69

VI. Distribution;

In organisms 72

In tissues 72

In cells 72

Polarity in secreting cells, during cell divi-
sion, relation to centrosome. perinuclear

condensations 72

Accumiilations in peripheral cytoplasm, in
ciliated epithelial cells, clumping and fu-
sion 76

VII. Amount:

In phylogeny 78

In ontogeny 78

In cytomorphosis (senescence) 78

In dividing cells 79

In different cell types 79

Significance of variations in the amount of

mitochondria 82

40

PAGE.
VIII. Chemistrj-:

Constitution 83

Variations in constitution 85

Reaction to Janus green 86

IX. Embryology:

Fertilization 94

Inheritance 96

Mitochondrial continuity 97

Organ-forming substances 99

Histogenesis 101

Fibrils 104

Plant plastids 113

Pigments 115

."-'ccretions 116

Fats 123

Other products 124

X. Physiology:

Secretion 129

Contraction 130

Irritability and conduction 130

Metabolism 131

Respiration 133

XI. Pathology:

Glandular system 135

Blood vascular system 136

Urino-genital system 136

Respiratorj', muscular, and supportive

systems 137

Nervous system 137

Regeneration 138

Fever 139

Acidosis 139

Tumors 139

XII. Discussion:

The general results of mitochondrial work. . 141

The possibilities of further study 144

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM/

By E. V. CowDRY.

INTRODUCTION.

Schultze defined the living substance, protoplasm, as being a glass-like, semi-
viscid material in which granules are embedded. We have been permitted to go a
step further, because we can now recognize among these granules a definite class,
which we call mitochondria and which, within a surprisingly short space of time,
have been described in the cells of all tissues, in the adult condition and in all stages
of development. They have been found in rapid succession in all organisms which
have been studied, from the protozoa to man, and from certain of the algse to the
highest of plants. It has become apparent that they are inseparable from living
protoplasm and that they yield place to no other visible constituent of living
matter in breadth of distribution. Their characteristic form (strongly suggestive
of bacteria), their lipoid properties, and their extraordinary sensitivity to certain
types of pathological change have become familiar. These and other discoveries
have stimulated interest in many quarters and have given rise to hastily conceived
and poorly supported theories and to much discussion, with the result that ideas
of cell structure have been, in large measure, recast.

Recognizing the importance of this new line of work, I have attempted to set
forth our present knowledge of mitochondra and their significance from the
standpoint of cell development and cell function, making use of unpublished
observations of my own and of the literature on the .subject. In a field so large it
has been necessary to choose and select, to elaborate some points and to leave
others almost untouched, so that many important contributions have been passed
without mention. At first the researches on mitochondria were purely descriptive,
but now they have taken an experimental turn, and it is this aspect of the subject
which I venture to emphasize.

From the beginning, the Department of Embryology of the Carnegie Insti-
tution of Washington has placed valuable apparatus at my disposal. The work
was carried on in the Anatomical Laboratory of the Johns Hopkins University, and,
during the summer of 1916, at the Marine Biological Laboratory, Woods Hole.
It has been completed through the action of the Peking Union Medical College in
giving me the freedom and the time necessary during the past year.

I. HISTORICAL REVIEW.

It is impossible to say who first discovered mitochondria, for, with the out-
burst of interest in cell granulations between 1870 and 1890, coincident with the
introduction of ajjochromatic lenses, mitochondria were observed and described
by many authors under very diverse names. Unhappily, however, this group of

'Contributions from the .\natoniical Laboratory, Peking Union Medical College No. 1.

41

42 THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

investigators failed to define them accurately and classified granules of totally
different nature, under the same heading, with the true mitochondria.

Chief among them was Flemming (1882, p. 77), who studied granules and fila-
ments in cells of many varieties in great detail, and, on the basis of his observations,
erected his celebrated "filar" theory of the constitution of protoplasm. We now
recognize among his "fila" our mitochondria of to-day, as well as other structures
of totally different character. Unfortunately, Flemming's work was limited by the
use of a mixture for fixation which contained a relatively large amount of acetic
acid, which dissolved the mitochondria in many of the cells which he studied.

Altmann (1890, p. 100) was able to go considerably further than Flemming
by the discovery- of a much superior fixative containing no acetic acid and capable
of preserving all the mitochondria. Unhappily, however, his technique was
still far from specific and brought to light many granulations other than mito-
chondria, like zymogen and fat, which he included with them under the general
heading of "Bioblasts." These "Bioblasts" he believed to be ultimate living
particles, or elementary organisms, existing in the form of colonies in all cells,
and he can not much be blamed for this mistake in view of the very real similarity
between mitochondria and bacteria. Nevertheless his theories deterred many
from the study of mitochondria.

F. and R. Zoja (1891, p. 237), following Maggi (1878, p. 326), made an elaborate
study of mitochondria under the heading of "Plastiduli fucsinofili" and arrived at
the interesting conclusion that they play a part in nutrition which approximates
surprisingly closely to our modern ideas. Their results are of special value, inas-
much as they paid particular attention to invertebrates, while Altmann confined
his observations to the cells of vertebrates. Others, about the same time, de-
scribed mitochondria under the headings "Cytomicrosomes," "Neurosomes,"
"Plasmosomes," " Plasmafaden," and so on (see table 1).

Another reason for the lack of interest shown in mitochondria during the
succeeding ten years is shown by a consideration of the technique. Virchow and
his followers in pathological cytology directed their chief attention toward the
nucleus; and biologists, dominated by the heredity problem, all looked in the same
direction. Consequently their aim in making up fixatives was to show nuclear
detail. For this purpose mixtures containing subHmate, alcohol, chloroform, or
acetic acid were employed because of their rapid penetration and their action
on chromatin. Now, these substances, unless certain precautions are taken,
destroy mitochondria; so that the more attention was focussed upon the nucleus,
the less chance there was for observation of mitochondria. Thus a vicious cycle
was produced and maintained until fixatives made up with a basis of formalin,
bichromate, or osmic acid were introduced.

The newer work on mitochondria may be said to begm with Benda's (1899(/, p.
397) study of them in spermatogenesis. He modified Flemming's fluid by reducing
the amount of acetic acid in it and devised a staining method by which the mitochon-
dria ma}' be colored with crystal violet, which, though not specific, has been of the
utmost service to mvestigators. Benda also introduced the term "mitochondria."

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. 43

In a review of this sort, one can not help being impressed with the existence of
certain definite landmarks which have determined the whole trend of subsequent
investigation and about which our ideas revolve. It can not be denied, for
instance, that Meves's generalizations and theories, unjustified though they may
be, have exercised a most stimulating effect upon the study of mitochondria. His
statement (1908, p. 845), prompted by the discovery of mitochondria in all
embryonic tissues, that all cellular differentiations are formed from mitochondrial
coming at a time when the origin of these differentiations had been more or less
explained to the satisfaction of cytologists without reference to mitochondria,
attracted world-wide attention. Much work was hastily done anew with the
most conflicting results, and even now uncertainty prevails in all branches of histo-
genesis. Similarh', his doctrine that mitochondria constitute in part the material
basis of heredity, supported by the discovery that they enter the egg on fertili-
zation, coming just when the chromatin hypothesis was receiving its strongest
support at the hands of Morgan and others, could not fail to attract attention.

We owe much to Regaud (1908rf, p. 720) and his school for supplying, through
skillful indirect methods, the first accurate information regarding the chemical
constitution of mitochondria, according to which they are made up of a com-
bination of phospholipin and albumin. This immediately clarified our ideas,
enabled us for the first time to form some estimate of their potentialities, and
served as a point of departure for many investigations of value. It is important
also to note that Regaud has thus brought the whole work on mitochondria into
line with the recent tendency among physiological chemists and pathologists to
become interested in the phospholij^ins, whereas formerly their whole attention was
devoted to the stud>' of proteins, being dominated by the tremendous impetus of
Eniil Fischer's work on protein synthesis, which attracted world-wide notice because
of the psychological factor involved in the supposed manufacture of living substance.

The introduction of the dye janus green opened up a new and most valuable
method of approach by making the study of mitochondria, specifically stained in
living cells, so simjjle and satisfactory. It afforded an excellent basis for expe-
rimentation, dispelled any lurking doubt of the existence of mitochondria in the
living condition, and sujiplied a means of detecting the artifacts produced bj- the
older methods of fixation and staining.

The adaptation of methods of tissue culture by Champy (1912, p. 987; 19136,
p. 188) and the Lewises (1914, p. 330) to the study of mitochondria bids fair, if
properly controlled, to give information of a type which can be obtained in no
other way. It makes possible the continuous study with the microscope of vital
processes going on in the cell. The Lewises in particular have devised methods
by which they are able to keep selected mitochondria under observation for com-
paratively long periods of time and to see just what they do. Yet the method has
its obvious limitations. In order to be most effective it should be used in con-
junction with the methods of cell dissection recommended by Kite and Chambers.

And finally, the imjwrtance of the recognition of the sensitivity of mitochondria
to pathological change becomes quite apparent when we remember that hereto-

44

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

fore the activities of cells in the investigation of disease and in the study of patho-
logical processes have been gaged almost exclusively by the appearance of their
nuclei. Now we have at our disposal another criterion of cell activity and of
cell injury, the mitochondria, which we have already found to be of great and
surpassing delicacy, and which respond, even before the nucleus, to injurious
influences. Furthermore, this indicator is cytoplasmic, and as the cytoplasm is
more intimately related to the environment than the nucleus, its study may yield
very valuable information. It is an entirely different kind of indicator from the
nucleus and we may confidently look to it to disclose facts which would never have
been revealed by the study of the nucleus alone. That mitochondria are destined
to play an important and conspicuous role in medical research, from now on, is
quite apparent from Goetsch's (1916, p. 132) recent work on toxic adenomata of
the thyroid gland. It may be said, by way of explanation, that all the work on
goiter in time past has been vitiated, at the outset, by the fact that we lack a reli-
able criterion for the activity of the gland. It has come to be an embarrassing
question to ask one to pick out from a number of thyroids, on the basis of the
histological a])i)earance, the one which was associated with the clinical symptoms
of hyperthyroidism. The height of the epithelium, the appearance of the nuclei,
and the amount of colloid are, at best, but poor aids in the dilenmia. Now Goetsch
has discovered that, in the cases which he has observed, the mitochondria are
enormously increased in number where there are symptoms of hyperthyroidism.
In other words, he has succeeded in correlating the perplexing clinical sj^mptoms in
the little-known condition of exophthalmic goiter, or Basedow's disease, with a
definite anatomical change in the gland itself (/. e., in the mitochondria). Many
other conditions lend themselves to work along these lines and the outlook is
promising.

Table 1. — The Terminology of Mitochondria.

Term.

•Author.

Remarks.

Apparato reticolare interno

Golgi (1898, p. 64)

Wrongly confused with mitochondria.

(apxw. a ruler.) Material forming attraction sphere,
astral rays, and spindle fibers. Wrongly confused
with mitochondria.

Same as mitochondria (Heidenhain, 1900, p. 527).

Produced by the action of the fixative upon homo-
geneous basophilic material. They are not mito-
chondria (Bensley, 1911, p. 362).

Products of the transformation of mitochondria
(Policard, 1912c, p. 458) ; see, however, Champy
(1914, p. 380).

Probably same as apparato reticolare interno.

(i8ios, life, and fiXaaro!, a grain.) A heterogeneous
class of granulation, partly composed of mito-
chondria,

.Analogous with mitochondria according to Perroncifo
(1910, p. 309).

Possibly the same as the apparato reticolare interno.

By definition mitochondria . arising from the cary-
osome.

An "organ-forming" substance containing mito-
chondria among other things.

Possibly the same as the nebenkern.

Rod-like mitochondria.

.Solution of mitochondria.

Boveri (1888, p. 746)

Hermann (1891, p. 586)

.Solger (1896, p. 248) . . .

Archoplasnia-schleifen

Bixtonnets (StJibchen)

Heidenhain (1874, p. 47). . . .

Kopsch (1902, p. 9.34)

.\ltmann

Binnennetz

Bensley (19106, p. 179)

Arndt (1914, p. 55)

Conklin (1905. p. 218)

Caudal chvmoplasni

Meves (1907a, p. 401)

Ronieis (1912, p. 139)

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.
Table 1 — Continued.

45

Term.

Author.

Remarks.

Chondriom . . ,

Meves (1907o, p. 403)

Benda (1899a, p. 382)

Benda (1898, p. 397)

Champy (1913a, p. 157)....
Benda'

Cytoplasmic content of mitochondria.

Thread-like mitochondria.

A feltwork of mitochondria.

By definition plast-fonuing mitochondria.

Shaft-like mitochondria.

A generic term to include mitochondria of all forms.

Sphere-like mitochondria.

The arrangement of granular mitochondria in

threads.
"Organ-forming" material which gives rise to chorda

and nervous system ; it naturally contains mito-
chondria.
A basophilic, iron-containing substance, possibly of

nuclear origin and quite distinct from mitochondria.
By definition pigment-forming mitochondria.
"Organ-forming" material containing of course

mitochondria.
Probably in part the same as the apparato reticolare

interno.
Cytoplasmic microsomes as contrasted with nuclear

ones. The term obviously includes mitochondria.
Lifeless cytoplasmic constituents.
Produced by transformation of mitochondria

(Faur^-Fremiet, 1913, p. 520).
(1) Peripheral cytoplasm often devoid of visible

granulations; (2) an "organ-forming" substance

(Conklin, 1905, p. 218).
More fluid part of protoplasm (Conklin, 1917, p. 357),

contains mitochondria.
(1) Deeper protoplasm often said to be nutritive by

contrast (Renaut and Dubreuil, 19066, p. 229);

(2) an "organ-forming" substance (Conklin, 1905,

p.218).
Possibly similar to chromidial substance.
Introduced to indicate supposed relation of mito-
chondria to ergastoplasm; see, however, Regaud

and Mawas (1909a, p. 462; 19096, p. 229).
Includes mitochondria as well as other structures.
Occur in the germinative layer of the epidermis and

are, according to Favre and Regaud (1910, p. 1138),

true mitochondria.
Are in part mitochondria according to Perroncito

(1910, p. 308).
These are mitochondria (Biondi, 1915, p. 227);

nevertheless the term has been applied to many

structures which are certainly not mitochondria.
The clear homogeneous ground substance. It does

not include mitochondria which may be separated

out by centrifuging.
(1) Same as nebenkern (Lewis and Lewis, 1915, p.

349); see, however, Duesberg (1910, p. 653); (2)

same as nucleus of Balbiani (Sjovall, 1906, p. 577).

Should not be confused with "idiosorae, " the

hypothetical carrier of heredity.
The term has been applied to mitochondria as well as

to other granulations in muscle-cells.
A product of the transformation of mitochondria

according to Saguchi (1913. p. 177).
Erroneously introduced to designate nuclear origin of

mitochondria.
Any energy-producing material.
(1) Same as chondriomitom, Prenant (1899, p. 429);

(2) more solid part of protoplasm, Conklin (1917,

p. 357).
A general terra, but those described by him in sperm-
atogenesis are true mitochondria.
Relation to mitochondria uncertain.
Lifeless cytoplasmic material.
(^ixp6s, small, and ffu^a, a body.) A most general

term, including of cour.se mitochondria,
(/iiroi, a thread, and xioSpos, a grain.) Fadenkornern,

thread granules.

Chondriomitom .... ....

C'hondrioplastes

Benda'

Benda'

Chondriotaxie

Giglio-To3 and Granata
(1908, p. 14)

Conklin (1905, p. 218)

Goldschmidt (1904, p. 124)

Asvadourova (1913, p. 293)
Conhlin (1905, p. 218)

Cajal (1908. p. 123)

Strasburger (1882, p. 479)

Chromidial substance

Chymoplasm . .

Conduits de Golgi-Holmgren

Cytoraicrosomes

Deutoplasra (deuteroplasm)

Dyctyosomes

Perroncito (1910, p. 315)..

Butschli

EndoDlasm

Ergastidions

Laguesse (1911, p. 276)

Flemming

Fila

Herxheiiuer

Formations luxtanucleaires

Meves (1897, p. 315)

Koelliker

Eberth

Karv'ochondria

WUdman (1913. p. 428)

Strasburger

Hanstein

Mitochondria

Benda (1898, p. .397; 1899a,
p. 382).

In the discussion of a paper by Van der Stricht (1904. p. 145).

46

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.
Table 1 — Conliiiued.

Term.

Author.

Remarks.

Duesberg (1907, p. 284) ...

A general term, same as ehondriom (Duesberg,

1908, p. 261).

Mitogels

Koltzoff (1906, p. 468)

Large mitochondrial masses formed bv confluence.

Mitome

Fleraming (1882, p. 77)....

Same as hyaloplasm, according to Faure-Fremiet

(1912a, p. 408).

Mitosols

Koltzoff (1906, p. 468)

(Mitochondrosols) l)y definition, mitneliondrial

droplets.

Mitosoma

Platner (1889, p. 718)

Said to be the same as the nelienkern.

Mvochondria

Jordan and Ferguson (1916,

Introduced to flesignate mitochondria, wliich were

p. 94).

thought to give rise to myofibrils.

Myoplasm ".

Conklin (1905, p. 218)

The "organ-forming" material giving rise to muscle,
which of coiu'se contains mitochondria.

Nebenkern

Butschli

Same as micronuclei and geschlechtkerne (Hertwig,
1893, p. 212): probably a mitochondrial product.

Nebenkorper

Hermann

It is, according to Hermann, quite different from the
nebenkern: the relation to mitochondria is un-
certain.

Neniatobla-sten (Nematoj)lastpn) . .

Zimmermann (1893, p. 215)

Some descriptions under this term apparently relate
to mitochondria, while others do not.

Neurosome

Held (1895, p. .S9G)

Some of Held's neurosomes are mitochondria, others
are not (Cowdry, 1913a, p. 487).

Nissl substance

Nissl

Same as chromidial substance (chromophile and
tigroid substances, etc.), but quite distinct from

mitochondria (Cowdrj-, 1911, p. 753; 1913b,

p. 311).

Parabasal bodies

These, according to Alexeieff (1917b. p. 501), are
homologous with mitochondria.

Paramitome

Flemming (1882, p. 77).. . .

(1) Same as paraplasm (Faure-Fremiet, 1912a, p.
408) ; (2) the more fluid part of protoplasm (Conk-
lin, 1917, p. 357).

P^ricarvoneme (perineme)

Renaut and Dubreuil

Mitochondria as they occur in certain connective-

(19066, p. 230).

tissue cells.

Plasmosomes

Arnold (1907, p. 640; 1913.

Mitochondrial nature doubtful (Duesberg, 1912,

p. 433).

p. 823).

Plasomes

Wiesner

(irXoafia, a thing formed, and aufia, a body.) Ultimate
vital units not to be confused with plastosomes.

Plastidulen

Maggi

Probably in part mitochondria.

Meves (1910a, p. 150)

histogenesis.

Plastochondriomiten

Meves (1910a, p. 150)

Rows of granules.

Plastoconten

Meves (1910a, p. 150)

Long rods.

Plastolysis

Ciaccio (1913a, p. 725)

Ciaccio (1913a, p. 725)

Loss of mitochondria.

Breaking up of mitochondria into granules and

vesicles.
Cieneral term to include mitochondria of all shapes.

Plastorhexis

Plastosomen

Meves (1910a, p. 150)

Praplastorhexis

Ciaccio (1913a, p. 725)

Pathological changes in mitochondria preliminary to
plastorhexis.

Protoplasme supfrieure

Prenant (1899, p. 421; 1910,

A general term including archoplasm, kinoplasm, and

p. 219).

ergastoplasm ; does not relate to mitochondria.

Pseudochromosomes

Heidenhain (1900, p. 520) .

Structures thought by him to be identical with
mitochondria.

Holmgren (1899, p. 139:
1903, p. 9).

Similar to apparato reticolare interno in some re-
spects.

Spheroplastes

Same as mitochondria (Faure-Fremiet, 1907, p. 524).

Spiremes

Nelis (1899, p. 102)

Same as canalicular apparatus.

Cesaris-Demel (1907, p. 11)

Trophoplasni . .

Strasburger

(1) Nutritive cytoplasm (Renaut and Dubreuil,

1906b, p. 229) ; (2) more fluid cytoplasm (Conklin,

1917. p. 357) ; it contains mitochondria among other
things.
Bv definition, juice canals which penetrate cell from

Trophospongium

Holmgren (1903, p. 9)

the exterior, evidently not synonymous with

apparato reticolare interno and quite distinct from

mitochondria.

Vegetative filaments . . .

Altmann

Certain types of bioblasts occurring in gland-cells and

identical with mitochondria.

Vermicules

Laguesse (1900, p. 5)

Certain filamentous mitochondria present in gland cells.

Vibrioides

Swingle (1898, p. 110)

Relation to mitochondria doubtful.

Same as Binnennetz (Sjovall, 1906, p. 561).

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. 47

II. NOMENCLATURE.

The terminology of mitochondria is unnecessarily comj)licated and confusing.
The confusion has resulted from incoordination and from hasty individual action
in elaborating new names, often only to discard them in a new paper in favor
of some other. Pome have attempted to convey information with respect to the
morphology of mitochondria, others with regard to their physiology, and still others
with respect to their colloidal chemistry. There has been no attempt to come
together in a friendly spirit and arrive at some agreement or compromise. It is to
be deplored that c^'tologA' should be so far behind gross anatomy in possessing no
official list of terms like those of the Basle Anatomical Nomenclature. That the
need for this is great may be seen by reference to table 1 (pp. 44-46).

The term "mitochondria" originated with Benda (1899a, p. 397), who introduced
it to designate certain " Fadenkornern " which he had been studj'ing in spermato-
genesis. It is derived from the Greek niros, a thread, and xovdpos, a grain.

It was soon discovered that mitochondria do not always occur in the form of
thread-granules and investigators grasped at the idea that they should be named
on the basis of their morphology. It was felt that the word "mitochondria"
should be apphed only in the sense in which Benda originally used it. Accordingly
Meves (1907a, p. 401) devised the term " Chondriokont " to describe the rod-like
forms and Benda, ^ almost simultaneously, came out with a whole list of terms
with which to describe the various forms of mitochondria: " Chondriomiten "
(threads), " Chondriorhabden " (shaft-like forms), " Chondriospharen " (spheres),
and "Chondriom" or " Mitochondrion! " (the cj^toplasmic content of mitochondria
of whatever form) .

The words " chondriorhabden " and "chondriospharen" were not favorably
received and were soon forgotten and investigators gradually drifted into the use
of the following nomenclature for mitochondria based purel}^ on morphology:
" chondriosomes " (a generic term to include all forms); " mitochondries " (granules);
" chondriocontes " (straight or curved threads); " chondriomites " (filaments of
granules); and "chondriome" (the cytoplasmic content of chondriosomes). But,
since all these forms grade by imperceptible transitions into each other, there was
much confusion, and it is a difficult matter to find two investigators who are in
entire agreement on the question of nomenclature.

As I have pointed out elsewhere (Cowdry, 1916a, p. 424), this system of
terminology based on morphology is entireh' superfluous in the light of recent w'ork.
We are coming to realize that the fundamental thing is the nature of the material
rather than the form which it assumes (see p. 66). The Lewises (1915, p. 353),
in the Uving cells of tissue cultures, were actually able to observe that mitochondria
are continually changing in shape by bending in various directions, by shortening
and thickening, by elongating, and by thinning, etc. They saw rods and tln-eads
change into granules, threads fuse to form networks, and many other alterations
in the morphology of mitochondria. In other words, the same material under
different conditions assumed different forms, as one would naturally expect.

'In the discussion of a paper by Van der Stricht (1904, p. 145).

48

THE MITOCHONDRIAL CONSTITUE>fTS OF PROTOPLASM.

Now, the advocates of the above-mentioned terminology beUeve that it is of
use because they think it convenient to have a special word to designate granular,
rod-like, and filamentous forms, and rows of granules; but this system of terminology
is not only cumbersome, confusing, and arbitrary, but it is also inadequate. That
it is cumbersome needs no explanation. It is confusing because individual inves-
tigators do not all use it in the same way (see table 2). Indeed, it is not an
easy matter to find two workers in agreement about it. It is arbitrary because it
leads one to make sharp and clear-cut distinctions between dilTerent forms of mito-
chondria where none exist. Instead of saying, for instance, that the mitochondria
in a cell are characterized by their great diversity of form, we must remark that
chondriosomes of this particular cell occur in the form of mitochondria, chondrio-
contes, and chondriomites, and we are obliged to speak of the cellular content of
chondriosomes as the chondriome. It is also inadequate because branching mito-
chondria— networks and circles and bleb-like sweUings and many other varieties —
also occur which are not provided for. Let us hope that no one will coin special
names for them.

Table 2.

Term.

Dubreuil
(1912, p. 127).

G. Arnold
(19125, p. 255).

Benda
(1914, p. 21).

Duesberg
(1917, p. 469).

Mitochondria

Chondriooontes

Chondriomites

Chondriosomes

Granules

General term

Granules.
Filaments.
Not used.
General term.

Not used.

Rods

Filaments

Not used

Filaments of grains

More or less voluminous
body, larger than pre-
ceding, of irregular form.

Homogeneous filament
General term

Cellular content

Chondriosome (p. 253) .

Cellular content

But the most pernicious systems of nomenclature arose several years later,
when investigators became deluded into thinking that they knew something about
the functional significance of mitochondria.

Koltzoff (1906, p. 468) has called the mitochondrial droplets " mitochondrosols"
(or, briefly, "mitosols"), and the larger masses which he believes they form by
confluence in the course of spermatogenesis he terms "mitogels."

The plastochondrial nomenclature of Meves is important in this connection.
It is based upon his theory (Meves, 1908, p. 845) that, with the specialization of
the embryo into different organs and tissues, primitively similar cells assume special
functions which find expression in characteristic structures or differentiations. All
these products, no matter how heterogeneous they may be, arise through the meta-
morphosis of one and the same elementary plasma constituent, the chondriosomes.
He writes, subsequently (1910a, p. 150), that when we consider the important role
that chondriosomes play in histogenesis we can speak (instead of chondriosomes)
of " Plastosomen " ("Plastochondrien," " Plastochondriomiten " or short chondrio-
mites, and "Plastoconten"). This terminology was accepted by all except those
who oppose, or are skeptical, of the theoretical considerations which inspired it.

Duesberg (1912, p. 598) first accepted this system of nornenclature and adopted
it, but later (1915, p. 35) rejected it for several reasons, but particularly in order to

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. 49

avoid any confusion between the term " plastosomes " and the " plasmosomes "
of Arnold. It may also be remarked that there is a possibility of confusing
the term " plasto.some " with the "plasmasonie," or plasma-staining nucleolus
as contrasted with the karyosome, as well as the "plasomes," or elementary vital
units of Weisner, and the word "chondriome" with "chondroma," a cartilaginous
tumor.

Still another series of names has been advocated, on the basis of. the plast-like
function of mitochondria ; for it is thought that the mitochondria are plast-f ormers
in plants and pigment-formers in animals. This view, so far as the plants are con-
cerned, has a good mass of evidence in support of it. We owe the term "chondrio-
plaste" to Champy (1913o, p. 157) and the term " chromochondries " to Prenant
(Asvadourova, 1913, p. 293).

Reference may be made to the term "karyochondria," which Wildman (1913,
p. 428) introduced to describe certain mitochondria which he believed to arise from
the nucleus in spermatogenesis in Ascaris. Arndt (1914, p. 55) for the same
reason devised the term "Caryosomochondrien."

Since it is difficult to bear in mind all these terms and their various shades of
meaning, we must make a more or less arbitrary selection of the one which appears
to be the least objectionable, for it would obviously be highly inadvisable to follow
the precedent and invent a new term. Priority is almost impossible to establish.
I have accordingly .selected the word "mitochondria" for the following reasons:
(1) Its introduction by Benda in 1899a (p. 397) marks the beginning of much of
the recent work on the subject; (2) ever since it has been, with but few exceptions,
in quite general use, especially in this country; (3) it is simple and does not commit
the user to any interpretation whatsoever.

Duesberg (1917, p. 469) opposes this position, as stated in my paper (1916a
p. 424), as follows:

"To this I take the liberty of making the following remarks: The term 'mito-
chondria,' in Cowdry's sense, is not in general use even in this country; if the nomenclature
must be based on morphology, one should, to be logical, reject 'mitochondria' as a general
term, for ' mitochondrium ' means granule and can not in consequence mean filament.
The analog>^ with the term 'cell' is not adequate, for, though this word is not appropriate,
it is, however, generally adopted; and it should not be forgotten that the invitation to use
a term in a different sense from its original meaning comes in this case precisely from
those who want to use 'mitochondria' as a general expression."

In answer I venture to point out :

(1) In 47 out of the 55 or more papers published in this country during the last five
years (1912-1917) the term "mitochondria" has been used in the general sense, that is to
say, in 85 per cent.

(2) The word "mitochondria" was introduced by Benda (1898, p. 397) and is derived
from the Greek filros, a thread, and x^«'5pos, a grain. In order to interpret it as granules
one must ignore the "mito." That mitochondria are not simply granules (as Duesberg
asserts) is also made perfectly clear by Benda himself (1899a, p. 382), who refers to them
continually as " Fade nkornern." The term "mitochondrium" is the Latin singular of the
Greek compound word. The correct Greek singular is mitochondrion.^

'I am indebted for this information to Professor Miller, of the Greek department, Johns Hopkins University.

50 THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

(3) The objection to the analogy with the term "cell" is not valid, since the word
"mitochondria" has been generally adopted in Europe as well as in this country.

(4) The invitation to use the "term in a different sense than its original meaning"
comes in this case from Duesberg if he says that " mitochondrium " means "granule," for
the reason aforesaid.

That the word "mitochondria" has been repeatedly used as a general expression,
both intentionally and unintentionally, is shown by the following citations :
Benda (1914, p. 20) states:

"In gelungenen Priiparaten nach meiner sowohl wie anderen Methoden erkennt
man innerhalb der Zellen die Mitochondrien in sehr verschiedener Form, Menge und
Anordnung. Als Grundforni betrachte ich die runde Gestalt, die in alien Zellarten gele-
gentlich vorkommt, in den Geschlechtszellen und den ersten Furchungszellen aber bei
weitem vorherrscht. An Stelle der runden Korner findet man auch langliche Kornchen
oder kiirzeste Stiibchen, die alle als gleichwertig mit den Rundkornern zu betrachten
sind. Besonders bei spiirlicher Kornermenge finden sich die Korner bisweilen unregel-
miissig im Zelleib verteilt und auch bei sehr grossen Anhiiufungen ist naturgemiiss eine
Anordnung schwer zo erkennen. Sonst ist aber das Charakteristische, was in ihrem
Namen von mir ausgedrlickt werden sollte, ihre Neigung zur fadeufcirmigen oder reihen-
formigen .\nordnung, wo sie wie Streptokokken erscheinen. Man findet sie ferner in mehr
homogenen gewimdenen Filden, die ge\V(")hnlich unregelmiissig segmentiert erscheinen
und weiter als starre homogene oder segmentierte Stiibchen, endlich als grossere homogene
Kugeln."

Favre and Regaud (1910, p. 1137) state:

"Les mitochondries (Benda) sont des organites du protoplasma qui, au contraire de
Tcrgastoplasma rare et contingent, ont ime existence absolument gencrale dans les cellules.
Ces organites se presentent tantot sous la forme de granulations non ordonnees, tantot
sous la forme de granulations alignees en series (on les appelle alors chondriomites) , tantot
sous la forme de filaments continus (on les appelle alors chondriocontes) , tantot enfin (mais
plus rarement) sous forme de corps volumineux de forme quelconque."

Duesberg (1907, p. 284) says:

"Unter obigem Titel sollen einige Mitochondrienstudien veroffentlicht werden. Ich
have mit Absicht diesen allgemeinen Ausdruck ' Mitoch()ndrialapi)arat' gewiihlt, well
(lieser iiber die Form dieser Gebilde keinen Vorbehalt einschliesst."

I am merely conservative in this and advocate nothing new. It may be
objected that the old word "mitochondria" does not adequately describe the
granulations as we now know them. This, to my mind, is only a very encouraging
indication of the healthy condition of our science. Herein lies the key to the whole
situation, for investigators have attempted to make the terminology keep pace
with the discoveries, by making it convey information of all kinds. If a similar
policy had been pursued with respect to the nucleus, or the cell itself, the results
would have been even more disastrous. The ([uestion of nomenclature is but a
very small part of the real problem.

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

51

III. LITERATURE.

VARIETIES OF CELLS.

For convenience I have reviewed the descriptive work on mitochondria in the
different types of cells of the body (with special reference to the condition in man,
given in italics) as well as in the tissues of liigher plants. I have usually noted the
most recent account, as well as the original reference, because experience has shown
that the most up-to-date descriptions are often of far greater value than the old
(and often vague) original records. Special care has been taken with the original
descriptions because they are very difficult to trace accurately by reason of the
fact that in them the mitochondria are invariably alluded to in a vague way under
misleading terms (for which see p. 44). Such a summary serves as a kind of
balance-sheet, and a glance at it will be sufficient to reveal the most conspicuous
gaps which still persist in our knowledge of mitochondria and which might otherwise
easily be overlooked (see also p. 144).

ANIMALS.

Eintliclial tissue:

Epidciiiiis, Favre and Rcfjaud (1910, p. 1138).

Hair follicles, Branca (1911, p. 5.59).

Meibomian gland-s, Altmann (1894).

Sebaceous gland.s, Altmann (1894); Nicolas: Regaiul,

and Favre {1912a, p. 201).
Sweat glands, Nicolas, Regaud, and Fawc {1012b, p. 191).

Ceruminous glands, .

Mammary glands, Altmann (1894, p. 159); Hoven
(1911, p. 321).

Glands of Moll, .

Lachrymal glands, Altmann (1894, p. 1.59); Sundwall

(i916, p. 202).
Oil glands of birds, Altmann (1894, p. 159).

Feathers, .

Harder's gland, Altmann (1894, p. 159).
Leydig's " Schleimzellen," Meves (19076, p. .565).
Teeth, Manca (1913, p. 121).
Alimentary tract:

Hypobranchial gland, Grynfeltt (1912a, p. 12).
Parotid gland, Altmann (1894, p. 159).

Ebner's gland, .

Submaxillary gland, Regaud and Mawas {1909b, p. 226).

Blandin's gland, .

Tonsil, Alagna (1911, p. 27).

Labial (buccal, molar) glands, .

Retrolingual gland, .

Esophagus, KoUmann and Papin (1914, p. 222).
Principal and accessory kibial glands, Fauj(5-Fremiet

(1910c, p. 3).
Stomach :

Parietal cells, Eklof (1914, p. 224). See, however,
Regaud (1908a, p. IS).

Chief cells, Altmann (1894, p. 157); Eklof (1914,
p. 22.3).

Goblet cells, Eklof (1914, p. 227).

Brunner's glands, Eklof (1914, p. 225).

Paneth cells, Eklof (1914, p. 225).
Pancreas:

Acinas cells, Altmann (1894, p. 1.58).

Islet cells, Bensley (1911, p. 368).
Hepato-pancreas, Guieysse-Klissier (1910, p. IS).

Alimentary tract — continued.
Liver, Altmann (1894, p. 155).

Bile ducts, PoUcard (1914, p. 023).
Respiratory tract:

Bowman's glands, .

Trachea and lung, Prenant (19116, p. 337); Meves and
Tsukaguchi (1914, p. 289).
Urino-genital tract :
Kidney (metanephros), Heidenhain (1874, p. 47);
Benda (18996, p. 331); Regaud (1908o, p. 15);

Policard (1915, p. 539).
Pronephros and mesoneplu-os
(1913/, p. 131).

Bladder, .

Urethra, .

Glands of Littr6, .

Spermatogenesis, v. Brunn
heading of "Korner."

in amphibia, Luna

(1884, p. 132) under

Interstitial cells, Benda (18996, p. 351).
Epididymis. Fuclis (1901, p. 328).

Cowper's glands, .

Seminal vesicles, Akatsu (1903, l). 566).

Prostate, Akatsu (1903, p. 566); Dominici {1913

p. 295).
Prepucial glands (glands of Tyson), Altmann

(1890, p. 99).

Corpus cavernosum, .

Corpus spongiosum, .

Ovary, Zoja brothers (1891, p. 259), under head of

"Phistidulen."
Insterstitial cells of ovary, d'Athias (1912, p. 448).
Egg cells. Van der Strichi {1905, p. 7).
Lutein cells, Levi (1913, p. 526); Corner (1914, p.

76).
Fallopian-tube glands, Altmann (1894, p. 159).
Uterus, Romeis (1913a, p. 9).

Vagina, .

Clitoris glands, Altmann (1890, p. 99).

Bartholin's glands, .

Placenta, cells of Langhans, Van Cauwenberghe,

confirmed by de Kervily (1916, p. 589).
Coccygeal gland. Altmann (1S90, p. 145).

52

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

AN IMALS— Continued.

Connective tissue:

Fibroblasts, Dubreuil (1913, p. 95).

Clasmatocytes, Dubreuil (1913, p. 87).

Plasma cells, Schridde (1005, p. 729); Dubreuil (1909,

p. 80); and Dubreuil and Fairre (WHb, p. 2J,).

Mast cells (basophile cells), mitochondria arc ab.sent

according to Galeotti (1895, fig. 18). (I have also

found that they are absent in mast ceils occurring

in the thymus, but present in those in bone marrow.)

Cartilage cells, Flemming (1882, p. 22); Dubreuil

(1911a, p. 1S9); Pensa (19136, p. 557).
Bone cells, Dubreuil (19106, p. 1101).
Osteoblasts, Dubreuil (19106, p. 1101; 1913, j). 119).
Osteoblasts (myeloplaxcs, giant multinucleated cells
etc.), cells of Bizzozero, Dubreuil (1910a, p. 72;
1913, p. 132).
Odontoblasts, Prenant (19116, p. 335).
Fat cells, Metzner (1890, p. 82); Dubreuil (1913, p. 98).
Muscular tissue:

Striated, Benda (1899a, j). 379); Regaud and Favre

(1909, p. 298).
Smooth, Benda (1899o, p. 379).
Cardiac, Regaud (19096, p. 426).
Purkinje, Mironesco (1912, p. 30).
Nervous tissue:

Neuroglia cells, Nageotte (1909, p. 826), Collin (1913a,
p. 179).

Sheath cells, .

Nerve cells, Altmann (1890, p. 52) ; Cou'dry (1914a, p.ll).
Ear, organ of Corti, Van der Stricht (1918, p. 6:i).
Eye:

Retina, Leboucq (1909, p. 576); Collin (19136,

p. 1358).
Lens, Maggiore (1912, p. 118).
Cornea, Prenant (19116, p. 333).
Nose:

Nasal mucous membrane, .

Olfactory bulbs, .

Organ of Jacobson, .

Motor end plate, .

Meissner's corpuscle, .

Pacinian corpuscles, .

Crandy's corpuscle, .

Herbst's corpuscle, .

Taste buds, .

Endothelium:

Artery, mitochondria frequently described incidentally.
Vein, mitochondria frequently described incidentally.
Lj'mphatic, ,

Mesothelium:
Peritoneum, -
Pericardium,
Pleura,

Syno^aal membrane, .

Plexus chorioideas, Hworostuchin (1911, p. 232).
Blood:

Erythrocyte, .

Young erythrocytes, Meves (1907(). p. -102).
Normoblasts, Meves (1907o, p. 402).
ErythroblMt, Meves (191 In, p. 495).
Lymphocyte, Schridde (1905, fig. 69).
Transitionals, Naegeli (1912, p. 175).
Lymphoblast, Schridde (1907, p. 231).
Polymorphonuclear neutrophile, Benda (1899a, p.

' 380): Coirdry (1914b, p. 27S).
Polymorphonuclear eosinophile, Benda (1899a, p.

38(J); Coirdry (191. ih, p. 279).

Polymorphonuclear basojihile, .

Neutrophile myelocyte, .

Eosinophile, .

Basophile, Dubreuil (191 In, p. 138).

Myeloblast, Schridde (1907, p. 231); see, however,

Klein (1910, p. 407).
Megakaryocytes, Prenant (1911b, p. .335); see, also,

Dubreuil (1910a, p. 73).
Polvkarvocytes, Dubreuil (1910a, p. 72).
Platelets, Cowdry (1914b, p. 275).
Macrophages, Alagna (1911, p. 37).
Ductless glands:
Thymus:

Small cells and epithelial cells, Salkind.

Hassal's corpiLScles, .

Thyroid, Galeotti, under the name of "fuchsinophile

granules," see Bensley (1916, p. 41).
Suprarenal gland, Mulon (1910a, p. 103), da Costa

(1916, p. 164).
Hypophysis, vmder heading of "Altmann's granules,"

Prenant, Bouin, and Maillard (1906, p. 52).

Epiphvsis, .

Parathyroid, Bobeau (1911, p. 183).

Carotid body, .

Tissue cultures:

Kidnev. Champv (1914, p. 312).

Thyroid, Champy (1912, p. 987; 19i:i6, p. 192; 1915,

p. 63).
Embrvonic tissues, Lewises (1914, p. 330); Levi

(19160, p. 101); Maximow (19166, p. 465).

PLANTS.

Reproduct ive system :
Thallophytes :

Unicellular forms: Saccliaroniyces. Endomyces.

Guilliermond (19146. fig. 10)'.
Conidia conidiophores : Penicillium, Guilliermond

(19146, fig. 10); Albugo, Lewitsky (1913, pi. 21).
Oospore oogonium, oogonium anlage.; Albugo,

Lewitsky (1913, pi. 21).
Zygospore, gametangia, suspensors: Sporodinea,

Moreau (191.5o, p. 171).
Ascospore, asci, ascogenoas hyphff, paraphyses;

Pustularia, Peziza, Aleuria, GuilUermon<l (19146,

figs. 12-15; 19i:5a. fig. A; 191:56, figs. ll-i:S).
Basidiospore, basidia, mother cells of probasides,

livmenial layer; Psalliota, Puccinia, Beauverie

(i9Ua, p. 798; 19146, p. 359).
Tdeutospore, Puccinia, Beauverie (19146, |). 3.59);

Phr.agmidium, Mme. Moreau (1914, p. 421).

Coleosporium, Mme. Moreau
Phi-agmidium, Mme. Moreau

Reproductive system — continued.
Thallophytes — continued.
Uredospore,

p. 421).
.^Jcidospore,

p. 421).
Bryophytes and pteridophytes :

Spore, sporogenous-t issue cells. tai)etum,

mother cells; Funaria, Sapehin (1915, p

Anthoceras, Scherrer (1913, pi. 20).
Sperm, spermatogenous-t issue cells, sperm moth<'r-

cells, antheridium mother-cells; Polytrichium,

Funaria, Bryum, Marchantia, Sapehin (1915,

pis. 22-26): .\nthoceras, Sclierrer (1913, pi. 20).
Egg, ventral-canal cell, ncck-cclls of archegonium;

Funaria, Bryum, Sapehin (1915, pi. 26); .\ntho-

cer.as, Scherrer (1914, pi. 20).

(1914,

(1914,

pore-

22);

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

53

PLXNTS— Continued.

Reproductive system — contimied.
Spermatoph>'tes :

Pollen, pollen mother-cells, tapetum; Nymphsea,
Meves (1904, p. 284); AsparagiLs, Lewitsky
(1910, pi. 17); Lilium, Cucurbita, Guilliermond
(1912a, fig. 6; 1914h, fig. 6).
Embryo sac, ovule, antipodaLs, synergids, embryo
tegument, nutritive cells; Lilium, Guilliermond
(19120, figs. 4, 5; 19146, fig. 13); Pensa (1911,
fig. 2).
Seed, barley, Ricinus, Guilliermond (1912o, fig. 7).
Vegetative system:

Filaments of alga> and fungi; \'aucheria, Achlya,
Rudolph (1912, pi. IS); Penicillium, Endomyces,
Guilliermond (1914b, fig. 10).
Apical cells of alga", Cystoseira, Nicolosi-Roncati

(19126, p. 144); Fucus, Le Touze (1912, pi. 9).
Primary meristem of root, dermatogen, poriblem,
plerome, root-cap; Asparagus, Lewitsky (1910,
pi. 17), Trianea, Sapehin (1915, pi. 11).
Primary merLstem of stem-tip, buds, dennatogen,
periblem, plerome; Elodea, Euphorbia, Sapehin
(1915, pLs. 10-11); Asparagus, Rudolph (1912, pi.

Vegetative system — continued.

18); Lewitskv (1910, pi. 17); Tropa>olum, Guillier-
mond (19120^ fig. 2); Tulip, Pensa (1911, figs. 6, 7);

Secondary meristem, cambium, Guilliermond (1912(1,
p. 3.58).

Epidermis of stem, root, leaf, flower, fruit ; Aspara-
gus, Trianea, Sapehin (1915, pis. 10, 11); Cheno-
podium, Juglans, Iris, Tulij). Cli\ia, Tropa?oluni,
Ciuilliermond (1917, p. 233; 19146, fig. 14).

Mesophyll cells, palisade cells, leaves, carpels;
Nereum, "haricot," Guilliermond (19146, fig. 11,
12; 1912a, pi. 18).

Leaf cells, stem-tip of mosses, Polytrichium, Sapehin
(1913, pi. 14).

Leaf cells of ferns; Aspidium, Scolopendrium,
Pensa (1911, figs. 3, 4, 5).

Protonenia cells of mosses, Funaria, Sapehin
(1913, pi. 14).

Secreting cells, Juglans, Guilliermond (1914n,
p. 566).

Hair cells. Cucurbita, Maximow (1913, p. 243).

Giant cells of Heterodera galls; Vitis, Ngmcc
(1910, p. 171).

DIFFERENT TYPES OF ORGANISMS.

It is often of interest also to know in just what forms, both animal and vegetable,
mitochondria have been described and by whom. I have prepared the following
summary with this end in view. The work of the older authors is very difficult to
interpret, especially that of Retzius, for he has described structures derived from
mitochondria in mature spermatozoa of hundreds of different forms. These bodies
do not possess the staining reactions of mitochondria, though they are formed
from them ; they are not soluble in acetic acid and may be demonstrated by ordinary
methods of technique. Strictly speaking, they are assuredly not mitochondria, but
since they develop from mitochondria (see page 101), it is difficult to say when they
cease to be mitochondria. For a description of them I venture to refer to the
excellent account of Duesberg (1912, page 615).

ANIMALS.

Protozoa :
Rhizopoda :

Amoeba gorgonia (?), Faure-Fremiet (1910a, p. .507).

Cochhopodium pellucidum, Faure-Fremiet (1910«,
p. 508).

Actinophrys sol, Faure-Fremiet (1910a, p. 509).

Amoeba proteus, Vonwiller (1915, p. 485).

Amoeba limax, L. and R. Zoja (1891, p. 242).

Amoeba chondrophora, Arndt (1914, p. 51).
Mastigophora:

Cryptomonas , Faure-Fremiet (1910a, p.

510).

Monadi, L. and R. Zoja (1891, p. 242).

Chilomonas paramoecium, Prowazek, according to
Faur^-Fremiet (1910a, p. 510).

Trypanosoma lewisi, Shipley (1916, p. 444).

Blastocystis enterocola, Trichomonas angusta,
Alexeieff (19166, p. 1074).

Giardia cunicula, anadenalis, miuis, and agilis.
Trichomonas batrachiorum and muris, Hexamastix
termitis, Octomastix jiarva and motelUe, Octo-
mitus intestinalis, Trichonympha agilis, Entri-
chomastix motellae, and Tetramastix, Alexeieff
(19176, p. 499).

Protozoa — continued.
Sporozoa :

Monocystis a.scidi;e, Hirschler (1914, p. 304).
Hsemogregarina sergentium, Viguier and Weber

(1912, p. 92).
Infusoria:

Vorticella, under heading of "Spheroblasts," by

Czermak, according to P'aure-Fremiet (1910a,

p. 478).
Vorticella convallaria, Faure-Fremiet (1910a, p. 523).
Epistylis flavicans, under heading of "Spheroblasts,"

by Greef, according to Faure-Fremiet (1910o,

p. 478).
Toxodes rostrum, Faur6-Frcmiet (1910a, p. 495).
Stentor, Nassula, Urostyla grand is, Faure-Fremiet

(1910a, p. 496).
Glaucoma piriformis, Faur6-Fremiet (1910o, p. 501).
Trichodinopsis paradoxa, CycIo.stoma elegans, Faur6-

Fremiet (1910a, p. .502).
Carchesium jjolypinum, .Spiro.stomum ambiguum,

Faur<:'-Fremiet (1910a, p. 504).

Campanella — . , Faur6-Fremiet (1910a, p. 506).

Parama'cium caudatum, Faur^-Fremiet (1910a,

p. 511).

54

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

ANI MALS— Con((>i.He(2.

Prutozoa — continued.
Infusoria — continued.

Trachelius ovum, Faure-Frcuuct (191()a, p. 512).

Strobilidium gyrans, Faure-Fremiet (1910a, p. 514).

Cjttarrocyclis Ehrenbergii, under heading of
"Chromidia," Gesa Entz, jr., according to
Faiu-e-Frcmiet (1910a, p. 514).

Tintinnidium inguilinuni, Faurc-Frcmict (1910a,
p. 514).

Opisthonecta lienncguyi, Faurc-Fremiet (1910o,
p. 516).

Trichodinopsis jjaradoxa, Faure-Fremiet (1910a,
p. 517).

Opercularia racemosa, Faurc-Fremiet (1910a, p. 519).

Oi)ercidarianotonecta;, Faure-Fremiet (1910a, p. 520).

Noctiluca miliaris, Faure-Fremiet (1910a, p. 524).

Opalina ranarum, Colopoda cucuUus, Stentor cu-
cuUiLS and polymorphus, Isotherica prostoma,
Entodinium bvu'sa and minimus, Diplodinium
maggii and cattanei, Butschlia parva, L. and R.
Zoja (1891, i)p. 243, 244).

Blantidium enterozoon, Benda (1899?;, p. 168).

Fabria salina, Condylostoma patens, Epiclintes am-
biguus, Strombidium sulcatum, Strombidium
stylifer, Faurc-Fremiet (1912a, p. 411).
Porifcra:

Spongilla fluviatilis, L. and R. Zoja (1891, p. 244).
Coelenterata :
Hydrozoa :

Hydractinia echinita, eggs, Beckwith (1914, p. 204).

Halistemma rubrum, Forskalia contorta, Phy-
sophora hydrostatica, Gleba hippopus, Praya
maxima, spermatogenesis, under heading of
" Cvtomicrosomes," Pictet, according to Faur6-
Fre'miet (1910a, p. 571).

Hydra vulgaris, L. and R. Zoja (1891, p. 244).
Scyphozoa :

Aurelia aurita, Tsukaguchi (1914, p. 117).
Actinizoa: .

Ctenophora:

Platyhelminthes :
TurbeUaria:

Plagiostoma girardi, spermatids, Weygandt (1907,

p. 271).
Planaria adherens, almost all cells of the body,

Korotneff (1909, p. 1010).
Trematoda:
Dicrocoelium lanceatum, spermatogenesis. Dingier

(1910, p. 698).
Cestoda: .

Nemertinea:

Nemathelminthes :
Nematoda:

Ascaris lumbricoides, male and female germ-cells,

Hirschler (1913, p. 353).
Ascaris megalocephala, gonocytes, Faurg-Fremiet
(1912b, p. 346) ;L. and R. Zoja (1891, p. 245).
Filaria papillosa, Meves (1915b, p. 38).

Acanthocephala: .

Chfftognatha: • .

Trochelminthes:

Rotifera: .

Dinophilea: .

Gastrotricha : .

MoUiiscoida:
Polyzoa:

Membranifera pilosa L., spermatogenesis, Bounevie
(1907, p. 578).

Molluscoida — conlinucd.

Phoronitta: .

Braehiopoda: .

Echinodermata:
Asteroidea:

Asterias glacialis, oocyte, Schaxel (1911, p. 346).
Asteracanthion tenuispinis, L. and R. Zoja (1891,
p. 250).

()l)hhu-oidea: .

Ecliinoidea :

Strongylocentrot iLS lividus, Ai-bacia pustulosa, Echi-
nas microtiiberculatus, Spha^rechinus granula-
ns, Spatangus purpureus, under heading of
"Cytomicrosomes," in spermatogenesis, Pictet,
according to Faure-Fremiet (1910a, p. 571).
Holothuroidea : Holotlnu-ia tubulosa and polii, oocytes,

Schaxel (1911, pp. 344, 347).
Crinoidea: Comatula mediterranea, L. and R. Zoja
(1891, p. 250).
Annulata:
Chffitopoda:

Nereis virens, nerve-cells, Cowdry (1914a, p. 11).
Serpula uncinata, Nais proboscidie, L. and R. Zoja

(1891, p. 248).
Tapinambis tegnixin, kidney, Mayer and Rathery

(1909, p. 335).
Lumbricus herculeus and tcrrestris. AUobophora

terrestris, kidney, Maziarski (1903).
Aricia fcetida, egg segmentation, Schaxel (1912, p.
402).

Myzostoinida: .

Gephyrea :

Archi-annelida: .

Hirudinea:

Hirudo medicinalis, L. and R. Zoja (1891, p. 248).

Arthropoda:
Crustacea:

Astacus fluviatilis, spermatogonia, and spermato-
cytes, Prowazek (1902, p. 448); L. and R. Zoja
(1891, p. 253).

Homaras \Tilgaris, egg-cells, under heading of
"Pseudo-chromosomen," F. Henschen (1904,
p. 16).

Leander adspersus, spermatogenesis, SpitschakofT
(1909, p. 26).

Menippe mercenaria, spermatogenesis, Binford
(1913, p. 156).

Gammarus pulex, spermatid, Koster (1910. p. 491).

Callinectes hastatus. Cancer boreaUs, Homarus
americanus, nerve-cells, Cowdry (1914a, p. 11).

Eupagurus prideau.xii and Eriphia spinifrons, under
heading of "Microsomes," by Grobben, accord-
ing to Faure-Fremiet (1910a, p. 551).

Maia squamado, Carcinus mcEnas, and Eupagurus
bernhardus, spermatogenesis, Labbc (1904,
p. xi).

Scyllarus arctus, Portunus corrugatus, Pagiu-us stri-
atus, and Galathea ■ squamif era, spermatogene-
sis, KoltzofT (1906, p. 565).

Galathea strigosa, Palinui'us vulgaiis, and Corrugiis
fortunatus, hepato-pancreas, Guieysse-Pelessier
(1910, p. 18).

Pandarus sinuatus, spennatogenesis, McClendon
(1910, p. 234).

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

55

ANIMALS — Continued.

On3i'luii)lioia:
Myriopoda:

Pachyiulus varius, spermatoeytes, Octf iiifier (]9()!»,

p. 593).
Julus terrestiis, eggs, Faiu-c-Fremiet (1908, p. 1057).
Lithobius forticatus and scolopendra, spermatogenpsis,

Faiire-Frpiniet (1910o, p. 582).
Geophilus longieornis and carpopliagus, Polyxenus
lagurus, Polyzoniiim geimanicum, Blaniulns gut-
tulatus, oocytes, Faure-Fieniiet (1910a, p. 592),
Arachnida:

Limulus polyphemus, eggs, Munson (1898, p. 111).
Argas miniatus, spermatogenesis, Casteel (1917, p.

650).
Ixodes reduvius, spermatocytes, Nordenskiold (1909,

p. 514).
Tegenaria domestica, egg-celLs, under heading of Nu-
cleus of Balbian, Van der Stricht (1898, p. 135);
L, and R. Zoja (1891, p. 253).
Euscorpius carpathicus L., Buthus eupeus, spermato-
genesis, Sokolow (1913, p. 398).
"Arizona scorpion," Wilson (1916, p. ,539).
Insecta:

Forficula auricularia, Valette St. George (1887), Blatta
gennanica, Periplaneta, spermatids, Ik-nnc. uy
(1904).
PezotettLx pedestris, Acridium SBgyptium, Psophus stri-
dulus, Decticus verrucosus, Locusta viridissima,
Buchner (1909, p. .361).
Hydrophilus piceus, L. and R. Zoja (1891, p. 2'4).
Pamphagas marnioratus, spermatogenesis, C!i?lio-Tos

and Granata (1908, p. 115).
Stenobothrus biguttulus, spermatogenesis, Gerard

(1909, p. .599).
Locusta viridissima, Gryllus campestris, spermato-
genesis, Henneguy (1904).
Gryllas assimilis, spermatid, Baumpartner (1002

p. 50).
Gryllus campestris, primary spermatocytes, Henneguy

(1904,).
Gryllotalpa vulgaris, sperm.T,togonia, Buchner (1909,
p. 336); Gryllotalpa borealis, sex-cells, Payne
(1916, p. 30.5).
Smerinthus populi, Pieris brassicse, Oxygia antiqua,
Porthesia similis, Pieris r.apje, Spilosoma lubrici-
peda, Euchelia jacobea>, Cossus Ugniperda, Bom-
byx lanestris and rubi, Arbaxus grossularia, sex-
cells, Gatenby (1917, p. 413).
Aphis saliceti, spermatogenesis, v. Baehr (1909 p

290).
Pyrrhocoris apterus, spermatozoa, Faure-Fremiet

(1910a, p. .548).
Notonecta glauca, spermatocytes, Pantcl and de

Sin^ty (1906, p. 128).
Notonecta insulata, irrorata, and undulata, spermato-
genesis, Browne (1913, p. 99).
Hydrometra lacustris, spermatogonia, Wilke (1907,

p. 687).
Euschistus, spermatogenesis, Montgomery (1911,

p. 776).
Pyg.Tra bucephala, spermatogenesis, Mcves (19(K),
p. .556).

Bombyx , sperniatogcnesi.s, Henneguy (1904).

Oryctes nasicornis, spermatogenesis, Prowazek.'

Insecta — continued.
Silpha carinata, spermatogenesis, N. Holmgren (1902,

p. 197).
Cybister roe.selii, spermatocytes, A'oinov (1903, p. 222).
Dytiscus marginalis and oireumcinctus, spermato-
gonia, Schafer (1907, p. 543).
.^pis meUifica, spermatogenesis, Meves (1007c, p. 478).
Vespa erabro and germanica, Meves and Duesberg

(1908, p. .581).
Camponotus herculeanus, spermatocj'tes, Liims (1908
p. 530).

Libellula , muscle, Holmgren (1909, p. 305).

Chorthippus curtipennis, sjjprmatogenesis, LewLs an<l

Robertson (1916, p. 113).
Trichio.soma lucorum. oogeneses, Govaerts (1913, p. 437),
Leptinotarsa decemlinc.ata, oogenesis, Hegner (1914
p. 417).
MoUusca:
Pelecypoda :

Venus mercen.aria, nerve-cells, Cottdry (1914a, p. 11).
Anodonta cellensis, muscle, Bruck (1914, p. 481).

Amphineiu-a: .

Gast ci'opoda :

Enteroxenes ostergcni, spermatogenesis, Bonnevie

(1904, p. 269).
Conus mediterraneus, Vermetus gigas, spermato-
genesis, Kuschakewitsch (1911, p. 532).
Fulgar canaliculatus, nerve-cells, Cowdry (1914a

p. 11).
Helix pomatia, .Acanthopsole rubrovittata, L. and

R. Zoja (1891, pp. 251, 253).
Murex medicinalis, hypobranchial gland, Grynfeltt

(19126, p. 263).
Helix hortensis, pomatia, and jilanorbis, spermato-
genesis, Benda (1899a, p. 377).
Columbella rustica, spermatogenesis, Schitz (1916,

p. 4.5).
Arion rufus, spermatogenesis, Faui'^-Fremiet (1910a,

p. 631).
Arion empiricorum, egg-eells. Lams (1910).
Cymbulia peronii, spermatogenesis, under heading of
"Microsomes," by Pictet, according to Faurd-
Fremiet (1910a, p. 571).
Cephalopoda :

Octopus vulgaris, kidney, under heading of "En-
claves intra-cellulaires," Mayer and Rathery
(1907, p. 41).
Loligo peaUi, nerve-cells, Cowdry (1914o, p. 11).
Chordata :

Adelochorda: ,

Urochorda: .

Ciona intestinalis, .Ascidia mentala, Molgula socialis,

Cynthia tatraedra, Cynthia morus, eggs, Loyez

(1909, p. 190); L. and R. Zoja (1891, p. 255).

Cynthia partita, embryos, Duesberg (1913, p. 463).

Cynthia microcosmus, oocytes, BluntschU (1904

p. 427).
Phallusia mammillata, fertilization, Meves (1913
p. 21.5).
C3'clostomata:

Petromyzon marinus and Ammocoetes branchialis,

nerve-cells, Mawas (lOIOo, p. 126).
Myxine glutinosa, primary spprmatocvtes, Schreiner
A. and K. E, (1905, p. 296).

'Arb. a. der. zool. Instit., Wien und Triest, 1902.

56

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

ANIMALS — Continued.

Chordata — continued.
Pisces:

Perca americana, Morone americana, Salmo irideus,
Fundulus heteroclitus, embryos, Duesberg (1917,
p. 465); Coregonus mura^na, germ-cells, Aunap
(1913, p. 454); Box salpa, Tinea vulgaris, Corti
(1913, p. 52); Anguilla vulgaris, interrenal, in
testine, Comolli (1913, p. 408).
Bream, pike, and chevasson, kidney, Policard and

Mawas (1906, p. 218).
Scyllium canicula, choroid plexus, Grynfeltt and Eu-
ziere (1913b, p. 104).

Salmo , eggs, Cermak (1902, p. 159).

Amphibia:

Pleurodeles waltlii, pancreas, Chaves (1912, p. 30).
Molge waltlii, pancreas, Chaves (1915, p. 46).
Rana fvLSca, esculenta, arvalis, mugiens, spermatogene-
sis, Broman (1907, p. 348).
Rana viridis, kidney, Policard (1905, p. 381).
Rana temporaria, L. and R. Zoja (1891, p. 256).
Triton alpestris and vulgaris, intestine, Champy

(1913a, p. 76).
Triton cristatus, L. and R. Zoja (1891, p. 256).
Bombinator igneus, intestine, Champy (1913a, p. 82).
Bufo vulgaris, kidney, Policard (1905, p. 381).
Necturus maculatus, Rana palustris, nerve-cells,

Cowdry (1914a, p. 2).
Axolotl, Bombinator igneus, spermatogenesis, Champy

(1913a. p. 64).
Tinea vulgaris, Hyla arborea, L. and R. Zoja (1891,

p. 256).
Rhacophorus, larva;, Saguchi (1913, p. 185).
Geotriton fuscus, pancreas, Levi (1912, p. 579).
Reptilia:

Clemmys marmorata, Lacerta muralis, oogonia,

Munson (1904, plate vii).
Tropidonotus viperinus, kidney, Regaud (1908a, p. 15).
Vipera latastei, pancreas, Chaves (1915, p. 46).
Pseudemys hieroglvphica, nerve-cells, Cowdry (1914a,
p. 2).

Reptilia — conlin tied.

Testudo grfpca, nerve-cells, Busacca (1913, p. 327).
Cistudo europeiP, muscle, Aime (1911, p. 270).
Platydaetylus muralis, Lacerta viridis and muralis,
L. and R. Zoja (1891, p. 256).
Aves:

Hirundo communis, Chelidon urbica. Passer domes-
ticus, Fringilla, Turdus merula, Parus major, Reg-
ulus cristatus, ovaries. Van Durme (1914, p 78).
GalliLS domesticus, embryos, Meves (1907a, p. 399).
Columba livia, nerve-cells, Cowdry (1912, p. 34).
Parus major, Muscicapa grisola, follicle cells, under
heading of "Pseudochromosomes," D'HoUander
(1902, p. 170).
Fringilla cannabina, oocytes, Dulzetto (1916, p. 217).
Sylvia atracapilla, Gallus domesticus, Emberyza cit-
riiu'lla, L. and R. Zoja (1891, p. 256).
Mammalia;

Erinaceus eiu'opeus, intestine, Corti (1913, p. 188).
Didelphys virginiana, spermatogenesis, Jordan (1911,

p. .58).
Phoca>na communis, under heading of "Kornerchen,"

Ballowitz (1907, p. 234).
Sus scrofa, blood-cells, Shipley (1915, p. 61).
Microtus subterraneus, pancreas, Chaves (1915, p. 46).
Canis familaris, egg-cells. Van der Stricht (1908, p. 3).
Vesperugo noctula, egg-celLs, under heading of "Pseudo-
chromosomes, " Van der Stricht (1902, p. 5).
Vespertilio miu-inus and blasii, Rhinolophus euriale,
Miniopterus schreiberii, ovaries, Levi (1913, p.
518).
Many apes, spermatogenesis, Moreaux (1909, p. 371).
Macaccus rhesus, nerve-celLs, Cowdry (1914n, p. 3).
Vespertilio serotinus, suprarenal. Da Costa (1913,

p. 128).
Cercopithecus caUitrichus and saba^us, eggs, d'Athias

(1915, p. 68).
LepiLs cuniculus, Mus musculus and decumanus,
Cavia cobaya, Felis domestica. Homo sapiens, L.
and R. Zoja (1891, p. 256).

PLANTS.

Myxomycetes: .

Schizophytes: .

Algse:
Cystoseira barbata, apical cell, Nicolosi-Roncati

(19126, p. 144).
Fucus serratus, apical cell, Le Touze (1912, plate 9).
Gastroclonium reflexum, Gigartina teedi, Lemanea
torulosa, carpospores and tetraspores, Nicolosi-
Roncati (1912a, p. 60 (?).
Vaucheria, filaments, Rudolph (1912, plate 18).
Coccomyxa solorina>, mitochondria absent, the Mor-
eaus (1916, p. 212).
Fungi:

Achlya, hypha", Rudolph (1912, plate 18).

Albugo bUti and Candida, hyphte and conidia, Lewitsky

(1913, pi. 21).
Aleuria cerea, asci, Guilhermond (1913a, p. 618).
Botrytis cinerea, Guilhermond (1913?, p. 1781).
Coleosporium senecionis, Mme. Moreau (1914, p. 421).
Endomyces magnusi, fibuliger, and albicans, Guilher-
mond (1913^1, p. 1781).
Entyloma ranucuh, Moreau (1914b, p. .538).
Penicilhvun glaucum, Guiliiermond (1914b, fig. 10).
Peziza leucomelas, asci, Guilhermond (1913f, p. 65).
Peziza catinus, asci, Guilhermond (1913a, p. 618).

Fungi — contin ucd .

Phragmidium subcorticum, telautospore, Mme.

Moreau (1914, p. 421).
Psalhota campestris, hypha>, basidia, and spores,

Beauverie (1914a, p. 800).
Puccinia malvacearura, hyphse, Beauverie (1914b,

p. .359).
Pustularia vesiculosa, asci, Guilhermond (1913o, p.

618; 1913b, p. 646).
Rhyzopus nigricans, young spores and filaments,

Moreau (191.5b, p. 143).
Saccharomyces cerevisise and ludwigi, cells, Guilher-
mond (1913s, p. 1781).
Sporodinia grandis, zygospores, Moreau (1914a, p. 347).
Bryophytes:
Anthoceras husnoti, several parts, Scherrer (1913,

pi. 20).

Bryimi , Sa])ehin (1915, pLs. 25, 26).

Funaria hygrometrica, leaves, Sapehin (1915, pis.

22, 23, 25, 26).
Polytrichium piliferum, Sapehin (1915, pis. 22, 25).
Pteridophytes:

Aspidium feli.x-mas, young leaf, Pensa (1911, fig. 3;

1912, pis, 27, 28).
Azolla, Mu-ande (1916, p. 369).

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

57

PLANTS— Cora/mi/cd.

Pteridophytes — coiiiinued.

Lycopodium inundatum, Sapehin (191.5, pi. 12).
Scolopendrium vulgare, leaf, Pensa (1911, figs. 4,5;

1912, pi. 27).
Selaginella marten-si, leaf, Lowschin (1914, p. 268).
Spermatophyles :

Allium cepa and porruni, root-tip, Due.sberg and

Hoven (1910, p. 99).
.\mpelopsLs vietchi, leaf, Guilliermond (1913/, i). 1(XX);

1914a, p. 566).
Amaryllis, ovary, Guilliermond (1912c, p. 888).
Anagallis arvensis, petal, Moreau (1914c, p. 502).
Asparagus officinalis, .several portions, Lewitsky (1910,
pi. 17), Rudolph (1912, p. 18), Sapehin (1915, pi.
10).
Asp.aragus sprengeri, buds, Guilliermond (1914o,

p. 566).
.\ucuba japonica, buds, Pensa (1911, fig. 1).
- Begonia, Camelia, Canna, buds, ovaries, and flower,
Guilliermond (1912c, p. 888; 1914a, p. .566).
Cannabis sativa, cotyledon, Lowschin (1914, p. 268).
Cerasus Laiu-ocerasus, bud, Guilliennond (1912a,

p. 365).
Chenopodium amaranticolor, leaf, Guilliermond (19146

figs. 10, 11).
Corylus, leaf, Lowschin (1914, p. 268).
Curcurbita pepo, hair, Maximow (1913, figs. 1-9).
Dahlia, flower, Guilliermond (1914a, p. 566).
Daucus carota, root, Guilliermond (19126, p. 411-

1914a, p. 566).
Echeveria glauca, epidermal and hypodermal cells,

Rudolph (1912, p. 619).
Elodea canadensis, several parts, Lewitsky (1911,

pi. 28).
Erythrina, pollen mother-eelb, Guilliermond (19r2c,

p. 890).
Euonymus japonicus, buds, Guilliermond (1912a

p. 365).
Euphorbia mvrsinitLs, raeristem of stem, ,Sapehin

(191.5, pi. ii).
Ficaria ranunculoides, root, stem, leaf, Guilliermond

(1913d, p. 439; 1914a, p. 566).
Ficus elastica, leaf, Guilliermond (1914a, p. 566).
Fritillaria imperialis, ovary and pollen, Orman (1912,

pis. 1, 3, 4).
Gladiolus, ovary, Pensa (1910, p. 326).
Helleborus fcetidus, tapetum, Nicolosi-Roncati (1910,

p. 109).
Hordeum, Guilliermond (1912rf, p. 86).
Hyacinthus orientalis, root, von Smirnow (19(X), p.

146).
Hydrangea vestita, winter buds, Pensa (1912, pk.

26, 28).
Ilex aquifohum, leaf, Guilliermond (1914a, p. 566).
Iris germanica, parts of flower, Guilliermond (1915,
p. 241).

Spermatophytes — continued.

Juglans regia, secreting hair-cells and voung leaf

Guilliermond (1914a, p. 566).
Liliuni candidum, ovule, Pensa (1911, fig 2' 1912

pi. 25).
Lilium croceum, Orman (1912, pi. 2),
Lilium martagon, ovary, Pensa (1910, p. 326).
Lilium rubrum, ovary, Guilliermond (1912c, p. 888).
Lupinus albus, Pen.sa (1912, pi. 27).
Lychnis dioica, hairs on stem, Moreau (1914c, p. .502).
Lyeopersicum pyriforme, fruit, Guilliermond (1914a

p. .566).
Mormordica elaterium, hair-cell, Zimmermann (1893

P- 21.5, fig. 2).
Nereum oleander, leaf, Guilliermond (1914a, p. 566).
Nympha^a alba, tapetum, Meves (1904, p. 284).
Oenothera biennis, .several parts, Sapehin (1915, pis.

Papaver rha-as, ovary, Pensa (1910, p. 326).

Phajus grandifolius, root, Guilliermond (1914a, fig. 6).

PhaseoliLs vulgaris, root tip, Duesborg and Hoven

(1910, p. 97).
Philodendron grandifolium, leaves, Guilliermond

(1914a, p. 566).
Pinguicula hirtiflora, gland-cells, Nicolosi-Roncati

(1912c, p. 184).
Pisum sativum, Duesberg and Hoven (1910, figs. 1-5).
Polygonum aviculare, stipule, Moreau (1914c, p. 502).
Populus tremula, leaf. Lowschin (1914, p. 268).
Quercus pedunculata, leaf, Lowschin (1914, p. 268).
Ricinus, Ricinus gibsoni, various parts, Guilliermond

(1914a, p. .566).
Rosa, leaf bud, Pensa (1913a, p. 81).
Rosa thea, ovary, Pensa (1910, p. 326).
Sedum telephium and reflexum, epidermis and hypo-
dermal cells, Rudolph (1912, p. 619).
Solanum tuberosum, ovary, Pensa (1910, p. .326).
Taxus baccata, bud, Pensa (1911, p. .528; 1912, pi. 25).
Tilia parvifolia, leaf, Lowschin (1914, p. 268).
Triticum, Guilliermond (lQl2d, p. 86).
Tradescantia discolor, root, Guilliermond (1914o
p. .566).

Tradescantia virginica, Duesberg and Hoven (1910.

p. 99).
Trianea bogotensis, hairs of root, Sapehin (1915, pi. 11).
Tropaeolum lobbeanum, portions, Guilliermond (1914a

p. .566).
Tulipa gessneriana and sylvcstris, ovule, etc Pensa

(1912, pis. 25, 28; 19il, figs. 6, 7).
Tulipa suaveolens, Guilliermond (1917, p. 233).
Vanillia planifolia, stem and leaves, Guilliermond

(1914«, p. 566).
Vicia faba, root meristem, Zimmermann (1893, p. 215).
Yucca filamentosa. ovary, Pensa (1910, fig. 4).
Zea mais, root, Guilliermond (1912d, p. 86).

58 THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. .

IV. TECHNIQUE.
EXAMINATION OF LIVING CELLS UNSTAINED.

There is no longer any question of the actual existence of mitochondria in liv-
ing cells. The}' may be studied in cells teased out in serum or physiological salt
solution, with direct illumination. It is possible to see them more easily in some
cells than in others on account of some variation in their refractive index or that
of the cytoplasm in which they are embedded. A remarkable phenomenon is fre-
quently seen in the pancreas and other tissues; the mitochondria, instead of being
visible with difficulty, flash out sharply and can be seen with the greatest clearness,
even though they are quite unstained. This may be due to a decrease in saturation,
because we are told that in substances like them outside the organism, a decrease
in saturation is accompanied by an increase in refractive index. Guilliermond
(1912o, p. 409) thinks that the high refractive index which the mitochondria some-
times possess results from the peculiar combination of lipoid with the albuminous
substratum of which they consist, which in his opinion may be disposed on the
surface, in the form of a fatty membrane.

Mitochondria may also be examined with the ultra-microscope (as well as by
ordinary dark-field illumination) in the manner suggested by Regaud. No sys-
tematic attempts have been made to study them with light of different wave-
lengths. One of the best ways to study mitochondria in living cells is by means of
the tissue-culture method as advocated by the Lewises (1915, p. 339).

VITAL STAINING.

It has been claimed by Tschaschin (1912, p. 304) that the mitochondria may
be stained vitally with isamine blue (= pyrrol blue); but Scott (1915, p. 835)
demonstrated that this was not the case by bringing to Ught the mitochondria in
these blue-stained cells with janus green. Nevertheless Levi (1916a, p. 107),
still more recently, has come to the conclusion that the mitochondria in the cells
of tissue cultures may be stained with pyrrol blue. This confusion is due, in the
first place, to mistakes in the identification of mitochondria, and, in the second
place, to a lack of any general consensus of opinion as to just what constitutes a
staining of mitochondria. Some investigators would speak of mitochondria as
being specifically stained when others would be loath to do so; some are content
with a mere tingcing of mitochondria, which can with difficulty be seen and which
others would overlook. My own experience is that almost any coloring matter
will stain the mitochondria if it is used in sufficient concentration, and I am not at
all sure that the controversy is a profitable one.

It is quite possible that in some cases the mitochondria are stained while in
others they are not. There is no a priori reason why the mitochondria should not
take up these vital dyes. In fact, such action would be in complete accordance
with Regaud's electosome theory (p. 118).

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. 59

SUPRAVITAL STAINING.

The most satisfactory dyes for this purpose are janus green B, janus bkie,
janus black I, and diethylsafranin (Cowdry, 1916a, p. 430). Their action is
similar and is described in detail on page 92. They may be applied by immersion
or injection.

IMMERSION.

The best results are obtained with blood (Cowdry, 19146, p. 267) as follows:

Janus green B should be employed in a concentration of about 1 : 10,000 in 0.85 per cent sodium-chloride solution.
A drop should be placed on a series of six or more slides. A small amount of freshly drawn blood is then
added to the dye and a cover-glass i.s immediately dropped on it. No attempt should be made to mix the
blood with the stain before covering.

The preparations should now be examined. Almost inmiediately one of them
will begin to show mitochondria, first in the lymphocytes and later in the granular
leucocytes. Soon the mitochondria will be stained in all of them. Under favorable
conditions they will last for several hours. Evaporation may be reduced by put-
ting a ring of vaseline around the edges of the cover-glass.

INJECTION.

This method is most satisfactory with the pancreas (Bensley, 1911, p. 304) and
salivary glands.

The animal is killed and janus green B is injected into the left ventricle or
aorta in a concentration of 1 : 10,000 of salt solution by gravity pressure. In order
to obtain a good penetration the return flow through the inferior or superior vena
cava, as the case may be, should be momentarily cut off by artery clamps. After
about 10 minutes' perfusion, small portions of the gland may be removed and
examined for mitochondria. When the desired intensity of staining has been
reached, the entire gland should be placed in salt solution pending examination.

fied)

FIXATION AND STAINING.
(1) Altmann's (1890, p. 27) anilin-fuchsin-picric acid method (slightly modi-

Fiialion:

(1) Fix small fragments of not more than 2 c. nmi. in equal parts of ."> per cent pot.assium bichromate and 2 per

cent osmie acid for 24 hours.

(2) Wash in water, 1 hour.

(3) Dehydrate in 50 per cent, 70 per cent, 95 per cent, and absolute alcohol 24 hours each.

(4) Half absolute alcohol and xylol, 3 hours.

(5) Xylol, 3 hours.

(6) 60° C. paraffin, 3 hours. Embed. Cut sections 4/1, and fix to slides by albumen-waler method.

.'^taitiing:

(1) Pass down through toluol, absolute alcohol, 95 per cent, 70 per cent, and 50 per cent alcohol, about .30 seconds

each, to aq. dest. in staining jars.

(2) Stain for 6 minutes in .\ltmann's anilin fuch-sin (anilin water 100 cc, acid fuchsin 20 gm.). The stain may

be poured onto the slide and the whole gently heated over a spirit lamp.

(3) Blot and differentiate by carefully flooding the section with a mixtuie of 1 part of .sat. ale. solution of picric

acid and 2 parts of aq. dest., added with a pipette. During this operation the color can be best seen against
a white background.

(4) Rinse rapidly in 95 per cent alcohol. Pass through several changes of absolute into xylol and mount in

balsam.

60 THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

In this way the mitochondria are stained a beautiful crimson color against a
bright yellow background. It is the oldest and in many respects the best of mito-
chondrial methods; but it has two disadvantages — the fixative penetrates badly
and the colors fade rapidly. Accordingly, neutral balsam or damar should be
used, the specimens should not be exposed to direct sunUght or to heat, and they
should be kept in a dry place.

This method has been greatly modified by Galeotti (1895, p. 466), Schridde (1905,
p. 696), Bensley (1911, p. 308), Freifeld,' Kull (1913, p. 153), Schirokogoroff (1913,
p. 523), Cowdry (19166, p. 30), Duesberg (1917, p. 469), and others.

Bensley proceeds as follows:

Fixation:

(1) Fix in acetic, osmic, bichromate mixture (2.5 per cent potassium bichromate 8 c.c, 2 per cent osmic acid

2 c.c, glacial acetic acid 1 drop) for 24 hours.

(2) Wash, dehydrate, clear, and embed (p. .^9), except that bergamot oil is substituted for xylol.

Staini?ig:

(1) Pass down to water.

(2) Dip in 1 per cent potassium permanganate about 1 minute.

(3) Rinse in 5 per cent oxalic acid same time and wash in water.

(4) Stain with anihn fuchsin as indicated (p. .iO).

(.5) Differentiate in a 1 per cent aqueous solution of methyl green.

(6) Rinse rapidly in 9.5 per cent alcohol. Pass through several changes of absolute into xylol and mount in
balsam.

The use of permanganate and oxalic acid corrects excessive mordanting with
the osmic and bichromate. It may sometimes be dispensed with. The methyl
green, which was first used in this way by Galeotti, is a much finer contrast stain
than the picric acid and is also more permanent. The precautions already men-
tioned against fading should be observed.

I make use of the following modification :

Fixation:

(1) Fix in Regaud's mixture (4 parts of 3 per cent potassium bichromate and 1 part of commercial formahn).

The commercial formalin may profitably be neutralized by saturation with magnesium carbonate. The
mixture may be applied by immersion or injection, the latter being recommended for large objects. It
should be changed every day for 4 days and be kept in an ice-box (though this is not essential). Mordant
for 8 days in 3 per cent potassium bichromate, changing every second day.

(2) Wash in running water over night.

(3) Dehydrate, clear, and embed as indicated (p. 59).

Staining:

(1) Pass slides to water as indicated (p. .59).

(2) 1 per cent potassium permanganate 30 seconds, but time must be determined experimentally.

(3) 5 per cent oxalic acid 30 seconds. Note. — Steps (2) and (3) may usually be dispensed with.

(4) Rinse in several changes of distilled water about 1 minute. Incomplete washing prevents staining with fuchsin.

(5) Stain in Altmann's anilin fuchsin made up as follows: Make a saturated solution of anihn oil in distilled

water by shaking the two together. Filter and add 10 gm. of acid fuchsin (Duesberg) to 100 c.c. of the
filtrate. The stain should be ready to use in about 24 hours. It goes bad in about a month. To stain,
dry the slide with a towel, except the small area to which the sections are attached; cover the sections on
the slide with the stain and heat over a spirit lamp until fumes, smelling strongly of anihn oil, come off;
allow to cool; let the stain remain on the sections about 6 minutes; return the stain to the bottle.

(6) Dry off most of the stain with a towel and rinse in distilled water, so that the only remaining stain is in the

sections. If a large amount of the stain is left it will form a troublesome precipitate with methyl green;
on the other hand, if too much stain is removed the coloration of the mitochondria will be faint.

(7) Allow a little 1 per cent methyl green, added with a pipette, to flow over the sections, holding the shde over a

piece of white paper, so that the colors may be seen. Apply the methyl green for about 5 seconds at first
and modify as required. This is the crucial point of the method.

(8) Drain ofT excess of stain, plunge into 9.5 per cent alcohol for a second or two. Then rinse in ab.solute alcohol,

clear in toluol, and mount in balsam.

Difficulties. — (1) The methyl green may remove all the fuchsin, even when
applied only for a short time. This is due to incomplete mordanting of the mito-

'For which see Naegeli (1912, p. 30).

THte MITOCHONDRIAL CONSTITUENTS OP* PROTOPLASM. 61

chondria by the chrome salts in the fixative. It nyiy be avoided by omitting
steps (2) and (3) , or by treating the sections with 2 per cent potassium bichromate
for a few seconds just before staining (as advised by Bensley). The action of the
permanganate and oxaUc is to remove the bichromate. (2) The fuchsin may
stain so intensely that the methyl green removes it imperfectly or not at all. This,
on the contrary, is due to too much mordanting. It may be corrected by pro-
longing steps (2) and (3). (3) Sometimes, after obtaining a^good differentiation,
the methyl green is washed out before the slide is placed in toluol, in which event
omit the 95 per cent and pass to absolute direct.

This fixative is a good penetrator, in which respect it is much superior to
Altmann's fluid or Bensley 's mixture; but it makes the tissues very brittle and
difficult to cut. The staining is satisfactory and unifornr.

(2) Benda's (1901, p. 155) crystal violet ahzarin method:

Fixation:

(1) Flemming's fluid, 8 days.

(2) Wash 1 hour, half pyroHgneous and 1 per cent chromic acid, 24 hours.
(.3) 2 per cent potassium bichromate, 24 hours.

(4) Wash in running water 24 hours, dehydrate, and embed in paraffin. Cut sections 5/i.

Staining:

(1) 4 per cent iron alum, 24 hours.

(2) Rinse in water and bring into an amber-colored solution of sodium sulphaUzarinate, made by adding a satu-

rated alcoholic solution to water, 24 houi's.

(3) Blot with filter-paper and stain in equal parts of crj-stal violet solution and water. (N. B. — The crystal

violet solution made of cone. sol. crystal \iolet in 70 per cent alcohol 1 volume, alcohol 1 volume, and anilin
water 2 volumes.

(4) The solution is warmed until the vajior arises and then allowed to cool for ,5 minutes.
(.')) Blot, then 30 per cent acetic acid 1 minute.

(6) Blot, plunge in absolute alcohol until but little more stain comes off, clear in xylol, and mount in balsam.

A useful modification is given in Meves and Duesberg (1908, p. 573). Success-
ful Benda preparations are excellent. The mitochondria are stained a deep violet
color against a rose background. They are also much more permanent than
Altmann preparations. Unfortunately the method is long, tedious, and difficult.

(3) Regaud (1910, p. 296) has used the iron-hematoxylin method of Heidenhain
after a large variety of fixatives, the best of which is his formahn and bichromate
mixture.

Fixation:

(1) Fi.\ in 3 per cent potassium bichromate 80 volumes, commercial formalin 20 volumes, for 4 days, changing

every day.

(2) Mordant in 3 per cent bichromate for 7 days, changing every second day.

(3) Wash in running water 24 hours, dehydrate, clear, embed, and section as indicated (p. .58).

Staining:

(1) Pass down to water as indicated (p. 59).

(2) Mordant in 5 per cent iron alum at 3.5° C. for 24 hours. Rinse in aq. dest.

(3) Stain for 24 hovu-s in hematoxylin made up as follows: Dissolve 1 gm. pure crystals of hematoxyhn in 10 c.c.

of absolute alcohol and add 10 c.c. of glycerin and 80 c.c. of distilled water.

(4) Differentiate in 5 per cent iron alum under microscope.

Note. — The crucial point in the technique is passing from the mordant to hematoxyhn. The slides
must be rinsed in distilled water, otherwise the iron alum will form a dense black precipitate in the stain.
On the ether hand, if they are rinsed too much, all the iron alum mordant will be removed. It is neces.sary
to strike the happy mean in which a darkening of the hematoxylin alone occurs. It is always difficult to
get good hematoxylin, and I find it best to keep on hand a ripe alcohohc solution.

This is the most permanent as well as the simplest of all mitochondrial stains.
It may be used in the damp climates of most of our marine biological laboratories,

62 THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

where the Altmann method and its modifications are useless. Unfortunately it
rarely gives good results with embryonic tissues; for these the older osmic acid-
containing fixatives arc best adapted. It is often possible to make use of material
fixed in the usual way with formalin by starting out with step (2). Moreover,
the preparations can be counterstained in a variety of ways (Cowdry, 1916a, p. 441).

(4) Dubreuil's (1913, p. 74) iron-hematoxylin method for blood-cells:

Fixation:

Take up the fluid to be examined in a pipette containinp several times its volume of 0.5 to 1 per cent solution
of osmic acid. Transfer to a centrifuge tube. A good fixation is obtained in about an hour. Then add
distilled water, centrifuge, and decant. The blood-cells remaining are shaken up with absolute alcohol and
then passed into a weak solution of celloidin. A di'op is allowed to spieafl on a slide, which, before complete
desiccation, is plunged into 80 per cent alcohol.

Staining:

The mitochondria in the cells are then stained with iron hematoxylin (p. 61).

(5) Bensley's copper chrome hematoxylin method:

Fixation:

(1) Flx in either .\ltmann's osmic bichromate mixture (p. .59) or in Bensley's acetic-osmic-bichromate fluid

(p. t)0) for 12 to 24 hours.

(2) Wash, dehydrate, clear, embed, and section (p. 59).

Staining:

(1) Pass down to water (p. 59).

(2) Sat. aq. copper acetate, 5 minutes.

(3) Wash in several changes distilled water, 1 minute.

(4) 0.5 per cent hematoxylin, 1 minute. (If the copper acetate has not been suflficiently washed out, a black

precipitate forms in the hematoxylin. The hematoxylin should be well ripened. It may be obtained l)y
dilution down from a 10 per cent alcoholic stock solution.)

(5) Rinse in aq. dest.

(6) 5 per cent neutral potassium chromate, 1 minute. (The sections should turn a dark blue-black color.

If they are only a light-blue shade, rinse in aq. dest., place again in the copper acetate, and carry through
as just described several times until no increase in color result=.)

(7) Wash in aq. dest. and return for a few seconds to the copper acetate in order to convert all the dye into the

copper lake.

(8) Wash in aq. dest.

(9) Differentiate under the microscope in Weigert's borax-ferricyanide mi.xture diluted with 2 volumes of

aq. dest.

(10) Wash 6 to 8 hours in tap water.

(11) Dehydrate, clear, and mount in balsam.

(6) Bensley's (1911, p. 308) neutral safranin method:

Fixatio7i:

(1) Fix in chrome subhmate (2.5 per cent potassium bichromate 100 c.c, mercuric chloride 5 gm.) for 24 hours.

(2) Wash, dehydrate, clear, embed, and section as indicated on page 59.

Staining:

(1) Preparation of stain: Add slowly sat. aa. sol. of the color acid, acid violet, to sat. .aq. sol. of the color-base,

safranin O, contained in a flask until a precipitate no longer forms. The point of neutralization may be
roughly determined by dropping a little of the mixture on filter-paper from time to time until the outside
red ring of safranin disappears and the whole blot takes on a neutral color. Filter. The filtrate should
be as nearly as possible colorless. Dry the precipitate on filter-paper for 12 hoiu-s, collect it, and make
a saturated solution of it in absolute alcohol.

(2) Pass sections down throug' two chanees of toluol and absolute alcohol in order to remove all traces of paraffin

or toluol, which might interfere with the staining. Then through 95 per cent, 70 per cent, and 50 per
cent to aq. dest. (Chrome and osmium fixed material must be bleached in permanganate and oxaUc acid
(vide p. 60), and sublimate-fixed tissues must be treated with Lugol's iodine solution for about 10 seconds
and washed in aq. dest.)

(3) Dilute the alcoholic stock solution of the dye with an equal \olume of aq. dest. and stain for from 5 minutes

to 2 hours.

(4) Blot quickly with several layers of filter paper.

(5) Plunge into pure acetone and pass immediately to toluol without waiting to drain.

(6) Examine under the oil immersion and if necessary differentiate in oil of cloves. If this is not sufficient, the

slide, after rinsing in absolute alcohol, may be instantaneously flooded with 95 per cent alcohol, and then
passed back through absolute alcohol to toluol.

(7) Wash in two changes of toluol and mount in balsam.

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. 63

Working on the same principle, a number of stains can be made up for mito-
chondria (Cowdry, 1913a, p. 485). Note also Bensley's neutral gentian method.

(7) Bensley's (1916, p. 47) brazilin-wasserblau method:

Fixation:

(1) Fix in Zenker's fluid, plus less acetic acid, 10 per cent formalin, 24 hours.

(2) Wash, dehydrate, clear, embed, and section (p. 59).

Staining:

(1) Pass down to water (p. 59).

(2) Iodize with Lugol's solution, 30 seconds.

(3) Stain in following solution several hours: Phosphotungstic acid, 1 gm.; aq. dest., 100 c.c; braziUn, 0.05 gm.

The brazilin is first dissolved in a small quantity of distilled water by the aid of heat and added to the
phosphotungstic-acid solution. Ripening may be accelerated by the addition of 0.4 c.c. of hydrogen pero.x-
ide, or of a few drops of a solution of soluble molybdic acid. The solution deteriorates with age and
should not be used after 3 days.

(4) Rinse in aq. dest. and place for 1 to 5 minutes in phosphotungstic acid, 1 gm.; wasserblau, 0.2 gm.; aq. dest.,

100 c.c.

(5) Wash rapidly in water, dehydrate in absolute alcohol, clear in toluol, and mount in balsam.

(8) Meves's (1905, p. 102) new Victoria green method: This method is intended
for red blood-cells which are simply stained in the fresh condition by the addition
of a 4 per cent iodic-acid solution to which a small quantity of new Victoria green
(malachite green) has been added.

(9) The methods of silver reduction employed by many ItaUan investigators
are essentially modifications of the original method of Golgi. They undoubtedly
reveal mitochondria in most cases, but one would hesitate to attribute any high
degree of specificity to them. For details see Veratti (1909, p. 34), Pensa (1910,
p. 326), Rina Monti (1915, p. 21), and Cajal (1915, p. 3).

For still other mitochondrial methods see Sjovall (1906, p. 563), Rubaschkin
(1910, p. 407), Koltzoff (1906, p. 384), Kingsbury (1911, p. 317), Rchultze (1911a,
p. 465), Maximow (1916a, p. 462), and others.

In unskilled hands the experimental error in some of these methods of technique
is sometimes very great, but it is a mistake to regard the methods as difficult.

In the examination of living cells it is of course essential that the medium should
be as nearly as possible isotonic. The presence or absence of Brownian movement
is a valuable criterion of the condition of the cells. Where it is very marked the
material should be discarded, because it indicates that an unusually large amount
of water has entered into the cells. Each particle, instead of being held in place
by the balanced action of many bombarding molecules, is subject to the action of
only a few, now on one side and now on the other, causing it to jump from place
to place. The smaller the particle the less chance there is of molecules on the
opposite side compensating. Large granules, on the other hand, offer a greater
surface, are more likely to be bombarded from all sides, and are thereby held in
place and do not show so much tendency toward exhibiting Brownian movement.

One of the most common and annoying sources of error is mechanical manipu-
lation of the fresh tissue before fixation. A mere crushing of the tissue with the
forceps will bring about the most astonishing and perplexing modifications in the
mitochondria. Allowing a surface film of the tissue to dry in air, as it stands on
the autopsy table, is another common blunder which changes the whole appearance
of the mitochondria. Osmotic changes in the tissues before fixation must also be

(54 THE MITOCHONDRIAL CONSTITtJENTS OF PROTOPLASM.

carefully guarded against. Merely keeping the tissue in salt solution will often
cause the mitochondria to assume the most bizarre and perplexing forms. The
tissues should be rjuite fresh, particularly the pancreas, in which autolysis takes
place very rapidly. But the need for fresh tissue has been so much exaggerated
that it has discouraged the study of mitochondria without reason. In the nervous
system, for example, where autolysis is slow, the mitochondria may be advantage-
ously studied 6 or even 12 hours after death. Material awaiting fixation should
be kept in a cool place and precautions taken against evaporation.

Poor penetration of the fixative is also well known to bring about definite
changes in the mitochondria. The mitochondria in the deeper parts of the tissue
become spherical and swell up. Indeed, when a fixative like the acetic-osmic-
bichromate mixture of Bensley is applied simply by immersion the action of each
of the ingredients may be seen. On the very surface the usual filamentous form
of mitochondria is maintained through the combined action of osmic acid, bi-
chromate, and acetic; a little deeper in, where the osmic acid has not penetrated,
the bichromate and acetic alone acting, the mitochondria are swollen up and
spherical; while in the central parts of the tissue the mitochondria have been com-
pletely dissolved through the ummodified action of the acetic acid. Simultaneous
action of the ingredients is satisfactory, but successive action is not. One comes
across varying degrees of artifact as one proceeds inward. Naturally, small pieces
of tissues alone can be used. They should be of not more than 4 cubic millimeters,
and it is better to have even smaller when the tissues are hard and firm and resist
the penetration of the fixative. It is unfortunate that investigators do not avail
themselves more generally of the method of applying the fixative by injection
through the blood-vessels. Naturally this can not be done with osmic acid on
account of the expense, but the mixtures of formalin and bichromate advocated
by Regaud will give excellent and uniform fixation in very tough and fibrous tissue
when applied by vascular injection. It is always necessary to wash out the blood
with salt solution before applying the fixative.

That the concentration of the fixative is also of great importance has been
shown by N. H. Cowdry (1917, p. 200), who finds that very dilute solutions of
formalin cause a swelling of mitochondria and concentrated ones a shrinkage. It
is hkely that most mitochondrial methods bring about a slight shrinkage of the
mitochondria.

Now, the fixative in the right concentration having come in contact with the
mitochondria, it becomes necessary to inquire whether it preserves them in their
true form without distortion. This is a difficult question to answer. In the
majority of cases with care and good fixation the original form of the mitochondria
is preserved. This was first shown with striking clearness by the Lewises (1915,
p. 347), who studied the living cells in tissue cultures and then actually watched
the process of fixation in them. When fixatives do alter mitochondria, they alter
them in a definite way. Filamentous mitochondria break up into granules and
spherules. Filamentous mitochondria are never formed by the fixative from
granules.

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. 65

Experience teaches that a great polymorphism of mitochondria in a preparation
of cells of the same type, acinus cells of the pancreas for instance, is usually the
result of experimental error before or during fixation. Such preparations should
be discarded unless the polymorphism occurs constantly with different fixatives or
can be seen in the living condition.

Another error which frequently creeps in is a variation in mordanting which
manifests itself in the subsequent staining reactions, in the interpretation of which
the mordanting must always be borne in mind. But perhaps the most important
variable is differentiation. Preparations assume an entirely different aspect with
the degree of differentiation. This is particularly true where the differentiator itself
is colored, as in the case of methyl green.

Let us bear in mind that not one of these methods is specific for mitochondria.
Janus green is the most specific, but it will stain other structures (p. 59) under
certain conditions. The iron-hematoxylin method of staining is certainly the least
specific of all, because it even colors chromatin in the same way as mitochondria.
Neither is the Altmann method specific. There are fewer objections to its fuchsin-
methyl green modification and to the Benda method, though neither of them can
be completely trusted.

Finally, a word as to the relative value of observations on fixed and fresh
material. It is the fashion now to insinuate that the older methods of fixation and
staining are more or less useless. This should be deplored. It may be said that the
value of fixation consists in the rapidity of its action. The cell is killed instanta-
neously and a condition resembling more or less closely that of the living cell at a
definite moment in the course of its normal functional cycle is preserved. On the
other hand, in the observation of living tissue, the much-prolonged time factor con-
stitutes a serious and unfortunately but little recognized source of error. A half
hour spent in studying mammalian tissues teased out in so-called isotonic media
gives abundant opportunity for experimental error to creep in. Nevertheless,
the study of fresh tissue contributes invaluable information as to the process which
is taking place, even though the process can not be regarded as entirely normal. The
other great and unique advantage which the observation of living tissue presents
is that it jiermits of the experimental modification of the various vital processes
going on in the cells. To sum up, the methods of fixation of mitochondria will never
be replaced by the observation of living tissues, but they will be greatly supplemented,
extended, and aided by it. The two must go hand in hand.

The essential thing about mitochondrial technique is the necessity of experi-
mentation. There is nothing really difficult about it, but it is too much to expect
to get good results after the first or second trial. For this reason attempts to shift
the responsibility onto the shoulders of untrained technicians almost invariably
result in mutual dissatisfaction.

66 THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

V. MORPHOLOGY.

MORPHOLOGY IN ORGANISMS.

We rarely meet with forms which possess mitochondria of pecuhar or dis-
tinctive morphology. In multicellular animals and plants, high up in the scale,
in which there is considerable division of labor among their constituent parts, all
forms of mitochondria are represented, though some may predominate. In uni-
cellular organisms, on the other hand, the mitochondria are sometimes granular,
sometimes filamentous, sometimes large and sometimes tiny, depending upon con-
ditions which are wholly obscure. Frequently we meet with all forms within the
compass of a single cell. Obviously they can never be used in a taxonomic way,
like the chromosomes, to distinguish between nearly related species. This is equally
true for animals and for plants.

MORPHOLOGY IN TISSUES.

The several tissues of the higher plants and animals possess mitochondria
of characteristic form; that is to say, in some of them filaments predominate, in
others granules, and so on, but in similar tissues of different animals they are much
the same. For instance, I find it difficult, even in the different classes of verte-
brates, to distinguish the spinal ganglion cells from each other on the basis of their
mitochondrial content alone. The mitochondria in the liver, pancreas, lung,
prostate, and other organs are characterized by the predominance of some definite
form, granule, rod, or filament in all nearly related animals. This general constancy
of mitochondria, where the function is similar, is, I think, of considerable importance,
because it must surely indicate that their morphology is a fundamental property
ingrained in the very organization of the cell in phylogeny and that it is not always
a passing trivial affair which varies from moment to moment.

Mitochondria are for the most part filamentous in certain nerve-cells (plate 1 ,
fig. 4), in gland-cells (plate 1, fig. 9), and others, as well as in most of the tissues of the
developing embryos of all forms. Indeed, filamentous mitochondria are among the
most common met with anywhere. The average length of the filaments varies in
different tissues : they are perhaps longest in secreting cells, like the pancreas, where
they may attain a length of 10 to 12 ii. Their diameter also varies (0.05 to 0.2 n)
in different localities, but in individual cells of the same tissue it is astonishingly
uniform. Within a single cell the length is variable and the breadth uniform.
They may be straight, curved, or even twisted, depending upon their surroundings.
They do not begin to taper toward their extremities, which end abruptly and
evenly. Sometimes their ends are somewhat swollen.

Filaments are never produced from granules through the action of the fixati\-e.
Some think that they are formed, in the living tissue, by linear coalescence of
individual granules; others, that they arise from single granules through growth by
accretion. Certain it is that in those cells where filaments are abundant all transi-
tions may be seen between them and granules, while filaments are often totally
absent in cells with granular mitochondria.

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. 67

There is a tendency among investigators to l:)elieve that filamentous mito-
chondria are the direct result of streaming movements in the cytoi)lasm. Certain
cases may be cited in support of this theory, like the outgrowing nerve-fiber, for
instance, in which the mitochondria are always filamentous. In gland-cells also
they are usually filamentous and stretch from the basement membrane toward the
lumen in the direction of the passage of materials from the blood-stream into the
gland-ducts. Again, in dividing cells, the mitochondria often stretch out in the
direction of separation. But we have other cases, which demand explanation, in
which the mitochondria are rod-Uke or even granular, in spite of the streaming
movements; and still other cells where the mitochondria are filamentous, though
the cytoplasm is to all intents and purposes stationary, or at any rate as motionless
as it ever is. Thus, the mitochondria are granular and rod-hke, not filamentous,
in living human polymorphonuclear leucocj'tes during amoeboid movement and in
the streaming protoplasm of certain plants. On the contrary, in sharp contrast
are cartilage-cells and bone-cells, where, so far as we know, the cytoplasm is rela-
tively quiescent, yet the mitochondria are filamentous. They remain thread-like
in the nerve-fiber throughout the whole life of the animal, even though there is no
longer a pushing-out of substance. Nicholson (1916, p. 332) has found that they
are constantly filamentous in the bodies of some nerve-cells, granular in others;
and that where they are filamentous they are more so in the peripheral cytoplasm
(where they are disposed parallel to the cell-wall) than in the deeper cytoplasm
about the nucleus. If their shape is conditioned by protoplasmic streaming, our
conclusion must be that some types of nerve-cells constantly exhibit it, that others
never do, and further that the little vortices are most powerful in the peripheral
cytoplasm, where the mitochondria constitute a clue to their direction. And this
is not the end of the interesting deductions which would follow. While some
workers are inclined to pin their faith in the real existence of such currents in nerve-
cells. Kite's' microscopic dissections with very fine glass needles have brought to
light the fact that their cytoplasm is very viscid and has the physical consistency
of a gel. Furthermore, Key's'- observation that it is difficult to alter the distribu-
tion of cytoplasmic materials in nerve-cells by centrifuging indicates the improb-
ability of any considerable protoplasmic streaming.

Rubaschkin's (1910, p. 428) idea that filamentous mitochondria are char-
acteristic of specialized cells and granular ones of embryonic, undifferentiated
cells has been negatived again and again; and Dubreuil's (1913, p. 137) view that
filamentous ones are indicative of rest and granular ones of rapid multiplication
by division is not borne out by recent work. According to the descriptions of
Moreau (1914b, p. 538) they are granular in the spores of the fungi in which the
cell activities are at a very low ebb, and X. H. Cowdry finds that they are filament-
ous in the more or less inactive cells of the radicle of the dried seed pea. We have
here two instances of mitochondria in quiescent protoplasm; in one they are granular
and in the other filamentous. Numerous other instances might be cited to show
that filamentous mitochondria are not indicative of rest.

' Public lecture at the Marine Biological Laboratory. Woods Hole. ' Personal communication.

68 THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

Filamentous mitochondria with bleb-like swellings are to be found in some secret-
ing cells, but not in all. The blebs are supposed to be the precursors of secretion.
In plant cells, starch, pigments, and other materials are unquestionably laid down
in them, while in animal cells they are said to give rise to granules of zymogen.
Many cells which form a definite secretion (plate 1, figs. 3 and 7) do not possess
mitochondria armed with these swellings. I refer also to mucous cells. On the
other hand, while filamentous mitochondria are exceedingly common, they never,
to the best of my knowledge, have blebs in cells which do not form some distinct
secretion. I have in mind particularly the nervous system, where mitochondria of
this kind are unknown. The swellings are certainly not the result of technique;
they represent a reaction on the part of mitochondria to some definite and specific
demand or condition of the cell.

In other localities mitochondria are rod-like. They vary in length and in
girth. They are long and very straight in the cells of the convoluted tubules of
the kidneys, where they apparently constitute the well-known batonnets of Heiden-
hain. Here also they are of very large diameter. vSmaller rod-like mitochondria
are often to be seen in muscle-cells between the fibrils, in some nerve-cells, in fact
in almost all tissues. Sometimes a swelling will develop at one end of a rod and in
this way a pear-shaped mitochondrion will be formed. Occasionally they are formed
through the action of the fixative upon mitochondria of the filamentous variety.
The fixation may also modify their size to a certain extent; it may cause them to
swell or to shrink, so that we must be on our guard.

Granular and dumb-bell-shaped mitochondria are of very common occurrence,
though they are not so widely distributed as the filamentous and rod-like forms.
They are quite spherical and resemble cocci rather than bacilli. It is sometimes
diflficult to distinguish between them and other cell inclusions on this account,
because these other materials occur in the form of droplets also, rarely if ever as
rods or filaments. Mitochondria of this variety are very abundant in some egg-cells
(plate 1 , fig. 6) . The absence of very minute granular mitochondria merging into the
invisible is of interest from the point of view of an origin de novo through condensa-
tion (page 98). The dumb-bells are usually interpreted as stages in the division
of single granules, but it is unnecessary to say that they may equally well represent
an approximation of two originally separate granules. Some fixatives will cause
filamentous mitochondria to break up into granules, and I have even known janus
green to bring about this change. Good fixatives which have poorly penetrated
the deeper layers of pieces of tissue regularly cause this granular degeneration of
mitochondria, so that the mitochondria become filamentous and gradually assume
their normal condition as one approaches the periphery of the block.

Mitochondrial networks are of comparatively rare occurrence, though they are
found normally in certain types of cells. They have been described in sper-
matogenesis, the Lewises (1915, p. 356) have seen them in tissue cultures, and I
have occasionally seen them in the acinus cells of the pancreas of mice stained with
janus green. They are of variable extent. They may consist of one or two meshes
or of several. They are often, though not always, arranged about the nucleus.

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. 69

Branching filaments are found in association with them and may constitute a
stage in their formation. The strands of the network are of the same girth as the
branching forms and of the individual filaments, except perhaps at the nodal points,
where they may be slightly enlarged. It is still impossible to say whether the net-
works arise through a fusion of many mitochondria or by outgrowth of separate
ones. They are difficult to preserve in fixed tissues by reason of the tendency of
fixatives to cause fragmentation. Where they are found in fixed preparations we
may be sure that they occurred in the living state. So far as we know at present,
they are not associated with any definite function on the part of the cell.

Large spherical mitochondria are usually on the border-line between true
rnitochondria and lipoid droplets. They represent a stage in the transformation
and may be found in any tissue where the change is taking place. They may be
seen to the best advantage in the nervous system, in the cells of the spinal ganglia,
where all stages in the metamorphosis may be made out. When the mitochondria
assume a diameter of about 3 or 4 m they begin to become more and more resistant
to acetic acid and clump together in the familiar lipoid accumulations. The ques-
tion is rather complicated, however, because large spherical mitochondria, of much
the same appearance, may be produced by poor fixatives as well as by good fi.xatives
which have penetrated badly. We must know the tissue. This warning is par-
ticularly applicable to large spherical mitochondria with clear centers and to ring-
shaped forms, which should always be regarded with suspicion and interpreted
with caution.

N. H. Cowdry (1917, p. 225) summarized his work on the morphology of
mitochondria as follows:

"Their morphology is identical in plants and in animals; they assume no forms in the
one which are not present in the other; they undergo similar variations in size and shape in
different tissues and in different cells in both. If it were possible to view mitochondria
dissociated from their environment, it would be impossible to decide whether they came
from plant or animal tissues, provided they did not contain starch, pigment, or some other
easily recognizable substance, to serve as a clue."

The little-known condition of chondriolysis is of interest. The mitochondria
lose their definite form, whatever it may be, and give rise to a diffuse deposit of
material staining with mitochondrial stains. It is of frequent occurrence in the
nervous system in normal conditions and is conspicuous in other organs in patho-
logical states (page 136). We should bear in mind the possibility that in all cells
individual mitochondria go into solution and gradually fade away and reappear
again, even though the change escapes our observation, except in special cases like
that of the nervous system. Something similar has been described in the supra-
renal by Mulon (1912, p. 36).

Speaking in a general way, all mitochondria have smooth and even outlines,
no matter how bizarre their forms may be. Each merges into the other through
imperceptible transitions, so that any terminology is more or less arbitrary. No
difference can be detected in the solubilities or in the staining reactions of the
different forms. A fundamental distinction may, however, be made between

70 THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

variations in length and variations in breadth. While the mitochondria in the
acinus cells of the pancreas, for instance, vary greatly in length, filamentous and
rod-like forms predominating, their girth is remarkably uniform, throughout the
whole tissue, in individual cells and in different parts of the same filament. The
ends of the filaments are never tapering. They are quite abruptly but smoothly
rounded off. Similarly in other tissues the diameter is fixed, but the length is not;
one is stamped on the cell through its organization, the other is not. We have
here two attributes, independently variable, which may perhaps be influenced in
different ways.

Evidently mitochondria increase in size by growth in length, probably by the
addition of material at their extremities. That is to say, when mitochondrial
substance is added the extension is always lengthwise, never lateral. It is equally
true that when foreign material, like starch, pigment, or fat is deposited within the
mitochondria, expansion is always provided for by increase in girth. There
must be some difference in the method of addition. At present no explanation is
available.

Much confusion would have been avoided if investigators could have come to
reaUze that the same material under different conditions may assume different
forms. The unfortunate part about it is that the conditions are almost wholly
unknown to us. But whatever they may be, it is altogether likely that they operate
in much the same way in plants and animals, because all forms of mitochondria
are represented in both.

We have abundant evidence that mitochondria are semi-fluid in consistency.
They flow together and fuse to form large droplets, under certain conditions, and
in the streaming protoplasm of plant-cells their form is continually changing in
response to currents and eddies in the stream. This semi-fluidity, together with
their lipoid-like properties, is responsible for their remarkably smooth and even
outUnes and rounded ends. It precludes rough excrescences, sharp angles, and
pointed ends.

The very pronounced property which they have of being able to take in sub-
stances from the surrounding cytoplasm and heap them up and concentrate them
in their interior is responsible for manj^ of the swellings and enlargements which
occur in certain situations in the course of the filaments as well as in the smaller
rods and granules.

It is hard to say how great a part osmotic factors play in molding the shape
of mitochondria. The mitochondrial substance is sometimes denser in the periph-
eral parts of the filaments and larger granules and bears some semblance to a
membrane. The Lewises (1915, p. 373) found that they could alter the form of
mitochondria by changing the osmotic pressure of the fluid bathing the cells in
their tissue cultures. With hypotonic solutions there was a marked increase in
size, with hypertonic a decrease. The osmotic pressure of the blood is very con-
stant and has been so, if Macallum's work is correct, for untold ages in evolution.
Since the same blood bathes tissues containing entirely' different mitochondria, it
seems that the osmotic pressure of the fluid about the cells must play a very minor

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. 71

role, if any, in determining the shape of the initochondria in the Uving organism.
The mere presence or absence of high osmotic tensions within the cells makes no
difference, because in certain plants pressures as high as 13 atmospheres are
developed, yet, as far as our observations go, the mitochondria difTer in nowise
from those in plant and animal cells at atmospheric pressure.

It is quite possible that variations in the acidity or alkalinity of their surround-
ings may affect their form. Bearing in mind the fact that the mitochondria are
supposed to be a combination of hpoid and albumin, we might conceive of acidity
acting upon their protein fraction, causing it to become hygroscopic and to swell.
But we must remember that oxidized phospholii)in has a much greater affinity for
water than the unoxidized (Mathews, 1915, p. 98). It is possible, then, that with
increase in oxidation initochondria will take up water and swell. But acidity
inhibits oxidation. It begins to look as if we had two antagonistic influences to
deal with. Acidity may cause the protein fraction of the mitochondria, which is
the smallest fraction, to take up water; but it will also prevent the oxidation of the
phosphohpin and prevent it from taking up water. So that, arguing along these
lines, one would expect the bchaA'ior of mitochondria to depend upon the relative
proportions of phosphohpin and albumin, or (more properly speaking) of protein,
in their composition. But here we meet with the same difficulty as in the case of
osmotic pressure. The astonishing neutrality of the organism would prevent this
factor, if it is one, from playing any great part.

We must simply acknowledge our ignorance and hope for the future. It may
be said, however, that the shape and size of the cell have no influence upon the
shape or size of the contained mitochondria; neither has the water-content or the
consistency of the cytoplasm.

We may conclude by saying that ^■ariations in the shape and size of mito-
chondria constitute by far the most delicate criterion of cell injury known to us.
Mitochondria react long before the nucleus, and their morphology is the first thing
to change, though it is soon followed by alterations in distribution and in amount.

72 THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

VI. DISTRIBUTION.
DISTRIBUTION IN ORGANISMS.

Mitochondria seem to be i:)resent in all organisms from man to the most lowly
protozoon and from the angiosperms to the fungi, but their existence is doubtful
in the myxomycetes, the schizomycetes, and most of the algie.

Some think that the mitochondrial substance is here present in solution, and
this idea is supported by the staining reactions of certain bacteria. It occasionally
happens that tissues prepared for mitochondria have been invaded by bacteria, in
which case the bacteria stain just like the mitochondria by the Benda method,
with iron hematoxylin and with fuchsin methyl green. I have found that large
bacilli contain granules which stain intensely and apparently specifically with
janus green. They resemble in distribution the so-called polar granules. Smaller
forms often stain diffusely.

Others consider the mitochondrial substance to be really absent. According
to them its function is carried on by the chloroplasts. The fact that mitochondria
diminish in number progressively with the development of chloroplasts in the
higher plants lends some support to this view. Mitochondria may even be entirely
absent when the full complement of chloroplasts is attained.

We know of no other exceptions in the distribution of mitochondria, but it
must be borne in mind that comparatively few organisms have been investi-
gated as compared with the magnitude of the phylogenetic series and excep-
tions may be brought to light at any moment.

DISTRIBUTION IN TISSUES.

Mitochondria occur in all the tissues of both plants and animals, with few
exceptions, wherever protoplasm is active; except, of course, in the plants already
mentioned. They are found in epithelial tissues, muscular tissue, bone, and all
others, except in the terminal stages of cytomorphosis ; and this (according to N. H.
Cowdry) is one of the greatest points of similarity between these granulations in
the plant and animal kingdoms ; that is to say, their progressive diminution and
final absence in the later stages of the life of the cell. It will be discussed in detail
on page 78.

DISTRIBUTION IN CELLS.

Within individual cells mitochondria are, in the vast majority of cases, dis-
tributed indifferently, without definite order, throughout the cytoplasm; but there
are some notable exceptions, examples of which may be seen b.y reference to plate 1.

The distribution of mitochondria with respect to polarity in secreting cells in
animals is most remarkable. In the acinus cells of the pancreas (plate 1, fig. 9)
the mitochondria are most numerous in the basal region next the basement mem-
brane. They are long and filamentous and are, in a general way, distributed parallel
to the long axis of the cell, streaming from the basement membrane toward the
lumen. A similar arrangement obtains in other glands like the vesicula seminaUs

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. 73

(plate 1, fig. 3). Here the polarity, or the direction of secretion, is proximo-distal—
that is to say, from the basement membrane in the direction of the lumen — and is
indicated by the arrow.

The mitochondria are, however, arranged entirely differently in the epithelial
cells of the intestine (plate 1, fig. 10). Champy (1911, p. 109) emphasizes the fact
that here they are sometimes gathered together at both poles of the cell, instead of
being heaped up only in the basal part, as in the pancreas. He interprets this as
meaning that the intestinal epithelial cells are polarized in two directions for
secretion and absorption. This is again represented by the arrows.

In the thyroid (plate 1, fig. 8) mitochondria are not most abundant in the part
of the cells next the basement membrane as in the pancreas, parotid, and other
glands, but are much more numerous in the distal region of the cell next to the
lumen and the colloid substance. Bensley (1916, p. 50), on other grounds, has
concluded that the original proximo-distal polarity of the thyroid cells has been
reversed; that the normal direction of secretion is toward the basement membrane,
blood-vessels, and lymphatics, instead of in the direction of the intrafollicular
colloid as has been generally supposed, and he points out that the mitochondria
are also reversed.

To sum up, in all glands where the polarity is proximo-distal (parotid, plate 1,
fig. 7), the mitochondria are accumulated in the proximal part of the cell; in cells
with double polarity (intestinal epithelium, plate 1, fig. 10), the mitochondria are
heaped up in the distal as well as in the proximal cytoplasm; and finally in cells
with reversed polarity (thyroid, plate 1 , fig. 8) , the mitochondria are gathered together
in the distal part of the cell instead of in the pro.ximal.

In view of these facts one is tempted to inquire whether the clumping of the
so-called batonnets of Heidenhain, which (according to Regaud) are derivatives
of mitochondria, in the proximal parts of the cells of the kidney tubules (plate 1,
fig. 1), means that the original proximo-distal polarity of these cells is maintained
and that secretion takes place from the basement membrane toward the lumen.

The question is, how universal is tliis apparent relation between mitochondria
and polarity? Are the mitochondria ever uniformly distributed in gland-cells
which are polarized? And, conversely, are they ever arranged in this way in cells
which are not polarized? It may be said that this accumulation of mitochondria
in the basal parts of the cells occurs in all zymogenic cells with proximo-distal
polarity which have been investigated in mammals. In addition to the examples
already cited the following may be mentioned: the cells of Paneth of the small
intestine, the chief cells of the stomach, the serous cells of Harder's gland, etc.

I know of no instance in which the mitochondria are uniformly distributed in
definitely polarized gland-cells. It is certainly true that in gland-cells which are
unpolarized, or are but slightly polarized, the mitochondria are more uniformly
distributed throughout the cytoplasm. Thanks to the work of Nicolas, Regaud,
and Favre (1912fl, p. 201), we know the relations of mitochondria inhuman sebaceous
glands where the cells themselves are bodily transformed into the secretion and
are passed out through the ducts. The cells (particularly the older ones) well

74 THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

along in the metamorphosis show no evidences of polarity and contain mitochondria
scattered uniformly throughout the cytoplasmic area. In the thymus gland the
cells are unpolarized and the mitochondria are not arranged in any definite way.
The so-called " rhagiocrine " cells of Renaut are unpolarized and contain mitochon-
dria quite uniformly distributed. Other instances might be cited. We must
regard secretion as a fundamental property possessed by all cells. Where they
speciaUze in it, and give up all their energies to it, and retain their primitive connec-
tions with the basement membrane, the polarity is the most pronounced and the
arrangement of the mitochondria the most striking. It is hard to say whether,
when cells become depolarized (if they ever do), the mitochondria lose their polar-
ized arrangement and become distributed uniformly. This has a very definite
bearing upon the problem of the relation of mitochondria to polarity in gland-
cells. It would seem that the carcinomata would constitute favorable material,
because in some of theni the secreting cells lose in a measure their power to secrete
and often with it their polarity and take on unusual powers of multiplication. At
such times one would expect the mitochondria to become redistributed uniformly
throughout the cytoplasm.

In plants secreting cells are not so generally polarized and the ones which are
polarized have not been investigated, so that we can obtain no help in this direction.

Child's (1915, p. 202) view that, in general, polarity is dependent upon the
axial gradient in metaboUsm, considered with the likelihood that mitochondria
themselves pro])ably play some part in metabolism, is perhaps not without signifi-
cance. There may be some relation here, and there may not — we can not tell.
Tashiro (1915, p. 112) has established, by measurements of the CO-, output, a
gTadient in metabolism of the nerve-fiber which corresponds with the polarity of
the nerve-cell. I have diligently searched in nerve-cells of many forms for some
relation between the distribution of mitochondria and dj'namic polarization and
have failed to find any. The polarity of egg-cells— that is to say, the plane in
which cleavage takes place — ^is apjmrently not related to the arrangement of mito-
chondria. Beckwith's (1914, p. 217) centrifuge experiments have shown that the
plane cuts the mass of mitochondria at any angle, sometimes into very unequal
parts, which is in full accord with Lillie's (1908, p. 907) previous work, showing
that in the egg polarity is dependent upon the organization of the ground substance
itself, not upon the arrangement of any visible granulations within it.

Before closing this subject it may be emphasized that, so far as I know, there
is no valid reason for the general assumption that all forms of polarity, the polarity
of secreting cells, the polarity of nerve and egg cell, etc., are referable to the same
fundamental cause, which some are inclined to do. An axial gradient in metabolism
and the distribution of mitochondria may be factors in one type (i. e., in gland-
cells) without having anything at all to do with the others. I am inclined to think,
however, that the truly remarkable distribution of mitochondria in secreting cells
is much more likely to be one of the results or manifestations of the polarity rather
than the cause of it. It may be entirely a question of mechanics, the mitochondria
being absent from the distal zone, owing to the pressure of secretion therein.

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

75

During cell division the mitochondria are distributed in much the same way
in plants as in animals. They persist during the whole process; they are absent in
the spindle area, whether a. definite spindle be formed or not; and they are divided
in approximately equal amounts between the two daughter-cells. But in minor
respects their distribution varies more in animals. I have reference, for instance,
to the mitochondrial palisade described by Benda (1902, p. 781) in Blaps, which
is (so far as I know) unknown in plants. In animal cells they are almost invariably
disposed in a radial fashion about the centrosome, but such a condition has, to the
best of my knowledge, not been described in plants. Tlois discrepancy may,
however, be due to the well-known absence of a typical centrosome in the angio-
sperms. Kingsbury (1912, p. 45) makes the interesting suggestion that (since
in the terms of Lillie's theory of cell division the centrosome is a negative center)
the mitochondria, being reducing substances, carry a positive charge and accumu-
late around the centro.some in order to discharge it.

In the cells which are not dividing
the mitochondria sometimes heap up _, — ^k\

about the centrosome and sometimes do
not. Their behavior may indicate whether
the centrosome is active.

Fig. 1. — Dividing cells in chick embrjo. Fig. 2. — Meristem and young and old cortical cells of the |iea.

A-E in mitosis and f-j in apparent amitosis. showing primary diffuse arrangement of mitochondria (a),

secondary condensation about the nucleus (b), and final
even distribution throughout the cytoplasm (c). (.Wter
N. H. Cowdry.)

Perinuclear condensations of mitochondria occur in both plants and animals.
In the early meristem of plants, generally, mitochondria are found indifferently
distributed in the protoplasm. They seem to approach and come in actual con-
tact with the nucleus, in which position they enlarge and form plasts which migrate
away from the nucleus and become distributed more or less evenly in the surround-
ing cytoplasm. Guilliermond has repeatedly described this migration and finds
that the mitochondria become more and more resistant to acetic acid during this
process of plast formation. Similarly, in the spermatogonia of certain animals
the mitochondria make their way to the nucleus and become so closely applied to
it that investigators have been deluded into thinking that they actually originate

76

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM,

Fig. 3. — Spermatogonium, spermatocyte, and spermatid
of mouse, showing (a) similar diffuse arrangement of
mitochondria, (b) secondary circumnuclear condensation,
(c) final even distribution in the cytoplasm. (After
N. H. Cowdry, 1917.)

from it. In the later stages of spermatogenesis they leave the nucleus, becoming
more resistant to acetic acid, as Regaud (1910, p. 298) has shown. Indeed, the
parallelism is very close (figs. 2 and 3). The Lewises (1915, p. 349) have observed
mitochondria journeying to the nucleus and back again in the living cells of tissue
cultures. Here their movements are much more rapid and no change in com-
position is evident. What their mission is we have no idea.

They also gather together in the peripheral cytoplasm, just beneath the cell-
membrane, especially in animal cells. This arrangement is very pronounced in
egg-cells and it has often been alluded to by Van der Stricht (1909, plate 1) and
his pupils. It is a general phenomenon for
which there must be some explanation.
After a time the mitochondria become
redistributed, just as in the perinuclear
condensations. Curiously enough, other
cells, like gland-cells, rarely if ever show
it. An interesting reversal of this condi-
tion is seen in certain pathological con-
ditions where the mitochondria quit the
peripheral cytoplasm and become heaped
up about the nucleus instead.

Both perinuclear condensations and peripheral condensations frequently occur
in one and the same cell, as in the ascidian eggs described by Loyez (1909, p. 192).

In ciliated epithelial cells the mitochondria are often permanently heaped up
in the region of the cytoplasm just beneath the ciliated border. I have found that
this is the case in the ciUated cells of the epididymis of the white mouse. This
fact, together with the familiar clumping of mitochondria about the axial filament
in the tail of the spermatozoon and the feeling that they are transformed into
myofibrils, led Benda to the conclusion that they play a part in the motor activities
of the cell.

The clumpi7ig and fusion of mitochondria to form other substances is to be
regarded as a very special instance of modifications in their arrangement. A dis-
cussion of this process logically falls under the heading of histogenesis. It may be
simply mentioned here that we come across it in the formation of the nebenkern,
the spiral filament, certain portions of the rods and cones of the retina, and other

structures.

It is unnecessary to go into a discussion of other minor variations in the
arrangement of mitochondria, dependent upon the deposition of substances in the
cytoplasm, upon j^ressure, and other obvious causes.

In all these journey ings of mitochondria to and fro, and in these transitory and
permanent condensations and fusions, not a shred of evidence can be seen that they
possess powers of independent motility like bacteria. The prevalent belief that
they do possess these powers seems to be simply a relic of the old conception of
Altmann that they are elementary organisms endowed with all vital properties,
just as the idea that in all cases they arise from other mitochondria by longitudinal
or transverse division persists under the guise of the misleading doctrine of mito-

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. 77

chondrial continuity. A very careful study made by N. H. Cowdry, in this labora-
tory, of mitochondria in the streaming protoplasm in many varieties of plant-cells
(where they may be easily followed unstained) failed to reveal any indication of
independent motility on their part. True, they dart from place to place in the
cytoplasm, but this may in almost all cases be referred to definite currents and
eddies in the stream. Moreover, I have examined mitochondria vitally stained
with janus green in human polymorphonuclear neutrophile leucocytes during
amoeboid movement and phagocytosis and I have likewise failed to detect any sign
of independent motility. The movements can not, however, be entirely explained
away on the basis of currents in the cytoplasm, because we have no reason to sup-
pose that such currents exist in the egg-cells. They seem to be almost purposeful.
They are entirely different from Brownian movement, though the mitochondria do
exhibit true Brownian movements when the cells take up water and when the
balancing action of bombarding molecules is upset. The consistency of the cyto-
pla.sm has little to do with it, because peripheral condensations occur in nerve-
cells as well as in egg-cells; in nerve-cells the cytoplasm is very viscid, in egg-cells
very fluid. The mitochondria may be easily thrown down with the centrifuge in
egg-cells, but not in nerve-cells.

It may be a question of adsorption. The Gibbs-Thomson principle tells us
that any process which diminishes free energy at an interface will tend to take place.
Now, the mitochondria are nothing less than minute particles of lecithin-hke ma-
terial in suspension, and we know that hpoids decrease surface tension, so that we
would naturally expect them to be heaped up at the nuclear and plasma membranes;
but it is difficult to explain their active migration to and fro. Perhaps we are deal-
ing with electrical adsorption. The mitochondria may carry a charge, but no
adsorption could take place without the presence of a charge of the opposite sign
upon the nuclear or plasma membrane, as the case may be, and of this we have no
information whatever. Here again the movements backward and forward are the
stumbling-block.

There is one more point for which a tentative explanation may be advanced.
It will have been noticed "that in animal cells rather more variations seem to be
met with in the arrangement of mitochondria than in plant cells. This may be
correlated in some way with the fact that animal cells are more generally polarized;
I mean polarized for irritability, conduction, secretion, contraction, and so forth,
properties which do not play so great a role in the life of plants where separate
regions of the cell are not so distinctly marked off.''^ The division of labor among
animal cells is greater than in plants. They have to perform a great variety of
functions under different conditions, so that in a single animal there is far greater
diversity of organization among its cells than in a single plant. This the mito-
chondria reflect.

From a practical point of view it is a very simple matter, with comparatively
little experience, to tell by the inspection of a cell whether the distribution of mito-
chondria within it differs from the normal, and this is another indicator of cell
activity and of cell injury which has received but scant attention.

IN. H. Cowdry, 1917, p. 215.

78

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

VII. AMOUNT.

AMOUNT OF MITOCHONDRIA IN PHYLOGENY.

What is true in the case of morphology holds also here. In man, as an
example of a multicellular organism, with great division of la}-)or among his cells,
variations in amount of mitochondria occur because some cells are best fitted to
l^erform their duties with much and others with little. From our general informa-
tion— we have no specific measurements — we can not say that the protoplasm of
higher animals differs from that of the lower ones in the amount of contained mito-
chondria. Neither is there any noticeable difference in amount between animals
and i)lants. Their ai)parent absence in the myxomycetes, schizomycetes, and most
of the algae has already been referred to on i)age 72.

AMOUNT OF MITOCHONDRIA IN ONTOGENY.

No definite measurements have been made. It may be said, however, that
in the very young embryo the cells usually (p. 81) contain approximately the
same amount of mitochondria. As development proceeds toward maturity the
different tissues become specialized and distinctive differences in the amount of
mitochondria often become apparent. It is possible, but improbable, that there
is any noticeable difference in the number of mitochondria in actively functioning
cells of young and old animals, but this contingency should be borne in mind until
decisive information is forthcoming. There is no reason to believe that mito-
chondria vary in amount with sex.

AMOUNT OF MITOCHONDRIA IN CYTOMORPHOSIS (SENESCENCE).

Mitochondria progressively decrease in number and finally disappear entirely
in the later stages of the life of the ceU.

iiii\

Fig. 4. — Meristem and parenchyma cells of the bean, show-
ing the progressive disappearance of mitochondria and the
formation of plastids containing chlorophyll. (After
Guilliermond, 1912, modified.)

Fig. .5. — Erythroblast, megaloblast, normoblast, and erythro-
cj-tes from bone marrow of a rabbit stained supravitally
with janus green, showing the parallelism between the
disappearance of mitochondria and the appearance of
hemoglobin in the form of a diffuse deposit. (After N. H.

Cowdry, 1917.)

I refer to the decrease in number of mitochondria in plant cells which runs
parallel to the formation of chloroplastids. It is said that when the plasts are
fully formed few if any mitochondria remain (Guilliermond, 1912a, pi. 17), and
these are mature and highly differentiated cells. In animals there is a similar
disappearance of mitochondria in the life cycle of red blood-cells. In the young,
nucleated forms, as they occur in the bone marrow, they are very abundant; but
they become less and less so as the cell differentiates. A few persist after the

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. 79

nucleus is lost, but in the senile forms, present in the circulation in man, mito-
chondria are entirely absent. In jilants this disapi)earance is associated with the
production of chloro])hyll, in animals with the formation of hemoglobin, two sub-
stances with strikingi}' similar chemical constitution; in both it is gradual and
progressive and runs parallel with an increase in the degree of differentiation and
with the age of the cell, general metabolism being diminished and special functions
being accentuated (see figs. 4 and 5).

This is but a single instance of a very widespread phenomenon which attracts
attention only in those cells which normally die and are replaced in large numbers,
collectively, in the life of the individual, hke the cells of the epidermis. It is not
without significance in any theory of senescence. Senescence is now thought to
result from excessive differentiation with the heaping-up of relatively inert mate-
rials in the cell, which clog the vital processes and proportionally diminish greatly
the volume of active cytoplasm. Child pictiu-es these substances as large col-
loidal complexes, and Burrows (1917, p. 339) thinks that they are retained by virtue
of their relative insolubility. Evidently, we have to deal also with a diminution
in the mitochondria, but whether it is the cause or the result of the condition we
can not tell.

AMOUNT OF MITOCHONDRIA IN DIVIDING CELLS.

It would appear that there must be some increase in mitochondria if the daugh-
ter-cells contain the normal amount and if their volume, taken together, exceeds
that of the parent cell. Apparently, the increase is progressive and proportional
to the volume of the cytoplasm. I have made (1914c, p. 102) a careful study of
1,000 dividing cells in cliick embryos and at no stage in the process is it possible to
say that the number of mitochondria in the cytoplasm is relatively greater than
in the neighboring cells of the same type. This is confirmed by the Lewises
(1915, p. 371).

Usually the mitochondria are distributed in approximately equal amounts to
the two daughter-cells, but the distribution is a haphazard one, depending only
upon the arrangement of mitochondria in the parent cell. An admirable instance
of unequal division is to be found in the cleavage of the ascidian eggs as described
by Duesberg (1917, p. 481). In my opinion the great differences in the amount
of mitochondria m the tissues of the embryo are determined by the physiological
condition (p. 82) of the cells rather than by whether or not many mitochondria
had been handed down to them from some remote cell ancestor.

AMOUNT OF MITOCHONDRIA IN DIFFERENT CELL TYPES.

Striking variations in the amount of mitochondria obtain, but unhappily
there has been but Uttle attempt made to distinguish between absolute and rela-
tive variations. The obstacle is particularly exasperating in gland-cells, and this
is just where the demand for information is most insistent, because the cycUcal
changes in the volume of secreting cells can not be ignored. In all tissues the
estimation of mitochondria should be carefully controlled by measurements of

80

THE MITOCHONDRIAL CONSTlTlTENTS OF PROTOPLASM.

volumes, especially in experiments. This difficulty is almost insurmountable in
muscle-cells, but it is comparatively negligible in nerve-cells.

Accordingly, Thurlow (1917, p. 37), working in this laboratory, selected nerve-
cells for study. She managed to enumerate the mitochondria with surprising accu-
racy by inserting in the ocular a glass disk on which a square of known dimensions
had been ruled. Using a 1.5 mm. apochromatic objective and sections of known
thickness, the number of mitochondria per cubic millimeter of cytoplasm was
easily calculated. Observations were confined to the cells of the nuclei of the
cranial nerves, because they may be most readily referred to the different func-
tional types. She carefully controlled her counts and found that the experimental
error was never more than 1.3 per cent. She found that there is a constant num-
ber of mitochondria per unit volume of cytoplasm in normal nerve-cells of the
corresponding cranial nerves of different white mice; further, that the constant
differs for cranial nerves of different types, so that certain groups of nerve-cells can
be distinguished by the number of mitochondria within them. The amount does
not depend upon whether the cells are sensory or motor in character. A casual
inspection of the sections showed at once that the mitochondria are most abundant
in the cells of the mesencephalic nucleus of the fifth nerve and least numerous in
the nucleus of the tenth nerve. The counts disclosed that there are 284,378,159
per cubic millimeter in the former and 178,210,313 per cubic millimeter in the latter.
She examined all the cranial nerves, this being only a case in point.

It is highly desirable that these studies should be
extended to other varieties of nerve-cells in the central
and peripheral nervous systems, and the study of late
embryos and young animals might tell us when these
remarkable differences in number become first manifest,
for we know that in undifferentiated nerve-cells the
mitchondria are fairly uniform in number. We want,
of course, to discover whether or not these differences
in amount arise at the time of functional maturity ; but
whichever way it turns out we shall have obtained a
much-needed clue to their significance. Furthermore,
this work supplies us with a new criterion of nerve-
cell changes which has the rare merit of being quantita-
tive, and, as such, may well be compared with the nucleus
cytoplasmic ratio which we owe primarily to Hertwig.
The quantity of mitochondria does not depend at
all upon their form. They niay be heaped up in the
condition of granules, rods, or filaments. Similarly,
when the mitochondrial content is reduced to a mini-
mum either form may predominate.

It may be argued that these counts of mitochon-
dria do not give us any real information about the cells
in question, because tissues, kidneys especially, which

Fig. 6

-Spinal ganglion cells of the
pigeon, showing (a) great increase in
the amount of mitochondria as con-
trasted with the average amount (b).

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

81

show no evidence of fatty substances on histological examination may neverthe-
less actually contain a large amount as revealed by chemical analysis, and vice
versa. The mitochondria are phosphatids, not neutral fat. It may also be urged
that the cell contains a large variety of phosphatids of different solubilities and
that our mitochondrial methods bring to light only those of a certain kind. This
is undoubtedly true, but it is not the whole story, because our technique reveals
all the phosphatids which occur in definite form and may be seen in the living cell,
so that we are studying a constant, not a variable thing.

A word of caution in connection with the interpretation which we may justly
place upon variations in the amount of mitochondria: In many pathological
conditions there is a deposition of lipoids within the cell, the so-called lecithin
metamorphosis. These must be carefully distinguished from true mitochondria. In
other cases deposits of neutral fat occur which may be detected by the fact that
they reduce osmic acid. Moreover, these substances almost invariably occur as
spherules, very rarely as filaments, as is usually but not always the case with the
mitochondria.

The total absence of mitochondria in normal actively functioning cells which
are not senile (p. 79) is of considerable interest, though it is in all probability
a phenomenon of very rare occurrence. Most of the examples recorded have,
on further examination, proved to be erroneous; for instance, the parietal cells of
the stomach, myeloblasts, the cells of malignant tumors, and others. I have never
succeeded in finding mitochondria in tissue mast-cells, but I am inchned to think
that this may simply be due to the acidophilic mitochondria being obscured by the
densely packed basophilic granulations. They are present, however, in blood
mast-cells. The cells in the kidney of
snakes, described by Regaud (1908a, p. 17),
and the cells of the glomeruU of embry-
onic human kidneys (PoUcard, 19120, p.
442; 12/, p. 12) have never been found to
contain mitochondria. Branca (1911, p.
559) has failed to find them in the in-
ferior or germinative zone of hairs. Other
cases of the apparent absence of mito-
chondria may be cited.

In neighboring cells of the same type,
which usually contain approximately the
same amount of mitochondria, we occasionally come across variations in amount
from cell to cell which are difficult to explain. One cell sometimes contains mito-
chondria in tremendous excess of the normal amount present in the cells on either
side of it. The dilTerence may be so striking as to lead us to think at first sight
that we are dealing with a different kind of cell altogether. In others there may be
a decrease. The decrease is,. I believe, often indicative of bad technique, just as
we have learned to attribute to the same cause the artificial polymorphism of
mitochondria in cells where they are normally alike. The increase I am unable

Figs. 7, 8, and 9. — Neural tube, myotome, and Wolffian
duct, showing cells apparently normal yet devoid of
mitochondria.

82 THE MITOCHONDRIAL CONSTIl UENTS OF PROTOPLASM.

to explain. It is very commonly met with in nerve-cells and in gland-cells (see
fig. 6). In the cells of chick embryos I find that a decrease in the mitochondrial
content is quite common (figs. 7, 8, and 9), but an increase above the normal
very rare.

SIGNIFICANCE OF VARIATIONS IN THE AMOUNT OF MITOCHONDRIA.

We can say at once that such variations do mean something, because even if
the mitochondria are as inert as iron filings their presence in such variable amounts
must surely exercise some influence upon the activity of the cells containing them,
and we have good reason to think that they are not chemically inert substances.

There are at least two series of observations to be explained. In the first
place, the association of abundant mitochondria with intense protoplasmic ac-
tivity. In cytomorphosis, for example, they are especially numerous in the active
stages in the life of the cell and they diminish with senility in both plants and ani-
mals. There is a sharp increase in mitochondria with regenerative activity, in
compensatory hypertrophy, and in many other conditions. In the second place,
there is a distinct reciprocal relationship between the amount of mitochondria and
the amount of fat. Where there are few mitochondria there is much fat, and
vice versa. Decreased oxidation favors the accumulation of fat and increased
oxidation favors its elimination, which suggests at once some connection between
the amount of mitochondria and oxidation; and their abundance in the active
stages of the hfe of the cell, where protoplasmic respiration is rapid, jioints to the
same conclusion. Further evidence of a convincing nature has been detailed
elsewhere (p. 134), and it seems probable that normal variations in the amount of
mitochondria are in some way dependent on variations in the respiration of the cells
containing them.

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. 83

VIII. CHEMISTRY.

THE CONSTITUTION OF MITOCHONDRIA.

It is an interesting and rather unusual occurrence, in the study of mitochondria,
for three independent Unes of investigation to yield similar results, yet Regaud
(1908d, p. 720), in the first place in the study of mammalian tissues, Faure-Fremiet
(1910a, p. 622), who worked on protozoa, and the botanist Lowschin (1913, p. 203;
1914, p. 269) have all arrived at the conclusion that mitochondria are, chemically,
a combination of phosphoUpin and albumin, which, in itself, speaks very strongly
in favor of the unity of the class of granules under consideration. The evidence is

briefly this :

(1) Mitochondria are almost completely soluble in alcohol, chloroform, ether,
and dilute acetic acid. They are rendered insoluble by chromization. They are
not doubly refractile and they do not stain with either Sudan III or Scharlach R.
They are only sometimes blackened with osmic acid.

(2) It is" said that part of the mitochondrial substance is not soluble in these
fat solvents and it is supposed that this portion is albumin (see also Bullard, 1916,
p. 26), for formahn and bichromate, which are used as fixatives for mitochondria,
are energetic coagulants of albumin. IVIillon's reagent is the only color-test for
protein which can be satisfactorily apphed to material in section (the xanthoproteic
reaction may also be used, but it is less satisfactory because it is more destructive).
I learn from Dr. R. R. Bensley that the mitochondria do not give a definitely posi-
tive Millon reaction in comparison with the strong Millon reaction which is given
by such cytoplasmic structures as the zymogen granules. Even if there were a
change in color in the mitochondria it might not be of sufficient intensity to be
appreciated in filaments of such extreme fineness as mitochondria (0.2 micron in
diameter) embedded in a colored cytoplasm. I have obtained no success with the
xanthoproteic reaction. Mitochondria do not give any of the color reactions of
polysaccharides.

(3) Artificial mitochondria have been made by Lowschin of lecithin in differ-
ent salt and albumin solutions (resulting in the formation of lecithalbumin), which
apparently present the same form and solubilities as true mitochondria. They
form granules, rods, and filaments which, he claims, multiply by division. He
embedded them in glycerin-gelatin, fixed them, and found that they stained in
the usual way by the various mitochondrial methods.

(4) The temperature solubiUty of mitochondria may also be significant. It
has been discovered by Policard (1912rf, p. 229) in the case of animal tissues and by
N. H. Cowdry (1917, p. 220) in plants that the mitochondria are soluble at a tem-
perature from 48° C. to 50° C, while the other parts of the cells remain practically
unaffected. Phosphatids have a low melting-point also.

(5) Apparently the specific gravity of mitochondria is somewhat greater than
protoplasm (Faure-Fremiet, 1913, p. 602). This is determined by the centrifuge
method. If they are thrown down they are said to be of high specific gravity.
If the protoplasm is in the physical condition of a "gel" rather than a "sol," as in

84 THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. .

the nerve-cell, the distribution of the mitochondria is unaltered by centrifuging
(KeyO- There is no reason to believe that the mitochondria themselves are
different. At any rate, where the method is applicable ({. e., in egg-cells) the
mitochondria are heavier than protoplasm, in which respect they conform to what
we know of phosphatids and differ sharply from oils and neutral fats, which rise to
the surface and float instead of being thrown down.

(6) Mitochondria act as solutes for various substances. They are often
pigmented and assume the most brilliant hues. Prenant (Asvandourova, 1913,
p. 293) has actually styled them " chromochondria " on this account. This solu-
tion of other materials in mitochondria is particularly frequent in plant cells. It
may or it may not be significant from the point of view of their constitution.

(7) There seems to be a certain correspondence between variations in the
histological picture of mitochondria and the variations in the phospholipin content
of the same organ on chemical analysis. Thus Mayer, Rathery, and Schaeffer
(1914, p. 612) have been able to alter the mitochondria experimentally in liver-
cells. In stages with more mitochondrial substance, chemical analysis showed an
increase in phosphorized lipoid; in stages with less, a diminution. Faure-Fremiet
(19126, p. 347) has extracted from the ovaries and testes of Ascaris a phosphatid
with properties identical with those of mitochondria in the cells of these organs.

(8) Russo (1912, p. 215) has apparently been able to increase the number of
mitochondria in the oocytes of the fowl by the injection of lecithin. R. Van
der Stricht (1911, p. 435) found that there are two different kinds of eggs in the
cat, one containing much vitellus and the other containing only a small amount;
and, further, that, following intraperitoneal injections of lecithin, the relative num-
ber of female offspring increased noticeably. In the normal condition 62 per cent
are males, while after treating in this way only 23 per cent are males. That is to
say, the administration of lecithin increases the amount of deutoplasm in the eggs,
increases the number of eggs with much deutoplasm as contrasted with those with
small amount, and in this way increases the percentage of females in the offspring.
While this is of great interest in the determination of sex, and will be discussed
in that connection, it is also of importance as an indication of a possible relation-
ship between the amount of mitochondria and the phosphatid lecithin. The re-
searches carried on about the same time by Whitman, and subsequently by Riddle,
in the determination of sex in pigeons, are in complete accord with these observa-
tions of R. Van der Stricht. Riddle (1916, p. 389), in summarizing the results
of a long series of studies, points out that, in the first place, the eggs of late summer
and autumn produce mostly females and that their yolks are larger than those of
the spring, which give rise chiefly to males; and secondly, that old, "overworked"
females tend to produce female offspring earlier and earlier in the season, and that
this, also, is correlated with larger egg-yolks. His chemical analyses showed that
the storage metabohsm is higher and the water-content lower in these female-
producing eggs than in those which give rise to males. The general conclusion,
of course, is that sex is conditioned by variations in rate of metabohsm, which is

'Personal communication.

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. 85

of interest when we remember that all our evidence points to the conclusion that
mitochondria are concerned in metabohsm. This work of Riddle seems to give us
another chance to correlate the amount of mitochondria demonstrated histologically
with the results of chemical analysis as well as with variations in physiological
behavior. It must be borne in mind that the deutoplasm and yolk are not neces-
sarily mitochondrial, but the substances out of which they are built up are phos-
pholipins and resemble the mitochondria very closely in some important respects.

VARIATIONS IN CONSTITUTION.

In some varieties of cells the constitution of the mitochondria apparently differs,
slightly but noticeably, from that of the mitochondria in other cells, though in
cells of the same kind their composition is very constant. It was at first noticed,
in a casual way, on the examination of preparations, that the mitochondria were
occasionally preserved in one kind of cell and were lost, or imperfectly fixed, in
others. While it is possible that this may be due to some difference in the cells
themselves, it is far more likely that we are really dealing with a true variation in
the solubility of the mitochondria. In the staining of fixed tissues, also, one occa-
sionally meets with differences in the reactions of the mitochondria to the stain.
Similarly, the mitochondria in different varieties of cells color with janus green
somewhat differently, some more easily than others. Many investigators have
also found that some mitochondria blacken or turn gray with osmic acid, although
others do not. But observations such as these are unsatisfactory in the extreme,
because they can not be controlled and because there is nothing quantitative about
them.

We owe our first detailed information to Regaud (1910, p. 295), who very
carefully compared the solubilities of mitochondria in the different cells of the testis
with respect to acetic acid. He found that there is a progressive increase in their
resistance to acetic acid as one passes from spermatogonia to spermatozoa. Some
years later Nicholson (1916, p. 336) applied the same methods of technique to the
central nervous system of the white mouse and found that there also the mito-
chondria in certain varieties of cells presented different and characteristic solubili-
ties in acetic acid. This will have to be done in other organs and with a large
variety of solvents before we shall be able to arrive at even a glimmering of the
variations in the constitution of mitochondria which occur in different types of
cells.

Variations in the constitution of mitochondria in the course of histogenesis
are often quite pronounced. We have all had the experience that technique
which gives good results when apphed to adult tissues is frequently very disap-
pointing for embryos. This may be due to factors other than a difference in the
mitochondria — to changes in the water-content, for example. Here we have no
detailed observations on the mitochondria themselves to fall back on; they are
much needed. If we exclude the cases of the transformation of the mitochondria
into other substances, I think that we may say that there are no very great differ-
ences between the mitochondria in these different stages of histogenesis. Cer-

86 THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

tainly the variations are not so great as are those between the mitochondria in
different varieties of cells. Fundamentally the problem is the same, because the
variations between different types of cells arise throughout the course of histo-
genesis.

In phylogenesis also we have to deal with certain variations in the constitution
of mitochondria, but here it is greater. The technique which we use has to be
specially adapted to almost every form that we study. There is undoubtedly
some difference in the mitochondria. It is not all a question of their surround-
ings. A case in point is that of Planaria, in which I have attempted, again and
again, to identify the structures which Korotneff (1909, p. 1010) has described
under the heading of "Mitochondria." I have found that they are satisfactorily
fixed in mixtures like acetic sublimate, which destroy mitochondria in mammalian
tissues; and, conversely, that the methods which I have been accustomed to use
for higher forms fail completely. This experience has impressed upon me the
danger wliich lies in arguments that since the mitochondria do something in one
type of cell thej' must necessarily do it also in another.

N. H. Cowdry (1917, p. 217) has made the first detailed comparison of the
microchemical reactions of plant and animal mitochondria. His general conclu-
sion (p. 225) is as follows:

"We have every reason to suppose that their chemical composition is much the same
in both plants and animals, but here our knowledge is for the most part supposition and
inference, since direct chemical analyses are obviously out of the question. Their com-
position, as indicated by solubility with respect to acetic acid, heat, and other reagents,
is certainly subject to similar variations in both."

We know very little of variations in the constitution of mitochondria in differ-
ent physiological conditions, though they probably occur. The only reference
known to me of work along this line is Policard's (1912c?, p. 229) on the tempera-
ture solubihty of mitochondria in kidney-cells, in which he makes the unqualified
statement that the temperature solubihty varies with the state of physiological
activity. This should be confirmed.

In conclusion, we may say that slight variations do occur in the chemical con-
stitution of mitochondria in different varieties of cells, in the course of histogenesis
and phylogenesis, and in different physiological states, though we do not know
their nature or extent. They are the exception rather than the rule, but they
must nevertheless be kept in mind when we venture to argue by analogy.

REACTION TO JANUS GREEN.

Michaehs (1899, p. 565), while making a detailed study of the beha\dor and
chemical nature of vital dyes, found that janus green stained certain filaments in
gland-cells specifically. The janus green was obtained from the Farbwerke Hoechst
Company. He called it diethylsafraninazodimethylanilin, and gave it the formula
shown in figure A. He was careful to emphasize that a slight alteration in the
composition of the janus green alters the specificity of the stain, for he found

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. 87

that dimethylsafraninazodimethylauilin did not stain the filaments specifically.
He found that, on reduction of the dye in the tissue itself or in the test-tube, the
dimethylanilin group splits off, leaving the red diethylsafranin. According to him
the diethylsafranin formed in this way does not stain the filaments specifically.

(C2 Hs) N

-N= N-/\

I J N(CH3l

Fig. a.

Laguesse immediately recognized the importance of Michaelis's discovery,
obtained some janus green from Griibler (not Hoechst), and used it in his inves-
tigations on gland-cells. It stained some structures which he called "vermicules"
and "ergastidions." His results, however, were not uniformly satisfactory, prob-
ably on account of the fact that he obtained the dye from the wrong source.

Accordingly, janus green was soon forgotten as a vital stain for mitochondria,
and it remained for Benslej' (using the janus green of the Farbwerke Hoechst Com-
pany) (1911, p. 304) to revive interest in the discovery of Michaelis

The reaction of mitochondria to janus green may be conveniently described
in three stages: the staining of mitochondria with janus green, the reduction of the
janus green with the formation of diethylsafranin, and the production of the leuco-
base.

If one injects the pancreas of a guinea-pig through the blood-vessels with a
solution of 1:25,000 janus green in 0.85 per cent sodium-chloride solution, the
mitochondria become intensely stained in the course of 15 minutes or more. The
other structures in the cell, like the nucleus and zymogen granules, remain lui-
colored unless the stain is applied in too great concentration for too long a time.
The staining is facilitated by exposing the pancreas to the air.

If, now, portions of the pancreas are mounted in salt solution on a slide and
are covered with a cover-glass, the dye is slowly reduced by the tissues, with the
formation of diethylsafranin (see, however, Michaelis, 1902, p. 101), which has
a bright pink color. The change first takes place in the more central parts of the
tissue remote from the oxygen of the air and surrounding salt solution and proceeds
slowly towards the periphery. It is hastened by ringing the preparation around
the edge of the cover-glass with vaseline, which excludes the air and prevents
evaporation. The faintly green-stained ground substance first changes to pink
before the intensely stained mitochondria are affected. The change passes like
a wave across the tissue from the center to the periphery. The mitochondria in
the part of the cell nearest the middle of the j^reparation change their color first.
Those in the remainder of the cell then follow suit. There is no evidence that
mitochondria of different sizes change at a different rate, or that different parts of

88 THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

the same mitochondrion differ in this respect. In this way the entire cell assumes
a pink color, but the mitochondria are most intensely stained.

The next change is a further reduction of the diethylsafranin to the leucobase.
The tissue bleaches. If the preparation is viewed with a low power, the perijihery
of the piece of tissue in contact with the surrounding fluid is colored green (stage 1) ;
nearer the center a band of pink is seen (stage 2); the center itself is color-
less; so that all stages in the reaction may be seen at one time. When the leuco-
base is once formed it is impossible to restain the mitochondria with either diethyl-
safranin or janus green.

The reaction of mitochondria to janus green is essentially similar in other
tissues. I have elsewhere (19146, p. 276) described it in detail in human lymiiho-
cytes, which may be contrasted with pancreas-cells, since they float freely in a
fluid medium.

All my attempts to make permanent preparations of these vitally stained
mitochondria, made on the basis of the fact that both the picrates and molybdates
of janus green are relatively insoluble in alcohol, proved futile.

Frozen sections may be made of the tissues stained with janus green, but the
freezing generally brings about a destruction of the mitochondria.

Certain conditions may be recognized which tend to inhibit the j anus-green
reaction. The mucus that is secreted by Planar ia and other invertebrates stains
intensely with janus green, but the mucus secreted by leeches does not prevent
staining of mitochondria in spermatogenesis. In this way a large amount of janus
green is removed from the solution, with the result that the concentration of the
solution in actual contact with the cells is much weaker than that originally applied.
Moreover, masses of mucus often surround the cells and form barriers against the
diffusion of the dye. Albumin acts in a somewhat similar way. When present in
sufficient amounts, as in the blood and tissue juices of Linmlus, Callinectes, etc.,
it becomes coagulated and stained by the janus green. It then presents a very
confusing picture which is very likely to lead the unwary astray. A high con-
centration of sodium chloride (3 per cent) or potassium nitrate (3 per cent) will
prevent the janus green from acting. The greatest obstacle, however, to the gen-
eral use of janus green as a vital stain for mitochondria in all cells is its very limited
power of penetration. For this reason the staining of mitochondria in intact brain-
cells is very diflRcult. Here, however, another factor enters in, namely, the rapid
reduction of the dye to its leucobases by the reducing action of the tissue itself,
so that when the skull is opened up the brain is found to be colored red instead
of blue. I have attempted to prevent this reduction by bubbling pure oxygen
through the stain as it was being injected, without much success. It does not help
matters to remove the skull-cap. Vasodilators are of no assistance in bringing
a larger volume of solution in contact with the cells.-

Conversely, other conditions favor the janus-green reaction. Of these, ready
permeability of the cells is by far the most important. It is highly desirable that
the janus green should be brought into immediate contact with the cells them-
selves. This indeed is the reason why the best results are always obtained in blood-

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. 89

cells and in protozoa, and in organs which may be stained by the injection of janus
green through the blood-vessels.

Janus green occasionally stains other structures in addition to mitochondria,
of which a few may be enumerated: cement substance between the cells of the
testis, lipoid inclusions, blepharoplasts, chromosomes, granules of islet-cells, etc.
Rut it is important to note that it is specific for mitochondria when used in great
dilution, 1 : 500,000, for instance.

The toxicity of janus green varies in different cells of the same animal and in
(lifTerent animals. The same is true in plants. Fresh and old solutions of the dye
have the same toxicity. Lewis found that in cultures of heart-muscle cells the
mitochondria stained with janus green while the muscle continued to beat. Ship-
ley (1916, p. 441) found that trypanosomes retained their motility for some time
after the mitochondria in them were specifically stained with janus green.

I have been able to stain the mitochondria with janus green in the polymor-
phonuclear leucocytes of man and several other vertebrates. Such leucocytes,
with their mitochondria stained, move around in an amoeboid fashion. The tip of
the pseudopod is generally free of both mitochondria and the specific granula-
tions. It is followed by a mass of mitochondria and granules. The nucleus gen-
erally comes last. Cells stained in this way engulf foreign particles and show no
deviations from their normal behavior. They will continue to do this for from
half an hour to an hour, depending upon the temperature of the warm stage, the
rate of evaporation, etc. Their rate of disintegration is not greatly accelerated by
the action of the stain.

When applied subcutaneously it remains localized at the point of injection.
Intraperitoneal injection causes death in a short time and intravenous injection
in a very few minutes. Feeding experiments gave no results and attempts to
increase the tolerance of protozoa to it proved futile.

To what is the toxicity due? There are a number of possibilities. It may be
due to the presence of some toxic material used in the manufacture, but rather
against this is the fact that different samples of janus green B possess the same
toxicity when applied to the same type of cell. The toxicity varies, however, mark-
edly in different kinds of cells, and it is not a question of penetration. The idea of
obtaining, in some way, a completely non-toxic janus green is an inviting one, but
janus green can never approximate to the azo dyes like trypan blue in this respect,
for the reason that it quickly undergoes chemical change in the body, with the
liberation of a number of potentially toxic substances.

Before taking up the specificity of the janus-green reaction, it will be of
interest to bring out a few more facts about janus green itself.

The name may or may not be associated with the Roman deity Janus, gen-
erally represented with two faces looking in opposite directions. The dye certainly
shows two colors, green and red, of quite opposite character.

According to IMichaelis (1902, p. 51), janus green is made by (he action of
nitric acid on safranin, one of the two amido groups being changed into the diazo
group. To this dimethylanilin is added according t(i the following eciuation:

90 THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

Diazo-diethylsafranin Dimethylanilin

N

(Cs H5)j N- \y\^\/ - N = NOH

+ /N - N (CHs)!

Fig. B.

Diazingriin (janus green) :

Fi<i. C.

Diethylsafraninazodimethylanilinchloride:

N

(CaHsJN- Ky\./

-N=N- r 1 N

i CI

(cHj),

Fig. D.

This is the janus green B of the Farbwerke Hoechst Company. It is the same as
"Diazingriin" and may be the same as "Halbvollgriin," but of this I am not cer-
tain. There are two other janus greens of different formula;, which makes three
varieties of janus green in all:

(1) Janus green (Griibler), safraninazodimethylanihnchloride:

N

HjN

N = N

N (CH,),

Fig. E.

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

91

(2) Janus green C (Farbwerke Hoechst Company), dimethylsafraninazodi-
methvlanilinchloride :

N

(CH4),N \y\/\y N = N

Fio. F.

N (CH,),

\y

(3) Janus green B (Farbwerke Hoechst Company), diethylsafraninazodi-
methylanilinchloride :

N

/^/\/\

(aH,),N

N = N

N(CH,),

\y

Fig. G.

The specificity of the reaction is shown by the fact that only tlie latter, janus
green B, the one originally recommended by Michaelis, will stain mitochondria,
though the others differ only in the substitution of an H., or (CH,,)-, in the place of
the (C2H5)2 group. The presence of the diethyl group in the safranin molecule is
evidently the determining factor. Compounds containing two ethyl groups are
more basic than those containing two methyl groups, and the compounds con-
taining two methyl groups are, in turn, more basic than those with two hydrogen
atoms alone, so that there is a decrease in basicity as we pass from the diethyl to
the dimethyl and to the dihydrogen. This may well e.xplain the difference in the
behavior of these janus greens toward mitochondria. One would expect to find
that janus green CJ, with the two methyl groups, would color the mitochondria
better than the janus green of Grtibler, which possesses only the H., group, but I
have failed to detect any difference between the two. Both of them occasionally
stain mitochondria, together with other cell structures, when used in relatively
high concentrations. Nevertheless, the difference in the actions of the janus green
B with the two ethyl groups and the other two dyes, together with the difference
in basicity between them, would suggest that the dye actually combines chemically
with the mitochondria and that the staining is not simply a process of selective
absorption. Our evidence, however, is too scanty to permit us to arrive at any
conclusion. The poor results obtained with some samples of the janus green are
probably due to admixtures of the first and second varieties.

92 THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

The properties of janus green B are: (a) in water, blue solution; (b) on addi-
tion of hydrochloric acid, soluble blue precipitate; (c) on addition of caustic soda,
black precipitate; (d) in concentrated sulphuric acid, olive-green solution, on dilu-
tion becoming green, then pure blue.

The azodimethylaniUn has but Uttle to do with the specificity of the reaction,
because the diethylsafranin alone will stain the mitochondria more or less specifi-
cally. Moreover, I have prepared the safranin from the janus green of Griibler,
the dimethylsafranin from the janus green G of Hoechst, and the diethylsafranin
from the janus green B of the same firm, and I find that the diethylsafranin alone
will stain the mitochondria.

The method of preparing the safranin is as follows :'

1. Make a saturated solution of janus green in distilled water in a flask.

2. Add a little zinc dust and a few drops of hydrochloric acid. The solution first assumes a

bright crimson color and then bleaches, the hydrochloride of the leueobase of the safranin
being fonned.

3. Filter. Shake the filtrate in air and thus reoxidize the leueobase.

4. Precipitate the dye by saturating the solution with sodium sulphate. It is often necessary

to use a little heat. A dark red precipitate is formed.

5. Filter. Collect the precipitate on the filter. Wash with a saturated solution of soduun

sulphate and drj- it.

6. Dissolve out the dye from the dried precipitate with absolute alcohol.

7. Filter and evaporate the filtrate to dryness.

8. Dissolve the dye in the required concentration in distilled water or in salt solution.

The diethylsafranin prepared in this way behaves in exactly the same fashion
as some pure diethyl safranin manufactured especially for me by the Farbwerke
Hoechst Company.

There is, in addition to these janus greens, a large series of other janus dyes,
of which janus blue G and R, janus gray B and BB, janus black D, I, II, and O,
and janus yellow G and R are of particular interest because they are safranin
derivatives, the others being dyes of other series.

Janus blue is diethylsafranin-B-naphthol and it stains mitochondria in hving
lymphocytes in a constant and specific fashion. It is inferior to janus green in
that it will stain mitochondria in these cells only in a dilution of 1 : 300,000, but
as an indicator of processes of reduction it is better than janus green, for the con-
trast between the blue of the dye itself and its red safranin base is more brilliant
than in the case of janus green. The marks G and R indicate, according to Schultz
(1914, p. 48), that the janus blue is made by two processes, from clematin (mark
G) and from safranin (mark R). It is worthy of note that this color contrast with
janus blue is particularly beautiful in the kidney, where the glomeruh may be
colored deep blue and the remainder of the tissue red; so sharp is the contrast that
the glomeruli in thick sections of the entire kidney may be easily counted with a
binocular. Janus green hkewise stains the glomeruh specifically.

Janus black I also stains mitochondria in living blood-cells specifically, but,
on exammation, I find that it is not a pure dye, but a mixtm-e of two substances,
diethylsafraninazodimethylanilin and a brown substance, the nature of which I

'Dr. R. R. Bensley, personal communication.

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

93

am unable to determine. Thus, the specificity of janus black is undoubtedly due
to the fact that it contains janus green as one of its ingredients.

I have isolated the diethylsafranin from janus blue, janus black, and janus
gray (I failed with janus yellow, which may not contain it), and they all stain
mitochondria, which is further evidence that the specificity of janus green depends
upon the diethylsafranin group. It may be said that the staining is favored by
the addition of azodimethylanilin to it, as in janus green; increased, though not
so much so, by adding B-naphthol; and altogether prevented by the addition of
other groups, as in janus gray.

I have made an attempt to compare the specificity of janus green for mito-
chondria with other dyes which investigators have made use of for the purpose of
staining them. I used living human lymphocytes in freshly drawn blood as mate-
rial. The results of the comparison are shown in table 3. The names of the dyes
are given in the left-hand column, the concentrations are noted above, and a few
notes are recorded on the right. Intense staining is designated #, a faint colora-
tion + , while the minus sign indicates that the mitochondria are entirely unaf-
fected.

Table 3. — Sj>ecificalion of vital strainn.

f-i

§

o
o

§

1

O

o
o

o"

O

S

I

I

o
o
o

g

Remarks.

Janus jtrcen B (Hocclist)

it

i?
+

+

+
+
+
+

+

IT

a

IT

+
+

+
(?)

TT

if

+

(?)

-

Gives a specific stain of mitochondria

only.
Acts in the same way l)ut rather more

slowly.
Same as janus preen B.
Stains mitochondria intensely and

Janus blue (Hocchst)

Janus black I (Hoechr^t)

Nill)lau B extra

+
+

nuclei diffusely.
Stains specific Rranulations more in-
tensely than mitochondria.
Much the same as N'ilhlau B extra.
Tinges mitochondria only.
Do.
Do.
Same, with diffuse coloration of

entire cell.

Tinges mitoclioiidria slightly.

Negative.

Do.

Do.

Do.

Janus green G (Hoochst)

Methylene blue med. (Hoechst) . . .

Bismarck brown (Grubler)

Methvl violet 5B (Grubler)

+

Brilliant Kresvlblau (Cirubler)

-

Dahlia (Grubler) ....

....

Pvrrolblau (Grubler)

-

94 THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

IX. EMBRYOLOGY.

FERTILIZATION.

Investigations on the behavior of mitochondria during the process of fertil-
ization are at the root of the theory that they constitute the cytoplasmic basis of
heredity. These studies have brought to light several facts:

(1) That in the multiplication of spermatogonia and spermatocytes in the
testis, the mitochondria are divided in approximately equal amounts between the
daughter-cells.

The mitochondria are usually divided en masse. There seems to be no gen-
eral provision for a qualitatively equal division of the mitochondrial substance in
any waj' analogous to the longitudinal splitting of the chromosomes. Wilson's
(1916, p. 539) discovery of the pecuUar behavior of the mitochondria in the scorpion
must be mentioned. He found that in the Arizona scorpion there is in spermatogen-
esis an accurate fjualitative distribution of mitochondrial substance. The mito-
chondria become condensed into a ring of material which "divides somewhat after
the fashion of a heterotype chromosome ring, each spermatid receiving exactly one-
fourth of its substance." In the Cahfornia scorpion, on the other hand, the process
is entirely different. "The ring is here absent, its place being taken by about
twenty-four separate hollow spheroidal bodies that show no evidence of division
at any time and establish no definite relation to the spindle, but are passively
segregated by the spermatocyte divisions into four approximately equal groups.
Each spermatid thus receives as a rule sLx, not uncommonly five, rarely seven of
these bodies." Now, the astonishing regularity of the process in the Arizona
scorpion recalls to mind the behavior of the chromosomes themselves and strongly
suggests that we are here deahng with a distribution of substances which may ac-
tually play a part in heredity. But the phenomena observed in the California scor-
pion lead irresistibly to a diametricallj' opposed conclusion, that the mitochondria
are not carriers of heredity, since their distribution to the spermatids seems to take
place in an irregular, haphazard way. One can not help remarking how fortunate
it is that the two forms were studied at once, because otherwise it would be diffi-
cult indeed to resist being led astray by the extraordinary behavior of the mito-
chondria in the first of them.

(2) That structures derived from mitochondria persist in the fully developed
spermatozoon. *

Benda (1897, p. 401) was the first to show that they become transformed into
the spiral filament of the sperm. This observation has been corroborated by many
authors. The mitochondria also form the "nebenkern." But the mitochondrial
substance is usually confined to the middle piece of the sperm. Here also the great
variations in the behavior of the mitochondrial material are recorded, and we notice
at once the absence of anytliing approximating the remarkable orderUness which
is so characteristic of the changes which the chromatin undergoes. Finally, the

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. 95

fact has not been given due emphasis, so far as I am aware, that this mitochondrial
substance which persists in the fully developed spermatozoon is very different from
the typical mitochondria occurring elsewhere. It is well known that the mitochon-
dria undergo definite chemical changes in the course of spermatogenesis. Regaud
(1910, p. 294) clearly showed that their resistance to acetic acid grows greater and
greater; in fact, the structures which they form, nebenkern and spiral filament,
were known and recognized long before the mitochondria in the earUer stages of
spermatogenesis were brought to hght. They can be stained by ordinary methods
of technique and are resistant to acetic acid. While it can not be denied that these
structures are actually developed from mitochrondria, it certainly requires a con-
siderable stretch of our definition of mitochondria to include them under the head-
ing of mitochondrial apparatus and to make the statement that mitochondrial
substance occurs in the fuUy developed spermatozoon.

(3) That in many cases these mitochondrial products enter the egg on fertil-
ization.

The observations of Aleves (19116, p. 709) on Ascaris megalocephala, confirmed
by Held (1912, p. 247), Romeis (19136, p. 166), and Faure-Fremiet (1913, p. 585);
those of Meves on Parechinus miliaris (1912, p. 102), Phallusia mamillata (1913,
p. 225), Filaria papillosa (19156, p. 58), and Mytilus edulis (1915c, p. 54); those of
Duesberg (1915, p. 41) on Ciona intestinalis; of Levi (1915, p. 488) on Vespertilio
murinus, and many others all show that this occurs.

It is to be noted, furthermore, what a very comprehensive list of animals has
been studied which is representative of many of the great divisions of the animal
kingdom. It is indeed a body of evidence which can not be well evaded. A few
years ago derivatives of mitochondria were known to pass into the egg in only a
few instances. Now we reahze that they usually do so.

(4) That these same mitochondrial products which enter the egg in this way
can be recognized up to a certain stage in the development of the resulting embryo.

Thus Meves (1912, p. 116) followed them through the first and second divi-
sions of the fertiUzed egg. Faure-Fremiet (1913, p. 586) studied the fate of the
male mitochondria in Ascaris. Levi (1915, p. 523) traced them into a 3-blastomere
stage in the Fallopian tube. It can easily be seen by a careful study of the illustra-
tions presented by these authors, and it is important to note, that these substances
behave in every way Like foreign bodies; they undergo no changes and seem to
exercise no influence upon the behavior of the other formed elements in the cyto-
plasm. In this they may be contrasted sharply with chromatin. It would be
interesting to compare, with respect to mitochondria, the subsequent development
of eggs fertilized artificially with some fertiUzed with sperm in the usual way.
Jacques Loeb (1916, p. 316) has succeeded in rearing to sexual maturity frog
embryos which were artificially fertilized. An examination of mitochondria dur-
ing spermatogenesis in these forms might afford the most valuable information.
If the mitochondrial content of the sperms differed from those of ordinary frogs
it would be of the greatest significance from the point of view of whether or not
the mitochondria are heredity carriers, which is the central theme of most of the
recent studies on mitochondria.

96 THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

INHERITANCE.

The doctrine that the mitochondria play a part in the transmission of heredi-
tary traits is based upon the necessity of admitting the existence of a cytoplasmic
heredity. The claim is based on the well-known experiment of Godlewski, which
showed that an egg (deprived of its nucleus) when fertilized with sperm of an-
other species retained certain maternal characteristics on development. In fact,
there is nothing new in the conception that there is such a thing as cytoplasmic
heredity. Jenkinson (1914, p. 152) and Conklin (1915, p. 176) freely admit it.
What is new is the view that mitochondria carry it.

Let us pause a moment in order to consider the claims of the advocates of this
view, which was first enunciated by Meves (1908, p. 849), though Van der Stricht
also guardedly made reference to the possibility that mitochondria may play a
part in heredity in 1908 (p. 4). They do not say or even hint (as the adherents
of the chromosome hypothesis have not refrained from doing, in the case of the
chromosomes) that the mitochondria constitute in any sense of the term the sole
basis of heredity. They believe, however, in general, that the mitochondria play
a part in the transmission of those characters which are cytoplasmic. Meves
(1908, J). 850) says that the nuclear characters are carried over by the chromosomes,
those of the plasma by the chondriosomes. This idea is beautifully supplemented
by Meves 's subsidiary hypothesis, according to which the mitochondria are trans-
formed into all products of cellular differentiation. In other words, here is a
material that can be seen to go through all the stages of development and actually
give rise to the peculiar and characteristic differentiations wliich are the hereditarj^
traits. The trouble is that the evidence that the mitochondria play a part in
histogenesis (p. 102) is no more convincing than that in favor of the view that they
constitute the material basis of heredity. Broman's working hypothesis is that
while the chromosomes are the bearers of the hereditarj^ qualities of the race and
species, the mitochondria carry those peculiar to the individual. It is interesting
to observe that Wilson (1914, p. 352) writes:

"Genetic experiment has already given some ground for the conclusion that definite
types of hereditary distribution may be immediately dependent upon elements containetl
in the protoplasm. Recent advances in our knowledge of the 'chondriosomes' or 'plasto-
somes' provide this conclusion with at least a possible cytological basis."

All through the work on heredity, as well as in the other biological problems,
the tendency is observed to affix the responsibility for the phenomena to structures
with definite form — with each new discovery of a morphological entity to become
optimistic and to think that we are well on the road toward a solution of the prob-
lem. The basic laws of heredity were appreciated by Mendel long before the dis-
covery of the chromosome; yet the chromosomes were seized upon with great
avidity. So it is with the mitochondria.

The adherents of the chromosome hypothesis in this country and elsewhere
are naturally opposed to this view that the mitochondria are even partial carriers
of heredity. It is said that the cases in which it has been shown that mitochondria

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. 97

pass into the egg on fertilization are exceptional and that the crucial cases are those
in which no mitochondrial substance passes into the egg. This Lillie believes to
be the case in Nereis. Mitochondria generally occur in the middle piece and tail
of the spermatozoon, though this is not always true. Lilhe (1912, p. 418) says
that "the middle piece and tail of the spermatozoon do not enter in the fertiUza-
tion of Nereis." He admits (p. 426) that "it is possible that the fixation granules
produced by the spermatozoon represent a cytoplasmic element." So that, until
new facts are discovered, through the use of mitochondrial methods of technique,
the case of Nereis does not offer an insurmountable barrier to the acceptance of
the view that mitochondria play a part in inheritance.

MITOCHONDRIAL CONTINUITY.

The question of mitochondrial continuity arises just as surely as the doctrine
of the permanence of the chromosomes, and the proof of it is every bit as unsatis-
factory, perhaps more so.

Duesberg (1912, p. 766) is in favor of the theory of mitochondrial continuity;
Beckwith (1914, p. 230), Chambers (1915, p. 291), and several others are against
it. It has taken strong hold on the botanists, GuilUermond (1912a, p. 398) com-
ing out strongly in favor of it and saying: "The mitochondria result from the
division of the preexisting mitochondria of the egg, none of them ever arise 'de
novo' in the cytoplasm." However, in certain plants some mitochondria are
large and others extraordinarily minute. It is possible that the large ones may
arise in this way by division of preexisting ones, but there is no evidence that the
small ones are formed only by the segmentation of filaments or rods, which are of
larger girth than they are. They probably arise de novo by condensation.

The idea that all mitochondria arise from preexisting mitochondria by divi-
sion is a relic of Altmann's (1894, p. 155) doctrine, Omne granulum ex granule, and
has persisted in our minds on account of the newly conceived idea that mitochondria
are concerned in heredity. Altmann thought that they were elementary organisms
endowed with a certain measure of individuaUty. It is possible that the truly
remarkable morphological resemblance which they bear to bacteria led him to
believe that they multiply in this way. This conclusion has been supported quite
recently by the growing tendency to regard mitochondria as plast-like in nature.
It is apparently quite true that mitochondria form plastids in some plants, and,
in view of the evidence at hand that the plasts exhibit a true genetic continuity,
some of them always arising from preexisting plastids by division, it was quite
natural to assume that the mitochondria themselves behaved in somewhat the same
way.

It has been generally believed for some time past that mitochondria multiply
by transverse division and perhaps also by longitudinal division. The evidence
for the latter method is meager and unsatisfactory and does not merit discussion.
With regard to transverse division, it must be said that this undoubtedly does
occur in some cases — for example, in the course of spermatogenesis of Vespa crabro,
as was found by Meves and Duesberg (1908, fig. 39). This has already been men-

98 THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

tioned. Faure-Fremiet (1910a, p. 527) has also forwarded strong evidence that
the mitochondria in certain infusoria multiply in this way. The multiphcation of
mitochondria in mammaUan tissues is a phenomenon exceedingly difficult to demon-
strate. As a possible source of error it should be borne in mind that the most com-
mon reaction on the part of filamentous mitochondria to unusual conditions is to
fragment — that is to say, to multiply by transverse division. This is a very defi-
nite reaction to injury, by no means a normal method of multiplication. Again,
in sectioned material it is impossible to say with certainty whether the appear-
ances observed mean an actual multiplication by division or simply an approxima-
tion of originally separate mitochondria, for mitochondria everywhere show a
tendency to clump together which is markedly enhanced if the tissue is in any way
injured. The only way to obtain definite information on the subject is to study
a process, rather than a process killed at a certain point, as in fixed material,
or even a perverted process, as in living cells teased out and growing in isotonic

media.

As against the doctrine of mitochondrial continuity, we have certain cells in
which the mitochondria appear to go into solution. I have found that this is the
case in the chromophile cells of the nervous system (Cowdry, 1916/), p. 41). There
have been, so far as I am aware, but few observations (Chambers, 1915^ p. 291) of
a de novo origin of mitochondria in the cytoplasm, but to my mind the burden of
proof is not that it occurs, but rather that it does not take place. It would surely
be arbitrary to assert that the phosphatid albumin complex (of which mitochondria
consist) always occurs in certain aggregates of definite size which are visible with
our present powers of the microscope and that these aggregates multiply by divi-
sion like independent organisms. In this connection we can only speculate, but it
certainly seems much more likely that the phosphatid may be deposited free in the
cytoplasm or upon an albuminous matrix, and that it subsequently grows by
accretion. It is quite possible that when it attains a certain size or shape it divides
for purely physical reasons; but in the absence of definite proof, one way or the
other, it seems to me highly j^robable that mitochondria are continually arising
in the cytoplasm de novo, and furthermore that perhaps this is the most important
method of multiplication.

It is, however, incumbent upon those who believe that the mitochondria are
carriers of heredity to demonstrate their continuity; and, further, the origin of
mitochondria de novo, if it does take place, would be very difficult to reconcile with
the view that they transmit hereditary traits.

The observations of Beckwith (1914, p. 216) on the eggs of Hydradinea echi-
nata are of particular interest in this connection. She found that the cytojilasmic
contents of the fertilized eggs on centrifuging separated into three layers— a layer
of oil, a clear zone, and a layer containing mitochondria and vitellus. The first
plane of segmentation cut these layers in various directions, resulting in planulas of
very diff'erent appearance. Beckwith separated the two first blastomeres of cen-
trifuged eggs and allowed them to develojx Some of these formed planulas, which,

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. 99

though they contained no mitochondria, nevertheless appeared normal, from
which she concludes that mitochondria are not necessary for differentiation. Dues-
berg (1915, p. 67) very aptly remarks that these experiments do not show that
mitochondria do not play a role in the differentiation of tissues, because no tissues
at this stage have yet become differentiated. He is not incUned to accept Beck-
with's statement that mitochondria are absent in the early stages of oogenesis and
that they do not appear until the vitellus is well formed (they form, Beckwith
behaves, de novo in the cytoplasm), which is totally at variance with all we think
we know of the mitochondria in oogenesis, as well as being in contradiction to the
observations of Tsukaguchi (1914, p. 117) on Aurelia aurita, an animal belonging
to the same class as Hydractinea echinata.

The greatest obstacle to the acceptance of the view proi)ounded by Benda
and Meves is our conception of the chemical nature of mitochondria. If it is true
that they are phosphohpins it is hard to regard them as carriers of heredity, even
though they may contain albumin also. It can not be denied that, chemically,
chromatin appears to be the best htted to play the part of heredity carrier. The
relative equality in the amount of chromatin between the male and female gametes
and the deficiency in the amount of the cytoplasm must mean something. Even
should it be shown that mitochondrial substance passes over in fertilization in all
animals, it may indicate nothing more than that a living portion of the sperm,
capable of metabohsm, enters the egg. It is a mistake, however, to arrive at a
hasty conclusion, because those who make the conservative statement that mito-
chondria play some part in heredity occupy just as secure a position as those, on
the other hand, who claim that chromatin is the sole heredity carrier.

In the higher plants it is well known that all the cells of the gametophyte
contain the x number of chromosomes and the cells of the sporophyte, or sexless
generation, contain the double number 2x, yet no distinction has been shown in
the mitochondria, which appear similar in every particular. Fvu-thermore, in the
aphides, or plant lice, there is also an alternation of generations, but a compari-
son of the mitochondria in the two has not been made.

ORGAN-FORMING SUBSTANCES.

Recent investigations on mitochondria in the early stages of the development
of Ascaris (Faure-Fremiet, 1913, p. 676) and of ascidians (Duesberg, 1915, p. 66)
throw a flood of new light upon our conception of the so-called "organ-forming sub-
stances" (Conkhn, 1905, p. 216).

Conklin (1905, p. 211) has discovered the fact that the cytoplasm of the egg
of Cynthia is structurally differentiated into three substances — a clear substance,
a yellow substance containing yellow pigment, and a gray substance containing
yolk; and that "the upper clear half of the egg gives rise to ectoderm; the cres-
cent of yellow jirotoplasm surrounds the posterior side of the egg just below the
equator and is later transformed into the muscle and mesenchyme of the larva;
the gray protoplasm occupies the remainder of the lower hemisphere and gives

100 THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

rise to the endoderm, to the chorda, and to the neural plate." So much for the
fate of the "ectoplasm," the "mesoplasm," and the "endoplasm," as he styles
them. That these substances form the organs in question Conklin (p. 217) has
shown lieyond the shadow of a doubt. In the absence of one of the substances,
the organ to which it would naturally give rise is not produced; conversely, each
substance develops, if it develops at all, into the i^arts which it would normally
produce. The portions of the egg which lack these substances form embryos
which lack the corresponding organs. From these three fundamental substances
he derives six (p. 218), viz, ectoplasm, endoplasm, myoplasm, chymoplasm, caudal
chymoplasm, and chordaneuroplasm.

Due.sberg (1915, p. 35) attacked the same problem in Ciona with new methods
which revealed the mitochondria and which showed that they occur in very differ-
ent amounts in the areas of cytoplasm described by Conklin. In the light of his
work it is evident that the myoplasm of ConkUn is simply an accumulation of
mitochondria, the gray protoplasm a region whose vitelline granules are particu-
larly numerous, and, lastly, the clear protoplasm nothing else than the fundamental
ground-substance of the egg, which contains but few mitochondria. Duesberg's
observations confirm the discoveries of Conklin as well as extend them. He found
that the areas form just exactly the organs which Conklin affirmed — that the yel-
low crescent which contains many mitochondria forms muscle, etc. But he does
not agree with Conklin's interpretation. To repeat, Conklin believes (p. 211) that
"all the principal organs of the larva in their definitive positions and proportions
are here marked out in the two-cell stage by distinct kinds of protoplasm." Dues-
berg (p. 60), on the other hand, is of the opinion that the different substances
in which Conklin believes do not exist. "The different appearances of different
regions of the egg and of the blastomeres depend not upon the existence of special
substances, but upon a special distribution of the elements figured in the ovoplasm, "
that is to say, of mitochondria, vitellus, pigment, etc. In other words, the regions
differ only in containing different proportions of the same substances ; none of them
possesses a special kind of substance. This interpretation, which I thoroughly
believe in, is also in accord with ConkUn's own observations. He writes (p. 212) :

"Although these different ooplasmic substances are chiefly locahzed in certain regions
of the egg, which give rise to certain portions of the embryo, this segregation is not quite
complete. Most of the clear protoplasm is found in the upper (ectodermal) half of the
egg, but some of it is also present in the lower half. Most of the yolk is found in the lower
(endodermal) half of the egg, but a little of it is found in the upper half. Almost all of the
yellow protoplasm is located in the mesodermal crescent, but a very small amount of it
is found around the nuclei of all the cells. Thus samples of these egg-substances are con-
tained in all the cells; nevertheless the segregation is so nearly complete that the clear,
the gray, the light gray and the yellow areas are marked out with the greatest distinctness."

Apparently there is a distinct relationship betwen the amount of mitochondria
and the amount of vitellus. For instance, the mitochondria are tremendously
abundant in muscle-cells and the vitelline granules few in number; in nerve-cells a
fair amount of both is present; while in the endodermal cells there is an enormous

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. 101

amount of vitellus and few if any mitochondria. A similar quantitative relation
between mitochondria and fat has been noted by many authors, and I have par-
ticularly in mind the observations of Goetsch (p. 13G). We know that decrease in
oxidation causes the deposition of fat and we suspect that mitochondria are actively
concerned in oxidation (p. 134). It is common knowledge that they are particu-
larly abundant in active cells, which must be respiring at a rapid rate. In other
words, the heaping-up of fat and the diminution in the mitochondria indicate
reduced oxidation; and conversely, the absence of fat and the abundance of mito-
chondria indicate an increase. Accordingly it is possible that the distribution of
these substances, as described by Conklin and Duesberg, is merely the visible mani-
festation of differences in the rate of oxidation in different parts of the egg and in
the different tissues of the embryo.

HISTOGENESIS.

The origin of the idea that mitochondria are concerned with histogenesis is
not difficult to trace. They occur in all embryonic cells. In early stages of
development they are the only formed elements in the cytoplasm. They are fila-
mentous in the mj'obla.sts and neuroblasts, and it is perfectly natural to think that
they become transformed into fibrils and other products of differentiation; but the
trouble is that the wa,ys of nature are not simple, that the obvious interpretation
is not necessarily the correct one. It also falls in line with the view that they con-
stitute the material basis of heredity (p. 98). Meves (1908, p. 845) writes that,
with the specialization of the embryo into different organs and tissues, primitively
similar cells assume special functions which find expre.ssion in characteristic struc-
tures or differentiations. All these products, no matter how heterogeneous they
may be, arise through the metamorphosis of one and the same elementary plasma-
cortstituent, the chondriosomes. Thi.s is a very sweeping statement, but even it
does not express the situation correctly, because claims are also made that most of
the products of the activity of speciahzed cells of the adult organism, like secretion
granules, are also formed by a chemical transformation of mitochondria.

The dominating influence of this dogma has made itself felt in many ways.
To cite a single instance, Hoven in 1910 arrived at the conclusion that mitochondria
are transformed into neurofibrils in the developing nerve-cell. This was gener-
ally accepted (Firket, 1911, p. 545; G. Arnold, 1912a, p. 289, and others). Fol-
lowing this line of reasoning, Hoven, (1910a, p. 478) and Meves (19106, p. 655)
concluded that the mitochondria are absent in adult nerve-cells after neurofibrillar
formation has ceased, and looked for them and failed to find them. And this in
spite of the fact that Altmann (1890, p. 52), Levi (1896, p. 180), Lobenhoffer (1906,
p. 491), Nageotte (1909, p. 827), and others had already clearly and precisely
figured and described them.

In order to get a true conception and perspective of what these claims really
mean, I have tabulated some of the structures which are said to be developed
through the transformation of mitochondria or under their influence.

102

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

List of sithslances supposed to he formed from miUichomiria.

Amyloplasts ( = leucoplasts).

Anthocyanin, Guillierniond (1913c, p. 478; 1913/, p.

1002).
Apparato reticolare (Binnennetz), Hoven (1910a, p. 479).
Autoplasts ( = chromoplasts).
Aqueous humor, Mawas (1909, p. 284).
Batonnets of Heidenhain, Policard (1912f, p. 457).
Carotine, Guilliermond (1912a, p. 389).
Cerebrospinal fluid, Grynfeltt and EuziSre (1912, p. 64).
Chloroleucites ( = chloroplasts).
Chlorophyll, GuiUiermond (1912a, p. 407).
Chloroplasts, Guilliermond (1912o, p. 412).
Chromoplasts, Guilliermond (19r2a, p. 412).
Ciliary apparatus, Saguchi (1917, p. 266).
Collagenic fibrils, Moves (1910a, p. 164).
Composes phenolique, Guilliermond (1913/, p. 1002).
Connective-tissue cell granulations, Renaut and Dubreuil

(1906b, p. 230).
Cuticular formations. Van der Stricht (1918, p. 36).
Cyanoplasts, Guilliermond (1913i, p. 1282).
Disque Q, Faure-Fremiet, Maver and Schaeffer (1910,

p. 75).
Eberth's intracellular structures, Saguchi (1913, p. 239).
Elaioplasts, Guilliermond (1912a, p. 405).
Eosinophile granulations, Meves (19106, p. 6.56).
Epidermal cell secretion, Saguchi (1915, p. 395).
Epidermal fibrils, Firket (1911, p. 537).
Fat (neutral), Altmann (1889, p. 94).
Fibrils of Herxheimer, Favre and Regaud (1910, p. 1138).
Filament h^licoldal, Porroncito (1910, p. 310).
Fucasin, Le Touze (1912, p. 33).
Glycogen, Arnold (1908, p. 365); Alexeieff (1916a;

19166, and 1917a).
Goblet ceU mucus, Grynfeltt (1913, p. 11).
Grains, chromophiles, Leplat (1913, p. 219).
Grains safranophiles, Regaud (1910, p. 342).
Granules de segregation, Renaut and Dubreuil (19066,

p. 230).
Hemoglobin pigment, Ciaccio (1911, p. 16).
Hemoglobin crystals, Policard (19r2a, p. 92; 19126,

p. 230).
Leucites ( = leucoplasts).
Leucoplasts, Lewitsky (1910, p. 542).
Lipoid in nerve-cells, Cowdry (1914a, p. 13).
Liposomes, Faure-Fremiet (19106, p. 538).

Macromitosome, Ciatenby (1917, p. 416).

Mammary-gland secretion, Hoven (1911, p. 325).

Melanin, Asvadourova (1913, p. 263).

Metachromatic corpuscles, Guilliermond (1913d, p. 439).

Mucorine, Moreau (1915a, p. 171).

Mucus (boules picriphiles), Grynfeltt (1913, p. 11).

Myofibrils, Benda (1899a, p. 379).

Nebenkern in spermatogenesis, Meves (1900, p. 585).

Neurofibrils, Meves (1907a, p. 403).

Ncurogha fibrils, Meves (1907a, p. 403).

Pancreatic zymogen, Hoven (19106, p. 349).

Parathyroid secretion, Bobeau (1911, p. 186).

Parotid secretion, Regaud and Mawas (19096, p. 220).

Pheoplasts, Nicolosi-Roncati (19126, p. 144).

Pigment, Prenant (1913, p. 926).

Poison-gland secretion, Torraca (1916, p. 326).

Pole-disc granules in chrysomelid eggs, Hegner (1914,

p. 419).
Prepucial-gland secretion, Altmann (1890, p. 99).
Prostate secretion, Akat.su (1903, p. 566).
Randreifen of blood corpuscles, Meves (1905, p. 103).
Retinal pigment, Luna (19136, p. 56).
Russell's bodies, Dubreuil and Favre (1914a, p. 373).
Sebaceous-gland secretion, Nicolas, Regaud and Favre

(1912a, p. 203).
Sheath of tail-thread of sperm, Meves (1900, p. 586).
Spiral filament of sperm, Benda (1897, p. 412).
Starch, Guilliermond (1912a, p. 412).
Submaxillary-gland secretion, Regaud and Mawas

(19096, p. 220).
Suprarenal pigment, Mulon (1913, p. 1026).
Suprarenal secretion, Mulon (19106, p. 873).
Sweat-gland secretion, Nicolas, Regaud and Favre

(19126, p. 191).
Tannin, Guilliermond, (191.'W, p. 437).
Test of foraminifera, Faure-Fremiet (1911, p. 120).
Trophoplasts ( = leucoplasts) .

Thyroid-gland secretion, Grynfeltt (1912c, p. 147).
Tyrosin, Asvadourova (1913, p. 263).
Tonofibrils, Saguchi (1913, p. 2:39).
Urea in the kidney, Oliver (1916, p. 318).
Venom, Grynfeltt (1913, p. 11).
ViteUus, Loyez (1909, p. 191).

Xanthophyllous pigment, Guilliermond (1913A, p. 1926).
Yolk, Coghill (1915, p. 349).

The term "transformation" has been used too freely and too loosely. In-
vestigators speak glibly of the transformation of mitochondria into other mate-
rials without stopping to think what it means. We can understand the transforma-
tion of a liquid into a gas, but we can not conceive of the transformation of oxygen
into carbon. The hkelihood of a transformation taking place depends upon the
difference in the properties of the original and the transformed substances; but
the differences in the properties of mitochondria and all these materials are usually
ignored. Accordingly, some of the statements involve chemical and physical im-
possibilities. We have good reason to suppose that mitochondria resemble phos-
phohpins, and it is therefore incredible that they should transform into hemoglobin
which contains iron, chlorophyll which contains magnesium, and the colloid of the
thyroid gland with its iodine. The iron, magnesium, and iodine can come only
from the cytoplasm. The question, however, is vastly comjilicated by statements,
apparently supported by fairly good evidence, to the effect that mitochondria consist,

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. 103

in addition to the phospholipin, of an albuminous substratum of some sort. Nev-
ertheless, in the vast majority of cases a true transformation is out of the question.

The idea of a transformation is often based upon the mere observation of sub-
stances within the mitochondria. The fallacy of this line of reasoning is evident,
for no one would say that because the cell contains iron or phosphorus the iron or
phosphorus is produced by a transformation of the substance of the cell. Sub-
stances unquestionably penetrate into the mitochondria from the cytoplasm, as
shown by the fact that in these supposed transformations of mitochondrial mate-
rial into something else there is always a distinct increase in size. For instance,
the mitochondria possessing bleb-like swellings in gland-cells are larger than those
without them, and mitochondria containing starch, fat, pigment, crystalloids,
and other materials are invariably greatly enlarged. In cases where granular
mitochondria expand to form vesicles careful observation will often show that there
has been little or no change in the absolute amount of mitochondrial substance; it
has simply become spread over a larger area. We may safely regard this imbibi-
tion from the cytoplasm as establishcxl, but how the materials are taken in is exceed-
ingly difficult to explain. It differs sharply, however, from the normal process
of growth because the expansion is lateral, while in growth extension is usually
longitudinal (p. 70). I incline strongly toward Regaud's eclectosome theory (IQOOa,
p. 919), according to which mitochondria play the part of plasts choosing and
selecting substances from the surrounding cytoplasm, condensing them and trans-
forming them in their interior into infinitely diverse products; but I would venture
to emphasize the fact that in all this the mitochondria may be acting in an entirely
passive manner as a vehicle, taking up materials by virtue of their phospholipin
constitution, or on account of physical forces acting on their surfaces, or for other
reasons, and that the optically homogeneous ground-substance of the cytoplasm
may be the active and essential agent in this as in so many other vital mani-
festations. No change of the mitochondrial substance need be involved. Sugar,
which is heaped uji in i^lasts, is certainly not formed through a transformation
of mitochondrial material, or of the plasts which contain it. They merely
act as containers, the foreign material being localized in certain regions of the
filament.

But in rare cases there is evidence of an actual change in the mitochondria
themselves, especially in the formation of fibrillar structures, in cornification, and
in other similar processes. We may conceive of this as taking place in several
ways: (1) by the addition of substances from the cytoplasm which enter into
close combination and become integral constituents of the mitochondria; (2) by
the mitochondria giving up to the cytoplasm certain of their normal constituents;
(3) by chemical dissociation which may or may not be followed by resyntheses.
Fat, lipoid, and other similar substances might be formed; but it is imj^ortant to
bear always in mind that the po.ssibility of a transformation diminishes in direct
proportion to the degree of dissimilarity between the mitochondria and the material
in question. It is for this reason that I am willing to entertain almost any alterna-
tive hypothesis rather than accept unqualified statements of the chemical trans-
formation of mitochondria into dissimilar substances.

104 THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

FIBRILS.
COLLAGENIC FIBRILS.

Meves's own work (1910a, ]:>. 164) on the histogenesis of collagenic fibrils may
now be briefly mentioned. His material consists of a series of preparations of the
growing tendon in the posterior extremity of chick embryos of from 6 to 19 days'
incubation. He fixed the embryos in a modification of Flemming's fluid, cut some
of the specimens in transverse, others in longitudinal sections, stained them with
iron hematoxylin, and counterstained with acid fuchsin. In this way the mito-
chondria were colored black and the fibrils red.

The illustrations of the preparations on which he l:)ases his contention are
beautifully shown in his second plate (Tafel III). They are arranged in two
series of increasing grades of differentiation, the uppermost of which is taken from
longitudinal sections and the lower from transverse sections. They show that the
mitochondria become accumulated in the peripheral parts of the cytoplasm in
stages during which the collagenic fibrils first appear. The mitochondria are fila-
mentous, but are not so long as the primitive fibrils, the ends of which he was
unable to observe. The figures show, in addition, that there is a decrease in the
number of mitochondria in the cells of later stages and that the mitochondria are
no longer most abundant in the peripheral parts of the cell. Meves's line of rea-
soning is instructive. He says (p. 164) :

"Was nun die Deutung der beschreibenen Bilder anlangt, so bleibt meiner Meinung
nach nichts anderes iibrig, als anzunehmen, dass die Bindegewebsfibrillen aus Chondrio-
conten hervorgehen. Die Chondrioconten werden zunachst epicellulJir. Sie andern dann
ihre chemische Beschaffenheit, indem sich ihre Substanz in eine solche umwandelt, welche
weder durch Eisenhametoxylin noch durch Fuchsin farbbar ist. Auf diesem Stadium
treten diejenigen von ihnen, welche in einer Reihe liegen, untereinander mit ihren Enden
in Verbindung. An der Bildung einer Fibrille deteiligen sich zahlreiche Zellen (alle die-
jenigen, denen ihr Verlauf fest anliegt), indem jede einen Fibrillenabschnitt liefert. Die
Fibrillen andern dann zum zweiten Mai ihre chemische Beschaffenheit, indem sie eine
intensive Farbbarheit fiir die CoUagenfarbstoffe gewinnen. Schliesslich werden sie von
den Zellen frei und kommen in den Spaltraumen zwischen ihnen zu liegen."

Before we enter upon a criticism of Meves's work it is necessary for us to
recognize, in all fairness, that he does not claim to have conclusively established
the transformation of mitochondria into collagenic fibrils, since, as he himself empha-
sizes, the steps which lead up to this conclusion consist of assumptions as well as of
positive evidence. It is important above all to note that, according to Meves,
the mitochondria (chondriocontes) are invisible, not staining with either iron
hematoxylin or fuchsin when they form the fibrils. This assumi)tion that they
are invisible when the most important stage of the whole process is taking place
disarms all criticism at the outset.

Meves makes his chain of evidence ("Kette der Beweise," p. 165) still more
fragile by asking the question : " If the chondriocontes have nothing to do with the
formation of fibrils, why then do they become epicellular?" I do not know whether
Meves means by the term "epicellular" that the mitochondria are actually out-

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. 105

side the cells, lying upon them, or that they are simply accumulated in the periph-
eral parts of the cytoplasm within the cell-membrane. The derivation of the
term and the appearance of certain mitochondria illustrated in his figure 25 seem
to support the first interpretation. The difficulty of asserting that mitochondria
in Meves's preparations are without the cell-wall is great, not only because the cell-
walls are not differentially stained, but also by reason of the mode of deposition of
collagen. It seems, therefore, that what Meves describes is simply a heaping-up
of mitochondria in the cytoplasm beneath the cell-membrane.

A similar peripheral arrangement of mitochondria in the cytoplasm has been
described in a whole host of conditions other than fibril formation and therefore is
without special significance in this connection (see p. 76).

The filamentous shape of the mitochondria might at first sight appear to be
indicative of a transformation into collagenic fibrils. This is, however, not the
case, because mitochondria, which are also thread-like, occur in the same stage
of development in cells which do not form fibrils.

Meves's other argument that the mitochondria decrease in number as the
fibrils form in the course of development would be valid only could it be shown
that the diminution in number was not brought about in some other way. In-
deed, it is only one instance of a very general phenomenon, that the mitochondria
grow fewer and fewer in all cells as they grow older. Meves, himself (1911a, p. 495),
has shown this to be the case in the later stages of the cytomorphosis of red blood-cells
of the guinea-pig; Firket (1911, p. 544) has demonstrated the same phenomenon
in the cells of the egg tooth of chick embryos; and Regaud and Favre (1912, p. 328)
have confirmed his results by their observations on the epidermis of man.

Meves's hypothesis of the role of mitochondria in connective-tissue fibril
formation can not, apparently, be reconciled with Baitsell's (1916, p. 754) recent
work on wound healing. Baitsell discovered that certain fibrils form, quite apart
from the cells, as a differentiation of a typical fibrin net in the coagulation tissue
between the cut surfaces. The cells all wander in later. The staining reactions
of the new fibrous tissue, formed in this way, appear to resemble in many ways those
of true connective tissue. Here we are dealing with the formation of fibrils Uke
those of connective tissue from a known substance, fibrin, iii the absence of cells,
and, since the mitochondria are alwa^^s intracellular, it is inconceivable how they
could be bodily transformed into the fibrils, as Meves claims.

It is interesting also to note that M. R. Lewis (1917, p. 56) has made a care-
ful study of the behavior of mitochondria and fibrils in cultures of subcutaneous
tissue of chick embryos. The fibers in the explanted piece were not observed to
grow either in length or bulk, but new fibrils arose as delicate lines in the exoplasm
of the cells and quite independently of the mitochondria. The cessation of growth in
the one and the initiation in the other may indicate some superficial, or perhaps
fundamental, difi"erence. No attempt was made to compare the microchemical re-
actions of the new-formed fibrils with the definitive connective-tissue fibrils in the
organism. The observations indicate the independence of mitochondria and afford
a plausible alternative hypothesis of the development of connective-tissue fibrils.

106 THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

It may even be said that Meves's observations, instead of proving that the
mitochondria are transformed into collagenic fibrils, indicate that the two are
quite distinct; for if there is. as he assumes, a change in the chemical constitution
of mitochondria so that they do not stain with either iron hematoxylin or fuchsin,
one would expect to see some evidence of this in his figures. But his figures
show that the mitochondria are characterized by the extreme uniformity of their
reactions to iron hematoxylin; they show no variabihty whatever. Again, if, as
he further assumes, there is another change in the chemical constitution of the
invisible fibrils (invisible by his own assumption) by which they acciuire an intense
affinity for collagen-staining dyes, one would look for some variation in the stain-
ing with the said collagen dyes. But the figures show that there is no variation
in the reactions of the fibrils. The third point, which may be justly urged, is his
last assumption that in later stages the fibrils are differentiated by virtue of a
formative activity. It follows that the theory of the origin of the fibrils from
mitochondria is aiiplicable only to a very limited stage in their formation and does
not fit the facts which he himself has observed relative to their formation in older
embryt)S.

The argument from analogy advanced by Duesberg (1912, p. 759), that the
formal proof of the role of mitochondria in the formation of myofibrils may be
regarded as indirect evidence of their participation in the develoi^ment of other
formed elements (collagenic fibrils and neurofibrils) cuts both ways; for it is etiually
true that the notable absence of evidence in favor of the formation of collagenic
fibrils by a transformation of mitochondria leads one to doubt a like origin of
myofibrils.

MYOFIBRILS.

Although Benda (1899rt, p. 379) and Meves (1907a, p. 402) were the first
investigators to claim that mitochondria became changed into myofil^rils, Dues-
berg (1909, 1). 126, and 1910, p. 647) has furnished the most complete evidence in
support of this contention. His material consists of a very complete series of
chick embryos of from 19 hours' to 10 days' incubation prepared by Benda's method
for mitochondria. He employed also, for control, subhmate acetic, alcohol, and
other fixatives, and the chloride-of-gold method of Ranvier. He studied both the
myotomes and the myocardium, but his most detailed work related to the former.
He found that at first the mesoblastic cells contain only typical mitochondria in
the form of granules, rods, and short, wavy filaments, and that in more advanced
stages of development filaments of the same girth and morphological characters
became more and more numerous and of greater and greater length, until they
began to show traces of differentiation into segments. The mitochondria and
the homogeneous filaments stained alike by the Benda method, but after the first
indications of segmentation appeared the staining reaction of the fibril began to
change, for they no longer stained as deeply as the mitochondria. His figures
indicate a marked decrease in the amount of mitochondria parallel with the differ-
entiation of fibrils. This is seen by a comparison of figure 14 with the succeed-
ing ones, figures 16, 20-23, 25, and 26. He arrived at the conclusion that the
mitochondria elongate and become transformed into myofibrils.

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. 107

Von Kurkiewicz arrived at the same conclusion concerning the mitochondrial
origin of the fibrils in the heart-muscle of the chick. Schultze also claims to have
confirmed Duesberg's contention regarding the role of mitochondria in myofibril
formation. Brlick (1914, p. 581) has described the mitochondrial origin of myo-
fibrils in Anodonta cellensis. Moreover, Leplat (1912, pp. 458 and 509) has studied
the development of fibrils in Mm. sphincter pupilla? and ciliaris of birds by the
application of the Benda method. He was able to observe all the stages in the
differentiation of the myofibrils which Duesberg described and he reaches the same
conclusion. The following additional investigators favor the doctrine of the trans-
formation of mitochondria into myofibrils: Lewitsky (1910, p. 539), Favre and
Regaud (1910, p. 1138), Hoven (1910a, p. 476), Prenant (1911a, p. 463), G. Arnold
(1912a, p. 289), Schiifer (1912, p. 193), Luna (1913f, p. 478), Jordan and Ferguson
(1916, p. 94), and others. Heidenhain (1911. p. 1086), Levi (1911, p. 191), and
Gurwitsch (1913, p. 123) are in the minority in that they do not subscribe to it.

The accuracy of the facts forwarded by Duesberg being beyond cavil, it
becomes necessary for us to determine whether they permit of any other interpre-
tation except that advanced by him. The nature of the unsegmented fibrils and the
significance of the fluctuations in the amount of mitochondria are important points.
Apparently the technique employed is not specific, for it colors the mitochondria
and the primitive fibrils in the same way, although they differ in their microchemi-
cal properties, because we find that the mitochondria are dissolved by fixatives
containing a concentration of acetic acid which in no way affects the myofibrils.
They also react differently to stains. If, therefore, differences of this nature are
not revealed by the technique employed, the possilMlity must be entertained that
the elongated homogeneous filaments described by Duesberg may differ inter se;
in other words, that we may be dealing with two different kinds of filaments which
may appear similar on account of the stain which is used, one of which is mito-
chondrial, the other a precursor of the definitive myofibrils which does not possess
the properties of mitochondria; so that Duesberg's investigations do not exclude
the possibility of the origin of myofibrils from material which is not mitochondrial.
In fact, there is some indication of the existence of non-mitochondrial precursors.
I refer, for instance, to the observations of Godlewski (1902, ix 149) and others,
according to which the myofibrils result from the confluence of small masses of
material, not through the elongation and transformation of a homogeneous filament.

Moreover, I have observed in my own preparations of chick embryos of 100
hours' incubation (stained with fuchsin and methyl green) very delicate green-
colored fibrils, side by side with others stained red, and still others beginning to show
traces of segmentation. I have also seen similar fibrils stained red with alizarin
in Benda preparations and light gray with iron hematoxylin. Morphologically
they resemble the filamentous mitochondria, but their staining reactions are entirely
different. Whether they give rise to the definitive fibrils or not I can not say.

Duesberg (1915, j). 59) has sujiplemented his discovery of the abundance of the
mitochondria in the myoblasts and their subsequent diminution during fibril for-
mation by important investigations on ascidians, where he found that the mitochon-

108 THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

dria are particularly numerous in the primordial muscle-cells, often being arranged
in chains, indicative, he thinks, of transformation into fibrils. Moreover, Romeis
(1913o, p. 10) has found a marked increase in mitochondria and transition forms
like those figured by Duesberg during the regeneration of muscle-cells in Triton.

Too much weight should not be placed in the presence-and-absence argument.
There may be other reasons than fibril formation for oscillations in the amount
of mitochondria. Romeis attributes the increase in mitochondria to the more
embryonic condition of the cell. It is very possible that it may depend upon
increase in oxidations (see p. 82). Even should there be some association between
the decrease in the number of mitochondria and the formation of fibrils, it does
not follow that the mitochondria themselves change into them. The arrange-
ment of mitochondria in chains may simply be the outward and visible sign of the
formation of fibrils between them from non-mitochondrial precursors.

The results thus far obtained with tissue cultures by Levi (1916c, p. 82) are dif-
ficult to reconcile with Duesberg's view, for Levi found that there was no relation-
ship whatever between mitochondria and the growth of fibrils in mesenchyme cells.

We hold, not without some justification, that mitochondria are chemically
a combination of phospholipin with a small fraction of albumin. Now we are
asked to believe that, at a certain stage in the development of the embryo, fila-
mentous mitochondria, which to all our solubility tests and staining reactions are
alike and show no variability, in three different locahties become chemically trans-
formed into three different materials. In the myoblasts they are said to change
into myofibrils, which contain tyrosin; in the neuroblasts they are supposed to
change into neurofibrils, the chemical nature of which is unknown; and lastly,
they are also said to form connective-tissue fibrils, which yield collagen, a protein
devoid of tyrosin. But the mitochondria do not contain tyrosin (Cowdry, 1916a,
p. 427). Where, then, does it come from? Certainly not from the mitochondria.
Other more diflficult questions must be asked and answered before we can bring
ourselves to beheve in the chemical transformation of mitochondria into myofibrils.

EPIDERMAL FIBRILS.

Firket (1911, p. 537) has investigated the role of mitochondria in the differ-
entiation of epidermal fibrils in the egg tooth and feathers of chick embryos. His
material consists of three series of chick embryos. The first, from 8 to 15 days'
incubation, was fixed in Bouin's fluid; the second, from 6 to 21 days, in Flemming's
fluid as modified by Meves; and the third, embryos from 6§ days until hatching, in
Benda's fluid. Preparations from the first were stained in safranin or iron hema-
toxyhn, either alone or followed by a counterstain of rubin, eosin, and orange G.
Some sections from the second and third series were also treated in this fashion,
although the majority were stained with iron hematoxylin or crystal violet.

Firket found (p. 540) a certain variability in the coloration of the fibrils after
fixing in Bouin's fluid and staining with iron hematoxyhn. The first ones to
appear were lighter colored and stained irregularly, whereas the completely differ-
entiated fibrils stained a darker uniform shade. He says (p. 544) that this "mon-

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. l09

iliforme" appearance leads one to suppose that the transformation into epidermal
fibrils takes place in several places in the substance of a single chondriokont. Ac-
cording to him, the fibrils are at first basophile, become acidophile, and are finally
masked, the whole cell assuming a homogeneous aspect. In the preparations which
he fixed in Meves's fluid and stained with either iron hematoxylin or krystallviolett,
both the mitochondria and the fibrils were stained. He writes (p. 542) that as
one examines the cells of more and more superficial layers of the corps muqueux
inferieur, one sees part of the chondriosomes elongate and assume an undulating
appearance; in the same cells other chondriosomes retain their primary dimensions.
The latter become less and less numerous as one approaches the corps muqueux
inferieur. 8oon almost all the chondriosomes have assumed the form of long
undulating filaments, which he says undoubtedly constitute the first-formed
epidermal fibrils. Since he found that the number of completely formed fibrils
in a cell greatly exceeds the original amount of mitochondria, he concludes that
the fibrils formed by the transformation of chondriosomes multiply by longitudinal
division, although neither he nor Branca observed it. He bases this conception
on two considerations: (1) that isolated epidermal fibrils are generally of finer
diameter than the chondriokonts; (2) that the division of other fibrillar formations
such as myofibrils and neurofibrils is admitted by most authors.

Duesberg's observations on the epidermis of the tadpole lead him to the same
conclusion that the epidermal fibrils arise by the transformation of mitochondria
(1912, p. 796).

Since the iron-hematoxylin method and the Benda method color the mito-
chondria and the completely formed fibrils (two very different structures micro-
chemically) alike, these methods of technique can not be regarded as suitable for
an investigation of this nature. More specific methods must of necessity be em-
ployed. It follows that Firket, in his series showing a parallelism between the
disappearance of mitochondria and the formation of fibrils, has had to rely solely
upon the diameter of the filaments to determine whether they are mitochondria
or differentiated fibrils. His position is therefore insecure, since he can not dis-
tinguish with certainty the structures between which he claims to show tran-
sitions. If we admit that this parallelism does exist, we find that it is capable of a
similar explanation to that advanced in the discussion of collagenic fibrils and
myofibrils, namely, that the amount of mitochondria is diminished because the
activity of the cells is lessened in the later stages of cytomorphosis.

The problem is rendered more difficult and deceiving because of the superficial
resemblance which obtains between the stauiing reactions of the mitochondria
and of the fibrils. Some of the fundamental differences which are said to exist
between them may be indicated :

Mitochondria. Fibrils.

1. Granules, rods, filaments. Long threads of finer diameter (Firket, p. ,'544).

2. Soluble in fluids containing an excess of acetic acid, Resistant to acetic acid in fixatives.

e. g., Zenker's fluid.

3. Destroyed by fixation in Bouin's fluid (Firket). Well preserved by Bouin's fluid (Firket).

4. Acidophile in pancreas after fixation in non-mordanting First basophile, becoming later acidophile (Firket,

fluids (Bensley, 1911, p. 362). p. 539).

110 THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

Firket's evidence is not conclusive, because he has not demonstrated the
existence of a complete series of transitional stages showino- tlie loss in the jiroiier-
ties of the fibrils. He has, however, gone further than Aleves did in connection
with the collagenic fibrils, or Duesberg in the case of the myofibrils, because he
succeeded in demonstrating a variability in the staining reaction of the first fibrils
to appear.

NEUROFIBRH.S.

Many iinestigators have touched on the question whether mitochondria play
a part in the differentiation of neurofibrils, but Hoven (lOlOo, p. 427) in partic-
ular, working with chick embryos, has furnished the most complete evidence in
favor of the view that mitochondria are actually transformed into them.

This interesting conception has been supiwrted by Meves, who originally
enunciated it (1907rt, p. 403), as well as by G. Arnold (1912fl, p. 288), and several
others to be mentioned subsequently; it has been rejected by Marcora (1911, p.
952), Levi (1911, p. 180), and Gurewitsch (1913, p. 126), while Duesberg (1912,
]). 745) has assumed a non-committal attitude with regard to it.

It is based upon the following statements :

(a) That the neurofibrils increase in amount as the mitochondria decrease, until
finally the adult condition is attained in which the neurofibrils are completely differ-
entiated and the mitochondria absent (Hoven, 1910o, p. 478; Meves, 19106, p. 655).

(6) That microchemical transitions exist between mitochondria and fibrils,
since the primitive neurofibrils may first be stained by mitochondrial methods,
then by both mitochondrial and neurofibrillar methods, and finally by the various
neurofibrillar methods of technique alone (Meves, 1908, p. 838; Hoven, 1910a,
p. 478, etc.).

(f) That morphological transitions also exist between mitochondria and
neurofibrils; according to Meves (1908, p. 838), chains of mitochondria are changed
into neurofibrils; according to Hoven (1910a, p. 475), the mitochondria form a
reticulum from which the neurofibrils are differentiated.

{d) That tlie development of myofibrils, connective-tissue fibrils, and the
fibrils in epithelial cells support this theory, since they, in a similar fashion, are
developed from mitochondria. This constitutes the argument from analogy (Meves,
1907fl, p. 403; Duesberg, 1910, p. 613; Meves, 1910a, p. 162; Firket, 1911, p. 545;
and Duesberg, 1912, p. 759).

The bearing of my own observations (1914^/) upon the statements upon which
the theory of the mitochondrial origin of the neurofibrils rests is as follows :

(a) My own findings are utterly at variance with the first argument, for I
can discover no decrease in the amount of mitochondria running parallel to the
formation of neurofibrils. Moreover, the statement that they are absent in the
adult condition is wholly unwarranted in view of the fact that several investigators
had already unquestionably seen mitochondria in adult nerve-cells (p. 101).

[b) The second statement postulates the existence of three distinct phases
in the development of the neurofibril, each of which is characterized by certain
microchemical properties. In the first stage the primitive neurofibrils may, it is

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. Ill

said, be stained by mitochondrial methods; in the second by both mitochondrial
and neurofilirillar methods; in the third by the various nevu-ofibrillar methods
of techni(iLie alone. I have found (and I have already described the fact) that
structures which we are accustomed to call neurofibrils may in truth be stained
by certain mitochondrial methods. I refer to the iron-hematoxylin method of
Meves, the Benda method, and the anilin fuchsin methylene-blue erythrosinate and
toluidin-blue methods of Bensley, but the staining is not specific and depends on
the degree of differentiation. A comparison of figures 21, 26, and 25 published
in my 1914d paper will be sufficient to show that this is true in the case of the
last-mentioned method. These three figures have been drawn from neighboring
sections of the same embryo of 100 hours' incubation, mounted on the same slide, and
stained with anilin fuchsin and toluidin blue. In the first figure, the differentia-
tion is practically nil, the mitochondria staining exactly the same color as the
neurofibrils; in the second figure it has been carried a little further, with the result
that the neurofibrils have lost their bright crimson color and have assumed a dull
red shade; wdiile in the last (fig. 25) the decolorization has been carried to an
extreme, so that the neurofibrils have lost all of the acid fuchsin and have become
stained with the differentiator, toluidm blue. It is to be noted that in these pro-
gressive stages of differentiation the initial affinity of the neurofibrils for the acid
dye (acid fuchsin), in which they resembled mitochondria, is gradually changed
to an affinity for a basic dye (toluidin blue), while the intensity of the coloration
of the mitochondria with the acid fuchsin remains unaltered. Furthermore, the
fact that the coloration of the neurofibrils by mitochondrial dyes is marked in adult
cells, which I have mentioned in a preceding contribution, should be taken into
consideration before regarding it as indicative of the existence of transitions
between mitochondria and primitive neurofibrils.

Let us now consider the statement that the primitive neurofibrils may be
stained by both the mitochondrial and the neurofibrillar methods (i. e., the second
phase). The completeness of the demonstration of mitochondria by the iron-
hemato.xyUn method depends upon the presence in the fixative of chromic acid,
osmic acid, and acetic acid, in suitable amounts, and on the mordanting action of
iron alum, while theu- complete absence in the neurofibrillar preparations is due
to the unmodified action of silver nitrate. The neurofibrils seem to have a special
affinity for silver nitrate, upon which all silver impregnation methods depend.
So it is extremely unhkely, especially in the absence of direct evidence, that so
widely divergent methods stain the same thing, namely, the primitive neurofibrils.
Of course it will be argued by the adherents of this theory that the Italian inves-
tigators have succeeded in demonstrating mitochondria by modified Golgi methods,
but there is a long step between this fact and proving that the mitochondria in a
certain specified stage in the developing nerve-cell may be stained interchange-
ably by mitochondrial and neurofibrillar methods. If this should be the case in
other stages when neurofibrils are not being formed, and in other tissues, it would
of necessity be deprived of the significance which investigators have been incUned
to attach to it.

112 THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

Finally, the neurofibrils are said to enter upon a third phase in their history
characterized by the loss of their affinities for mitochondrial dyes. I have never-
theless failed to find any conclusive evidence that the neurofibrils change their
chemical properties after their first formation. My failure may be due to the
unstandardized condition of the neurofibrillar methods of technique which still
prevails. In any case the burden of supplying the evidence rests with those who
make the statement. If the neurofibrils are formed by a chemical transformation
of mitochondrial substance into neurofibrillar material, one would expect to find
variations in the effects of fixation and in the staining properties of mitochondria
during their formation. I have shown that the exact converse obtains. Both
the solubility of mitochondria in acetic acid and the staining reactions of mito-
chondria in the cells of the neural tube in which neurofibrils are being actively
formed remain remarkably uniform and constant. Moreover, these properties
apparently differ in no wise from those of mitochondria in the neural tube before
the formation of neurofibrils or from the mitochondria in other embryonic cells.
It is evident, therefore, that the facts do not justify the statement that microchem-
ical transitions exist between mitochondria and neurofibrils.

(c) With respect to the evidence for morphological transitions I would state
that I have failed to confirm Meves's contention that chains of mitochondria are
transformed into neurofibrils. Mitochondria are sometimes oriented end to end,
and one may often observe very long filamentous forms. It is a very far cry from
a hnear arrangement of mitochondria or from long filamentous mitochondi'ia to
neurofibrils. This is manifested, among other things, by the fact already men-
tioned, that there is nothing pecuharly distinctive about the morphology or the
arrangement of mitochondria in the cells of the neural tube during neurofibrillar
formation; they are ahke indistinguishable, on the basis of their morphology and
distribution, from the mitochondria in the cells of the neural tube in stages prior
to the differentiation of neurofibrils, and from the mitochondria occurring in other
embryonic tissues both before, contemporaneous with, and after the development
of neurofibrils. Therefore, on the ground of the shape and cytoplasmic arrange-
ment of mitochondria, there is just as much evidence for the formation of neuro-
fibrils in structures derived from mesoderm and endoderm as there is in the case of
the neural tube.

(d) The value of the argument from analogy has already been made plain
by the preceding discussion of the development of other fibrillar structures.

It seems clear that the neurofibrils are not formed by a transformation of
mitochondria. They are elusive structures. They can not be seen in the living
cell, or with the aid of vital dyes; neither can they be dissected out (Kite^. There
is every reason to believe that they do not occur in the living cell in the form in
which we see them in our silver preparations. They are formed of material quite
different from the Nissl substance, mitochondria, or canalicular apparatus, though
we neither know what it is nor the factors concerned.

'Dr. Kite, personal communication.

THE MITOCHONDRTAL CONSTITUENTS OF PROTOPLASM. 113

It has been claimed that many other fibrillar structures are formed from mito-
chondria. Saguchi (1913, p. 239), for example, has made a careful study of the
development of the tonojibrils and intracellular structures of Eberth in the epidermal
cells of batrachian larvae and concludes that they arise through a chemical trans-
formation of mitochondria. His figures show a very definite association of the
mitochondria with the fibrils and all morphological gradations between the two
are represented. His evidence, however, for a chemical change (p. 241) is not
conclusive, for it does not exclude the possibility that the substances making up
the fibrils arise elsewhere and are deposited within the mitochondrial filaments
without change of the mitochondrial material. The fibrils of Herxheimer are
considered by Favre and Regaud (1910, p. 1138) to be true mitochondria. Meves
(1907a, p. 403) has suggested that the neuroglia Jibri Is are hkewise developed from
mitochondria, but has forwarded no evidence in substantiation.

In concluding this discussion of the histogenesis of fibrils it may be remarked
that it is somewhat illogical to suppose that substances of such diverse chemical
constitution are all formed through the transformation of a single substance, a
phospholipin, combined perhaps with a small fraction of albumin. The coUagenic
fibrils on boiling yield gelatin, a protein devoid of tyrosin; the neurofibrils are of
unknown composition, and we even doubt their existence in the living cell; certain
of the epidermal fibrils contain a keratin-like material, and our chemical acquaint-
ance with the neuroglia fibrils is of the slightest. It seems far more likely that the
fibrils are, from the outset, different from one another— that they are formed from
different materials rather than from the same material. They may be formed
through a condensation of substances in the cytoplasm, either in the form of minute,
perhaps ultramicroscopic, particles which tend to be arranged in rows following
lines of stress or strain in the cell and which naturally fuse together, end to end, in
accordance with the law of least surfaces; or, it is entirely conceivable that there
may be only a single center of condensation which grows and enlarges by the addi-
tion of more material by accretion, something like a crystal. In connection with
the first of these alternatives we know that Unes of stress do exist in living proto-
plasm, for the fibrils are deposited at a time in development when growth activi-
ties are greatly pronounced, when cellular migrations are common, and changes
in the form of various parts of the body quite frequent. The fully formed fibrils
undoubtedly correspond, in position, to fines of stress and strain. We know that
certain cytoplasmic elements are subject to orientation along such lines. I refer,
particularly, to the arrangement of material in lines about the centrosomes, to the
arrangement of granules in the rootlets of ciUa?, to the characteristic deposition
of material in bone, etc.

PLANT PLASTIDS.

By far the most convincing evidence in favor of a participation of mitochondria
in histogenesis, through an actual chemical transformation of their substance, is
to be found in the botanical literature. In fact, these newer methods of mito-
chondrial technique strike at very important problems, for they have a definite
bearing upon the origin of all plastids.

114 THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

Lewitsky (1910, p. 542) and Pensa (1910, p. 325) deserve credit for opening
up this important field almost simultaneously. Lewitsky studied mitochondria
in growing asparagus tips. He claims to have discovered that the leucoplasts arise
from them, and advances the general conclusion that the chondriosomes (mito-
chondria) in plant cells are to be regarded as formative granules, just as they are
supposed to be in animal tissues. In other words, he extends Meves's important
generalization ip. 101) to plant tissues. Pensa arrived at essentially similar results,
and the problem has since been flooded with contributions from all quarters.

We have to consider the formation of leucoplasts and starch; of chloroplasts
and chlorophyll; of chromoplasts and pigments of almost infinite varietj^; and of
the elaioplasts, which elaborate fats. Let us analyze the facts observed and see
whether they permit of any other interpretation than that of direct chemical trans-
formation. It can not be denied that there is a very definite topographical rela-
tion between the mitochondria and the deposition of these substances, for they
are actually laid down within them, but it is too much to say that we are here deal-
ing with an actual chemical transformation of the mitochondrial substance. Yet
this is just the claim that Guilliermond (1912r(, p. 394) makes. He has shown that
filaments which possess these swellings are chemically different from the other mito-
chondria in the cell, because their solubilities in certain fixatives are different ; but
it does not follow, nor is there any reason to suppose, that further alterations occur
in the constitution of the mitochondria by which they become changed into starch,
pigments, chlorojihyll, and so on. Guilliermond (1912^, p. 408) points out at length
that Hoppe Seyler and others look upon chlorophyll as toeing a combination of
lecithin and other substances, and that it accordingly resembles mitochondria
quite closely. The other pigments and the crystalline substance "carotine" differ
widely from mitochondria chemicall}-. To my mind, the facts observed, which no
one would question, demonstrate nothing more than that the said topographical
relationship between mitochondria and the formation of these materials exists.
These substances are simply dejiosited. or accumulated, or heaped up within the
mitochondria, which serve as a convenient and suitable vehicle, by virtue of their
chemical and physical properties. There is good reason to believe that some of
the substances, like the pigments, may be soluble within them. They may also
act as condensers, as Regaud believes. But what I want to make absolutely clear
is that this does not necessarily involve any chemical change whatever in their
constitution.

The fact that mitochondria diminish in number, pari patisu, with differentiation
in plant cells as well as in animal cells, does not mean that they change chemically
into the products of differentiation, as some investigators have tacitly assumed,
for this decrease in amount of mitochondria is probably associated with a decrease
in the rate of metabolism which we know occurs with differentiation and senility.
The evidence now at hand that the mitochondria are concerned with metabolism
is given in detail on page 131. There is nothing to indicate that the mitochondrial
substance about the granule of starch, or the i:)igment, as the case may be, is chem-
ically transformed into something else as it decreases in amount. It may simply

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. 115

become resorbed and ^o into solution in the cytoplasm, for we have good reason
to believe that the mitochondria do go back into solution (p. 98). These are not
all the points that must be cleared up before we can look upon this question of the
actual chemical transformation of mitochondria into plastids as being definitely
settled. The work which has lieen done so far is extremely suggestive, but it is
not conclusive.

The evidence seems to point to the conclusion that the single chloroplasts in
some of the algae do not arise from mitochondria, because mitochondria are absent,
so that they would constitute an important exception to the general rule. It is
thought, moreover, that these single plastids perform a similar function to that of
those in the adult cells of higher plants, where the mitochondria are either absent
or greatly reduced in number. This, by Sapehin, is brought forward as evidence
against the view that mitochondria are transformed into plasts. It is necessary
to find out whether the chlorojilasts in animals, some varieties of Faramocciuin
for example, are formed from mitochondria in the same way that has been claimed
in plants. Information on this point is urgently needed, because it would tell us
whether animal mitochondria are capable of performing the same feats which have
been ascribed to vegetable mitochondria, and give us an idea of the degree of
resemblance of mitochondria in the plant and animal kingdoms. Varying degrees
of chlorophyll production, caused by regulating the illumination, should be studied.
It would be interesting to note whether phosphatids outside the body are capable
of entering into close combination with starch and chlorojihyll. And, finally,
experiments might be devised to show whether animal mitochondria and plant
mitochondria, in species devoid of chlorophyll, are able to pick up and condense
starch and chlorophyll when they are brought into intimate contact with them,
which would have a very important bearing upon this question of the transforma-
tion of mitochondria into plastids.

PIGMENTS.
HEMOGLOBIN.

Ciaccio (1911, p. 16), in October, arrived at the conclusion that hemoglobin
is formed under the influence of mitochondria on the basis of his observation that,
in the rabbit, not only the basophilic erythroblasts but also those provided with
hemoglobin, and the erythrocytes, just before they enter into the circulation, con-
tain typical mitochondria. In October, also, Meves (1911a, p. 495) published
identical results relating to the red blood-cells in the bone marrow of the guinea-pig.

A few months later, Schridde (1912, p. 516) attempted to go further than
Ciaccio and Meves by claiming that the mitochondria diminish in number in
direct proportion to the increase in the hemoglobin, and, on the basis of this, con-
cluding that the plastosomes (mitochondria) are the formers of the hemoglobin.

This was followed almost immediately by a second paper by Ciaccio (19136,
p. 393), in which he says that there is no foundation for Schridde's opinion. He
holds to his original view that the part played by the mitochondria is quite indirect.
His chief and only objection appears to be that the acid fuchsin, which J^chridde

116 THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

employed, colors the hemoglobin intensely and may mask or hide the mitochondria
(plastosomes) ; but Ciaccio ignores the fact that the iron-hematoxylin stain, which
he himself used, acts in m.uch the same way.

It is quite evident that our information is not sufficiently clear-cut. The
technique upon which the above-mentioned observations are based is inadequate.
There has been no attempt made to estimate quantitatively either the mitochondria
or the hemoglobin, and in spite of this the old i?resence-and-absence argument is
made use of. Fixed preparations were alone studied. Now that it is not only
possible, but easy, to stain the mitochondria specificall}^ with janus green in these
living blood-cells, and to count them, it ought not to be difficult to devise some
quantitative colorimetric way of estimating the hemoglobin and to obtain decisive
results. The information at hand does not show that the mitochondria exercise
even an indirect influence upon the formation of hemoglobin. There is no topo-
graphical correspondence between the mitochondria and the deposition of chloro-
phyll. The formation of hemoglobin from mitochondria alone involves chemical
impossibilities. It involves a change from a phospholipin into an entirely differ-
ent material. Ninety-four per cent of the hemoglobin consists of the protein
globin, but it also contains the coloring matter hematin. It is accordingly very
much like a nucleoprotein in nature, and nothing could be further removed from
the mitochondria. We ask ourselves where can the iron come from? Certainly
not from a phosphohpin. Pathology may help us. In chlorosis the characteristic
thing is a diminution in the amount of hemoglobin in proportion to the number of
red blood-corpuscles. There is either a deficiency in the formation of hemoglobin
or an increase in its destruction. At any rate, the red blood-corpuscles contain
less than the normal amount, and it would be interesting to find out whether
there is a corresponding fluctuation in the number of mitochondria in the precur-
sors of these erythrocytes; for if a relation exists between the mitochondria and the
hemoglobin production it might be possible to detect it in this little-known condi-
tion. The mitochondria should also be studied in hemochromatosis and in other
disturbances of a similar character. It is interesting to note that PoUcard (19126,
p. 230) has discovered that mitochondria form the matrix in which hemoglobin
crystals are deposited in the liver of animals defibrinated by the process of Magendie.
Mitochondria, however, are not associated with the deposit of other crystalloids
(d'Athias, 1915, p. 68), except perhaps in plants.

SECRETIONS.
THVROID-GLAND SECRETION.

Grynfeltt (1912c, p. 147) was the first to suggest the formation of the colloid
substance through either a direct or an indirect transformation of mitochondria.
He thinks that this is quite probable in view of our general knowledge of the role
of mitochondria in gland-cells and in consideration of some observations which
he has made upon the new-born dog. He found appearances which seemed to
him to indicate that the mitochondria near the surface of the apical zone of the
thyroid cells undergo certain modifications and transform into clear spherules,

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. 117

having some of the reactions of the colloid substance. But he admits that his
observations are not sufficiently numerous to permit him to arrive at any definite
conclusion. It is difficult to conceive of the direct transformation of a phosphoUpin
into a compound containing relatively large quantities of iodine.

Goetsch's (1916, p. 132) recent work on toxic adenomata of the thyroid gland
is of very great interest in this connection, for he found that there is a great
increase in the amount of mitochondria parallel with the appearance of the chnical
symptoms of hyperthyroidism. This may be interpreted by the adherents of the
transformation hypothesis to mean that the epithelium is secreting more rapidly,
that the mitochondria are increased in number for this reason, and that this is
evidence that they are actually transformed into the secretion, just as Grynfeltt
supposes. But this explanation is taking a good deal for granted. True, the
hyperthyroidism may be due to an increase in amount of a single secretion which
is produced normally, yet there is some evidence for the alternative assumption
that the thyroid secretion is polyvalent and not univalent. Moreover, it is possi-
ble that the symptoms of hyperthyroidism may result from an increased rate of
metabolism on the part of the thyroid epithelium and the consequent liberation in
excess of products of this heightened metabolism rather than of normal secretion.

PARATHYROID SECRETION.

The evidence in favor of the mitochondrial origin of tliis secretion is still less
satisfactory and convincing. Bobeau (1911, p. 186) bases his conclusion entirely
upon a correspondence in the distribution of the mitochondria and certain lipoid-
like droplets in the cells. He demonstrated the mitochondria by mitochondrial
methods which he does not specify and the Hpoid by the method of Ciaccio. He
assumes that the lijwid droplets constitute a precursor of the secretion, or rather
the secretion itself, and that they arise by a swelling-up of the mitochondria.

CEREBROSPINAL FLUID.

Though Hworostuchin (1911, p. 232) was the first investigator to supply us
with an accurate description of the mitochondria in the choroid plexus, it remained
for Grynfeltt and Euziere (1912, p. 64) to make an attempt to discover their rela-
tion to the formation of the cerebrospinal fluid. They found three types of cells
in the choroid plexus of mammals: (1) striated cells, containing many long fila-
mentous mitochondria generally running from the base of the cell toward its distal
portion; (2) vesicular cells, filled with small vesicles possessing clear centers sur-
rounded by a peripheral layer of stainable substance; (3) vacuolated cells, crowded
with droplets of variable dimensions. They concluded that these three represent
different stages in the same process of secretion. According to them the mito-
chondria enlarge to form the vesicles, the vesicles change into the vacuoles, and
the vacuoles discharge their contents into the ventricular system and thus form
the cerebrospinal fluid.

PoUcard (1912e, p. 430), in a short paper iniblished only a few weeks later,
says that he is unable to accept this interpretation; for he is of the opinion that
we have to deal with two processes, not with one. He found all stages between

118 THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

the mitochondria and the small lipoid vacuoles, but he could not discover any
relation whatever between these and the large vesicles which invade the whole
cell; so that he is unable to agree with Grynfeltt and Euziere concerning the
origin of the cerebrospinal fluid.

Ciaccio and Scaglione (1913, p. 167) question the evidence in favor of this
interpretation. They call attention to the fact that Grynfeltt and Euziere used
for the most part the choroid plexuses of horses killed in the abattoir. They say
that horses of this kind are usually old and in poor condition and that for this
reason results based upon them are unreliable.

Grynfeltt and Euziere (1913fl, p. 198) have not met Policard's criticism, but
they have answered that of Ciaccio and Scaglione by a careful study of animals of
different kinds killed in a variety of ways. They found that cells of these three
varieties occur in many forms of mammals, and they also discovered that bleeding
and the administration of pilocarpin increases greatly the number of cells contain-
ing the large, clear vesicles. They advance this as evidence that the vesicles con-
stitute in truth a stage in the formation of the cerebrospinal fluid. They have also
(19136, p. 101) extended their studies to the selachian Scyllium canicula among
the lower vertebrates. The choroid plexus cells in this animal have a particularly
well-developed striated border, which enabled them to make a detailed study of
the fate of the large clear vesicles. They observed them pass through the plasma
membrane and break for a moment the regularity of the striated border before
discharging into the ventricular cavity. This, they urge, is strong evidence in
support of the vesicular theory of the secretion of the cerebrospinal fluid and of
their general contention of the role played by mitochondria in its formation. They
do not, however, throw any further light upon the crucial question of the existence
of transitions between the mitochondria and the vesicles. In a still more recent
paper they make a general review of the whole problem, but contribute nothing
new to the discussion.

My own studies lead me to agree with Policard, that Grynfeltt and Euziere
may be dealing with more than one process and that we have no suflftcient reason
to believe that transitions occur between the mitochondria and the vesicles. Even
should such transitions occur, we would want to know definitely whether or not
the vesicles do form the cerebrospinal fluid.

.SECRETION OF THE PAROTID AND SUBMAXILLARY GLANDS.

Regaud and Mawas (19096, p. 220) have studied the mitochondria in the
sero-zymogenic cells of the parotid of the ass and the human submaxillary. They
found that there is a quaUtative relationship between the amount of mitochondrial
substance and of zymogen; where there is a large amount of mitochondrial sub-
stance there is little zymogen, and vice versa. Between these two extremes there
is a complete series of gradations, and they believe these represent stages in
secretion. They describe transitions between the mitochondria and the zymogen
granules in the form of sphericMl Ixxlies of variable size and staining reaction
embedded in the suljstance of the mitochondrial filaments, and they formulate

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. 119

the following theory of secretion: The mitochondrial filaments fix the substances
which the cell takes up from the blood. At one or more points in the course of
each filament there is an accumulation and elaboration of these materials, and at
these points the filament swells up into spherules to which one can apply the name
of "plastes." Each plast gives rise to a grain which matures and grows little by
little. Usually before the grain has acquired its definitive size and colorability the
mitochondrial filament in which it is embedded becomes paler and can no longer
be seen. The plast or grain is then set free in the protoplasm. At the moment
of excretion there is a dissolution and the product passes in the dissolved state
through the cell-wall. By this ingenious and plausible hypothesis they overcome
the difficulty of assuming that the mitochondria form the secretion through actual
chemical transformation.

F.4.T DROPLETS IN SEBACEOUS GLANDS.

Nicolas, Regaud and Favre (1912a, p. 203), working on human tissues, have
done little beyond confirming Altmann's observation of vesicular mitochondria
in the cells of sebaceous glands. Arguing from analogous appearances in other
tissues, they suggest that these give rise to the droplets of fat, but they do not
attempt to describe any transitions between the two and advance no evidence in
support of their suggestion.

SECRETION OF S^^■EAT GLANDS.

Here also Nicolas, Regaud and Favre (19126, p. 191) have studied the rela-
tions of mitochondria to secretion. They have found a relationship between the
number of mitochondria and the number of secretion granules, but they have failed
to discover any indication of the granules developing within the mitochondria as
in the parotid and submaxillary glands. They emphasize the fact that in sweat
glands the mitochondria are small and usually granular, and they say that even if
the granules did arise within them, as they believe to be the case, it would be diffi-
cult to observe it.

SECRETION OF MAMMARY GLANDS.

Hoven (1911, p. 325) has made a careful study of mitochondria in resting
and lactating mammary glands. He has arri\'ed at the conclusion that they play
a part in the formation of the different constituents of the milk. According to
him, the mitochondria break up into granulations, some of which transform into
grains of secretion, which give rise to the casein and the sugar; others transform
into fat.

SECRETION OF THE PROSTATE.

Akatsu (1903, p. 566) gave the first clear-cut description of mitochondria,
under the heading of "Altmann's granules," in the cells of the prostate. He ex-
pressed the opinion that they gave rise to granules of secretion. More recently
Dominici (1913. p. 295) has worked over the entire question and has found that
the mitochondria vary cjuantitatively with the activity of the cell. He thinks,
however, that they play an indirect part in the formation of the secretion and that
they are not directly transformed into it. He claims that De Bonis (1907, p. 14)

120 THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

simply saw lipoid droplets in the epithelial cells, and I am inclined to share his
opinion on the basis of my own prostate preparations.

SECRETION IN THE VENOM CELLS OF MUREX TRUNCULUS.

Grynfeltt (1913, p. 11) found that the venom cells in the hypobranchial gland
of Murex contain large granules of secretion antecedent (which stain intensely with
picric acid and are therefore called "Boules picrophiles ") as well as large and con-
spicuous mitochondria. He claims to have traced a genetic relationship between
the mitochondria and the secretion granulations. He found large mitochondria
which stain more faintly than the rest with the crystal violet in Benda's stain, and
large mitochondria with a central core of material staining orange with alizarin
just as the secretion granulations do. He also found that the masses with more .
of the central yellow-staining material possessed less peripheral coating of mito-
chondrial substance, and vice versa. He interprets these observations as follows:
In the process of secretion certain mitochondria, arising through the fragmenta-
tion of chondriocontes, increase in size. Their central part undergoes a chemical
transformation by which it loses its coloration with crystal violet and takes the
orange color of the alizarin. This tint, at first very pale, increases more and more
in intensity until finally it is identical with that of the secretion granules. He con-
cludes that the mitochondria play the part of plasts and he regards the product
as chemically a transformed mucus and the cells as goblet cells. His interpreta-
tion does not necessarily follow from his observations. On page 113 I have dis-
cussed in detail the theory that the mitochondria are plast-formers. Suffice it
here to say that he has presented no evidence that the mitochondrial substance
itself changes chemically into the material of the secretion antecedents. It may
be, as I say, that the secretion is formed in the surrounding cytoplasm and is accu-
mulated in the mitochondria on account of its solubility in them or for some other
reason. The mitochondria may be entirely passive in the process. Neither are
we justified in sajdng that the disapi)earance of the mitochondrial substance indi-
cates that it is transformed into the secretion, for it may well be that it goes back
into solution as in the parotid and submaxillary glands. Furthermore, we would
like to know something about what the steps are in the chemical transformation
of a phospholipin into a mucin, and it is just this that makes one so skeptical. We
have not the right to say, in any of this work on mitochondria, that a substance
of totally different character is formed by a chemical transformation of their sub-
stance, because it is altogether impossible to dissociate the mitochondria from their
surroundings. These attempts to make definite statements are futile. We can
hope only to make approximations to the truth and should bear in mind that the
apparently homogeneous ground-substance of the cytoplasm probably plays the
most important part in secretion as well as in all other cytoplasmic activities.

PANCREATIC ZYMOGEN.

Hoven (1910f>, p. 349) has furnished detailed evidence in favor of the par-
ticipation of mitochondria in the formation of secretion granulations in the pancreas.
He records the presence of little swellings in the course of the mitochondrial fila-

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. 121

ments which he calls "plastes," and which stain like the mitochondria and secre-
tion granules with iron hematoxylin, crystal violet, and acid fuchsin. These, he
believes, change into secretion granules, since he has observed all stages of transi-
tion between the two. He is of the opinion (contrary to Regaud) that we are here
deahng with an actual transformation of mitochondrial material.

Mislawsky (1911a; 19116, p. 505), however, on the basis of very similar ob-
servations, finds no evidence of the direct transformation of mitochondria into
secretion granulations. Schultze (19116, p. 258) has also studied the mitochon-
dria and zymogen granules in the frog, but does not commit himself regarding
their genetic relationship. Champy (1911, p. 122) experimented with secretin and
found that intermediary forms between mitochondria and secretion granules are
more numerous during secretory activity. Laguesse (1911, p. 277) maintains
that his ergastidions (mitochondria) play the part of elaborators in the produc-
tion of pancreatic zymogen. CI. Arnold (19126, p. 268) claims that the zymogen
gramdes are formed through a maturation of mitochondria, and Chaves (1915,
p. 67) speaks of a physiological transformation.

Key (1916, p. 215) has experimented with the pancreas of the toad. He
stimulated the gland by the injection of pancreatic secretion in some cases and of
pilocarpin in others. In some experiments the injections were repeated at regular
intervals for several days. He found that while this caused a discharge of zymogen
granules, the mitochondria were not exhausted, but in some cases seemed actually
to increase in length. He was unable to detect any difference between certain
bleb-like swellings, which mitochondria possess in almost all secreting cells, and
the other parts of the mitochondrial filaments, and he concluded from this that
the blebs do not contain zymogen granules. Furthermore, the absence of reciprocal
changes in the amount of mitochondria with variations in the cytoplasmic content
of zymogen granules led him to believe that the zymogen granules are not formed
directly from the mitochondria.

Scott (1916, p. 249), working in this laboratory, studied the effect of experi-
mental phosphorus poisoning upon the pancreas of the mouse. He discovered
that slight poisoning brings about a change in the mitochondria only, which lose
their bleb-like swellings and their filamentous shape. The nucleus and zymogen
granules show no changes. The interesting thing is that in a mild case of this sort
the formation of zymogen is not interfered with. Indeed, in much more severe
intoxications unmistakable evidences of further formation of zymogen granules
are seen. Whole cells are frequently found crowded with them. This means that
the mitochondria do not participate through their bleb-like swellings in the produc-
tion of zymogen, because zymogen continues to be formed long after these swellings
disappear. It can not come from the swellings, because there is none. This is
another strong blow to the doctrine that the mitochondria transform into secretion
granulations.

URINARY SECRETION.

The question of the relationship between changes in the mitochondria and
variations in urinary secretion is a difficult one. It is complicated by the presence

122 THE MITOCHONDUIAL CONSTITUENTS OF PROTOPLASM.

of the so-called "hatonnets" of Heidenhain, which resemble mitochondria in some
respects but differ altogether in others (Policard, 1910, p. 225). In many cases
it is hard to tell to which investigators refer. The "batonnets" have been known
for years, and much careful work was done on their possible relation to secre-
tion before the more elusive mitochondria were discovered. Another difficulty is
encountered in establishing a normal from which to work. Analogy is a pitfall
here, because there are such marked differences in the kidneys of different forms.
Certain snakes and amphibians are good material because the cells, unlike those
of adult mammals, contain characteristic secretion granulations like other glands.
Above all it is necessary to distinguish sharply between physiological changes and
pathological lesions; the latter will be considered on page 137.

Benda (1903, p. 127) makes no reference to any relationship between the
mitochondria and the formation of secretion except to suggest that by contrac-
tion they draw the proximal and distal ends of the cell nearer together and thus aid
in the expulsion of the secretion. To this Policard (1905, p. 382) rightly objects.
Modrakowski (1903, j). 230), at an earlier date, described definite changes in the
Altmann's granules (mitochondria) in experimental diuresis and suggests that they
may act as condensers in the formation of secretion, but his illustrations do not
show them. The distinctive features of mitochondria in the different segments of
the urinary tubules have been studied by Regaud (1908c, p. 1145) and others.

Regaud (1909f, p. 1035) was the first to claim that the mitochondria play a
definite part in the formation of the secretion. He worked with the kidneys of
snakes and found that where the mitochondria are abundant the secretion granules
are few, and vice versa. He records the gradual formation of the secretion granules
in the substance of the mitochondria in precisely the same way as in the salivary
glands.

Policard (1910, ]). 272), however, working with the frog, after experimenting
in many ways, remarks on the fixity of the mitochondria and describes no marked
variations in them depending upon the quantity and character of the urine, though
they respond very readily by fragmentation to ]3athological changes. He has
described also (1912c, p. 450) the transformation of nutochondria into certain
granulations of unknown nature in the developing human kidney. Fahr (1914,
p. 120) has been able to stain secretion granulations and mitochondria differentially
in the rabbit's kidney and believes that there is no relation between them.

More recently, Oliver (1916, p. 318) has studied the modifications which the
mitochondria undergo in experimental diuresis. The "batonnets" in the cells
of the proximal convoluted tubules, where the urea is secreted, lose their rod-like
form and occur simply as rows of granules. 8he found that the urea also appears
in the form of granules wliich are likewise arranged in rows and suggests that the
secretion of urea is bj- means of the mitochondrial batonnets which act as condensers.
It may be remarked that the rows of urea granules do not necessarily correspond
with the rows of mitochondria; they may simply alternate with them. It may
also be worth while to inquire whether phosphatids, of which we believe mito-

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. 123

chondria to be composed, are able, outside the body, to pick up, condense, and
concentrate urea. We desire further information on these and other points before
passing tentative judgment on the possible role of niitochondria.

FATS.

The relation between vitellus, neutral fat, lipoid, and myelin droplets and
mitochondria is undoubtedly very intimate, for we have good reason to suppose
that mitochondria are themselves, at least in part, phosphatids, and phosphatids
are made up of phosphoric acid, fatty acid, glycerol, and some nitrogenous base
like chohn. The exact chemical relationship of yolk spherules to ovovitellin is
little understood, but it seems clear that, though ovovitellin is more of the nature
of a nucleoalbumin, it contains, nevertheless, a large amount, some say as much
as 25 per cent, of the phosphatid lecithin. In recent years interest has become
focussed on these phosphorized hpoids and there is an ever-growing demand for
accurate information regarding them. The difficulties presented to those who
attempt to study them within individual cells are verj' obvious and should be
kept in mind during the subsequent discussion.

There are indications in the oogenesis of almost all organisms which may be
taken to mean that the mitochondria are either partially or totally transformed
into deutoplasmic substances, hkc vitellus. The same is true in certain cases of
spermatogenesis where the male sex-cells are very large, approaching the eggs in
structure. The relation between mitochondria and vitelline granules has also
been studied in amphibian embryos undergoing metamorphosis.

The oft-cited observations of Loyez (1909, p. 191) are generally regarded as
constituting the most con\'incing evidence of the direct transformation of mito-
chondria into vitelline globules. She was able to distinguish the following stages
in the development of the eggs of Ciona intestinalis: (1) in the youngest oocytes
only a few granulations may be seen about the germinative vesicle and in the periph-
eral cytoplasm; (2) in a more advanced stage the mitochondria are disposed in
linear series, in strings, throughout the cytoplasm; (3) the individual mitochondria
become spherical, increase in size, and in their interior present a central and clearer
area. Finally these globules fuse together, increase still further in size, and be-
come the definitive vitelline sjiheres. She remarks, further, that in other closely
related ascidians, like Cynthia tetraedra and Cynthia morus, the vitelline globules
are not formed from mitochondria, but arise quite independently from them.
These results are partially confirmed by Govaerts (1913, j). 415), working with
insects, but he has not been able to demonstrate a direct transformation of mito-
chondria into vitelline spherules.

The clumping of mitochondria and their fusion to form fatty masses which
enter into the composition of the vitellus have been described by Henneguy in the
oogenesis of Pynhorcis apterus, and by Faure-Fremiet (1910a, p. 548) in both the
oogenesis and the s[)ermatogenesis of Lithobius forficatus, but they do not specify
the properties of the said fatty masses.

124 THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

Van Durme (1914, p. 118) has gone into the question in detail from the mito-
chondrial point of view in birds. He describes the origin of vitelline granules by
a direct transformation of mitochondria and by the initial deposition of vitelline
substance in a clear vacuole and its subsequent growth.

Coghill's observations (1915, p. 349) on the relation of mitochondria to yolk
are perhaps the most interesting^ since he alone worked with living cells. His
investigations point to the conclusion that the change takes place in the reverse
direction, that is to say, from yolk to mitochondria. He studied with great care
yolk-laden cells of amphibian embryos which he stained vitally with janus green.
He observed structures actually originating from the surface of the yolk globules,
by a chemical transformation of their substance, which migrated into the rest of
the cytoplasm, where they became indistinguishable from the other mitochondria
in the cell. Their morphology, their reactions to janus green, and their staining
properties when fixed were identical with those of true mitochondria. From this
he concluded that the mitochondria arise from the yolk. Without subscribing to
this view, it may be said that these results are not so much at variance with those
of other workers as they might at first sight appear to be; for many reactions are
reversible and there is no reason why this should be an exception to the rule.
Then again, the amphibian embryo is an entirely different tissue from the avian
or mammalian egg and there is no reason why the processes going on in it should
be identical.

The possible relationship of mitochondria to the formation of neutral fat has
been under discussion since the time of Altmann (1889, p. 94), Metzner (1890, p.
82), and others; but Dubreuil (19116, p. 264) has given the most detailed as well as
the most recent account of it. The relationshij) of mitochondria to droplets of
hpoid in nerve-cells I have already discussed (1914o, p. 13). Reference should
also be made to the work of Azzi (1914, p. 7).

OTHER PRODUCTS.

The problem of the origin of the external segment of the rods and cones of the
retina is a very intricate one, but a few words of description will serve to make it
a little clearer. It is well known that the intimate structure of the rods and cones
is essentially the same, though their form differs. The constitution of the outer
segment of each is alike, so far as our present methods reveal it. It is very dense,
especially about the periphery, which is bounded by a sort of envelope. It fre-
quently exhibits a transverse striation, just as if it were made up of superposed
disks. Authorities are not agreed as to whether this occurs in the living condition
or whether it is sim{)ly the result of the technique employed (see Mawas, 19106,
p. 115). A distinct longitudinal striation has been described by Ranvier in
batrachians, but it does not seem to obtain in other forms. The core of this outer
segment contains, according to Leboucq (1909, p. 597), an axial filament which
is attached to a centrosome in its most proximal part.

There is also a close resemblance between the structure of the inner segment
of the rods and of the cones, but it is somewhat more complicated than the outer
segment, is less dense, and possesses more fluid cytoplasm. There is no hint of

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLAglVt. 125

transverse striation, but the longitudinal striation is very marked. It is difficult
to analyze. Apparently several factors enter into it. In the first place, there
are the fibers of Mliller, which do not belong to the cell itself. The exoplasmic
prolongations of the membrana liniitans externa give a striated appearance. The
so-called baskets ("corbeilles") also contribute to it. These are supportive. The
cell-membrane itself seems, according to the best accounts, to be striated. Mawas
(19106, p. 117) emj^hasizes, in the rabbit and in man, a striation which is intra-
cytoplasmic and wliich is due to the pecuhar arrangement of the mitochondria in
a peripheral sheet just beneath the cell-membrane. This, however, Leplat (1913,
p. 220) failed to observe in birds. Leboucq (1909, ]). 594) gives a detailed descrip-
tion of a system of intracytoplasmic filaments arising near the centrosome at the
base of the external segment, at first spreading out in the inner segment to form
the "ellipsoid" or "Fadenapparat," then condensing into a single filament, in
the rods, and into a bundle of three or four filaments in the cones and running
toward the nucleus. The neurofibrils are also to be reckoned with in connection
with this longitudinal striation; little is known for certain about them. Small
droplets of oil, "Olkligeln," occur in this inner segment in the early stages of devel-
opment and may persist in the fully differentiated state. Small quantities of pig-
ment are occasionally observed. Nothing is known of the Golgi apparatus in the
rods and cones in the adult condition, though Cajal (1915, p. 21) has described and
figured them in the early stages of development. Now that we have an idea of
the elements, we can proceed with the problem at hand. It is the envelope of the
external segment of the rods and of the cones that is said to be formed by a trans-
formation of mitochondria.

Leboucq (1909, p. 593) was the first to advance this view on the basis of his
observation that it stained intensely violet (Uke mitochondria) by the Benda
method of coloration, but his paper deals more with the centrosomes and their
transformations than with the mitochondria and is merely suggestive.

Mawas (19106, p. 114) studied the mitochondria in the fully developed retinae
of a number of vertebrates, including man (1910c, p. 113). He did not use any
embryos. He calls attention to the fact that the external segment stains intensely
with iron hematoxylin, hke the myelin of peripheral nerves, and blackens with osmic
acid, like fat (Ranvier). In addition, he confirms Leboucq's and Magitot's observa-
tions that it stains with mitochondrial dyes. It is soluble in alcohol and xylol.
These are, he says, the reactions of mitochondria, and he therefore looks upon the
external segment of the rods and of the cones as an example of diffuse mitochondrial
material present in protoplasm without any structural differentiation whatever.

Leplat (1913, p. 219), extending some already published work, attacked the
same problem in the developing chick with mitochondrial methods with a view
to determining exactly the part played by the mitochondria. He described the
heaping-up of mitochondria (first around the base of the axial filament in the
external segment) and their subsequent extension along its whole length. These,
he believes, form the envelope (p. 218) and he attempts to draw a close analogy
between this process and the grouping of mitochondria around the axial filament

126 THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

in the spermatozoon, which to my mind does not add any strength to his argument.
He found that there is a chemical change in addition to the change in position of
the mitochondria; for these mitochondria, which he styles "grains chromo])hiles,"
are more resistant to insufficient fixation than the mitochondria in the rest of the cell.
Levi (1914, ]). 199) comes out definitely against this hypothesis, and Dues-
berg (1912, p. 752), on the ground of his own observations on the retinse of chick
embryos, is inclined to reserve judgment on the question. According to Levi,
the external segment is a cuticular formation and is not mitochondrial in nature,
and he cites some earlier work, which he already published in 1901, in support of
■this conclusion. He worked on Triton larvae and his observations do not tally
with those of Leplat, yet both may be correct, because we can not take the posi-
tion that the retinal elements in widely different forms originate in precisely the
same way. We must try not to assume an altogether ultracritical attitude. The
evidence presented appears to be fairly conclusive that, in the bird at least, the
envelope of this external segment is mitochondrial in origin, for it is certainly
lipoidal and chemically resembles mitochondria to some extent. We can not regard
it as a direct transformation, however.

THE "RANDREIFEN" OF AMPHIBIAN RED BLOOD-CORPUSCLES.

Meves (1905, p. 103) advanced the view that this fine peripheral network,
just within the cell-membrane, results from the coalescence of individual mito-
chondria. He based this conclusion upon the similarity which he found in the
staining reactions of this network and the mitochondria by his iodic-acid and
malachite-green method. This is to be regarded merely as a suggestion until the
material has been worked out more carefully with adequate methods of technique.

EOSINOPHILIC GRANULATIONS IN LEUCOCYTES.

According to Ehrlich, the granulations which he described in blood-cells are
a true product of the secretory activity of the cells themselves. Meves (19106,
p. 656) concludes that these granules, just like those of gland-cells, arise from
mitochondria, and he forwards the additional argument that the eosinophile cells
of the salamander contain few if any mitochondria, the assumption being that the
mitochondria have all been transformed into the granules. It is unnecessary to
point out how loose this reasonmg is. I have studied the mitochondria in living
human polymorphonuclear leucocytes stained vitally with janus green and have
found no indication at all of transitions between the mitochondria and the specific
granulations. In fact, they are entirely distinct, (1) on the basis of the high
refractive index of the eosinophile granules, the low refractive index of the mito-
chondria; (2) the large size and spherical shape of the granules, the small size and
rod-like shape of the mitochondria; and (3) the lack of coloration of the granules
and the intense specific staining of the mitochondria. Moreover, I have examined
the eosinophihc myelocytes in the bone marrow of the guinea-pig, both vitally
stained with janus green as well as in fixed and stained preparations, without find-
ing any trace of a transition between the mitochondria and the granules. Neither
have I been able to find any indication of a transition between the mitochondria
and the neutrophilic or basophilic granulations in man.

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. 127

GLYCOGEN FORMATION IN THE LIVER.

Arnold (1908, p. 365^ claims that mitochondria play a part in the origin of
glycogen, but does not submit any evidence. Fiessinger and Lyon Caen (1910,
p. 454) conclude that glycogen is formed between the mitochondria, not in them.

GRANULATIONS IN CONNECTIVE-TISSUE CELLS.

The whole Lyon school, headed by Renaut, look upon cells of the connective-
tissue variety as gland-cells, and they have attempted to correlate the mitochondria
with the processes of secretion, which they believe to go on in them. They differ
from most investigators in including lymphocytes under this heading.

Renaut and Dubreuil (19066, p. 230) call attention to the following observa-
tions: (1) that the perineme (i. e., the mitochondrial apparatus) is rudimentary
in young forms like the small rhagiocrine lymphocytes; (2) that it develops more
and more in direct measure as the cell becomes older and its secretory activity
grows ; (3) that it decreases Uttle by Uttle as the cell ages and its secretory activity
declines. In other words, they relate the mitochondria to the act of secretion;
but they say nothing about transitions between the mitochondria and the secre-
tion granules, which they style "grains de segregation." Tliis is an important
omission.

A few years later Dubreuil (1913, p. 134) brought forward more detailed
observations of the same nature, bearing on the same problem. He beUeves that
all the connective-tissue cells in the embryo are secreting, and he has found that
they all contain abundant mitochondria. In the adult he considers the secreting
function to be relegated to the round and mobile connective-tissue cells and to
the elasmatocytes, the fixed connective-tissue cells being quiescent. He discov-
ered that the mitochondria are very numerous in the former and rare in the latter.
He cites his observations on fat-cells as a .second example. While the mitochondria
are few in connective-tissue cells destined to undergo fatty metamorphosis, they
are enormously augmented in the earlj' stages of differentiation, and, when the cell
has accumulated the maximum amount of secretion in the form of fat, another
change takes place and the mitochondria disappear almost completely. The cells
of the lymph and serous fluids furnish, he believes, still another instance of parallel-
ism between the amount of mitochondria and the secretory activity. In direct
proportion as they increase in size, departing from the true lymphocyte type, their
secretory properties increase; and at the same time the mitochondria also increase
and soon form a dense layer about the nucleus. In addition to this, he thinks
that young cartilage and bone cells secrete and for this reason contain many more
mitochondria than the older ones. Lastly, he has observed that, on inflammation,
fixed connective-tissue cells, which he supposes have lost their ability to secrete,
begin to secrete again at a very rapid rate; and that, coincident with this, their
mitochondria become just as abundant as in the actively secreting connective-
tissue cells of the embryo. He admits that a demonstration of the direct trans-
formation of mitochondria into products of segregation in large mononuclear cells,
connective-tissue cells, cartilage and bone cells, has not been made, but claims that
we can nevertheless believe in such a transformation because it occurs in many

128 THE MITOCHONDRIAL CONSTITtTENTS OF PROTOPLASM.

other gland-cells. This conception is also in accord with the theory that the mito-
chondria play the part of " eclectosomes " (p. 103).

Favre and Dubreuil (1914, p. 91) have, still more recently, attacked the same
problem in plasma cells. Here they find a similar relation between the mitochondria
and the "grains de segregation." Where there are many mitochondria there are
few "grains de segregation," and vice versa. They make no reference to morpho-
logical or chemical transitions between the two.

Their general conclusion of the participation of mitochondria in the secretory
activities of connective-tissue cells is open to criticism along several lines. In the
first place, we know next to nothing about the said secretory activities. We can
not measure them or obtain any clue to the properties of the secreted substances.
Yet we can not say that these cells do not secrete because, in all probability, all
cells give off materials into the surrounding fluids. Consequently their state-
ments that the secretory activity is high or low must be accepted with caution,
for the mere presence of a large number of "grains de segregation" does not suffice.
Their increase in number may be due, not to an increase in the rate of their produc-
tion, but to a decrease in the rate of their elimination. We may be dealing in some
cases with retention pictures. Neither do we know definitely that these "grains
de segregation" represent the secretion. So much for the secretion itself and
for their statements of variations in its amount,- for their argument is none other
than the old one of presence and absence and of reciprocal relations.

In the second place, investigators will question thek genetic series, particu-
larly with reference to the lymphocytes; because if the lymphocytes do not trans-
form into the other connective-tissue cells, as they beUeve, there is nothing signifi-
cant about their small content of mitochondria as contrasted with the large amount
in the connective-tissue cells, wliich they suppose to have assumed the abiUty to
secrete. That is to say, they have not established beyond question the steps in
their series, in which they claim that there is a relation between the amount of the
mitochondria and the secretory activity, the weak link in the chain being the
lymphocyte, wliich many people look upon as a fully differentiated blood-cell.

Lastly, the absence of transition forms between the mitochondria and the
"grains de segregation" can not fail to escape attention. None of these inves-
tigators claims to have observed them, and we must therefore reserve judgment
with regard to the part played by mitochondria in the formation of the "grains
de segregation" and in the production of the unknown secretion of connective-
tissue cells.

CILIARY APPARATUS.

Saguchi (1917, p. 265) has made a careful study of the development of the
ciUary apparatus. He describes first an accumulation of mitochondria (chondrio-
contes) between the nucleus and distal cell border, a subsequent migration toward
the distal cell border and a transformation into rod-Uke corpuscles arranged in rows.
These emit, at successive periods, short initial cilia, wliich gradually lengthen until
the definitive ciha are formed. He is careful not to assert that there is a chemical
transformation of the mitochondrial substance into that of the cilia.

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. 129

X. PHYSIOLOGY.

There is a very widespread belief that mitochondria play an active part in
cellular acti\dty, though the nature of their behavior is almost wholly obscure.
We must approach a problem so difficult with caution and take stock of the possi-
bilities before we attempt to arrive at any conclusion. Now, vital processes, or
life phenomena in the cells, which are the ultimate structural units m our bodies,
are naturally divided into two great .groups — those which are fundamental, being
common to all cells, and those which are special, representing the peculiar duties
which certain highly differentiated, older cells have learned how to perform in the
course of their development.

Among the fundamental activities of protoplasm we are accustomed to group
metabolism, respiration, irritability, growth, and reproduction. Young cells pos-
sess all of them as contrasted with the older, more mature cells of later stages.
These older, fully differentiated cells can no longer reproduce their kind, but never-
theless contain mitochondria, so that we may safely eliminate reproduction from
consideration. Irritability is the power of being able to respond in some way to
variations of any kind in the environment. The stimulus varies, but the mitochon-
dria are constant. Thej^ are no more numerous in nerve-cells in which irritability
is developed to an extraordinary degree than in other cells. This leaves for con-
sideration only metabolism and respiration (which is really a special phase of
metabolism), though some would be inclined to include the "mnemonic factor"
with these basic activities of life. Mathews (1915, pp. 68 and 587) refers "memory"
to the cells themselves, and gives us an inkling of the possible chemical reactions
upon which it may be based, though psychologists tell us that it is due to the
interaction of reflex arcs rather than being a manifestation of the life of individual
cells. It is highly probable that mitochondria are concerned in growth.

The differential activities, on the other hand, are secretion in gland-cells,
contraction in muscle-cells and ciUa, and irritability and conduction in nerve-cells.
As a matter of fact they are merely the fundamental activities of living material,
enhanced and intensified, for we must admit that even the most embryonic cells
are capable of giving off substances, that is to say, of secretion.

SECRETION.

Ever since the work of Altmann there has been a tendency to assume that the
mitochondria play a part in the formation of secretion granules. This tendency
blossomed out vigorously under the influence of the more recent theoretical con-
siderations of Meves and Regaud. Meves (1908, p. 845) thought that the mito-
chondria constituted, in part, the material basis of heredity and he believed that
they become chemically transformed into all kinds of cellular differentiations like
secretion granules, fibrils, etc. The working hypothesis of Regaud (1911, p. 685)
was a little different. He at once perceived the difficulty of assuming that the
phosphatid albumin mitochondrial complex becomes chemically altered into a
large variety of different materials and advanced his "eclectosome " theory,

130 THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

according to which the mitochondria play the part of plasts, choosing out and
selecting materials from the blood-stream and the cytoplasm, condensing them
and changing them, in their substance, into infinitely diverse products. Chemical
substances are thus supposed to be drawn in from the outside, not to be formed
through a transformation of mitochondrial material. Regaud thus resolves the
problem largely into one of permeability. His conception is, as he himself points
out, essentially a modification of the celebrated lipoid membrane theory of Over-
ton, the chief difference being that, according to Regaud, the lipoid substance
is said to be distributed throughout the whole cytoplasmic area in the form of
mitochondria instead of being confined to a layer on the surface of the cell.

Specific cases of the alleged development of secretion granulations from mito-
chondria have been discussed under the heading of "Histogenesis" (p. 116).

CONTRACTION.

The view that mitochondria are directly concerned with the motor activities
of cells is of historic interest only. Benda (1903, p. 127) advanced it upon the
basis of the following considerations: (1) similarity in the microchemical reac-
tions of the dark bands of striated muscle and mitochondria; (2) the grouping of
mitochondria about the axial filament of the developing spermatozoon; (3) the
heaping-up of irdtochondria about the roots of ciUa in ciliated cells. Benda later
(1914, p. 25) modified his conception by making the statement that mitochondria
are concerned with the development of the motor organs of the cell, like myofibrils.

Holmgren's (1909, p. 307) observations on changes in mitochondria in mus-
cular fatigue are of special interest here. He made ingenious experiments with
dragon-flies, holding them by the thorax between finger and thumb and allowing
them to continue beating their wings furiously for different lengths of time. He
found that granules which stained in a typical way with the Benda method and
with iron hematoxylin underwent definite alterations, depending upon the severity
of the fatigue. The granules changed their position, diminished in number, and
stained less intensely. These findings are certainly suggestive, to say the least,
but they have never been confirmed. They evidently merit further attention.

Faure-Fremiet (in discussing the paper by Regaud and Mawas, 19096, p. 235)
records the observation that the mitochondria, gathered about the contractile
filament in the peduncle of vorticella, are unchanged at the moment of contraction.
Regaud (19086, p. 209) observed that the mitochondria in the ciliated cells of the
urinary tubules of cold-blooded vertebrates are few in number, are distributed
without apparent order, and have no relation to the cilia. Finally, Shipley (1916,
p. 444) discovered that there is no relation between the motility of Trypanosoma
lewisi and the amount of mitochondrial substance within them.

IRRITABILITY AND CONDUCTION.

Curiously enough, the jiroblem of mitochondria in this connection has so far
been untouched. This may perhaps be due to the reluctance of investigators to
even admit their existence in nerve-cells (see p. 101). Accordingly, it is a very
pertinent question to inquire whether there are alterations in the mitochondria in

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. 131

nerve-cells in muscular fatigue. Strongman (1917. p. 169), working in this labora-
torj', has undertaken to answer this. White mice were selected for the experi-
ments because they are the smallest mammals which can conveniently be used
in the laboratory and because more is known of the qualitative (Thurlow, 1917,
p. 37) and quantitative (Nicholson, 1916, p. 329) variations in mitochondria in
their nervous system than in that of any other mammal. They were fatigued
by the very simple method which Professor Tamao Saito uses, of letting them swim
in water until they are exhausted. It was soon discovered that they swim more
actively when the water is sUghtly agitated and is raised to body temperature:
otherwise they soon learn to float and refuse to exercise

The experiments were controlled in the usual manner by using only mice of
known age and by examining, in exactly the same way, an unexercised mouse of
the same htter for comparison. Five experiments of this kind were made, the mice
swimming 1 to 2 hours continuously before they were completely exhausted.
Larger mammals will swim for a day or more before exhaustion. Experience
showed that young mice 25 to 30 days old are more suitable than adults, because
they are more easily tired.

In each of the five experiments the fatigued mouse and the control mouse rom
the same litter were chloroformed. They were then fixed by the injection of a
formaUn and bichromate mixture in accordance with the method advised by Cow-
dry (1916a, p. 30). The brains were then removed, mordanted, dehydrated, and
cleared together in the same bottle, and they were embedded in the same block
of paraffin. Sections, cut 4 microns in thickness, from the fatigued and from the
control, were mounted on the same slide so as to avoid variations in the staining.
The .sections were then stained with fuchsin and methyl green, the fuchsin color-
ing the mitochondria crimson and the methyl green staining the Nissl substance
a bluish-green color. Preparations were made of the cortex, the cerebellum, and
the spinal cord, in this way, from each of the five experiments.

The net result of Strongman's five experiments, with their controls, was to show
that mitochondria are surprisingly constant in nerve-cells. A fair degree of fatigue,
as well as a certain amount of fright, brought about no constant changes in them.
It is quite possible that more prolonged exhaustion, if it can be induced, may lead
to definite and precise alterations. It is perfectly clear that these results are in
complete accord with those which Key and Scott obtained with the pancreas,
because they show that in the nerve-cell, also, the mitochondria are not directly
concerned with specialized activities. They do not play a part in conduction any
more than they do in secretion, which strongly points to the general conclusion
that they are concerned in some fundamental process common to all cells.

METABOLISM.

Here, as in all other fields of investigation, oliservation has preceded experi-
mentation. Mitochondria have been carefully studied in phylogeny as well as in
ontogeny. Their wide distribution is amazing. It has already been pointed out
(page 72) that they occur from man to the most lowly protozoon and from the

132 THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

angiosperms to the fungi, though their existence is doubtful in the myxomycetes,
schizomycetes, and most of the algse. They are apparently identical in both plants
and animals (N. H. Cowdry, 1917, p. 225). They are indeed as characteristic
of the cytoplasm as chromatin is of the nucleus. They differ .slightly in composi-
tion just as chromatin does. The fact that they are most abundant in the active
stages of life of the cell and decrease in number as the cell becomes old and senile
is not without significance. We find them in the egg and in all the tissues of the
developing embryo — from the very beginning (when the cells have no definite
specialized activities) to the later stages and adult life, when each has assumed
its own pecuhar duty of secreting or contracting, or conducting, as the case may
be. In other words, their presence in the absence of speciaUzed activities indicates
that, in these early days of development, they are either inert or else play a part
in the fundamental vital processes. The conclusion is obvious; and since they do
not differ in any noteworthy particular in the later stages, the assumption is justi-
fied that here also their function is a basic generalized one.

The meager and unsatisfactory yet direct experimental evidence at hand
seems to support this view. Thus Romeis (1913c, p. 12) has found that mitochon-
dria are very numerous in actively regenerating tissues; Busacca (1915, p. 232)
found that they decreased in number with fatigue in the cells of the retina stimu-
lated with intense light; Homans (1915, p. 12) associated the number of mitochon-
drial filaments with an increased activity of islet-cells in experimental diabetes;
Policard (1910, ]). 284) showed that there was an increase in the number of
\ mitochondria in kidney-cells on administration of phloridzin, and so on.

These statements relate, however, only to the general impression given by
the study of sections. There has been no attempt to distinguish, in a clear-cut
way, between absolute and relative fluctuations in mitochondrial content. The
observations have not been controlled by a careful estimation of cell-volumes.
Thurlow (1917, p. 37) has been the first to realize these discrepancies. She has
established a definite mitochondria cytoplasmic rate in the nerve-cell, just as
Hertwig years ago measured the nucleus c}'topla.smic ratio (see p. 80).

There have been no carefully checked observations on qualitative changes
in mitochondria with cell activity, although we have abundant evidence that the
solubilities of mitochondria vary. Of course, the changes in form of mitochondria
have been subjected to careful scrutiny, but so far they have yielded httle of value.
The observations of Holmgren (1909, p. 308) on the changes in mitochondria in
muscular fatigue are quahtative in a sense and very interesting, but they have
never been confirmed.

Furthermore, processes of metabohsm, as well as of respiration, are, as one
would naturally expect, very easily modified in pathological states; hence the
sensitivity of mitochondria to pathological change. Scott (1916, p. 243) discov-
ered that the mitochondria are the first of all the cell constituents of the pancreas
to become altered in experimental phosphorus poisoning. Moreover, the fact
(which is now emerging from the numerous recent pathological studies on mito-
chondria) that mitochondria in different types of cells respond in much the same

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. 133

way to dififerent varieties of injurious influences — in other words, that there is
nothing specific in the reactions of mitochondria to pathological change — is also
in accord with the prevalent conception that they take part in a type of activity
common to many cells, at which the noxious influences strike.

Relying on this confessedly inadequate information, investigators generally
are inclined to believe that mitochondria participate in some of the processes
involved in cell metabolism. Coghill (1915, p. 350) is rather more specific, for he
relates mitochondria to the constructive side of metabolism. But the term "metab-
olism" is very vague. Its meaning has changed from Michael Foster's original
definition. It is now used to designate a multitude of chemical reactions from the
taking in of foodstuffs to the elimination of waste products, including even cellular
oxidations and reductions. We have metabolism of carbohydrates, fats, and
proteins to deal with. Mitochondria occur in organisms entirely independently
of their diet. They are just as abundant in herbivors as in carnivors; and many
of them are contained in green plants, which obtain their nitrogen from nitrates,
their hydrogen from water, and their carbon from the air. Where metabolism
varies in kind, they apparently do not vary. They can hardly be a stage in the
utihzation of a specific substance taken in in the food, much less a waste product.
They must be built up in the jihylogenetic scale in a variety of different ways to
serve a common function. We naturally inquire whether they are reserve products
or whether they play an active part in cell metabolism.

The experiments of Russo (1912, p. 203) would seem to indicate that mito-
chondria are reserve food products. He claims that their number may be increased
in the ovarian eggs of chickens fed with lecithin. His conclusions receive some
indirect support from the work of R. Van der Stricht (see p. 84), but his experi-
ments have never been repeated, in spite of the great interest which attaches to
them. We are very much in need of information along these lines and it seems
that experiments on the effect of inanition and of different diets on the mitochon-
dria in mice might yield valuable results. The monograph by Champy (1911,
p. 146) on the behavior of mitochondria in intestinal epithehal cells would serve as
a good point of departure.

Regaud (1911, p. 685) thinks that mitochondria play actively the part of
plasts, choosing and selecting material from the surrounding protoplasm and from
the blood-stream, condensing them and converting them, in their substance, into
diverse products. He likens them to the hypothetical side-chains of Ehrlich, and
his idea is an extension of the famous lipoid membrane theory of Overton ; in other
words, he believes that they function in constructive metabolism. They may
act as a sort of vehicle or medium in which chemical reactions take place.

RESPIRATION.

Kingsbury (1912, p. 46) was the first to suggest that mitochondria play an
important part in iirotf)plasmic respiration. He says that osmic acid, potassium
bichromate, and formalin, which are the chief ingredients of mitochondrial fixa-
tives, are preeminently oxidizers and that their efficiency depends on the presence

134 THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

of reducing substances in the cytoplasm. These he beheves to be the mitochon-
dria on account of their hpoid characteristics.

But it remained for Mayer, Rathery and Schaeffer (1914, p. 619) to furnish
detailed evidence in support of this view. They point out: (1) That mitochondria
are chemically well adapted to function in oxidations and reductions; they are
phosphatids, and phosphatids have the power of auto-oxidation (Mathews, 1915,
p. 97), though some (Bayhss, 1915, p. 592) tliink that they do not possess it to any
appreciable degree; according to Mayer, Rathery and Schaeffer, mitochondria
contain unsaturated fatty acids with ethylidene groups. These have a great
affinity for oxygen and oxidize themselves to aldehyde. (2) That agents which
attack lipoids (Uke alcohol, ether, and chloroform among the anesthetics) at the same
time cut down the respiratory oxidations. (3) That mitochondria are present in all
cells and that respiration is the most fundamental property of Uving matter, etc.

Still more recently the Lewises (1915. p. 393) have arrived at much the same
conclusion from their studies on mitochondria in tissue cultures. It is in accord
also with my own observations on the staining of mitochondria with janus green
and related dyes (1916a, p. 429).

XI. PATHOLOGY.

Much more work has been done on mitochondria in pathological conditions
than is generally reahzed. It has been done from many points of view; it is widely
scattered, and some of it is very difficult of access. Moreover, it is not entirely a
new development, as many people suppose. We have to do with two mitochondrial
Uteratures, an old and a new. It is the old one which is so frequently ignored, an
outgrowth of Altmann's remarkable researches at Leipzig from 1880 to 1890; it
flourished for a while but was soon choked by the active criticism which his views
excited, for he thought that his "bioblasts" were the final ultimate Uving particles
embedded in protoplasm which he considered to be hfeless.

The central thought which underhes all the recent work on mitochondria in
pathological conditions is the conviction that we now have at our disposal a new
criterion of cell activity and cell injury. We do not know it or understand it, but
it has been proven over and over again in the last few months to be of great and
surprising deUcacy, for it responds (even before the nucleus) to injurious influences;
and it has the rare merit of being cytoplasmic. We may expect environmental
changes to act on it which would make no impression at all upon the nucleus. It
is convenient to our hands in all cells and there is no knowing what story it will tell
when we ask it.

In the study of mitochondria we are at once, and very forcibly, reminded of
bacteria. One has a feeling, in looking over mitochondrial preparations for the
first time, that aseptic precautions should have been taken in removing the tissues.
We can easily forgive Altmann for thinking that they were elementary organisms,
for the similarities between them and bacteria are really remarkable. Their form

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

135

is granular, like cocci; or rod-like, like bacilli; or filamentous, like certain vibrios;
while the tendency for mitochondrial granules to become arranged in rows is strongly
suggestive of streptococci. They are lipoidal and they agglutinate in the most
remarkable way, under certain conditions, just as bacteria do. Fortunately, they
may be easily distinguished from bacteria by their staining reactions (particularly
to janus green), by their occurrence in almost all cells, by their behavior, and by
their lack of independent motility. When less was known about mitochondria it
was not so easy to identify them as it is now. In consequence, we have to be
on the lookout in the hterature for descriptions of mitochondria under the head-
ing of "Bacteria and intracellular parasites" and vice versa}

For convenience I have indicated below the work which has already been
done on mitochondria in pathology, because the space will permit of detailed dis-
cussion of onl}^ the more important contributions:

Adenoma of thyroid, Goetsch (1916, p. 1.32).

Anemic necrosis, Israel (1891, p. .310).

.\sphy.\iation in tissue culture, Champy (1914. p. "20).

Autolytic changes, Dannehl (1892, p. is.'i).

Beri-beri, Clark (1914, p. 92).

Carcinoma:

Beckton (1909, p. 191).

PorcelH-Titone (1914. p. 237).

Fa\Te and Regaud (1913, p. 688).

Lubarsch (1897, p. 640) and others.
Cicatrization in tissue culture, Champy (1914. p. 3GS).
Cloudy sweUing in kidney, Schilhng (1894, p. 470).
Diabetes, Homans (191.5, p. 16).
Diphtheria toxin, d'Agata (1913, p. 443); Dibbelt (1914,

p. 119).
Diuresis, Policard (1910, p. 272).
Edema of prepuce, Regaud and Favre (1912, p. 330).
Epithelioma, G. .\rnold (1012a, p. 283) ; Favre and Re?aud

(1913, p. 688).
Fatty degeneration, kidney, Ophtils (1907. p. 136).
Fatty infiltration, .\ltmann (1889, p. 94).
Feltphanrosis, nerve-cells, Biondi (191.5, p. 224).
Calls (nematode) in plants, Nemec (1910, p. 166).
Hemoglobinuria, kidney, Barratt (1913, p. .566).
Hypertrophic tonsils and adenoids, .\lagna (1911, p. 27).
Inanition, Russo (1910, p. 173).

Inflammsition, connective tissue, Dubreuil (1913, p. 1.38).

Irritation of cortex, Collin (1914, p. .592).

Liver ovtirpation, effect on mitochondria in kidney,

Policaid (1910, p. 245).
Myelogenous leukemia, Klein (1910, p. 407).
Pernicious anemia, spleen, Shipley (1915, p. 75).
Phosphorus poisoning, Lubarscli (1897, p. 639); Scott

(1916, p. 237); Mayer, Rathery and SchaefTer (1914,

p. 607), etc.
Pigmentation, Barratt (1913, p. 566).
Poi.scns, various, Mayer, Rathery and Schaeffer (1914,

p. 607).
Polypus of nose, Benda (1899a, p. 380).
Prostatitis, Dominici (1913, p. 295).
Radium, Beckton and Russ (1911, p. 99).
Regeneration, Romeis (19i:3f, p. 1); Torraca (1914a,

p. .5.39: 19146, p. 4.59).
Rodent ulcer, Regaud and Favre (1912, p. 329).
Sarcoma treated with X-ray, Regaud (in discussion of

Laguesse) (19126, p. HI).
Scharlach R, skin proliferation, Barratt (1913, p. .566).
Suprarenal gland after ovariectomy, Mira (1912, p. 43).
Toxins, Mayer, Rathery and Schaeffer (1914, p. 609).
Tubercle giant cell, Champy (1911, ]). 1.54).
Uranium nephritis, Oliver (1916, p. 306).
X-ray, Beckton and Russ (1911, p. 99).

GLANDULAR SYSTEM.

Most of the work has been done on glands because of the ease with which they
can be observed and experimented with. The two outstanding contributions are
those of Homans on diabetes and of Goetsch on diseases of the thjToid.

Homans (1915, p. 16) produced various degrees of diabetes in the dog by
removal of portions of the pancreas and by feeding with carbohydrate food. He
found that the mitochondria in the B cells of the islands of Langerhans become
accentuated, fuse to form droplets, and finally disappear in stages of activity,
exhaustion, and degeneration of the tissue, while the other cells show no changes

'The relations, if any, whioh are affected lintwecn niitoihondrla and thi' invading l>artrria have never been stuliel.
Yet it is niilikply that twn materials of similar size and shapn and lipoilal iiroportitM woiil I not act on each other in sonic
way. Since mitochondria undoulilcdly aeKlutinatc, it is possilile that llicy may lie associated in intraecUiilar bacterial
agglutination. The difficulty of staining them lioth differentially would not be groat and it woiil<l lie interesting to compare
the active invasion of virulent bacilli with the phagocytosis of nonvirulent or dead organisms.

136 THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

whatsoever. According to Homans, this remarkable association of changes in
the B cells with the condition of diabetes is causal, and indeed this would seem
to be the most likely interpretation, though possibly they may be simply the result
of the condition. In any case the work constitutes a definite contribution to our
knowledge of diabetes on the one hand, and of mitochondria and cell physiology on

the other.

Goetsch's work on the thyroid (1916, p. 132) is of particular interest. He
has studied colloid goiters and lias found that there is an actual correlation between
the chnical symptoms and the amount of th>Toid tissue presenting evidence of
thyroid activity by the abundance of mitochondria throughout. That is to say,
in cases of colloid goiter, without any symptoms of hyperthyroidism, the mito-
chondria are generally reduced in number to the point of being almost entirely
absent. Those present are smaller than normal and there is a considerable
increase in the amount of fat. In the cases of adenoma with hyperthyroidism, the
cells of the adenoma are enormously rich in mitochondrial substance in the form
of granules and short, thick rods, and contain no fat. In every instance of exoph-
thalmic goiter the thyroid gland has been found to contain a great abundance
of mitochondria, much in excess of the normal amount and with little or no fat.
In fact, the correlation of the chnical manifestations of hyperthyroidism and the
mitochondrial picture is such that Goetsch is justified in saying that the two are
definitely related. His conclusions are based upon the study of approximately
125 cases of thyroid disease in the human.

BLOOD VASCULAR SYSTEM.

The blood offers a unique opportunity for the study of mitochondria because
they can so well be seen in the living cells by staining with janus green as well as
in fixed smears stained by appropriate methods. Nevertheless, investigators have
contented themselves with merely describing the mitochondria in the normal cells.
Considerable attention has been paid to the mitochondria as possible indicators
of the genetic relationship of blood-cells to each other. Blood from cases of anemia
and leukemia has been studied simply because it contains- cells which are normally
restricted to the bone marrow and which can not be obtained under normal condi-
tions without an operation or early autopsy. The question of the part played by
mitochondria in the development of the specific granulations in leucocytes has
been touched on (p. 126), but nothing whatever has been done on the changes in
the mitochondria in blood diseases in man or in different experimental conditions
in animals.

URINO-GENITAL SYSTEM.

The kidney has been very carefully stucUed and the part played by the mito-
chondria in the formation of urine carefully worked up, for wliich see page 126.
Little, however, has been done on pathological changes in the kidney.

Takaki (1907, p. 250) has made a study of autolytic changes by Altmann's
method. It is questionable whether Pizzini's (1908, p. 108) observations relate

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. 137

to mitochondria at all. Enderlen (1908, p. 208), Hirsch (1910, p. 168), and De Gia-
como (1911, p. 223) have all studied compensatory renal hypertrophy and record
concurrent increase in granulations which are probably mitochondria. There is
also a general consensus of opinion to the effect that mitochondria fragment on
the approach of degeneration. The work of Cesa-Bianchi (1909, 1910) and
Hjelt (1912, p. 207) is important in this connection.

Though much has been done on mitochondria in cloudy swelling, particularly
in the kidney, little if anything has been added to the excellent account of mito-
chondria under the heading of "Altmann's granules" given by SchilUng (1894,
p. 478) and usually ignored. Schilling produced cloudy swellings in rabbits by
ligation of the renal vein and found that the mitochondria lose their charac-
teristic staining reaction and serial arrangement and decrease in number. He
rightly maintains that the mitochondria are quite distinct from the albuminous
granules of cloudy sweUing, and he attributes the disappearance of mitochondria
to the approach of degeneration. His conclusions have been confirmed by Lubarsch
(1897, p. 631) and many others. Recent hterature has been reviewed by Ernst
(1914, p. 81).

Dominici (1913, p. 295) has studied the relationship of mitochondria in normal
and hypertrophied human prostates and has concluded that they are not directly
concerned in the formation of the secretion.

RESPIRATORY. MUSCULAR. AND SUPPORTIVE SYSTEMS.

The respiratory, muscular, and supportive systems have not been studied
to any extent. We merely have the observations of Dubreuil (1913, p. 134) on
inflammation in connective tissue already referred to on page 127.

NERVOUS SYSTEM.

Luna (1913^, p. 415) in a brief note records the behavior of mitochondria
(plastosomes) in transplanted ganglia. They first swell up into large granules,
and he claims that the cells must maintain their vitality owing to the presence
in them of well-developed mitochondria. He also experimented by cutting periph-
eral nerves and found that the mitochondria in the corresponding gangUon cells
lose their regular distribution, increase in volume, take on a more intense stain
with hematoxylin, and finally disappear entirely. Clark (1914, p. 92) observed
that the mitochondria are surprisingly constant and show no changes in experi-
mental beri-beri, and Biondi (1915, p. 232) has made a study of the relations of
mitochondria in autolyzing nervous tissue.

G. F. McCann (1918, p. 36) has made a study of mitochondria in the spinal
ganglion cells of monkeys in experimental poUomyelitis. She has found that the
mitochondria are surprisingly resistant, occurring even in those cells which no longer
contain typical Nissl substance. Evidently, therefore, experimental poliomyelitis,
like experimental beri-l)eri, does not modify to any great extent the vital processes,
whatever they may be, in which the mitochondria are concerned. This is impor-
tant because it brings us face to face with the fact that, while the mitochondria

138 THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

are extraordinarily sensitive to some pathological changes, they are equally
resistant to others. It is unsafe at present to hazard even a tentative explana-
tion. Why this should he remains one of the great problems of pathology.

The small amount of work on the pathology of mitochondria in the nervous
system is not due to a lack of interest, but rather to a misunderstanding and exag-
geration of the difficulties involved. Experience has shown that the older osmic-
acid-containing fixatives are almost useless for the study of the central nervous
system on accovuit of the large numlier of medullated and non-medullated fibers
offering insurmountable barriers to their penetration. It is not generally appre-
ciated that this difficulty may be in large measure overcome by the application of
Regaud's mixture, which consists of formalin and potassium bichromate. Ex-
cellent results may also be obtained by mordanting tissues in bichromate which
have been previously injected in the ordinary way with formalin.

It is important to note also that the mitochondria in the nervous system do
not imdergo autolytic changes immediately after death as they do in gland-cells.
It is by no means necessary that the tissue be fixed while the body is still warm;
6 or 8 hours after death is often soon enough.

While we are still ignorant of the normal appearance of mitochondria in the
human brain, isolated observations of their relations in pathological conditions
are of but little value. Great difficulties attend the ])roduction of lesions in the
brains of experimental animals, without surgical intervention, because the changes
in the other organs often prove fatal before the nervous s^ystem has been affected;
but in a field so promising these difficulties will probably be speedily overcome.
Mitochondria, being phospholijnns, differ sharply from the Nissl substance, which
is nucleoprotein in nature. Accordingly, the}" will serve as clues to a different
type of activity and will yield valuable information upon the question of nerve-
cell physiology. From their chemical constitution also it is probable that on dis-
integration they may elaborate cholin, and inasmuch as organic diseases of the
nervous system can be separated from functional neuroses by the formation of
chohn in the one and not in the other (Halliburton, 1907, p. 74), it is ciuite likely
that a study of mitochondria may afford a cytological basis of distinction between
these two great groups of nervous diseases.

REGENERATION.

Romeis (1913«, p. 10) has made a study of regeneration in the tails of Triton.
In the connective-tissue cells the mitochondria become greatly increased in num-
ber and he finds all the stages in the formation of fibrils reported by Meves. The
same is true in muscle-cells, and the mitochondria also increase in niunber in the
glands of the skin and in bone. He believes that with regeneration the cells
become more embryonic with resultant increase in mitochondria and that the various
structural differentiations are formed by the mitochondria and the cytoplasm acting
together.

Torraca (1914o, p. 539; 19146, p. 459; 1916, p. 326) has made a number of
contributions to the subject, working on cartilage, striated muscle, and glands.

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. 139

He also describes an increase in mitochondria in the active stage of regeneration
and is of the opinion (with Ronieis) that the mitochondria participate directly in
the formation of secretion.

It is accordingly interesting to note that Oliver (1916, p. 307) describes a dis-
tinct decrease in the number of mitochondria in regenerating kidney-cells in chronic
uranium poisoning, but this discrepancy may be due to studying a different stage
in the regenerative process.

FEVER.

The single observations of Policard (1912d, p. 229) on the temperature solu-
bility of mitochondria are perhaps significant. He found that exposure for 30
minutes to a temperature of from 47° to 50° C. dissolved the mitochondria in
kidnej'-cells without affecting the appearance of the nuclei. It is well within the
bounds of po.ssibility that a prolonged or intermittent temperature of say 40° C,
as in a high fever, may bring about a solution or chemical alteration of mitochondria
in some cells of the body. This is, at any rate, an interesting thought in connection
with Welch's (1888, p. 403) belief that fatty degeneration in heart-muscle is in some
way associated with high fevers.

ACIDOSIS.

Thus far no account of mitochondria in acidosis has appeared. Now, it is
common knowledge that mitochondria are very sensitive to acids. It is also well
known that one of the first manifestations of acidosis is a marked inhibition of
the respiratory oxidation of the cell (Mathews, 1915, p. 247). If there is anythng
in the theory that mitochondria function in processes of oxidation and reduction
it is possible that these two facts may be related. Let us remember also the
dyspnoea in acidosis. Moreover, mitochondria resj^ond to a wide range of noxious
influences by sweUing up before going into solution, which might well be due to
the effect of increased H-ion concentration upon their protein fraction, causing it
to become hygroscopic and to swell. The affinity of injured cells for basic anilin
dyes is probably due to a swing of the reaction in them- toward the acid side.

TUMORS.
The older jjathologists did not touch on the question of mitochondria in tumor-
cells. Veratti (1909, p. 34) and Beckton (1909, p. 182) independently, in the
same year, jHiblished their researches on tumors. Veratti applied the uncertain
Golgi method to cells of a transplantable mouse carcinoma and brought to light
filamentous structures which appeared to be mitochondria. Beckton's techni([ue
was a little better, though he relied entirely upon the old, original Altmami's
method, but his conclusions were startling and stimulated a great deal of interest.
He thought (1909, ]). 191) that the granules of Altmann (mitochondria) were
absent in the cells of malignant tumors and that malignant growths could be dis-
tinguished from benign in this way. Bensley (1910«, p. 81) at once grasped the
possible importance of this assertion and proceeded to test it out. His results
were entirely at variance with those of Beckton, for he found that as a matter of

140 THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

fact mitochondria are just as abundant in rapidly growing tumors as in those of
the benign variety. This has been confirmed by G. Arnold (1912a, p. 283), Favre
and Regaud (1913, p. 688), Porcelli-Titone (1914, p. 237), and all the others who
have worked with malignant growths.

A number of purely descriptive papers have appeared on mitochondria in
different varieties of tumors, chefly in human tissues taken at operation or autopsy.
No results of great importance have been achieved (see p. 135). Let us hope that
the work will take an experimental turn. Much remains to be done. One is
tempted to entertain, as a working hypothesis, the view that mitochondria may
serve, in a measure, as indicators of the effect of X-rays, radium, and other so-
called therapeutic agents on tumor cells. The isolated notes on mitochondria in
human tumors are all very well to begin with, but we must have detailed and
comprehensive accounts of mitochondria in some particular type of tumor in
animal cells before experiments can be profitably commenced. We want to know
the whole story of mitochondria from the origin of the tumor to the end. Another
prerequisite is some accurate information upon the effect of X-ray and radium upon
normal cells, because it appears that the results of Beckton and Russ (1911, p.
105) require confirmation.

It would be interesting to compare the mitochondria in the cells of the crown
gall in jjlants with cancer cells in man to ascertain whether they present any points
of sunilarity or dissimilarity. The degree of resemblance between the two is of
prime importance in view of Erwin F. Smith's (1917, p. 277) discovery that crown
gall is caused by an infection wdth a specific organism and his suggestion that
cancer in man is Ukewise infectious. Moreover, tumors offer a new and attractive
field for the study of the behavior of mitochondria in histogenesis. Nobody has
even touched on the relations of the myofibrils in myomata, and of the connective-
tissue fibrils in fibromata to mitochondria, to say nothing of the fascinating
problems presented by the rarer types of neuromata. If there is such a thing as de-
differentiation in tumors, a study of the process might go far toward clearing up
the whole problem of the part played by mitochondria in normal differentiation,
for many chemical reactions are reversible; if the mitochondria form the large
variety of chemical substances which it has been claimed they do in normal differ-
entiation, then it is quite reasonable to suppose that they may be reformed from
these same materials, dedifferentiation taking place.

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. 141

XII. DISCUSSION.
THE GENERAL RESULTS OF MITOCHONDRIAL WORK.

(1) It constitutes a definite addition to our knowledge of the fundamental
structure of Living material, for we have found in them a definite and concrete
class of cell granulation presenting even less variation than the chromatins and
distributed through almost all Uving matter.

(2) The discovery of their fatty, phosphatid nature (just when physiological
chemists are becoming interested in fats, whereas formerly their chief attention
was devoted to the study of proteins, owing to the inspiration of Emil Fischer's
work on protein synthesis), is a coincidence of some importance because it makes
probable a rapprochement between cytological and chemical work, a new and
promising point of contact having been established.

(3) The fact that they do serve, in some cases, as a basis for cell classification
may prove to be of great value when followed up. This is particularly true in ques-
tions of cell genealogy. The differences, however, in the appearance of mitochon-
dria in cells of different type are usually but sUght, and this is not to be wondered
at in view of the generality of their distribution in all protoplasm and in view also of
the probabihty that they play their part in some fundamental activity common to
many cells, not in highly speciaUzed functions Ukely to differ from cell to cell.

(4) In embryology they have excited the greatest interest. The radical claims
concerning their role in histogenesis have forced the reinvestigation of the entire
field. There is much difference of opinion, but it can be safely said that they are
associated in the formation of certain substances like fat, Upoid, pigment, and per-
haps also secretion granulations. While it seems clear that they do not form these
substances by direct chemical transformation, the exact part which they play is
a mystery and will probably remain so for some time to come. In inheritance also
much has been done, but no good experimental evidence has been found in support
of the supposition that they act as carriers of heredity.

(5) With regard to the part which they play in the physiology of the cell, we
have learned only that it is fundamental rather than specialized. We suspect
that they are concerned either directly or indirectly with processes of metabolism
or protoplasmic respiration, and this is as far as we can go.

(6) The extraordinary sensitivity of mitochondria to pathological change
has already proved of the greatest value in experimental medicine. It has been
found that, in some cases, they are much more sensitive than the nucleus. Prepa-
rations made by the old methods designed to give nuclear detail show no trace
whatever of the more subtle changes which may be brought to light by the newer
mitochondrial technique. Furthermore, since the mitochondria are so different from
the nucleus in every respect, it is not surprising that they have proved themselves
to be indicators of an entirely different type of activity. It is in fact this quality
which has been of such great service in the study of exophthalmic goiter (p. 136).

(7) In botany the study of mitochondria has greatly advanced our knowledge
of the development of plastids. It has been found that the plastids arise from the

142 THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

mitochondria; in other words, that mitochondria are associated in the production
of starch and chlorophyll and a great variety of other substances.

(8) And finally, their intensive study in so many cells of both animals and
plants has resulted in a well-marked movement toward the study of the whole cyto-
plasm and many facts of importance, not directly related to mitochondria, have
been brought to light. It has forced, for instance, a complete readjustment of
our conception of protoplasm.

(a) Flemming's filar theory.— According to this hypothesis, jirotoplasm con-
sists fundamentally of a homogeneous ground-substance in which fibrils are em-
bedded. The relatively dense and refractile fibrils he called the mitome and the
watery fluid between them the paramitome. The idea has been much modified
by himself and others. Now that we are in a better position to understand his
fila, we reaUze that they are a very heterogeneous group of structures. Many of
them are artifacts and others mitochondria. He included under the same head-
ing such widely different structures as spindle fibers. This mere separation of
protoplasm into mitome and paramitome helps but Uttle.

(6) Altmann's bioblast theory.— Mtmaim observed (in many varieties of cells)
minute granular rod-like and filamentous structures which he took to be elementary
organisms. He thought that they existed in the form of colonies in cells and that
they multiplied by di\-ision. He formulated the statement omne granulum e
granulo. According to him they constitute the vital hving substance as contrasted
with the lifeless inert ground-substance containing them, in token of which he
called them "bioblasts." Recent work on the bioblasts has robbed them of
all their mystery. They are in reahty a heterogeneous class of cell granulations
Uke Flemming's fila, no more living than the rest of the cytoplasm and comprising
mitochondria for the most part, but some fat, pigment, and secretion antecedents
in addition. We can see now Altmann's many mistakes as well as appreciate the
element of truth in his conception, for mitochondria are in reahty almost universal
constituents of protoplasm.

(c) Fromann's reticular theory.— In terms of tliis hypothesis all protoplasm
consists, in the last analysis, of a relatively dense reticulum of fine threads, some-
times called spongioplasm, and of a more fluid material in the interstices called
hyaloplasm. Contractility of the threads has been invoked to explain movement.
Others think that they constitute the origin of fibrillar structures, like myofibrils.
Recent studies in cell dissection by Kite and Chambers have failed to reveal the
existence of such a reticulum and our knowledge of mitochondria is incompatible
with its existence, for we find that they move freely from place to place in the
cytoplasm without let or hindrance. Nevertheless the conception persists in our
text-books in the form of misleading diagrams which should be eliminated as
quickly as possible.

(d) Butschli's foam theory.— Butschli afforded strong experimental evidence
in support of the alveolar theories of the structure of protoplasm, according to
which the continuous fundamental substance is composed of alveolar walls and
alveolar contents. The foam structure is sometimes visible in the living condi-

^^^ tui

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. 143

tion and sometimes it is not. Such a formation would develop great siu"face ten-
sion. Butschli is to be credited for the first attempt to explain the phenomena of
mitosis in terms of physics and chemistry. Apparently he was right in explaining
certain kinds of amoeboid motion on the basis of surface tension, though he beUeved
that the motion took place through the local enlargement of the alveoli. In spite
of many criticisms this theory is quite stimulating and helpful in the study of cell
physiology.

These theories are not so clear-cut as they appear to be. Many amendments
and subsidiary hypotheses with all grades of meaning have been introduced which
we have not time here to consider. The attempts to generalize have not been
fruitful. We constantly meet with filamentous, granular, and net-like appearances
in protoplasm, but they are transitory and superficial. The theories go so far
and no further. In my opinion they do not even touch on the main point at issue.
They do not in any way help us to understand the nature of vital processes or the
special phenomena of polarity and bilaterahty in the cell, and the reason is not far
to seek. It is because cytologists are usually versed in the use of the microscope
and fail to realize that in protoplasm the most important things are the things
unseen. Accordingly^, attention has been paid to only the visible constituents
and the rest have been ignored. It must be admitted that cytologists as a whole,
with present-day equipment and training, are not fitted as the biochemists are
for the study of the most fundamental of problems, the nature of hfe.

Vital phenomena are totally incomprehensible unless there exists some struc-
tural organization in protoplasm. Cellular polarity and bilaterahty must depend
on it. Cells are the unit structures in our bodies, but each and every one of them
a complicated and highly organized unit. Each is a httle factory which quickly
rings about chemical changes, possible only in rare instances outside of the body,
slowly with the aid of considerable temperature and pressure and much compli-
catedfcnachinery. That the cytoplasm is organized locally just as the great fac-
tory is organized in space is evident from the fact that when it is thoroughly mixed
life is no longer possible. It has been proved over and over again that this
organization does not reside in the visible constituents of the protoplasm, because
their distribution can easily be altered without modifying either the polarity or
the bilaterahty, as the case may be, or disturbing to any great extent the vital
processes going on. Mitochondria, pigment, and secretion granulations, fat, hpoid,
and all the other formed bodies are relatively unimportant. It is to the optically
homogeneous ground-substance that we must look, and our microscopes will help
us not at all. We must extend our conceptions to include a morphology of the
ultramicroscopic and invisible; otherwise we fail.

The intensive study of the mitochondrial constituents of protoplasm has
brought us an important point of contact with recent advances in chemistry. In
cytology as in physiology the mechanistic philosophy is the only fruitful one.
The old giant molecule or biophore hypothesis of Ehrlich is being rapidly discarded
and we are beginning to entertain the entirely opposite view that vital phenomena
are due to the orderly interaction of relatively simple substances, often of inorganic

144 THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

nature. The growing interest in ionization, hydrogen-ion concentration, the role
of inorganic salts, of calcium, and so on, all point in this direction. Instead of
looking upon the more soUd constituents as the most important, we now regard
them as the least. Attention is being directed toward the phosphohpins, which
occm- in protoplasm as a diffuse invisible deposit as well as in the form of mito-
chondria. According to Mathews (1915, p. 88), they are the most important
substances in living matter, "for they are found in all cells, and it is undoubtedly
their function to produce, with cholesterol, the pecuhar semifluid, semisoUd state
of protoplasm. This latter holds much water in it, but does not dissolve. Indeed,
it might be said that the phosphatids with cholesterol make the essential physical
substratum of living matter."

That the substratum is more or less fluid is shown by the free movement of
granules and other visible constituents. While many vital phenomena depend
on enzymes and we know that they diffuse but slowly, so that it is not essential
to assume the existence of compartments to localize their action, there must be
some organization in protoplasm of an exceedingly labile sort. It is Ukely that
this organization depends upon the pattern of colloidal structure, upon phase
differences and transitory and permanent membranes, which make possible the
combination and separation of chemical reactions, the orderly sequence of which
is at the root of all vital phenomena; and it is this plastic framework which enables
the cell to perform its proper functions, in the same way as the bones, comiective
tissue, membranes, and so on, permit of integration and division of labor in the
body itself and control the form and function of a wonderful mechanism handed
down from antiquity.

THE POSSIBILITIES OF FURTHER STUDY.

Obviously the investigation of mitochondria is rapidly progressmg beyond the
purely descriptive stage. The mere discovery of mitochondria in some new^'genus
or species excites but Uttle interest, because we have a very shrewd suspicion that
they are present there anjnvay. It would be a great mistake, however, to assume
that nothing further remains to be done. We know nothing whatever of mito-
chondria in any of the peripheral sense-corpuscles (p. 52), and many other prob-
lems suggest themselves. In descriptive work we can profitably let nature be the
experimenter and select those forms which aid in the solution of definite questions.
For instance, we can direct our attention toward the effect of environmental and
other conditions on mitochondria.

Temperature.— It has been found that mitochondria go into solution when the
temperature of the tissue containing them is raised for a few minutes to 45° or 50° C,
which is most suggestive with regard to their chemical constitution, the effect of
burns, the pathology of fever, and other questions. In this connection the study
of succulent plants, Uke Sempervivum, whose internal temperature is said to reach
about 52° C, in response to changes in the environment (Jost, 1907, p. 44), as com-
pared with others growing in cold climates and in the depth of winter, would afford
a new avenue of approach. The flora of hot springs should also be studied.

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. 145

Atmospheric pressure. — The effect of variations in atmospheric pressure on
mitochondria has not been studied on account of the technical difficulties; for this
reason a comparison of their relations in alpine and deep-sea fauna might yield
interesting results.

Osmotic pressure. — The effect of variations in osmotic pressure on mitochondria
might be studied by comparing the mitochondria in fresh-water forms with those
inhabiting the most concentrated brine. In plants it is not difficult to find a great
variation in osmotic pressure.

Acidity. — One of the most characteristic properties of mitochondria is their
solubility in mixtures containing even a small amount of acetic acid (0.5 to 2.5
per cent). It would be interesting to study their relations in vinegar eels, which
normally five in an environment containing 4 per cent of acetic acid. The sali-
vary glands of the rock-boring mollusc Dolium galea, which secretes sulphuric acid
in a concentration of 4 or 5 per cent (see Bayliss, 1915, p. 359), and the acid-forming
cells of plants would also repay investigation. If Macallum (1908, p. 628) and
others are right (which seems, however, unlikely) in assuming that the hydro-
chloric acid of the gastric juice is formed in the parietal cells, then these cells must
contain hydrochloric acid in much higher concentration than the juice (0.3 to 0.45
per cent), and Regaud's (1908a, p. 18) assertion that they are devoid of mitochon-
dria becomes of vital importance.

Water-content. — The effect of variations in the water-content upon the form of
mitochondria has never been determined. For this purpose the Scyphozoa, with a
water-content of 99 per cent, and the Trochebninthes, which can survive prolonged
desiccation (Parker and Haswell, 1897, p. 309), offer an excellent opportunity.

Hibernating animals take no water, though they continue to excrete urine.
All of the water which is absolutely necessary for the continuance of their vital
processes is metabolic, being produced by oxidation of the proteins, fats, and carbo-
hydrates of the tissues through respiration (Babcock, 1912, p. 170). The general
slowing-up of metabolism, the drowsiness and sleepiness, is in all {irobability due
to a reduced water-content, with consequent retardation of all chemical reactions.
The fact that the fats yield more water than either the proteins or the carbohydrates,
in the case of some fats even more than their own weight of water, would seem to
indicate the possibility of there being some change in the mitochondria (which
are, themselves, phosphoUpins containing glycerol and fatty acids among other
things). One would expect a diminution. The well-known occurrence of other
cytological variations in nerve-cells during hibernation would also indicate that a
study of mitochondria in this condition might give valuable results. Further-
more, since mitochondria are present in almost all the tissues of all animals, we
must entertain the possibihty that they may be in part the source of metabolic
water in general. For this reason investigations into their relations in the common
clothes moth, Tinea pellionella, desert animals like serpents and prairie-dogs, and
sea-birds which have no opportunity to drink fresh water, should be undertaken.

Respiration. — Still more recently it has been claimed that the chief function
of mitochondria is protoplasmic respiration (p. 133). There seem to be several
ways by which this hypothesis can be tested:

146

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

(a) We may inquire whether there is any relationship between the number of
mitochondria and the respiratory exchange of different organisms. Table 4 is
a portion of a table compiled by Krogh (1916, p. 148) relating to certain insects.
It shows a tremendous variation from as much as 82 to 1.45 calories per kilogram
per hour. Yet a comparison of the mitochondria in the two has not been made.

Table 4. — Metabolism of cold-blooded animals at about 20° C.

Name.

Weight.

Temper-
ature.

Calories.

Determi-
nation.

Authority.

0.1

0.01

0.09

1.0

"C.
18
20
20
20
20
20
20

1.45

2.5

82.0

24 0

15.0

6.2

3.5

Oo

O2
O2

O2

Thunberg.
Slowtzoff.
Parlioii.

Do.
Battelli and Stern.

Do.
Regnault and Reiset.

Apis mellifica ! . .

(b) Similarly for the oxygen consumption of tissues (Bayliss, 1915, p. 612) :

Lungs, 0.01.5 c.c. per gram per minute.
Submaxillary gland, 0.027 to 0.089 c.c. per minute.
Suprarenal gland, 0.045 c.c. per minute.

The oxygen consumption of the submaxillary gland is about twice that of
lung-tissue and the suprarenal four times. Roughly speaking, it is true that the
mitochonch-ia are relatively less abundant in the lungs than in the submaxillary
gland, but I do not think that there is any great difference between the suprarenal
and the submaxillary, at least in the mouse.

Ehrlich has made a study of the oxygen saturation of organs by another and
less satisfactory method. According to him the organs may be divided into three
groups (quoting from Bayhss, 1915, p. 595):

"1. Those of high 'oxygen saturation,' in which indophenol blue is not reduced, such
as the grey matter of the brain, the heart and some other muscular organs.

"2. Those which reduce indophenol blue, but not alizarin blue. Such are the greater
number of the tissues, smooth muscle, most voluntary muscles and secreting glands.

"3. Those which reduce even alizarin blue— lungs, liver, fatty tissue, Harderian gland.' '

Personally, however, I have been unable to find any correspondence between
the amount of mitochondrial substance and these figures of oxygen saturation.

(c) Within single cells also attempts have been made to measure oxidations.
For example, R. 8. Lillie (1913, p. 247) finds that "in frogs' blood-corpuscles
the formation of indophenol by the intracellular oxidation of a mixture of alpha-
naphthol and dimethyl-para-diamino-benzene takes place most rapidly in the imme-
diate neighborhood of the nuclear and plasma membranes. The conditions at the
surfaces of these structures are thus |)articularly favoral)le to rapid oxidations."
One might expect to find condensations of mitochondria in these localities, but I
have carefully examined mitochondria vitally stained with janus green in frogs'
red blood-corpuscles and I have found that they are distributed more or less uni-
formly throughout the cytoplasm.

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM. 147

(d) Or we may ask whether, when we modify experimentally the rate of oxida-
tion, there is any change in the mitochondria. Warburg (1910. p. 313) found that
he could double the oxygen consumption of sea-urchin eggs by the addition of
small amounts of sodium hydroxide to the sea-water. Adrenaline increases the
oxygen consumption of tissues and cyanide inhibits it. Israel's (1891, p. 334)
account of changes in the bioblasts of Altmann (mitochondria) in the cells of the
kidney following experimental ligation of the renal artery deserves confirmation
and extension. He found that they reacted very quickly — that is to say, in the
course of several hours. Champy's description of the solution of mitochondria
through asphyxiation in the center of pieces of tissue, remote from the surround-
ing oxygen, grown in serum, may have more bearing upon the question.

(e) It would be profitable also to inquire into the condition of mitochondria
in intestinal worms which are anaerobic and which normally live in the absence of
oxygen. Ascaris is a good example. According to Mathews (1905, p. 333):

" The only difference between anaerobic and aerobic respiration is that the anaerobic
protoplasm is so powerful a reducing agent that it is able to drive hydrogen out of the
water, thus oxidizing itself without the aid of atmospheric oxygen to act as a depolarizer.
iErobic protoplasm being less powerfully reducing, requires the presence of more or less
oxygen to take care of the hydrogen. The difference between these different kinds of
protoplasm is exactly the difference between metallic sodium and metallic iron."

If this be true we would expect the mitochondria to be of unusual abundance;
but Bayliss (1915, p. 611) does not seem to subscribe to it. As a matter of fact,
judging by the work of Romeis (1913a, p. 9) and others, the mitochondria in As-
caris do not seem to be pecuUar. Leeches can live for upwards of ten days in the
ab.sence of oxygen, yet the illustrations of Grynfeltt (19126, p. 263) of their mito-
chondria do not seem to show an\' noteworthy differences from those of other
annelids. The cells of the gas-bladder of fishes, wliich actually secrete oxygen
(Woodland, 1911, p. 225), constitute another field for study.

(/) And finally, in the condition of acidosis, there is an inliibition of the respir-
atory oxidation. It may be induced experimentally in a variety of ways. In
some preliminary experiments which I have made with rabbits by poisoning with
illuminating gas I have failed to detect any alteration in the mitochondria of the
lymphocytes in the circulating blood stained w^ith janus green, and I think that one
would expect to find changes in them earlier than in the other tissues.

While the hypothesis that the mitochondria are concerned in protoplasmic respi-
ration is very attractive and meets some of the reciuirements, it should nevertheless
be tested experimentally in many directions before it can be unreservedly accepted.

But we can hope for a more accurate analysis along purely experimental lines
by altering the rate and condition of the various vital processes and noting the
effect, if any, upon the mitochondria. We must bear continually in mind the
difficulty of dissociating functionally between the mitochondria and their environ-
ment. A change in the mitochondria does not necessarily mean that they are
concerned in the process in question, for they may be purelj- passive agents, the
change in their appearance being entirely due to some alteration in the protoplasm

148 THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

in wliich they are embedded— some variation in fluidity, refractive index, or elec-
trical state, for instance. And conversely, the absence of any noticeable reaction
on the part of the mitochondria does not exclude the possibiUty of their participa-
tion. Interpretation is extraordinarily difficult. The methods of tissue culture,
if rigidly controlled, may prove of service.

The most pressing need, however, is for further knowledge of the chemistry
of mitochondria, for, when chemical facts come in at one door, superstition inva-
riably vanishes at the other. Here also the difficulties seem almost insurmount-
able, because we have had to rely upon indirect and roundabout methods. If it
were possible to make a direct chemical analysis of mitochondria it would place
the whole work, once and for all, upon a secure foundation, but doing this involves
very special training and it is hard to see the way. It might be possible, as I have
already suggested, to partly separate out mitochondria by means of the centrifuge
and then collect them for analyses by some of the recently devised methods of
cell dissection. It would seem that some such procedure might be of use in the
analysis of many other cytoplasmic constituents. Should it prove feasible it would
have a most illuminating effect upon the whole question of the constitution of
protoplasm.

EXPLANATION OF FIGURES OF PLATE.

Figures 1, 3, 5, 6, 10, and 13 have been drawn from preparations fixed in Regaud's formalin and bichromate
mixture and stained with fuchsin and methyl green. Figure 7 was drawn from material fixed in the same way
but stained with iron hematoxylin. In making the drawings Zeiss apochromatic objective 1.5 mm., compensating
ocular 6, and camera lucida were used giving a magnification of l,.50O diameters.

Fia. 1. Kidney-cells of white mouse. ,.,,.., ui i ii •. u j • ■ *i, • u i

Fia. 2. Cells of trapezoid nucleus of white mouse, remarkable for the large block-hke mitochondria in the peripheral
cytoplasm (after Nicholson, 1916). , ■ , ,

Fig 3. Vesicula seminalis of white mouse. Note the absence of blebs on the mitochondrial filaments and the
stages in the production of the secretion.

Fig. 4. Large anterior horn nerve-cell of white mouse (after Nicholson, 1916).

Fia. 5. Prostate of white mouse containing very minute mitochondria.

Fig. 6. Ovarian egg of white mouse containing for the most part granular mitochondria.

Fig! 7. Serous celk of parotid of mouse with very fine rod-Uke mitochondria.

Fig. 8. Thyroid vesicle (after Bensley, 1916). . , , . , ^,

Fig 9. Pancreas (after Scott, 1916), with typical bleb-like swellings on the mitochondrial filaments.

Fig lo! Inte.stinal epithelium of white mouse with bipolar arrangement of mitochondria.

Fig u! I..arge cell of mesencephahc nucleus of the fifth nerve with small cell of locus coeruleus adjacent (after Nichol-
son, 1916). Note the difference in the mitochondria.

Fig. 12. Large pyramidal cell of hippocampus (after Nicholson, 1916).

Fig. 13. Thvnius of white mouse, showing large mitochondria in the small round cells, tiny mitochondria in the
epithelial cells, and an apparent absence of mitochondria in the mast cell.

COWDRY

v':f.''C:

>:'

A

'''•'■ ^ ^fe

i^i-

•

•

-i--

^ —/

\,

-1

I

*" ■—^<r 'Xiff. ."i^** -L-9 " N

10

.... : j*>;
>• 'V,

• /'I -^ **• ,* • (■'•,•.,■, .SE

r-^x"

— i- iis***^*fe

%lUr^

i$:^''^ ■■</

0 ^^'i^/yo/'^-^Vvv^

BIBLIOGRAPHY.

AiMfi, Paul, 1911. Bandes intercalaires et bandes de
contraction. Bibl. Anat., t. 21, pp. 263-272
(see text, p. 56).

AKATsn, 1903. Mikroskopische Untersuchung der Se-
kretionsvorgange in den Samenblasen. Pfliiger's
Archiv, Bd. 96, pp. .555-571 (pp. 51, 102, 119).

Alagna, a., 1911. Sulla presenza di formazioni mito-
condriali negli element! constitutivi delle tonsiUe
palatine normali ipertrofiche e delle vegetazioni
adenoidi. Arch. f. Zellforsch., Bd. 7, pp. 27-
40 (pp. 51, 52, 135).

Alexeieff, a., 1916a. Mitochondries chez quelques
protistes. Mitochondries glycoplastes et adi-
poplastes. Caractfires g^neraiLX des mitochon-
dries. C. rend. Soc. Biol., t. 79, pp. 1076-
1079 (p. 102).

1916b^ Mitochondries chez quelques protistes.

Mitochondries glycoplastes. C. rend. Soc. Biol.,
t. 79, pp. 1072-1075 (p. .53).

— 1917a. Sur les mitochondries a fonction gly-

coplastique. C. rend. Soc. Biol., t. 80, pp. 510-
512 (p. 102).

19176. Nature mitochondriale du corps para-
basal de-s Flagell^s. C. rend. Soc. Biol., t. 80,
pp. 499-502 (pp. 46, .53).

1917f. Mitochondries et corps parabasal chez

les Flagelles. C. rend. Soc. Biol., t. 80, pp.
358-360.

Altmann, R., 1889. Ueber die Fettumsetzungen im
Organismus. Arch. f. Anat. u. Phys., Anat.
Abt., Suppl. Bd., pp. 86-104 (pp. 102, 124, 135).

1890. Die Elementarorganismen und ihre Bezie-

hungen zu den Zellen. Leipzig, Veit Co., 145
pp. (pp. 42, 44, 51, 52, 59, 101, 102, 13.5, 142).

1894. Die Elementarorganismen und ihre Bezie-

hungen zu den Zellen. 2. Aufl. Leipzig, Veit
Co., 160 pp. (pp. 46, 51, 97).

Arndt, a., 1914. Ueber generative Vorgiinge bei
Amoeba chondrophora, n. sp. Arch. f. Protisten-
kunde, Bd. 34, pp. 39-.59 (pp. 44, 49, 53).

Arnold, George, 1912a. On the condition of the epider-
mal fibrils in epithelioma. Quart. Jour. Micr. Sci.,
vol. 57, pp. 283-299 (pp. 101, 107, 110, 135, 140).

19126. The role of the chondriosomes in the cells

of the guinea-pig's pancreas. Ai-ch. f. Zellforsch,
Bd. 8, pp. 2.52-271 (pp. 48, 121).

Arnold, J., 1907. Plasmosomen, Granula, Mitochon-
drien, Chondriomiten und Netzfiguren. Anat.
Anz., Bd. 31, pp. 640-648 (p. 46).

1908. Supravitale Farbung Mitochondrien ahn-

Ucher Granula in den Knorpelzellen nebst Be-
merkungen iiber die Morphologic des Knorpel-
glykogens. Anat. Anz. Bd. 32, pp. 361-366
(pp. 102, 127).

1913. Das Plasma der somatischen Zellen im

Lichte der Plasmosomcn-Granulalehre und der
Mitochondrienforschung. Anat. .\nz., Bd. 43,
pp. 43.3-462 (p. 46).

AsvADonROVA, Nina, 1913. Recherches sur la forma-
tion de quelques cellules pigmentaires et des
pigments. Arch. d'Anat. micr., t. 15, pp. 153-
314 (pp. 45, 49, 84, 102).

Aunap, E., 1913. Ueber die Chondriosomen der Gono-
cyten bei Knochenfischen. Anat. Anz., Bd. 44,
pp. 449-459 (p. 56).

Azzi, Azzo, 1914. Ueber das Verhalten der Chon-
driosomen bei der fettigcn Entartung. Centralbl
f. allg. Path. u. path. Anat., Bd. 25, pp. 7-13
(p. 124).

Babcock, S. M., 1912. MetaboUc water; its production
and role in vital phenomena. Univ. of Wisconsin
Agric. Exper. Stat. Res. Bull. No. 22, pp. 87-
187 (p. 14.5).

voN Baehr, W. B., 1909. Die Oogenese bei einigen
viviparen Aphididen und die Spermatogenese
von Aphis saUceti, rait besonderer Beriick-
sichtigung der Cliromatinverhaltnisse. Arch. f.
Zellforsch, Bd. 3, pp. 269-334 (p. 55).

Baitsell, G. a., 1916. The origin and structure of a
fibrous tissue formed in wound heaUng. Jour.
Exper. Med., vol. 23, pp. 739-756 (p. 105).

Ballowitz, E., 1907. Zur Kenntnis der Spermien der
Cetaceen. Arch. f. mikr. Anat., Bd. 70, pp.
227-237 (pp. 44, 56).

Barratt, J. O., 1913. Changes in chondriosomes occur-
ring in pathological conditions. Quart. Jour.
Micr. Sc, vol. 58, pp. 553-566 (p. 136).

Baumgartner, W. J., 1902. Spermatid transformations
in Gryllus assimiUs, with special reference to the
nebenkern. Kansas Univ. Sci. Bull., vol. 1,
pp. 47-74 (p. 55).

Bayliss, W. M., 1915. Principles of General Physiology.
London, Longmans, Green & Co., 850 pp. (pp.
134, 14.5-147).

Beauverie, J., 1914a. Sur le chondriome des Basidio-
mycdtes. C. rend. Acad. Sci., t. 158, pp. 798-
800 (pp. .52, 56).

19146. Sur le chondriome d'une Ur<5din6e: le

Puccinia malvacearum. (Reun. biol. Nancy.)
C. rend. Soc. Biol. t. 76, pp. 359-361 (pp. 52, 56).

Beckton, H., 1909. On granules in cells of normal
tissues and new growths. Arch. Middlesex
Hosp., vol. 15, pp. 182-191 (pp. 135, 139).

and Russ, S., 1911. The effect of radium emana-
tion on Altmann's granules. Arch. Middlesex
Hosp., vol. 23, pp. 59-105 (pp. 135, 140).

Beckwith, C. J., 1914. The genesis of plasma-structures
in the egg of Hydractinia echinata. Jour. Morph.,
vol. 25, pp. 189-252 (pp. .54, 74, 97, 98).

Benda, C, 1897. Neue Mitteilungen iiber die Histo-
genese der Saugetierspermatozoen. Verh. d.
physiol. Ges., pp. 406-^14 (pp. 94, 102).

1898. Ueber die Spermatogenese der Verte-

braten und hoheren Invertebraten. 2. Die
Histiogenese der Spermien. Verh. d. physiol.
Ges., pp. 393-398 (pp. 45, 49).

' The bibliography is not complete, for only the papers cited in the text are given. Dueaberg (1912) has given a oom-
ploto list of references to 1911.

149

150

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

Benda, C, 1899(7. Weitere Mitteilungen iiber die Mito-
chondria. Verh. d. Physiol. Ges., pp. 376-383
(pp. 42, 45, 47, 49, 52, 55, 102, 106, 135).
1899b. Weitere Beobachtunpen iiber die Mito-
chondria und ihr Verhaltnis zu Secretgranula-
tionen, nebst kritischen Bemerkungen. Verh. d.
physiol. Ges. (pp. 51, 54).

1901. Die Mitochondriafiirbung und andere

Methoden zur Untersuchung der Zellsubstanzen.
Verh. d. anat. Ges., pp. 155-174 (p. 61).

1902. Die Mitochondria. Ergebn. d. Anat. u.

Entw. (Anat. Hefte, 2. Abt), Bd. 12, pp. 743-
781 (p. 75).

1903. Die Mitochondrien des Nierenepithels.

Verh. d. anat. Ges., pp. 123-129 (pp. 122, 130).

1914. Die Bedeutung der Zelileibstniktur fur

die Pathologie. Verh. d. deutsch. path. Ges., pp.
5-42 (pp. 48. 50, 130).
Bensley, R. R., 1910a. On the so-called Altmann's
granules in normal and pathological tissues.
Trans. Chicago Path. Soc, vol. 8, pp. 78-83
(p. 139).
19106. On the nature of the canaUcular appara-
tus of animal cells. Biol. Bull., vol. 19, pp. 179-
194 (p. 44).

1911. Studies on the pancreas of the guinea-pig.

Amer. Jour. Anat., vol. 12, pp. 297-388 (pp.
44, 51, .59, 60, 62, 87, 109).

1916. The normal mode of secretion in the

thyroid gland. Amer. Jour. Anat., vol. 19,
pp. 37-54 (pp. 52, 63, 73, 148).
BiNFORD, Raymond, 1913. The germ-cells and the
process of fertilization in the crab, Menippe
mercenaria. Jour. Morph., vol. 24, pp. 147-
200 (p. 54).
BioNDi, G., 1915. Ueber die Fettphanerosis in der
Nervenzelle. Virchow's Archiv, Bd. 220, pp. 222-
233 (pp. 135, 137).
BLtJNTSCHLi, H., 1904. Beobachtungen am Ovarialei
der Monascidie Cynthia microcosmus. Morph.
Jahrb., Bd. 32, pp. 391-449 (p. 55).
BoBEAU, G., 1911. Mitochondries et lipoides dans les
glandules parathyroides du cheval. C. rend.
Ass. Anat., pp. 183-187 (pp. 52, 102, 117).
BoNNEViE, K., 1904. Zur Kenntnis der Spermiogenese
bei den Gastropoden (Enteroxenes ostergeni).
Biol. Centralbl., Bd. 24, pp. 267-274 (p. 55).

1907. Untersuchungen uber Keimzellen. II.

Physiologische Polyspermie bei Bryozoen. Jena
Zeitschr. f . Naturw., Bd. 42, pp. 567-598 (p. 54).
BOVERI, Th., 1888. Zellen-Studien. Jena Zeitschr.

f. Naturw., Bd. 22, pp. 686-882 (p. 44).
Branca, A., 1911. Sur la structure du poll. Jour, de
I'Anat et de la Phvs., t. 47, pp. 54.5-600 (pp.
51, 81).
Broman, J., 1907. Ueber Bau und Entwickelung der
Spermien von Rana fusca. Arch. f. mikr. .\nat.,
Bd. 70, pp. 330-360 (p. 56).
Browne, Ethel N., 1913. A study of the male germ-
cells in Notonecta. Jour. Exper. Zool., vol. 14,
pp. 61-121 (p. 55).
Bruck, a., 1914. Die Muskulatur von .\nodonta
cellensis Schrot. Zeit. f. wiss. Zool., Bd. 110,
pp. 481-619 (pp. 55, 107).
VON Brcnn, a., 1884. Beitrage zur Kenntnis der
Samenkorper und ihrer Entwickelung bei den
Siiugetieren uml Vogeln. .Arch. f. inikr. ,\nat.,
Bd. 23, pp. 108-132 (pp. 4.-), 51;.

Buchner, p., 1909. Das accessorische Chromosom in
Spermatogenese und Ovogenese der Orthopteren.
Arch. f. Zelforsch., Bd. 3, pp. 33.5-430 (p. .55).
Bullard, H. H, 1916. On the occurrence and physio-
logical significance of fat, etc. Amer. Jour.
Anat., vol 19, pp. 1-35 (p. 83).
Burrow s, M. T., 1917. Some factors regulating growth.

Anat. Rec, vol. 11, pp. 335-339 (p. 79).
BuSACCA, A., 1913. L'apparato mitochondriale nelle
cellule nervosc adulte. .\rch. f. Zellforsch., Bd.
11, pp. 327-339 (p. .56).

1915. SuUe modificazione dei plastomi nelle

cellule del epitclio pigmentato della retina sotto
I'azione dell' luce e dell' oscurita (Tabola 11).
Ricerche fatte nel Lab. di Anat. normale della
Univ. di Roma., vol. 18, pp. 217-237 (p. 132).
Cajal, S. R., 1908. Les conduits de Golgi-Holmgren du
protoplasma nerveux et le r<?scau p6ricellulaire
de la membrane. Trab. Lab. Invest, biol. Univ.
Madrid, t. 6, pp. 123-136 (p. 45).

1915. .\lgunas variaciones fisiologicas y patolo-

gicas del aparato reticular de Golgi. Trab. Lab.
Invest, biol. Univ. Madrid, t. 12, 101 (pp.
63, 125).
Casteel, D. B., 1917. Cytoplasmic inclusions in male
germ-cells, etc. Jour. Morph., vol. 28, pp.
643-686 (p. .55).
Cermak, N., 1902. Die Mitochondrien des Forelleneies.

Anat. Anz., Bd. 20, pp. 158-160 (p. 56).
Cesa-Bianchi, D., 1909. Experimentelle Untersuch-
ungen iiber die Nierenzelle. Frankfurt Zeitschr.
f. Path., Bd. 3, p. 461 (p. 1.37).

1910. Contributo aUa conoscenza della anato-

mia e della fisio-patologia renale. Internat.
Monatsschr. f. Anat. u. Phys., Bd. 27, pp.
89-186 (p. 137).
Cesaris-Demel, H., 1907. Studien uber die roten
Blutkorperchen mit den Methoden der Fiirbung
in frischem Zustande. Folia ha?m., Bd. 4,
Suppl., pp. 1-32 (p. 46).
Chambers, R., 1915. Microdissection studies on the
germ cell. Science, n. s., vol. 41, pp. 290-293
(pp. 43, 97, 98).
Champy, C, 1911. Recherches sur I'absorption intes-
tinale et le role des mitochondries dans I'absorp-
tion et la s6cr^tion. Arch. d'Anat. micr., t. 13,
pp. 55-170 (pp. 121, 133, 135).

1912. Sur les ph^nomfenes cytologiques qui s'ob-

servent dans les tissus cultives en dehors de
r organisme (tissus epith^Uaux et glandulaires).
C. rend. Soc. Biol., t. 72, p. 987 (pp. 43,'.52, 147).

1913a. Recherches sur la spermatogenese des

Batraciens et les elements accessoires du testicule.
Arch, de Zool. exp. et. g6n., t. 52, pp. 13-304
(pp. 45, 49, 56).

19136. La dediff^renciation des tissus cultives

en dehors de I'organisme. Bibl. Anat., t. 23
pp. 184-205 (pp. 43, 52).

1914. IV. Le rein. Arch, de Zool. exp. et g6n.

t. 54, pp. 307-386 (pp. 44, 52, 135).

1915. V. La glande thyroide. Arch, de Zool

exp. et gen., t. .55, pp. 61-79 (p. 52).
Chaves, P. R., 1912. Notes sur I'ergastoplasme. Bull
Soc. Portugaise Sci. Nat., t. 6, pp. 29-33 Tp
.56).

1915. Sobre a cellula serosa pancreatica. Arch

de Anat. e de .\nthroi)., vol. 4, pp. 1-131 (pp
56, 121).

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

151

Child, C. M., 1915. Senescence and rejuvenescence.

Univ. of Chicago Press. 481 pp. (p. 74).
Ci.\ccio, C, 1911. Sui niitochondri degli elementi

linfoidi e mieloidi. Pathologica, vol. 3, pp.

13-17 (pp. 102, 11.5).

1913a. Zur Physiopathologie der Zelle. Cent-

ralbl. f. allg. Path. u. path. Anat., Bd. 24, pp.
721-727 (pp. 40, 11.5).

19136. Les plastosomes des elements de la

s^rie heraoglobiniquo. Folia ha-m., pp. 391-393.

CiACcio and Scaglione, 1913. Beitrag zur cellularen
Physiopathologie der Plexus chorioidei. Ziegler's
Beitrage z. path. Anat., Bd. .55, pp. 131-167
(p. 118).

Clark, Elbert, 1914. Regeneration of niedullated
nerves in the absence of embryonic nerve fibers
following experimental non-traumatic degenera-
tion. Jour. Comp. Neur., vol. 24, pp. 61-111
(pp. 135, 137).

CoGHiLL, G. E., 1915. Preliminary studies in intra-
cellular digestion and assimilation in amphibian
embryos. Science, n. s., vol. 42, pp. 347-350
(pp. 102, 124, 133).

Collin, R., 1913a. Les mitochondries de la cellule
nfvToglique k expansions longues ct les grnnu-
lations lipoides de la substance grise des centres
nerveux chez I'homme. C. rend. Assoc. Anat.,
pp. 178-186 (p. 52).

19136. Les mitochondries du cylindraxe, des

dendrites et du corps des cellules ganglion-
naires de la ratine. C. rend. Soc. Biol., t. 74,
pp. 1358-1360 (p. 52).

1914. Sur les mitochondries extraneuronales

dans I'c^corce cerebrale irrit^e. C. rend. Soc.
Biol, t. 67, pp. 591-593 (p. 135).

CoMOLLi, A., 1913. Ricerche istologiehe suU' inter-
renale dei Teleostei. Arch. Ital. di Anat. e di
Embriol., t. 11, pp. 377-408 (p. 56).

CoNKLiN, E. G., 1905. Organ-formins substances in the
eggs of ascidians. Biol. Bull., vol. 8, i)p. 205-
230 (pp. 44-46, 99).

■ — 1915. Heredity and environment. Princeton

University Press, pp. 533 (pp. 45, 96).

■ — 1917. Effects of centrifugal force on the struc-
ture arid development of the eggs of Crepidula.
Jour. Exper. Zool., vol. 22, pp. 311-420 (pp.
45, 46).

Corner, G. VV., 1914. The corpus luteum of pregnancy,
as it is in swine. Carnegie Inst, of Wash. Con-
tributions to Embryology, Xo. 5, pp. 71-94
(p. 51).

C'onTi, A., 1913. Studi suUa minuta struttura della mu-
cosa intestinale di Vertebrati, etc. Arch. Ital.
di Anat. e di Embriol., t. 11, pp. 1-189 (pp. 55,
56).

CowDiiY, E. v., 1911. A preliminary report on the
structure of nerve-celLs. Jour. Nerv. and Ment.
Dis., vol. 38, p. 753 (p. 46).

1912. Mitochondria and other cytoplasmic

constituents of the spinal ganglion cells of the
pigeon. Anat. Rec, vol. 6, pp. 33-38 (p. 56).

1913a. The relations of mitochondria and other

cytoplasmic constituents in spinal ganglion cells
of the pigeon. Internat. Monatschr. f. Anat. u.
Phys., Bd. 29, pp. 473.504 (pp. 46, 63).

19136. Demonstration speciales No. 6. Les

mitochondries dans les cellules des ganglions
spinaux trait^'s par la m6thode de Bensley. Bibl.
Anat., t. 23, p. 311 (p. 46).

CowDRY, E. v., 1914a. The comparative distribution

of mitochondria in spinal ganglion cells of

vertebrates. Amer. Jour. Anat., vol. 17, pp. 1-29

(pp. 54, .55, .56, 102, 124).
19146. The vital staining of mitochondria, etc.

Internat. Monatschr. f. Anat. u. Phys., Bd. 31,

pp. 267-286 (pp. 52, 59, 88).
1914c. The relations of mitochondria in cells

multiplving by mitotic and amitotic division.

Anat. Rec, vol. 8, pp. 102-103 (p. 79).

WH'V. The development of the cytoplasmic

constituents of the nerve-cells of the chick.
I. Mitochondria and neurofibrils. Amer. Jour.
Anat., vol. 15, pp. 389-429 (pp. 110, 111).

1916a. The general functional significance of

mitochondria. Amer. .Jour. Anat., vol. 19, pp.
423-446 (pp. 47, 59, 62, 108, 134).

19166. The structure of chromophile cells of the

nervous system. Carnegie Inst, of Wash.
Contributions to Embryology, No. 11, pp. 27-
43 (pp. 60, 98, 131).

CowDRY, N. H., 1917. A comparison of mitochondria
in plant and animal cells. Biol. Bull., vol. 33,
pp. 196-228 (pp. 64, 69, 72, 77, 83, 86, 132).

d'.\gat.\, G., 1913. Sulla genesi del grasso e suUe modi-
ficazioni dell' apparato mitocondriale nell' in-
tossicazione difterica. Internat. Monatschr. f.
Anat. u. Phys., Bd. 29, pp. 443-459 (p. 135).

DA Costa, C, 1913. Recherches sur I'histophysiologie
des glandes surr^nales. Arch, de Biol., t. 28,
pp. 111-191 (pp. 52, 56).

1916. Sur la developpement des capsules sur-

ri-nales du chat. Bull. Soc. Portugaise des Sci.
Nat., t. 7, pp. 161-171.

d'Athias, 1912. L'appareil mitochondrial des cellules
interstitielles de I'ovaire du Murin. C. rend.
Soc. Biol., t. 73, pp. 448-449.

1915. Cristalloides dans I'oeuf de Cercopithecus

caUitrichus et de Cercopithecus sabaeus. Bull. Soc.
Portugaisedes Sci. Nat. ,t.7,pp.67-76 (pp.56, 116).

Dannehl, p., 1892. Ueber die cadaverosen Verander-
ungen der Altmannschen Granula. Virchow's
Arch., Bd. 128, pp. 48.5-496 (p. 135).

De Bonis, V., 1907^ Ueber die Sekretionserscheinungen
in den Driisenzellen der Prostata. Arch. f.
Anat. u. Phys., Anat. Abt., pp. 1-16 (p. 120).

DiBBELT, 1914. Genese der Epithelveriinderungen in
der Niere bei experimenteller Diphtheric, ein
Beitrag zur Pathologic der Zelle. Verh. d.
deutsch. path. Ges., pp. 114-118.

DiNGLER, M., 1910. LTeber die Spermatogenese des
Dicrocodium lanceatum Stil. et Hass (Distomum
lanceolatum). Arch. f. Zellforsch, Bd. 4, pp.
672-712 (p. 54).

DoMiNici, 1913. SuUe formazioni mitocondriali e sui
granuli di secrezione nella prostata del cane e
prostata umana ipertrofica. Folia urol., Bd.
7, pp. 295-303 (pp. 51, 119, 135, 137).

DuBREuiL, G., 1909. 1. Origine, destinee et appareil
mitochondrial des Plasmazellen du grand Epi-
ploon du Lapin. C. rend. Soc. Biol., t. 67, pp.
80-82 (p. .52).

1910a. Mitochondries des ost^oclastes et des

cellules de Bizzozero. C. rend. Soc. Biol., t.
69, pp. 71-73 (p. .52).

19106. L'appareil mitochondrial (pErin^me, mito-
chondries, chondriochontes), dans la lipn^c rellu-
laire allant du lymphocyte k la cellule osseuse. C.
rend. Soc. Biol., t. 68, pp. 1100-1102 (p. 52).

152

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

DUBREUIL, G. 1911a. Le chondriome des globules
blanrs mononucl^es et des cellules connectives,
cartilaprineuses, et osseuses chez les mammifferes.
C. rend. Assoc. Anat., pp. 130-145 (p. 52).

19116. Transformation directe des mitochon-

dries et des chondriocontes en graisse dans les
cellules adipeuses. C. rend. Soc. Biol., t. 70,
pp. 264-278 (p. 124).

1912. La mitochondrie forme la plus apte k la

multiplication des 616ments du chondriome. C.
rend. Assoc. Anat., pp. 127-133 (p. 48).

1913. Le chondriome et le dispositif de 1'

activit<5 s6cr6toire, etc. Arch. d'Anat. micr., t.
11, pp. 53-151 (pp. 52, 62, 67, 127, 135).
DUBRE0IL and F.wre, 1914a. Nature et signification
des corps de Russell. C. rend. Soc. Biol., t. 77,
pp. 372-374 (p. 102).

1914b. Chondriome de-s Plasmazellen. C. rend.

Soc. Biol., t. 77, pp. 24-26 (p. 52).
DoESBERG, J., 1907. Der Mitochondrial-Apparat in den
Zellen der Wirbelthiere und Wirbellosen. Arch,
f. mikr. Anat., Bd. 17, pp. 284-296 (pp. 46, 50).

1908. Sur I'existence de mitochondries dans

I'oeuf et I'embryon d'Apis mellifica. Anat. Anz.,
Bd. 32, pp. 261-265 (p. 46).

1909. Ueber Chondriosomen und ihre Ver-

wendung zu Myofibrillen beim Huhnerembryo.
Verb. d. anat. Ges., Bd. 34, pp. 123-126 (p. 106).

1910. Les chondriosomes des cellules em-

bryonnaires du poulet, et leur role dans la genese
des myofibrilles avec quelques observations sur
le d^veloppement des fibres musculaircs stri(5es.
Arch. f. Zellforsch., Bd. 4, pp. 602-671 (pp. 45, 46,
106, 110).
1912. Plastosomen, "Apparato reticolare in-
ferno," und Chroraidialapparat. Ergebn. d.
Anat. u. Entw. Gesch. (Anat. Hefte, 2. Abt.),
Bd. 20, pp. 567-916 (pp. 46, 48, 53, 97, 106, 109,
110, 126).
1913. Plastosomes et "organ-forming sub-
stances" dans I'ceuf des ascidiens. Bull, de la
classe des sci. Acad. roy. Belgique., No. 5, pp.
463-474 (p. 55).

1915. Recherches cytologiques sur la fdcondation

des ascidiens et sur leur d^veloppement. Car-
negie Inst, of Wash. Contributions to Embry-
ology, No. 8, pp. 35-70 (pp. 48, 95, 98-100, 107).
1917. Chondriosomes in the cells of fish em-
bryos. Amer. Jour. Anat., vol. 21, pp. 465-^96
(pp. 48, 49, 55, 60, 79).
DuESBERG, J., and H. Hoven, 1910. Observations sur
la structure du protoplasme des cellules v6g6-
tales. Anat. Anz., Bd. 36, pp. 93-100 (p. 57).
DuLZETTO, F., 1916. Contributo alia conoscenza della
struttura della granulosa e dell' origine dei
materiali deutoplasmici nell'oocite degU Uccelli
(Fringilla cannabina L.). Arch. ital. di Anat.
e di Embriol., vol. 15, pp. 193-217 (p. 56).
Eklof, H., 1914. Chondriosomenstudien an den
Epithel- und Driisenzellen des Magen-Darm-
kanals und den Oesophagus- Drtisenzellen bei
Saugetieren. Anat. Hefte, 1. Abt., Bd. 51,
pp. 1-227 (p. 51).
Enderlen, 1908. Beitrag zur Nierenchirurgie. Deutsche

Zeitschr. f. Chir., Bd. 41, p. 208 (p. 137).
Ernst, P., 1914. Die Bedeutung der Zellleibstruktur
ftir die Pathologie. Verb. d. deutsch. path.
Ges., pp. 43-103 (p. 137).

F.^HR, 1914. Zur Frage der sogenannten hyalintropfigen
Zelldegeneration. Verb. d. deutsch. path. Ges.,
pp. 119-122 (p. 122).
Faur^-Fremiet, E., 1907. Mitochondries et sph^ro-
plastes chez les Infusoires cili^s. C. rend. Soc.
Biol., t. 62, pp. 523-525 (p. 46).

1908. Evolution de I'appareil mitochondrial

dans I'ceuf de Julus terrestris. C. rend. Soc.
Biol., t. 64, pp. 1057-1058 (p. 55).

1909. Sur les reactions de quelques mitochondries.

C. rend. Acad. Sc, t. 149, pp. 163-166.

1910a. Etude sur les mitochondries des Proto-

zoaires et des cellules sexuelles. Arch. d'Anat.
micr., t. 11, pp. 457-648 (p. 53-55, 83, 98, 123).

1910b. Mitochondries et liposomes. C. rend.

Soc. Biol., t. 68, pp. 537-539 (p. 102).

1910c. Les glandes labiales des Hydrocorisis.

C. rend. Assoc. Anat., pp. 1-4 (p. 51).

1911. Le role des mitochondries dans I'^limina-

tion du fer chez les rhizopodes ar^nac^s. C.
rend. Soc. Biol., t. 70, pp. 119-136 (p. 102).
1912a. fttudes cytologiques sur quelques In-
fusoires des marais salant du Croisic. Arch.
d'Anat. micr., t. 13, pp. 402-480 (pp. 46, 54).

1912b. Sur la constitution des mitochondries des

gonocvtes de I'Ascaris megalocephala. C. rend.
Soc. Biol, t. 72, pp. 346-347 (pp. 54, 84).
1913. Le cycle germinatif chez I'Ascaris megalo-
cephala. Arch. d'Anat micr., t. 15, pp. 435-
757 (pp. 45, 83, 95, 99).
Faur6-Fremiet, E., Andre Meyer, et Georges
ScHAEFFER, 1910. Sur la microchimie des corps
gras. Application S, I'^tude des mitochondries.
Arch. d'Anat. micr., t. 12, pp. 19-102 (p. 102).
Favre and Dubrecil, 1914. Grains de segregation des
Plasmazellen. C. rend. Soc. Biol., t. 77, pp.
89-91 (p. 128).
Favre, M., and Cl. Regaud, 1910. Sur la nature des
fibres d'Herxheimer ou filaments basaux de
repiderme. Lyon m^d.. No, 22, pp. 1132-1138
(pp. 45, 50, 51, 102, 107, 113).

1913. Sur le.s formations mitochondriales dans

les cellules n^oplastiques des epith§Uomes de la
peau et des muqueuses dermopapillaires. C. rend.
Soc. Biol., t. 74, pp. 688-692 (pp. 136, 140).
Fiessinger and Caen, 1910. Les alterations du chon-
driome chez les mammifferes. C. rend. Soc. Biol.,
t. 68, pp. 454-4.55 (p. 127).
FiRKET, J., 1911. Recherches sur la genese d&s fibrilles
epidermiques chez le poulet. Anat. Anz., Bd.
38, pp. 537-549 (pp. 101, 102, 105, 108, 110).
Flemming, W., 1882. Zellsubstanz, Kern and Zelltei-
lung. Leipzig, F. C. W. Vogel, 424 pp. (pp. 42,
46, .52, 142).
FncHS, H., 1901. Ueber das Epithel im Nebenhoden
der Maus. Anat. Hefte, 1. Abt., Bd. 19, pp. 313-
347 (p. 51).
Galeotti, G., 1895. Ueber die Granulationen m den
Zellen. Internat. Monatschr. f. Anat. u. Phys.,
Bd. 12, pp. 461-.557 (pp. 52, 60).
Gatenby, J. B., 1917. The cj'toplasmic inclusions of the
germ-cells. Quart. Jour. Micr. Sci., No. 247,
pp. 407^63 (pp. 55, 102).
Gerard, P., 1909. Recherches sur la spermatogenese
dhez Stenobothus biguttulus. Arch, de Biol.,
t. 24, pp. .54.3-624 (p. 55).
de Giacomo, a., 1911. Suir ipertrofie compensatona e
sui fenomeni cellulari nei reni, etc. Internat.
Monatschr. f. Anat. u. Phys., Bd. 28, pp. 208-
232 (p. 137).

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

153

GiGLio-Tos, E., and L.Granata, 1908. I. Mitochondrl
nelle cellule seminali raaschili di Pamphagus
marmoratus (Burra). Biologica, vol. 2, No. 4,
pp. 1-11.5 (pp. 4.5, .50).

GoDLEWSKi, E., 1902. Entwickeluug des Skelett- und
Herzmuskelgewebes der Siiugetiere. Arch. f.
mikr. .Ajiat./Bd. 60, pp. 111-156 (p. 107).

GoETSCH, Emil, 1916. Functional significance of mito-
chondria in toxic thyroid adenomata. Johns
Hopkins Hosp. Bull.", vol. 27, pp. 129-133
(pp. 44, 100, 117, 135, 136).

GoLDsc'HMiDT, R., 1904. Der Chromidialapparat
lebhaft funktionicrender Gewebszellen. Zool.
Jahrb., Abt. f. Anat. u. Ont., Bd. 21, pp. 41-
134 (p. 45).

GoLui, C, 1898. Sur la structure des cellules nerveu.ses.
Arch. Ital. de Biol., t. 30, pp. 60-71 (p. 44).

GoVAERTs, P., 1913. Recherches sur la structure de I'
ovaire des insectes. Arch, de Biol., t. 28, pp.
347-444 (pp. .55, 123).

Grynfeltt, E., 1912a. Sur la presence de chondrio-
somes dans les cellules de la glande hypobranch-
iale de Murex trunculus. Bull. Acad, des Sci.
et Lettr. Montpellier, pp. 12-17 (p. 51).

1912b. Sur I'appareil mitochondrial des cellules

glandulaires de la glande hypobranchiale de
Murex trunculus. C. rend. Soc. Biol., t. 72, pp.
261-263 (pp. 55, 147).

1912c. Note sur le chondriome des cellules

epith^hales de la glande thyroide. Bull. Acad,
des Sci. et Lettr. MontpeUier, pp. 143-148
(pp. 102, 116).

1913. Sur la genese des boules picriphiles dans la

glande hypobranchiale de Murex trunculus.

Bull. Acad, des Sci. et Lettr. Montpellier, p. 11

(pp. 102, 120).
Grynfeltt and Euziere, 1912. Recherches cytolo-

giques sur les cellules (?pith^hales des plexus

choroides de quelques mammiferes. C. rend.

Assoc. Anat., p. 64 (pp. 101, 117).
1913a. Recherch&s sur les variations fonction-

nelles du chondriome des cellules des plexus

choroides chez quelques mamniiferes. C. rend.

Assoc. Anat., pp. 197-205 (p. 118).
19136. Note sur la structure de IV'pithc^Iium

des toiles choroidiennes et I'excretion du liquide

c6phalo-rachidien chez le scyllium. C. rend.

Assoc. Anat., pp. 101-111 (pp. 56, 118).
Guieys,se-Pelis.sier, A., 1910. Etude des mitochondries

de I'organe ent^rique des crustac^s d6capodes.

Bibl. Anat., t. 20, pp. 18-24 (pp. 51, 54).
Guilliermond, a., 1912a. Recherches cj^ologiques sur

le mode de formation de I'amidon et sur les

plastes v6g(?taux, etc. Arch, d'anat. micr., t.

14, pp. 309^28 (pp. .53, 57, .58, 78, 97, 102, 114;.
19126. Sur le mode de formation du pigment

dans la racine de carotte. C. rend. Acad. Sci.,

t. 1.5.5, pp. 411-J14 (p. .57).
1912c. Sur les mitochondries des organes sexuels

des v6g^taux. C. rend. Acad. Sci., t. 1.54, pp.

888-891 (p. 57).
I9l2d. Nouvelles remarques sur I'origine des

chlorleucites. C. rend. Soc. Biol., t. 73, p. 86

(p. 57).
1913a. Sur les mitochondries des champignons.

C. rend. Soc. Biol., t. 74, pp. 618-621 (pp. .52, 56).
19136. Nouvelles observations sur le chon-
driome de I'asque de Pustularia vesiculosa. C.

rend. Soc. Biol., t. 75, pp. 646-649 (pp. 52, .56).

Gdilliebmond, a., 1913f. Quelques remarques nou-
velles sur la formation des pigments anthocya-
niques au sein des mitochondries. C. rend. Soc.
Biol., t. 7.5, pp. 478-481 (p. 101).

1913d. Nouvelles remarques sur la signification

des plast&s de W. Schimper. C. rend. Soc. Biol.,
t. 75, pp. 437-440 (pp. 57, 102).

1913p. Sur le role du chondriosome dans 1' Elabo-
ration des produits de reserve des champignons.
C. rend. Acad. Sci., t. 157, pp. 6.3-65 (p. .56).

1913/. Nouvelles recherches cytologiques sur la

formation des pigments anthocj'aniques. C.
rend. Acad. Sci., t.' 157, pp. 1000-1002 (pp. 57,
101, 102).

1913^. Nouvelles observations sur le chon-
driome' des champignons. C. rend. Acad. Sci.,
t. 1.56, pp. 1781-1784 (p. ,56).

1913^. Sur la formation de I'anthocyane au

sein des mitochondries. C. rend. Acad. Sci.,
t. 1.56, pp. 1924-1926 (p. 102).

1913i. Sur I'etude vitale du chondriome de 1'

6piderme des pEtales d'Iris germanica et de son
Evolution en leuco- et chromoplastes. C. rend.
Soc. Biol., t. 74, pp. 1280-1283 (p. 102).

1914a. Nouvelles remarques sur les plastes

vEgEtaux. Anat. Anz., Bd. 46, pp. .566-574

(pp. .53, .57).
19146. fitat actuel de la question de revolution

et du role physiologique des mitochondries.

Rev. gen. de Botanique, t. 26, pp. 129-149;

182-208 (pp. 52, 53, .56, 57).
1915. Nouvelles observations vitales sur le

chondriome des cellules Epidermiques de la fleiu-

d'Iris germanica. C. rend. Soc. Biol., t. 78,

pp. 241-245 (p. 57).
1917. Recherches sur I'origine des chromoplastes

et le mode de formation des pigments du groupe

des xanthophylles et carotines. C. rend. Acad.

Sci., t. 164, pp. 232-2.35 (pp. 53, 57).
GuREwiTSCH, A., 1913. Vorlesungen liber allgemeine

Histologie. Jena, Gustav Fischer, 345 pp. (pp.

107, 110).
Haludurton, W. D., 1907. Die Biochemie der peri-

pheren Ner\'en. Ergebn. d. Phys., Bd. 4, pp.

1-88 (p. 1.38).
Hegner, R. W., 1914. Studies on germ cells. I. The

history of the germ cells in insects with special

reference to the Keimbahn-determinants in

animals. Jour. Morph., vol. 25, pp. 375-510

(pp. 55, 102).
Heidenhain, R., 1874. MikroskopLsche Beitrage zur

Anatomic und Physiologie der Nieren. Arch. f.

mikr. Anat., Bd. 10, pp. 1-.50 (pp. 44, 51).
Heidenhain, M., 1900. Die Centralkapseln und

Pseudochromosomen in den Samenzellen von

Proteus, sowie tiber ihr Verhaltnis zu den Idio-

zomen. Chondromiten und Archoplasmaschleifen

usw. Anat. Anz., Bd. 18, pp. 513-550 (pp. 44,46).
1911. Plasma und Zelle. Jena, Gustav Fischer,

1110 p. (p. 107).
Held, H., 1895. Beitrage zur Struktur der Nervenzellen

und ihrer Fortsatze. Er.ste Abhandlung. Arch.

f. Anat. u. Phys., Anat. Abt., pp. 396-416

(p. 46).
1912. Ueber den Vorgang der Befruchtung bei

Ascaris megalocephala. Verh. d. anat. Ges.,

pp. 242-248 (p. 95).
Henneguy, 1904. Les Insectes, Paris (quoted from

Duesberg, 1912, p. 576) (p. 55).

154

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

Henschen, F., 1904. Zur Struktur der Eizello gewisser
Crustaceen und (iastropoden. Anat. Anz., Bd.

24, pp. 15-29 (p. 54).

Herm.\.nn, F., 1891. Beitrag zur Lehre von der Entste-

hung der karyokinetischen .Spindel. Arch. f.

mikr. Anat., Bd. 37, pp. 569-586 (p. 44).
Hertwig, 1893. Die Zelle und die Gewebe. Jena,

Gustav Fischer, 296 pp. (p. 46).
HiRSCH, C, 1910. Experimentell-anatomische Unter-

suchungen an der Xierenzelle. .\nat. Hefte,

1. Abt., Bd. 41, pp. 131-172 (p. 137).
HiRSCHLER, J., 1913. Ueber die Pla.smastrukturen

(Mitochondrien, Oolgischer Apparat u. a.) in den

Geschlechtszellen der Ascariden. .\rch. f. Zell-

forsch, Bd. 9, pp. 351-398 (p. .54).
• 1914. Ueber Phi-smast rukt uren (Golgischer Ap-
parat, Mitochondrien u. a.) in den Tunicaten-,

Spongien- und Protozoenzellen. .Vnat. An?.., Bd.

47, pp. 289-311 (p. 53).
HjELT, K. .J., 1912. I^eber <lie Mitochondria in den

Epithelzellen der gewundenen Xierenkanidchen

bei der Einwirkung einiger Diuretika (Kaffein

und Theocin). Virchow'.s Arch., Bd. 207, pp.

297-313 (p. 137).
d'Holl.\nder, F., li)02. Le noyau viteUin de Balbiani

et le.s pseudochroinosomes chcz les oiseaux.

Verb. d. anat. Ges., pp. lOS-171 (p. 56).
Holmgren, E.. 1899. Zur Kenntnis der .Spinalganghon-

zellen von Lophius piscatorius. Lin. Anat.

Hefte, 1. Abt., Bd. 12, pp. 7.5-151 (p. 46).
1903. I'eber die "Saftkanalchen'' der Leber-

zellen und der Epithelzellen der Nebenniere.

Anat. Anz., Bd. 22, pj). 9-14 (p. 46).
1909. .Studien iiber die stofflichen Veriinderungen

der quergestreiften Mu.skelfasern. Skandin, Arch.

f. Phys., Bd. 21, pp. 287-314 (pp. .55, 130, 132).
HoLM(iRE.\, X., 1902. I'eber den Bau der Hoden und

die .Spermatogenese von Silpha carinata. Anat.

Anz., Bd. 22, (p. .55).
H0.VIANS, .J., 1915. \ study of experimental diabetes in

the canine and its relation to human tliabetes.

.Jour. Med. Res., vol. 33. pp. 1-51 (p. 132, 135).
HovEN", H., I91O0. .Sur rhistogene.se du sy.steme

nerveuse peripherique, etc. .\rch. de Biol., t.

25, pp. 427-491 (pp. 101, 107. 110).

19106. Contribution a I'etude du fonctionne-

ment des cellules glamlulaires. .\nat. Anz., Bd.

37, pp. 343-351 (pp. 102, 120).
— ;- 1911. Du role <lu chontiriome dans I'elaboration

des produits de secretion de la glande manunaire.

Anat. Anz., Bd. 39, pp. 321-326 (pp. 51, 102, 119).
HwoRosTUCHiN, W., 1911. Zur Frage Uber den Bau des

Plexus chorioideus. Arch. f. mikr. Anat., Bd.

77, pp. 232-244 (pp. .52, 117).
Israel, O., 1891. Die anami.sche Xekrose der Xier-

enepithelicn. Virchow's Archiv, Bd. 123, pp.

310-334 (pp. 136, 147).
Jenkinso.v, J. W., 1914. On the relation between the

structure and the development of the centrifuged

egg. Quart. .Jour. Micr. Sci., vol. 60, pp. 61-

1.58 (p. 96).
.Jordan, H. E., 1911. The spermatogenesis of the

opossum (Didelphys virginiana), with special

reference to the accessory chromosome and the

chondriosomes. Arch. f. Zellforsch., Bd. 4,

pp. 41-86 (p. .56).
, and Ferouso.v. 1916. A text-book of histology.

New York, D. Appleton & Co., 799 pp. (pp. 46,

107).

JosT, L., 1907. Lectures in plant physiology, translated.

Clarendon Press, (p. 14.5).
DE Kervily, M., 1916. Le chondriome des cellules de

Langhans du placenta humain. C. rend. Soc.

Biol., t. 79, pp. 589-590 (p. 51).
Key, J. A., 1916. On the relation of mitochondria to

zymogen granules. Anat. Rec, vol. 10, pp.

21.5-216 (p. 121).
IviNGSBURY, B. F., 1911. The histological demonstration

of lipoids. Anat. Rec, vol. 5, pp. 313-318 (p. 63).
1912. Cytoplasmic fixation. Anat. Rec, vol. 6,

pp. 39-49" (pp. 75, 1.33).
Klein, .S., 1910. Ueber die sog. Schriddeschen Granula.

Folia ha;m., Bd. 9, 1. 11, pp. 406-410 (pp. ,52, 135).
KoLLMANN, M., and L. Papin, 1914. Etude sur la

k^ratinisation, I'epithflium cornd de I'cesophage

de quelques mammif^res. Arch. d'Anat. micr.,

t. 16, pp. 193-260 (p. 51).
KoLTzoFF, X. K., 1906. Studien iiber die Gestalt der

Zelle. I. Untersuchungen iiber die Spermien

der Decapoden als Einleitung in das Problem

der Zellengestalt . Arch. f. mikr. Anat., Bd. 67,

pp. 364-571 (pp. 46, 48, 54, 63).
JvoPSCH, F., 1902. Die Darstellung des Binnennetzes

in Spinal-Ganglienzellen und anderen Ivorperzel-

len mittelst Osmiumsiiure. Sitzungsb. d. k. preuss.

Akad. d. Wiss., Bd. 40, pp. 92!»-935 (p. 44).
IvoROTNEFF, A., 1909. Mitochondrien, Chondrio-

miten und Faserepithel der Tricladen. Arch. f.

mikr. Anat., Bd. 74, pp. 1000-1016 (pp. 54, 86).
KosTER, H., 1910. Morphologic und Genese der

Spermatozoen von Gammarus pulex. Zool.

Anz., Bd. 35, pp. 490-496 (p. .54).
Krogh, A., 1916. The respiratory exchange of ani-
mals and man. Monographs on Biochemistry,

London, 173 pp. (p. 146).
KuLL, H., 1913. Eine Modification der Altmann-

schen Methode zum Farben der Chondriosomen.

Anat. Anz., Bd. 45, pp. 1.53-157 (p. 60).
IvuscHAKEwiTSCH, S., 1911. Ueber die Entwicklung der

Spermien bei Conus mediterraneus Brug. und

Vermetus gigas, Biv. Biol. Centralbl., Bd. 31,

pp. 530-.537 (p. 55).
Labue, a., 1904. Jja maturation des spermatides et la

constitution des spermatozoides chez les Crust-

ac<>s decapodes. Arch. Zool. Exp., vol. 11,

X'otes et Revues, pp. i-xiv (p. 54).
Laguesse, E., 1900. Sur les paranuclei et le mc^canisnie

probable de I'elaboration dans la cellule pan-

creatique de la salamandre. 13th Cong. Internat.

de Med., Sect. Hist, et Embr., Paris, pp. 1-9 (p. 46).
1911. Ergastoplasme et chondriome dans les

cellules seeretantes sereuses. Bibl. Anat., t. 21,

pp. 273-286 (pp. 45, 121).
1912a. Methode de coloration vitale des chon-
driosomes par le \ert Janus. C. rend. Soc. Biol.,

t. 73, pp. 1.50-153.
19126. Sur I'apparition de la substance amorphe

et des premieres fibrilles dans les tendons. C.

rend. Assoc. Anat., pp. 110-112 (p. 135).
Lams, H., 1908. Les divisions des spermatocytes chez la

foui-mi (Camponotus herculeanus) . Arch. f.

ZeUforsch., Bd. 1, pp. .528-.537 (p. 55).
1910. Recherches sur I'ceuf d'Arion empiri-

corum (Fer). Acad. roy. de Belg. Classe d.

sci. Mto., 144 pp. (p. 55).
Leboucq, G., 1909. Contribution a I'etude de I'histo-

genese de la retine chez les Mammiferes. Arch.

d'Anat. micr., 1. 10, pp. 5.55-605 (pp. 52, 124, 125).

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

155

Leplat, G., 1910. La sperniiogenese chez Ic chat.
Arch, de Biol., t. 25, pp. 401-426.

1912. Recherches sur le d^veloppement et la

structure de la membrane vasculaije de I'a'il des
oiseaux. Aich. de Biol., t. 27, pp. 403-524(p. 107) .

1913. Les plastcsomes des cellules visuelles et

leur role dans le dilTerenciation des cones et des
batonncts. Anat. Anz., Bd. 45, pp. 215-221
(pp. 102, 125).

Le TorzE, M. H., 1912. Contribution a I'etude histolo-
(jique des Fucacees. Rev. gen. de Botanique,
t. 24, pp. 33 (pp. .")3, 56, 102).

Levi, G., 1896. Contributo alia fisiologia della cellula
nervosa. Riv. di Path. \erv. e Ment., vol. 1,
pp. 169-180 (p. 101).

1911. Sulla presunta partecipazione dei con-

driosomi alia differenziazione cellulare. Arch,
ital. di Amit. e di Embriol., vol. 10, pp. 168-195
(pp. 107, 110).

1912. I condriosomi nelle cellule secernenti.

Anat. Anz., Bd. 42, pp. 576-592 (p. 56).

1913. Note citologiche suUe cellule somatiche

deir ovaio dei Mammiferi. Arch. f. Zellforsch.,
Bd. 11, pp. 51.5- .5.56 (pp. 51, 56).

1914. Uiteriori studi suUo sviluppo delle cellule

visive negli Anfibi. ,\nat. Anz., Bd. 47, pp.
192-199 (p. 126).

1915. II comportamento dei condiiosomi dur-
ante i piu precoci periodi dello svilujipo dei
Mammiferi. Arch. f. Zellforsch., Bd. 13, pp.
471-524 (p. 95).

1916a. La costituzione ilel protoplasma studiata

su cellule viventi coltivate in vitro. Arch. d.
Fisiologia, vol. 16, pp. 101-112 (pp. 52, .58).

1916b. II ritmo e le modalita della mitosi nelle

cellule viventi coltivate "in vitro." Arch. ital.
d. Anat. e d Embriol., vol. 15, pp. 243-264.

1916c. Differenziazione "in vitro" di fibre da

cellule fiiesenchimali e loro accrescimento per
movimento ameboide. Monit. zool. ital., pp.
77-84 (p. 108).

Lewis, M. R., 1917. Development of connective-tissue
fibers in ti.ssue cultures of chick embryos. Car-
negie Inst, of \\'a.sh. Contributions to Em-
bryology, No. 17, pp. 45-60 (p. 105).

Lewis, M. R., and W. H., 1914. Mitochondria in tissue
cultures. Science, n. s.. vol. .39, pp. 330-333 (pp.
43, 53).

1915. Mitochondria (and other cytoplasmic

structures) in tissue cultures. Amer. Jour. Anat.,
vol. 17, pp. 339-401 (pp. 45, 47, 58, 64, 68, 70,
76, 79, 134).

Lewis, M. R., and \V. R. B. Robertson, 1916. The mi-
tochondria and other structures observed by the
tissue culture method in the male germ-cells of
Chorthippus curtipennis, Scudd. Biol. Bull.,
vol. 30, pp. 99-125 (p. 55).

Lewitskt, G., 1910. Ueber die Chondriosomen in
pflanzlichen Zellen. Ber. d. deutsch. bot. Ges.,
Bd. 28, pp. .538-546 (pp. .53, .57, 102, 107, 114).

1911. Die Chloroplastenanlagen in lebenden und

fixierten Zellen von Elodea canadensis. Ber. d.
deutsch. bot. Ges., Bd. 29, pp. 697-703 (p. 57).

^— — 1913. Die Chondriosomen als Secretbildner bei
den Pilzen. Ber. d. deutsch. bot. Ges., Bd. 31,
pp. 517-.528 (pp. 52, .56).

Lilue, F. R., 1908. A contribution towards an experi-
mental analysis of the karyokinetic figure.
Science, n. s., vol. 27, pp. 907-908 (p. 74).

LiLLiE, F. R., 1912. Studies of fertiUzation in Nereis, III
and IV. ,Iour. Exper. Zool., vol. 12, pp. 413-
477 (p. 97).

LiLLiE, R. S., 1913. The formation of indophenol at
the nuclear and plasma membranes of frog's
blood corpuscles antl its acceleration by in-
duction shocks. .lour. Biol. Chem., vol. 15, pp.
237-248 (p. 146).

LoBENHOFFER, W., 1906. Ueber die Ergebnisse der
Altmann-Schriddeschen Farbenmethode beim
Zentralnervensystem. Arch. f. mikr. .\nat., Bd.
68, pp. 491-.500 (p. 101).

LoEB, J., 1916. The .sex of parthenogenetic frogs. Proc.
Nat. Acad. Sci., vol. 2, pp. 313-317 (p. 95).

LmvscHiN, A. M., 1913. Myelinformen und Chondrio-
somen. Ber. d. deutsch. bot. Ges., Bd. 31, pp.
203-209 (p. 83).

1914. Vergleichende experimental-cytologische

Untersuchungen, etc. Ber. d. deutsch. bot.
Ges., Bd. 32, pp. 266-270 (pp. .56, 57, 83).

LoYEZ, M., 1909. Les premiers stades de la vitello-

genese chez quelques Tuniciers. C. rend. Assoc.

Anat., pp. 189-195 (pp. 55, 76, 102, 123).
LuB.\RSCH, O., 1897. Fettdegeneration und Fettin-

filtration. Ergebn. d. allg. Path. u. path. Anat.,

Bd. 3, pp. 631-651 (pp. 1.35, 137).
Li"N.\, E., 1913a. Lo sviluppo dei plastosomi negli

anfibi. Anat. Anz., Bd. 45, pp. 19-21.

19136. Nuove ricerehe sulla biologia del con-

drioma (Condriosomi e pigmento retinico).
Anat. Anz., Bd. 43, pp. .56-.5S (p. 102).

1913r. .Sulla importanza dei condriosomi nella

genesi delle miofibrille. .\rch. f. Zellforsch.,
Bd. 9, pp. 4.58^78 (|). 107).

1913(/. Ricerehe sulla biologia dei condriosomi.

Condriosomi e pigmento retinico. Arch. f.

Zellforsch., Bd. 10, pp. 34.3-358.
191.3c. I condriosomi nelle cellule nervose. Nota

preventiva. Anat. .\nz., Bd. 44, pp. 142-144.
1913/. Sui fenomeni di plastorexi e di plastolisi

riscontrabili iiel processo di involuzione del pro-

nefro negli .\nfibi. Nota preventiva. Monit.

zool. ital., pp. 131-1.33 (p. 51).
1913(7. SuUe modificazioni dei plastosomi delle

cellule nervose nel trapianto ed in seguito al taglio

dei nervi. Anat. Anz., Bd. 44, pp. 413-415

(p. 137).
Macallum, .\. B., 1908. Die Methoden und Ergebnisse

der Mikrocheniie in der biologischen Forschung.

Ergebn. d. Phys., Bd. 7, pp. 5 2-6.52 (p. 14.5J.
Maggi, L., 1878. I plastiduli nei ciliati ed i plastiduH li-

beramente viventi. .\tti. Soc. Sci. Nat., Milano,

t. 21, pp. 326-330 (p. 42).
Magoiore, L., 1912. L'apparato mitochondriale nel

cristallino. Ricerehe Lab. di Anat. norm. Univ.

Roma, vol. 16, pp. 117-119 (p. 52).
Manca, p., 1913. Sulla presenza di condrioconti nelle

cellule degli abbozzi dentarii. Nota preliminare.

Monit. zool. ital., pp. 121-127 (p. 51).
Marcora, F., 1911. Ueber die Histogenese des Zentral-

nervensystems mit besonderer Riicksicht auf die

innere Struktur der Nervenelemente. FoUa

neuro-biol., Bd. 5, pp. 928-960 (p. 110).
Mathews, A. P., 1905. A theory of the nature of

protoplasmic respiration and growth. Biol.

Bull., vol. 8, pp. 331-346 (p. 147).
1915. Physiological chemistry. 1040 pp. New

York, William Wood & Co. (pp. 71, 129, 134,

139, 144).

156

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

Mawas, J., 1909. La structui-e de la ratine ciliaire et la

s6cr6tion de I'humeui' aqueuse. C. rend. Assoc.

Anat., pp. 282-285 (p. 101).
— — — 1910a. Sur la structure des cellules nerveuses

ganglionnaires de la moelle amy^linique des

Cyclostomes. C. rend. Acad, des Sci., t. 150,

pp. 126-127 (p. 55).
1910b. Etude cytologique et physiologique sur

la ratine ciliaire des Mammif^res. Arch. d'Anat.

micr., t. 12, pp. 103-176 (pp. 124, 125).
1910c. Notes cytologiques sur les cellules

visuelles de I'homme et de quelques mammifSres.

C. rend. Assoc. Anat., pp. 113-118 (p. 125).
Maximow, a., 1913. Ueber Chondriosomen in lebenden

Pflanzenzellen. Anat. Anz., Bd. 43, pp. 241-

249 (pp. 53, 57).
— 1916a. Sur les m^thodes de fixation et de colora-
tion des chondriosomes. C. rend. Soc. Biol., t.

79, pp. 462^63 (p. 63).
1916b. Sur la structure des chondriosomes. C.

rend. Soc. Biol., t. 79, pp. 46.5-466 (p. 52).
Mater, A. and Rathery, 1907. Etudes sur le corps

fungiforme du poulpe. Jour, de I'Anat. et de

Phys., t. 43, pp. 25-47 (p. 55).
1909. Histophysiologie du rein. Jour, de 1'

Anat. et de Phys., t. 45, pp. 321-338 (p. 54).
Mayer, A., Rathery and Schaeffer, 1914. Les granula-
tions ou mitochondries de la cellule h^patique.

Deuxienie Partie. Jour, de Phys. et de Path.

gto., t. 16, pp. 607-622 (pp. 84, 134, 135).
Maziarski, S., 1903. Recherches cytologiques sur les

organes segmentaries des vers de terre. Arch.

polon. sci. biol. et m^d., t. 2, pp. 1-81 (p. 54).
McCann, Gertrude Fisher, 1918. A study of mito-
chondria in experimental poUomyeUtis. Jour.

Exp. Med., vol. 27, pp. 31-36 (p. 137).
McClekdon, J. F., 1910. Further studies on the game-

togenesis of Pandarus sinuatus. Arch. f. Zell-

forsch., Bd. 5, pp. 229-234 (p. 54).
Metzner, R., 1890. Ueber die Beziehungen der Granula

zum Fettansatz. Arch. f. Anat. u. Phys., Anat.

Abt., pp. 82-96 (pp. 52, 124j.
Mbves, F., 1897. Zellteilung. Ergebn. d. Anat. u. Entw.

(Anat. Hefte, 2. Abt.), Bd. 6, pp. 284-397 (p. 45).
1900. Ueber den von la Valette St. George ent-

deckten Nebenkern (Mitochondrienkorper) der

SamenzeUen. Arch. f. niikr. Anat., Bd. .56, pp.

552-606 (pp. 55, 102).
1904. Ueber das Vorkommen von Mitochondrien

bezw. Chondriomiten in Pflanzenzellen. Ber. d.

deutsch. bot. Ges., Bd 22, pp. 284-286 (pp. 53, 57).
1905. Ueber die Wirkung gefarbter Jodsaure

auf die roten Blutkorperchen der Amphibien.

Anat. Anz., Bd. 26, pp. 97-103 (pp.63, 102, 126).
1907a. Ueber Mitochondrien bezw. Chondrio-

konten in den Zellen junger Embryonen. Anat.

Anz., Bd. 31, pp. 399-407 (pp. 44, 45, 47, .52, 56,

102, 105, 110, 113).
19076. Die Chondriokonten in ihrem Ver-

haltnis zur Filarmasse Flemmings. Anat. Anz.,

Bd. 31, pp. 561-569 (p. 51).
1907c. Die Spermatocytenteilungen bei der

Honigbiene (Apis mellifica L.), nebst Bemer-

kungen iiber Chromatinreduktion. Arch. f.

mikr. Anat., Bd. 70, pp. 414-491 (p. 55).
■ — ■ 1908. Die Chondriosomen als Trager erblicher

Anlagen. Cytologische Studien am Hiihner-

embryo. Arch. f. mikr. Anat., Bd. 72, pp.

816-867 (pp. 43, 48, 96, 101, 110, 129).

Meves, F., 1910a. Ueber Strukturen in den Zellen des
embryonalen Stutzgewebes sowie iiber die Ent-
stehung der Bindegewebsfibrillen, insbesonders
derjenigen der Sehne. Arch. f. mikr. Anat., Bd.
75, pp. 149-208 (pp. 46, 48, 102, 103).

1910b. Zur Einigung zwischen Faden- und

Granulalehre des Protoplasma. Beobaehtungen
an weissen Blutzellen. Arch. f. mikr. Anat.,
Bd. 75, pp. 642-658 (pp. 101, 102, 110, 126).

1911a. Gesammelte Studien an den roten

Blutkorperchen der Amphibien. Aich. f. mikr.
Anat., Bd. 28, pp. 465-540 (pp. 52, 105, 115).

19116. Ueber die Beteiligung der Plastochon-

drien an der Befruchtung des Eies von Ascaris
megalocephala. Arch. f. mikr. Anat., Bd. 76,
pp. 683-714 (p. 95).

1912. Verfolgung des sogenannten Mittelsttickes

des Echinidienspermiens im befruchteten Ei
bis zum Ende der ersten Furchungteilimg. Arch,
f. mikr. Anat., Bd. 80, pp. 81-123 (p. 95).

1913. Ueber das Verhalten des plastomatischen

Bestandteiles bei der Befruchtung des Eies von
Phallusia mammillata. Arch. f. mikr. Anat.,
Bd. 82, pp. 215-260 (pp. 55, 95).

1915a. Was sind die Plastosomen? II. Bemer-

kungen zu dem Vortrag von C. Benda. Die
Bedeutung der Zelleibstruktur fiir die Patho-
logie. Arch. f. mikr. Anat., Bd. 87, pp. 287-308.

19156. Ueber Mitwirkung der Plastosomen bei

der Befruchtung des Eies von Filaria papillosa.
Arch. f. mikr. Anat., Bd. 87, 2. Abt., pp. 12-46
(pp. .54, 95).

1915c. Ueber den Befruchtungsvorgang bei der

Miesmuschel (Mytilus edulis L). Arch. f.
mikr. Anat., Bd. 87, 2. Abt., pp 47-62 (p. 95).

Mbves, F. and J. Duesberq, 1908. Die Spermato-
cytenteilungen bei der Hornisse (Vespa crabro
L.). Arch. f. mikr. Anat., Bd. 71, pp. 571-587
(pp. 5.5, 61, 97).

and R. Tsukaouchi, 1914. Ueber das Vorkom-
men von Plastosomen im Epithel von Trachea
und Lunge. Anat. Anz., Bd. 46, pp. 289-292.

MicHAELis, L., 1899. Die vitale Farbung, eine Dar-
stellungsmethode der ZeUgranula. Arch. f. mikr.
Anat., Bd. ,55, pp. 558-575 (pp. 86, 87).

1902. Einfiihrung in die Farbstoffchemie fiir

Histologen. Berlin, S. Karger, 156 pp. (pp.
87, 89).

MiRA, M. Ferreiba, 1912. Sur I'^tat des capsules
suir^nales chez les animaux ovariectomis6s.
Bull. Soc. Portugaise Sci. Nat., t. 6, pp. 34-46
(p. 135).

MiRANDE, M., 1916. Observation sur le vivant de la
formation cytologique de I'anthocyanine. C.
rend. Acad. Sci., t. 163, pp. 368-371 (p. 56).

MiBONBSco, T., 1912. Le chondriome du r^seau de
Purldnje du coeur. C. rend. Soc. Biol., t. 72,
pp. 30-31 (p. .52).

MiSLAWSKY, A. N., 1911a. Demonstration von Chon-
driosomen in Pankreaszellen. Verh. d. anat.
Ges., pp. 221-222 (p. 121).

19116. Beitriige zur Morphologie der Driisen-

zelle. Ueber des Chondriom der Pankreaezelle
einiger Nager. Anat. Anz., Bd. 39, pp. 497-505.

MoDRAKOWsKi, 1903. Weitcre Beitriige zur Nieren-
funktion. Pfluger's Arch., Bd. 98, pp. 217-
232 (p. 122).

TFre MITOCHONDRIAL CONSTlTUtlNTS OF PROTOPLASM.

157

Montgomery, T. H., 1911. The spermatogenesis of an

hemipteron, Euschistus. Jour. Morph., vol. 22,

pp. 731-821 (p. .55).
Monti, R., 1915. I condriosomi e gli apparati di Golgi

nelle cellule nervose. Arch. ital. di Anat. e di

Embriol., vol. 14, pp. 1^5 (p. 63).
MoREAU, F., 1914a. Sur la formation decorpuscules mdta-

chroraatiques dans les mitochondries granuleuses.

C. rend. Soc. Biol., t. 77, pp. 347-349 (p. .56).
— • 19146. Sur le chondriome d'une ustilagin^e,

Entyloma ranunculi (Bonorden) Schra'ter. C.

rend. Soc. Biol., t. 77, pp. 53S-539 (pp. 56, 67).
1914c. Sur I'origine de I'anthocyane dans les

divers organes des v^gdtaux. C. rend. Soc. Biol.,

t. 77, pp. 502-503 (p. 57).
1914d. Le chondriome et la division des mito-
chondries chez les Vaucheria. Bull, de la Soc.

bot. de France, t. 61.
1915a. Sur la formation de cristalloides de

mucorine au sein des mitochondries. C. rend.

Soc. Biol., t. 78, pp. 171-172 (pp. 52, 102).
19156. La division des mitochondries et ses

rapports avec les ph^nomenes de s6cr6tion. C.

rend. Soc. Biol., t. 78, pp. 143-144 (p. 56).
MoREAU, Mme. F., 1914. Les mitochondries chez les

urddin^es. C. rend. Soc. Biol., t. 76, pp. 421^22

(pp. 52, .56).
MoREAU, F., and Mme. Moreau, 1916. Sur le chon-
driome d'une algue verte, Coccomyxa solorinae

Chod. C. rend. Soc. Biol, t. 79, pp. 211-212

(p. 56).
MoREAUX, R., 1909. Note sur la spermiogentee du

Macaque. C. rend. Soc. Biol., t. 67, pp. 369-

371 (p. 56).
MuLON, P., 1910a. La mdthode des mitochondries (de

Benda) appUqu6e k la corticale surr^nale du

cobaye. C. rend. Soc. Biol., t. 68, pp. 103-105

(p. 52).
19106. Sur les mitochondries de la surr^nale.

C. rend. Soc. Biol., t. 68, pp. 872-873 (p. 102).
1912. Note sur la capsule surrenale du Mouton.

Bibl. Anat., t. 22, pp. 30-36 (p. 69).
1913. Du role des hpoides dans la pigmentogenfese.

C. rend. Soc. Biol., t. 74, pp. 1023-1027 (p. 102).
MuNSON, J. P., 1898. The ovarian egg of Limulus.

Jour. Morph., vol. 15, pp. 111-211 (p. 55).
— ■ 1904. Researches on the oogenesis of the

tortoise, Clemnys marmorata. Amer. Jour.

Anat., vol. 3, pp. 311-348 (p. 56).
Naeqeli, O., 1912. Blutkrankheiten und Blutdiag-

nostik. Leipzig, Veit & Co., 719 pp. (p. 52).
Naqeotte, J., 1909. Mitochondries du tissu nerveux.

C. rend. Soc. Biol, t. 66, pp, 825-828 (pp. 52, 101).
Nblis, C, 1899. Un nouveau detail de structure du

protoplasme des cellules nerveuses (6tat spir6-

mateux du protoplasme). BuU. de la classe des

Sci. de I'Acad. Belg., pp. 102-125 (p. 46).
NfiMEC, B., 1910. Das Problem der Befruchtungs-

vorgiinge und andere zytologische Fragen.

Berhn, 532 pp. (pp. 53, 135).
Nicholson, N. C, 1916. Morphological and micro-
chemical variations in the mitochondria in the

cells of the central nervous system. Amer. Jour.

Anat., vol. 19, pp. 329-349 (pp. 67, 85, 131, 148).
Nicolas, J., Cl. Regaud, and M. Favre, 1912a. Sur

lea mitochondries des glandes sdbac^es de I'homme

et sur la signification g^n^rale de ces organites du

protoplasma. C. rend. Assoc. Anat., pp. 201-

205 (pp. 51, 73, 102, 119).

Nicolas, J., Cl. Reqaud, and M. Favre, 19126. Sur

la fine structure des glandes sudoripares de

I'homme particuherement en ce qui concerne les

mitochondries et les phdnomdnes de s^cr^tion. C.

rend. Assoc. Anat., pp. 191-200 (pp. 51, 102, 119).
NicoLOSi-RoNCATi, F., 1910. Formazioni mitochon-
drial! negli element! sessuale maschili dell'

Helleborus fcetidus. Rend. Acad. Sci. Napoli,

Ser. 3, t. 16, pp. 109-119.
1912a. Formazioni endocellulari nelle Rodo-

ficee. Bull. Soc. Bot. Ital, pp. 59-62 (p. .56).
19126. Genes! dei cromatofori nelle Fucoidee.

Bull. Soc. Bot. Ital, pp. 144-149 (pp. 53, 56, 102).
1912f. Contribute alia conoscenza cito-fisio-

logica delle glandule vegetaU. Bull. Soc. Bot.

Ital, pp. 186-193 (p. 57).
NoRDENSKiOLD, E., 1909. ZuT Spermatogenese von

Ixodes reduvius. Zool Anz., Bd. 34, pp. 511-

516 (p. 55).
Oettinger, R., 1909. Zur Kenntnis der Spermato-
genese be! den Myriopoden: Samenreifung und

Samenbildung be! Pachyiulus varius, Fabre.

Arch. f. Zellforsch., Bd. 3, pp. 563-626 (p. 55).
Oliver, J., 1916. A further study of the regenerated

epithelium in chronic uranium nephritis. An

anatomical investigation of its function. Jour.

Exper. Med., vol. 23, pp. 301-322 (pp. 102, 122,

135, 139).
OPHiJLS, W., 1907. Altpiann's granules in granular and

fatty degeneration in the kidney. Proc. Soc.

Exper. Biol, and Med., vol. 4, pp. 135-137 (p. 135).
Orman, E., 1912. Recherches sur les diff^renciations

cytoplasmiques (ergastoplasme et chondriosomes)

dans les v^g^taux. I. Le sac embryonnaire des

Liliac^es. Cellule, t. 28, pp. 363-444 (p. 57).
Pantel, J., and R. de Sin^ty, 1906. Les cellules de la

lign^e male chez le Notonecta glauca. Cellule,

t. 23, pp. 90-303 (p. 55).
Parker, T. J., and W. A. Haswell, 1897. Text-book

of Zoology, vol 1. London, Macmillan & Co.,

779 pp. (p. 145).
Payne, F., 1916. A study of the germ cells of Gryllo-

talpa borealis and Gryllotalpa vulgaris. Jour.

Morph., vol, 28, pp. 287-327 (p. 55).
Pensa, a., 1910. Alcune formazioni endocellulari dei

vegetal!. Anat. Anz., Bd. 37, pp. 325-333

(pp. .57, 63, 114).
1911. Ancora di alcune formazioni endocelluiari

dei vegetaU. Anat. Anz., Bd. 39, pp. 520-532

(pp. .53, 56, 57).
1912. Osservazioni di morfologia e biologia

cellulare ne! vegetal! (mitocondri, cloroplasti).

Arch. f. Zellforsch., Bd. 8, pp. 612-662 (pp.

.56, 57).
1913a. Condriosomi e pigment© antocianico

nelle cellule vegetaU. Anat. Anz., Bd. 45, pp.

81-90 (p. 57).
19136. La struttura della cellula cartilaginea.

Arch. f. Zellforsch., Bd. 11, pp. 557-.582 (p. 52).
Perroncito, a., 1910. Contribution a I'^tude de la

biologic cellulaire. Arch. ital. de Biol, vol. .54,

pp. 307-345 (pp. 44, 45, 102).
PizziNi, B., 1908. Ueber die Sekretionserscheinungen

in der Nierenzelle bei der Diurese. Internat.

Monatsschr. f. Anat.u. Phys., Bd. 25, pp. 108-148

(p. 136).
Platner, G., 1889. Ueber die Bedeutung der Rich-

tungskorperchen. Biol Centralbl, Bd. 8, pp.

718-720 (p. 46).

158

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

PoLicARD, A., 1905. Sur les formations mitochonHriales

du rein des vert^br^s. C. rend. Soc. Biol., t. 57,

pp. 380-382 (pp. 56, 1*22).
1910. Contribution a I'^tude du mfcanisme,

etc. Arch. d'Anat. niicr., t. 12, pp. 177-288

(pp. 122, 132, 13.i).

1912a. Role du chondriome dans la formation

des cristaux intracellulaire.s de la cellule hdpa-
tique. C. rend. Soc. Biol., t. 72, pp. 91-93
(pp. 101, 102).

1912b. Sur le role du chondriome dans le forma-
tion des cristaux intraprotoplasmique d'h^mo-
globine dans la cellule hepatique. Bihl. Anat.,
t. 22, pp. 226-230 (pp. 102, 116).

. 1912c. La cytogerifese du tube urinaire chez

I'homme. Arch. d'Anat. niicr., t. 14, pp. 429-
468 (pp. 44, 81, 86, 122).

1912d. Sensibility des chondriosomes aux 616va-

tions de temperature. C. rend. Soc. Biol., t. 72,
pp. 228-230 (pp. 83, 139).

1912e. Sur quelques points de la cytologic des

plexus choroides. C. rend. Soc. Biol., t. 73, pp.
430-4.32 (p. 117).

1912/. Recherches histophysiologiques sur les

premiers stades de la secretion urinaire. Arch.
d'Anat. micr., t. 14, pp. 1-40 (p. 81).

1914. Recherches histophysiologiques sur les

voies biliaires intrahepatiques. Jour, de Phys.

et de Path, g^n., t. 16, pp. 623-637 (p. 51).
1915. Chondriocontes et fibrilles plasmatiques

dans les cellules du tube urinaire des Batraciens.

Anat. Anz., Bd. 47, pp. .539-.543 (p. 51).
PoLicARD, A., and J. Mawa.s, 1906. Le canaUcule

urinaire des T^leosteens. Bibl. Anat., t. 15, pp.

215-221 (p. 56).
PoRCELLi-TiTONE, F., 1914. Der Mitochondrienapparat

der Geschwulstzellen. Beitr. z. path. Anat. u. allg.

Path., Bd. 58, pp. 237-249 (pp. 135, 140).
Prenant, a., 1899. Sur le protoplasma sup^rieur (archi-

plasme, kinoplasme, ergastoplasme). Etude

critique. Jour, de I'Anat. et de la Phys., t. 35,

pp. 408-466 (pp. 45, 46).

1910. Les mitochondries et I'ergastoplasme.

Jour, de I'Anat. et de la Phys., t. 46, pp. 217-
285 (p. 46).

1911a. Problemes eytologiques ggn^raux sou-

lev^s par I'etude des cellules musculaires. Jour,
de I'Anat. et de la Phvs., t. 47, pp. 450-524

(p. 107). ....
19116. Preparations relatives aux mitochondries.

C. rend. Assoc. Anat., pp. 333-337 (p. 51, 52).
1913. Sur I'origine mitochontkiale des gi-ains de

pigment. C. rend. Soc. Biol., t. 74, pp. 926-929

(p. 102).
Prenant, Bouin and Maillard, 1906. Traite d'Histo-

logie. T. 1 , 977 pp. Paris, Masson & Cie. (p. 52) .
Prowazek, 1902. Ein Beitrag zur Krebsspermato-

genese. Zeitschr. f. wiss. Zool., Bd. 71, pp. 445-

456 (pp. 54, .55).
Rkgaud, Cl. , 1908a. Sur les formations mitochondriales

de diverses especes cellulaires. C. rend. Assoc.

Anat., pp. 15-19 (pp. 51, 56, 81, 145).
1908b. Sur les mitochondries des cellules cihees

du tube urinaire. Ont-elles une relation avec la

fonction motrice de ces cellules? C. rend. Soc.

Biol., t. 65, pp. 206-208 (p. 130).
1908c. Variations des formations mitochon-
driales dans les tubes a cuticule stride du rein.

C. rend. Soc. Biol., t. 64, pp. 114.5-1147.

Regaud, Cl., 1908d. Sur les mitochondries de I'dpithehum
seminal. C. rend. Soc. Biol., t. 65, pp. 718-
720 (pp. 43, 83).
■ — 1909a. Attribution aux "formations mitochon-
driales" de la fonction g<^n6rale d'extraction et
de fixation (''lectives, exerc<'>e par les cellules
vivantes sur les substances dissoutes dan,s le
milieu ambiant. C. rend. Soc. Biol., t. 66, pp.
919-921 (p. 103).
1909b. Sur les mitochondries des fibres muscu-
laires du caur. C. rend. Acad. So., t. 149, pp.
426-428 (p. 52).
1909c. Participation du chondriomes h la forma-
tion des grains de segregation dans les cellules des
tubes contournes du rein (chez les ophidiens et
les amphibiens). C. rend. Soc. Biol., t. 66, pp.
1034-ia36 (p. 122).

1910. Etude sur la structure des tubes semini-

feres, etc. Arch. d'Anat. micr., t. 11, pp. 291-
433 (pp. 61, 76, 85, 95, 102).

1911. Les mitochondries. Rev. de Med., t. 31,

pp. 681-699 (pp. 129, 133).
Regaud, Cl., and M. Favre., 1909. Demonstration des
mitochondries dans les fibres musculaires striees.
C. rend. Assoc. Anat., p. 298 (p. 52).

1912. Nouvelles recherches sur les formations

mitochondriales de I'epiderme humain, a I'etat
normal et pathologique. C. rend. Soc. Biol., t.
72, pp. 328-331 (pp. 105, 135).
Regaud, Cl., and J. Mawas, 1909a. Ergastoplasme et
mitochondries dans les cellules de la glande
sous-maxillaire de I'homme. C. rend. Soc. Biol.,
t. 66, pp. 461-463 (p. 45).
19096. Sur la structure du protoplasma (ergasto-
plasme, mitochondries, grains de segregation)
dans les cellules sero-zjinogenes des acini et dans
les cellules des canaux excteteurs de quelques
glandes salivaries de Mammiferes. C. rend.
Assoc. Anat., pp. 220-237 (pp. 45, 51, 102, 118).
Regaud, Cl.. and A. Policard, 1913. Sur la significa-
tion de la retention du chrome par les tissus en
technique histologique au point de vue des Upo-
ides et des mitochondries. I. Fixation "morpho-
logique" et fixation "de substances." C. rend.
Soc. Biol., t. 74, pp. 449-451.
Renavt and Dubreuil, 1906a. Sur les cellules rhagio-
crines Ubres du liquide des diverses sereuses. C.
rend. Soc. Biol., t. 60, pp. 34-37.

19066. Les cellules connectives de la lignee rhagio-

crine. Bibl. Anat., t. 15, pp. 222-242 (pp. 45,
46, 102, 127).
Retzks. G., 1914. Was sind die Plastosomen? Arch,
f. mikr. Anat., Bd. 84, Abt. 1, pp. 175-214.
Riddle, O., 1916. Sex control and known correlations
in pigeons. Amer. Naturalist, vol. 50, pp. 385-
410 (p. 84).
Romeis, B., 1912. Beobachtungen liber Degenerat ions-
erscheinungen von Chondriosomen. Arch. f.
miki-. Anat., Bd. 80, pp. 129-170 (p. 44).

1913a. Ueber Plastosomen und andere Zell-

strukturen in den Uterus, Darm- und Muskel-
zellen von Ascaris megalocephala. Anat. Anz.,
Bd. 44, pp. 1-14 (pp. 51, 108, 138, 147).

1913b. Beobachtungen iiber die Plastosomen

von Ascaris megalocephala wiihrend der Em-
bryonalentwicklung unter besonderer Beriick-
sichtigung ihies Verhaltens in den Stamni- und
Urgeschlechtszellen. Arch. f. mikr. Anat., Bd.
81, Abt. 2, pp. 129-172 (p. 95).

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

159

RoMEis, B., 1913c. Das Verhalten der Plastosomen bei
dor Regeneration. Anat. Anz., Bd. 45, pp. 1-19
(pp. 132, 13.i).

RuBASCHKiN, W., 1910. Chondriosonien iind DiiTeren-
zierungsproze,sse bei Siiugetierenibryonen. Anat.
Hefte, 1. Abt., Bd. 41, pp. 399-431 (pp. 63, 67).

Rudolph, K., 1912. Chondriosonien und Chromato-
phoren. Beitrag zur Kritik der Chondrioso-
mentheorien. Ber. d. deutsch. bot. Ge.s., Bd.

30, pp. 60.S-629 (pp. 53, .56, 57).

Russo, A., 1910. Sui mutamenti che subiscono i
mitochondri ed i mateiiali deutoplasmiei dell'
oocite di Coniglia in diversi periodi di inanizione.
Arch. f. Zellforsch., Bd. 5, pp. 173-181 (p. 136).

■ — ■ 1912. Aumento dei granuli proplasmatici nell'

oocite delle Coniglie iniettate con lecitinia, loro
diminuzione nelle Coniglie digiunanti e loro
natura lipoide e niitochondriale. Arch. f.
ZeUforsch., Bd. 8, pp. 203-216 (pp. 84, 133).

Saguchi, S., 1913. Ueber Mitochondrien (Chondrio-
konten) und niitochondriale Strange {=sog.
Eberthsche intrazellulare Gebilde) in den Epi-
dermiszellen der Anurenlarven neb.st Bemer-
kungen iiber die Frage der Epidermis-Cutis-
grenze. Arch. f. mikr. Anat., Bd. 83, Abt. 1,
pp. 177-246 (pp. 45, 56, 102, 113).

1915. Ueber Sekretionsercheinungea an den

Epiderniiszellen von Aniphibienlarven nebst
Beitragen zur Frage nach der physiologischen
Degeneration der Zellen. Mitteilungen d. med.
Fak. d. Univ. z. Tokyo, Bd. 14, 299-415 (p. 102).

1917. Studies on cihated cells. Jour. Morph.,

vol. 29, pp. 217-279 (pp. 102, 128).

Samssonow, N., 1910. Ueber die Beziehungen der
Filarmasse Flemmings zu den Faden und Kornern
Altmann's nach Beobachtungen an Knorpel-
Bindegewebs- und Epiderniiszellen. Arch. f.
mikr. Anat., Bd. 75, pp. 635-641.

SapShin, a. a., 1913. Ein Bewiss der Individualitiit
der Plastiden. Ber. d. deutsch. bot. Ges., Bd.

31, pp. 321-324 (p. 53).

1915. Untersuchungen iiber die Individualitat

der Plastide. Arch. f. Zellforsch., Bd. 13, pp.
319-398 (pp. 52, 53, 56, 57).

ScHATER, E. A. 1912. Text-book of microscopic
anatomy. 739 pp. London, Longmans Green
& Co. (pp. 46, 55, 107).

ScHAFER, F., 1907. Spermatogenese von Dytiscus. Zool.
Xahrb., Abt. f . Anat., Bd. 23, pp. 535-586.

ScHAXEL, J., 1911. Plasmastrukturen, Chondriosomen
und Chromidien. Anat. Anz., Bd. 39, pp. 337-
353 (p. 54).

1912. Versuch einer cytologischen Analyse

der Entwicklungsvorgange. I. Die Geschlechts-
zellenbildung und die normale Entwicklung von
Aricia foetida Clap. Zool. Jahrb., Abt. f. Anat.
u. Ontog., Bd. 34, pp. 381-472 (p. .54).

ScHERRER, A., 1913. Die Chromatophoren und Chon-
driosomen von Anthoceros. Ber. d. deutsch.
bot. Ges., Bd. 31, pp. 493-500 (pp. 52, .56).

Schilling, C, 1894. Das Verhalten der Altmann'schen
Granula bei der triiben Schwellung. Virchow's
Archiv., Bd. 135, pp. 470-479 (pp. 135, 137).

ScHiROKOOOBOFF, J. J., 1913. Die Mitochondrien in den
erwachsenen Nervenzellen des Zentralnerven-
systems. Aniit. Anz., Bd. 43, pp. 522-524 (p. 60).

ScHiTZ, v., 1916. Sur la spermatogenfese chez Co-
lumbella rustica L. Arch, de Zool. exp. et g€n.,
Notes et Revue, t. 56, pp. 32-47 (p. 55).

ScHREiNER, A. and K. E., 1905. Ueber die Entwicklung

der mannUchen Geschlechtszellen von Myxine

glutinosa. Arch, de Biol., t. 21, pp. 183-314

(p. .5.5).
ScHRiDDE, H., 1905. Beitrage zur Lehre von Zellkorne-

lungen. Die Kornelung der Plasmazellen. Anat.

Hefte, 1. Abt., Bd. 28, pp. 691-768 (pp. 52, 60).
1907. Myeloblasten, Lyniphoblasten und lym-

phoblastLsche Plasmazellen. Ziegler's Beitrage z.

path. Anat., Bd. 41, pp. 22.3-238 (p. 52).
1912. Untersuchungen iiber die Bildung des Ham-

oglobins. Anat. Anz., Bd. 42, pp. 514-517 (p. 115).
ScHULTZ, T., 1914. Farbstofftabellen. Berlin, Julius

Springer (p. 92).
ScHULTZE, O., 1911a. Ueber die Anwendung der Os-

miumsaure und eine neue Osmiumhematoxylin-

methode. Zeitschr. f. wiss. Mikr., Bd. 27, pp.

465-475 (p. 63).
— 19116. Ueber die Genese der Granula in den

Driisen-Zellen. Anat. Anz., Bd. 38, pp. 257-

265 (p. 121).
Scott, K. J., 1915. The relation of mitochondria to

granules of the vital azo dyes. Science, n. s.,

vol. 41, pp. 834-835 (p. 58).
Scott, W. J. M., 1916. Experimental mitochondrial

changes in the pancreas in phosphorus poisoning.

Amer. Jour. Anat., vol. 20, pp. 237-253 (pp. 121,

132, 13.5).
Shipley, P. G., 1915. The mitochondrial substance in

the erythrocytes of the embryo pig. Folia

ha?m., Bd. 20, pp. 61-86 (pp. 56, 135).
1916. The vital staining of mitochondria in

Trypanosoma lewisi with janus green. Anat.

Rec, vol. 10, pp. 439-446 (pp. 53, 89, 130).
Sjovall, E., 1906. Ein Versuch, das Binnennetz von

Golgi-Kopsch bei der Spermato- und Ovogenese

zu homoIogLsieren. Anat. Anz., Bd. 28, pp. 561-

579 (pp. 45, 46, 63).
v. Smirnow, a. E., 1906. Ueber die Mitochondrien und

den Golgischen Bildungen analoge Strukturen in

einigen Zellen von Hyacinthas orientalis. Anat.

Hefte, 1. Abt., Bd. 32, pp. 146-153 (p. 57).
Smith, E. F., 1917. Embryomas in plants. Johns

Hopkins Hosp. Bull., vol. 28, pp. 277-293 (p.

140).
SoKOLOw, I., 1913. Untersuchungen iiber die Sperma-
togenese bei den Arachniden. Arch, f . Zellforsch.

Bd. 9, pp. 397-431 (p. 55).
SoLGER, B., 1896. Ueber den feineren Bau der

Glandula submaxillaris der Menschen. Festschr.

f. C. Gegenbaur, Bd. 2, pp. 180-248 (p. 44).
Spitschakoff, 1909. Spermien und Spermiohistogenese

bei Cariden. Arch. f. Zellforsch., Bd. 3, pp.

1-43 (p. 54).
Strasburger, E., 1882. Ueber den Theilungsvorgang

der Zellkerne und das Verhaltniss der Kern-

theilung zur Zelltheilung. Arch. f. mikr. Anat.,

Bd. 21, pp. 476-.590 (p. 45).
Strongman, B. T., 1917. A preliminary experimental

study on the relation between mitochondria and

discharge of nervous activity. Anat. Rec, vol.

12, pp. 167-171 (p. 131).
SnNDWALL, J., 1916. The lachrymal gland. Amer.

Jour. Anat., vol. 20, pp. 147-236 (p. 51).
Swingle, W. T., 1898. Two new organs of the plant

cell. Bot. Gaz., vol. 25, p. 110 (p. 46).
Takaki, K., 1907. Ueber die Stabchenstruktur der

Niere. Arch. f. mikr. Anat., Bd. 70, pp. 245-

265 (p. 137).

160

THE MITOCHONDRIAL CONSTITUENTS OF PROTOPLASM.

Tashiro, S., 191.5. On the nature of the nerve-impulse.

Proc. Nat. Acad. Sci., vol. 1, pp. 110-114 (p.

74).
Thurlow, M. DeG., 1917. Quantitative studies on

mitochondria in nerve-cells. Carnegie Inst, of

Wash., Contributions to Embryology, No. 16,

pp. .37-44 (pp. 80, 131, 132).
ToRRACA, L., 1914a. II comportamento dei condriosomi

nella rigenerazione dei muscoli striati. Aich. f.

Zellforseh., Bd. 12, pp. .539-5.52 (pp. 1.3.5, 138).
1914b. Alcune osservazioni sui condriosomi delle

cellule cartilaginee neUa coda del tritoue rigene-

rante. Anat. Anz., Bd. 45, pp. 4.59—474 (pp.

135, 138).
1916. Alcune osservazioni suUa rigenerazione e

sul processo di secrezione delle glandole velogene

delle Triton cristatus. Arch. ital. di Anat. e di

Embriol., vol. 15, pp. 283-330 (pp. 102, 138).
TscHASCHiN, S., 1912. Ueber vitale Farbung der Chon-

drio.9omen in Bindegewebszellen mit Pyrolblau.

FoUa h»m., Bd. 14, pp. 29.5-307 (p. 58).
TsuKAGUCHi, R., 1914. Ueber die feinere Struktur des

Ovarialeies von Aurelia aurita L. Arch, f, mikr.

Anat., Bd. 85, Abt. II, pp. 114-123 (p. 51, .54, 99).
Valette St. George, A., 1887. Zellteilung und

Samenbildung bei Forficula auricularia. Festschr.

f. Koelliker (p. 55).
Van der Stricht, O., 1898. Contribution k I'^tude du

noyau vitelUn de Balbiani dans I'oocyte de la

femme. Verh. d. anat. Ges., pp. 128-139 (p.

55).
1902. Les "pseudochromosomes" dans Toocyte

de chauve-souris (V. noctula). C. rend. Assoc.

Anat., pp. 1-60.
1904. La couche vitellogene et les mitochondries

de I'oeuf des Mammifdres. Verh. d. anat. Ges.,

pp. 138-145 (p. 45).
1905. La structure de I'oeuf de chauve-souris

(V. noctula). C. rend. Assoc. Anat., pp. 6-12

(p. 51).
1908. La structiu'e de I'a'uf de chienne et la

gen&se du corps jaune. C. rend. Assoc. Anat.,

pp. 1-9 (pp. .56, 96).
1909. La structiu-e de I'ccuf des mammifferes

(chauve-souris, Vesperugo noctula). Troisi^me

partie. Acad. roy. Belgique. Classe des sci.

M6m. 2e. s(5rie, t. 2, 173 pp. (p. 76).
1918. The genesis aod structure of the mem-

brana tectoria and the crista spiralis of the coch-
lea. Carnegie Inst, of Wash., Contributions to

Embryology, No. 21, pp. 57-85 (pp. 52, 102).
Van der Stricht, R. 1911. Vitellogenfese dans

I'ovule de Chatte. Arch, de Biol., t. 26, pp.

365-481 (p. 84).

Van Durme, M., 1913. Du role des mitochondries dans

la genese de I'ovoplasme. Ann. et Bull. Soc.

MM., Gand, n. s., t. 4, pp. 270-278.
1914. Nouvelles recherches sur la vitellogen^se

des oeufs d'oiseau.x, etc. Arch, de Biol., t. 29,

pp. 71-200 (pp. 56, 123).
Veratti, E., 1909. Sulla fine struttura delle cellule di

alcuni tumori. Boll. Soc. Med. Chir. di Pavia,

vol. 23, pp. 34-15 (pp. 63, 139).
Viguier, G., and A. Weber, 1912. Les formations chro-

midiales et mitochondriales de I'Ha'mogregarina

sergentium Nicolle, chez le Gongylus oceUatus.

C. rend. Soc. Biol., t. 73, pp. 92-93 (p. 53).
VoiNov, D. N., 1903. La spermatogenese d'4t6 chez le

Cybister roeselii. Arch, de Zool. exp. et g^n.,

pp. 173-260 (p. 55).
VoNwiLLER, P., 1915. Die Spharoplasten von Amoeba

proteus. Anat. Anz., Bd. 48, pp. 485-488

(p. 53).
Warburg, O., 1910. Oxydationen in lebenden Zellen

nach Versiiphen am Seeigelei. Hoppe-Seylers

Zeitschr. f. physiol. Chem., Bd. 66, pp. 305-

340 (p. 147).
Welch, W. H., 1888. Cartwright lectures on the

general pathology of fever. Medical News, vol.

52, pp. 393^05 (p. 139).
Weygandt, C, 1907. Beitrage zur Kenntnis der

Spermatogenese bei Plagiostoma girardi. Zeit-
schr, f. wiss. Zool., Bd. 88, pp. 249-290 (p. 54).
Wildman, E. E., 1913. The spermatogenesis of Ascaris

megalocephala with special reference to the two

cytoplasmic inclusions, the refractive body and

the mitochondria: their origin, nature and role in

fertihzation. Jour. Morph., vol. 24, pp. 421-450

(pp. 45, 49).
Wilke, G., 1907. Die Spermatogenese von Hydro-

metra lacustris L. Jena. Zeitschr. f. Naturw.,

Bd. 42, pp. 669-720 (p. 55).
Wilson, E. B., 1914. The bearing of cytological re-
search on heredity. Proc. Roy. Soc, Series B,

vol. 88, pp. 333-3.52 (p. 96).
1916. The distribution of the chondriosomes to

the spermatozoa in scorpions. Science, n. s.,

vol. 53, pp. .539 (pp. 55, 94).
Woodland, W. X. F., 1911. On some experimental tests

of recent views concerning the physiology of gas

production in teleostean fishes. Anat. Anz.,

Bd. 40, pp. 225-242 (p. 147).
Zimmermann, a., 1893. Sammel-Referate aus dem

Gesammtgebiete der Zellenlehre. Bot. Centralbl.,

Hefte 3, p. 215 (pp. 46, 57).
ZojA, F. and R., 1891. Intorno ai plaetiduh fuosinofili

(bioblastuli dell' Altmann). Mem. 1st. Sc.

Milano. Tomo 16, pp. 237-270 (pp. 42, 51,

53-56).

CONTRIBUTIONS TO EMBRYOLOGY, No. 26.

THE DEVELOPMENT AND REDUCTION OF THE TAIL AND OF THE
CAUDAL END OF THE SPINAL CORD.

By Kanae Kunitomo,

Nagasaki Medical School, Nagasaki, Japan.

Four plates, two text-figures

161

CONTENTS.

PAGE.

Introduction 163

Material antl methods 166

Serial description of embryos 168

Development and reduction of the tail 182

Chorda dorsalis 188

Development of the caudal end of spinal cord. . . 189

162

PAGE.

Abnormahties of caudal end of spinal cord .... 193

Spinal ganglia 193

Sympathetic ganglia 193

Summary ' 194

Bibliography 196

Description of plates 197

THE DEVELOPMENT AND REDUCTION OF THE TAIL AND OF THE CAUDAL

END OF THE SPINAL CORD.

By Kanae Kunitomo.

INTRODUCTION.

A great deal of literature has been published from time to time dealing with
the question of whether the human embryo at a certain stage of its development
has an actual tail — that is, a structure homologous with the tail of other mammals —
and with the persistence of such tail in after life. In biogenetic investigation this
is a subject of great interest (Darwin and Haeckel). It was from the assumption
of the occurrence of a tail in the human embryo, which was based upon one of
Ecker's figures (Icones Physiologic*, 1851-59), that Darwin drew one of his
arguments for the descent of man from a race of tailed ancestors. Von Kolliker
(1884) asserted that the human embryo has a tail-like process at its caudal
end which was not, however, recognized by him as a "true" tail. He described
it as eine spitze Schwanzartige Verlangerung. Ecker (1851-59) referred to it as
Schwanzformige Korperende, and stated that it contained only the notochord and
caudal end of the spinal cord, and was converted finally into a coccygeal tubercle
(Steisshocker) projecting caudalward. Rosenberg (1876, 1899), who investigated
the subject from a morphological standpoint, opposed the theory that the caudal
appendage in the human embryo was homologous with the tail of other mammals,
as he could not discover any part of the axial skeleton in the caudal projection.
He believed, therefore, that the latter must be concerned more with the develop-
ment of the spinal cord which he found embedded in its dorsal side. Ecker (1880),
on the other hand, held it to be a true tail, even though it contained no vertebral
primordia, and from his conclusions on the subject interest and discus.sion were
revived. His (1880rt), who coincided in general with Ecker, publi.shed his theories
under the heading "Besitzt der menschUche Embryo einen Hchwanz?" in his
Anatomic menschhcher Embryonen. He recognized a tail-hke appendage in his
younger specimen (4 mm.), but did not regard it as a true tail. In older embryos,
in which the primitive vertebrae had developed into cartilaginous tissue, he found
that one or two vertebrae entered into the root of the tail. This portion he desig-
nated as the vertebral tail. The remainder contained only notochord and medullary
cord (caudal filament) and was therefore called non-vertebral tail. In none of his
specimens did he find more than the normal number of vertebrae, 34. He states:
"Die Embryonen A and B haben sonach eine achte Schwanze-anlage, die aber
ausserordenthch kurz ist und jedenfalls nicht iiber zwei Segmentlangen umfasst."
The opinions of Ecker and His may be summarized as follows :

1. The term tail refers only to that portion of the embryo which projects beyond the
cloaca.

163

164 DEVELOPMENT AND REDUCTION OF THE TAIL

2. In younger specimens (8 to 15 mm.) the tail appears as a free, pointed projec-
tion from the cloaca, directed caudo-dorsally.

3. The tail consists of two portions — that containing vertebra; and that without ver-
tebrse. The latter contains only chorda dorsalis and medullary tube. In time this portion
disappears, the medullary tube atrophying and the chorda becoming converted into a knot.

4. The vertebral portion persists for a while, appearing later as a coccygeal prominence
in the caudal region — coccygeal tubercle; then it, too, disappears.

Keibel (1891) published in an important paper his findings in regard to the
development of the caudal gut in 4, 8, and 11.5 mm. embryos. The existence of
this structure, which forms a small canal or cell-strand at the caudal end of the
body axis, he regards as irrefutable evidence of a tail primordium. He found the
gut to be longer in the 8 mm. embryo than in the others. He asserts (page 378)
that in this stage the caudal gut extends through the whole length of the tail, and
apparently at this time attains its maximum length. Tliis author defines the
line of demarcation between the tail and the body in two ways: (1) he designates
as the tail the caudal portion beyond the attachment of the pelvic joint; (2) in the
younger embryos, in which the primordia of the legs have not yet appeared, he
defines the first 8 trunk segments as the cervical segments, the next 12 as the
dorsal, the next 5 as the lumbar, the next 5 as the sacral, and the remaining seg-
ments as caudal vertebrae. These he found were usually 6 in number. The last
one he called the mesodermic remnant and regarded it as one segment, although it
was two or three times as long as those cranial to it.

Braun (1882) published his observations on the development and reduction
of the embryonic tail among mammalia, having at his disposal a great number of
specimens. As a rule he found a caudal filament at the extreme end of the tail
in the mammals that he studied, and therefore beUeved this structure to be of
general occurrence, and probably true also of the human tail. On the other hand,
Ecker and His, who studied the same condition in human embryos, did not con-
sider the two exactly homologous. Braun classified the two portions of the tail
as internal and external, and subdivided the latter into vertebral and non-vertebral
tail, the caudal filament being part of the latter. Waldej-er (1896) takes excep-
tion to this division into internal and external tail, as he does not believe the former
is a tail. Rodenacker (1898) uses the terms canda aperta and cauda occulta instead
of internal and external tail. Unger and Brugsch (1903) give the following results
of their investigations:

1 . In the reduction of the tail the caudal vertebrae fuse to form the last vertebra.

2. The caudal filament represents the remnant of the tail-bud and contains a branch
of the middle sacral artery.

3. In the reduction of the tail two processes are concerned: (a) the formation of the
caudal tubercle; (6) the formation of the coccygeal tubercle. The first is the reduced
tail; the second is formed by the bulging of the caudal end of the vertebral column. This
is due to the fact that the growth of the vertebra is more rapid than that of the skin and
spinal cord.

4. The connective tissue contained in the caudal filament develops into the caudal
ligament.

s

AND OF THE CAUDAL END OF THE SPINAL CORD. 165

Regarding these changes they state (p. 100) :

"Wandelt sich dann durch starkeres Wachstum der Kaudalwirbelsaule der Schwanz-
hdcker in den Steisshocker iim, so wird durch den Zug der Haut, deren Wachstumrichtung
der der Steisswirbel entgegengesetzt ist , der Schwanzfaden von der Achse der Kaudalwirbel
entfernt und mit der Haut auf warts mitgenommen (cfr. Embryo Dii. 4^2 cm.). Durch
dieses Aufwartsriicken des Sehwanzfadens wird aber dieser seines Rlickenmarks beraubt,
d. h. er ist reduziert. Seinen Inhalt stellt nun ein Gewebe vor, das in Form von Bindege-
websziigen mit der kaudalen Flache des letzen Kaudalwirbels verbunden ist, und das aus dem
Mesodermrest des friiheren Sehwanzfadens hervorgegangen ist. Da auch hier diesen
Bindegewebsziigen und dem Schwanzfadenrest die Endaste der inzwischen durch
Bildung des Steisshockers sehr reduzierten Arteria sacraHs media zukommen, so konnen
wir sie als einen immerhin wesentUchen Rest der urspriinghchen Schwanzanlage bezeichnen.
Diese Bindegewebsziige sind das lig. caudale; sie schhessen auch den urspriingUch im
Schwanzfaden sich befindHchen kaudalsten Teil des Ruckenmarks ein, in dem sich spater
die 'vestiges coccygiens' von Tourneux und Hermann fs. o.) oder 'kaudalen Riicken-
marksreste' entwickeln."

Our knowledge concerning the development of the caudal end of the spinal
cord is very Umited, especially as regards its early stages. My aim, therefore,
has been to study the early development of this part of the spinal structure and to
follow the histological changes it undergoes in adaptation to later topographical
conditions. Before reporting my investigations, however, I would refer to some
of the writers who have preceded me in this field of study. Clarke (1859), in his
well-known study of the spinal cord, pictures the ventriculus terminalis as seen
in sections of the cord of the ox (plate xxiii, fig. 21). He regarded this structure
as a persisting remnant of the lower end of the sinus rhomboidalis, which in other
mammals is usually limited to the lumbar enlargement. Krause (1875) discovered
the ventriculus in the spinal cord of the human embryo and describes it as per-
sisting in adults as a rudimentary organ. He suggests that in the embryo, by
means of its ciliated cells, it serves in the maintenance of the circulation of the
contained cerebral spinal fluid. Tourneux and Hermann (1887), who studied the
caudal end of the spinal cord in the human embryo, discovered the remnant of the
neural canal in the caudal region and called it vestiges medullaires cocajgiens. Tney
describe in detail the process of reduction of the caudal end of the spinal cord.
Argutinsky (1898) discussed the morphology of the ventriculus tcrminaUs in older
fetuses and newborns, and classified it in three divisions — upper, middle, and
lower. Von Kolhker, Ecker, His, and others reported that in younger embryos the
spinal cord extends to the extreme end of the tail. Brugsch and linger briefly
summarized their investigations on the ventriculus terminaUs in the human embryo
as follows (p. 232) :

"Kurz gesagt stellt der V. t. also eine konische Erweiterung des Centralkanals im
unteren Ende des Conus medullaris und im Anfange des filum terminale vor, dessen oberer
weiter Abschnitt meistens Ausbuchtungen besitzt. Der untere Abschnitt endigt blind
im filum terminale."

In describing this structure they make the following divisions: (1) an upper,
wider part, which is continuous with the central canal of the more cephalic part

166 DEVELOPMENT AND EEDUCTION OF THE TAIL

of the spinal cord, and which forms an irregular, evaginated space in the conus
meduUaris; (2) an under part, which gradually narrows toward its caudal end and
terminates blindly in the filum terminale.

The various investigations of the occurrence of tails among adults, children,
and newborn infants have given rise to a great deal of discussion. Bartels (1884)
published an exhaustive study of the occurrence of tails among the human race.
Other pubhcations on the subject have appeared from time to time, notably bj^
Virchow (1884), Oskar Schaeffer (1892), Pyatnitski (1892), Dickinson (1894), Berry
(1894), Kohlbriigge (1898), Watson (1900), and others. Harrison (1901) describes
the histological structure of a large, well-developed tail which was removed from a
child six months old. He states:

"Two weeks after the birth of the child the tail was 4.4 cm. long; at the age of two
months it had grown to 5 cm., and at six months, when it was removed, it had attained
a length of 7 cm., showing altogether a fairly rapid rate of growth. The most remarkable
characteristic of the tail was its movability. * * * Beneath the skin the main bulk of the
tail was made up of areolar tissue containing much fat. Blood vessels, nerves, and striated
muscle fibers are embedded in this mass. There is no trace of anything like the medullary
cord or of notochordal tissue."

More recently similar observations have been made by Brugsch (1907), Kon-
stantiowitsch (1907), and Schwarz (1912).

MATERIAL AND METHODS.

The material upon which this study is based consists of 44 specimens in the
Carnegie Collection of human embryos, Baltimore. They range from 4 to 125
nun. CR length, and a table of them, with their respective measurements, is shown
herein. The specimens, for the greater part, had been carefully preserved in 10
per cent formahn and dehydrated in alcohol. The smaller ones were embedded in
paraffin, the larger ones in celloidin, and most of them were cut in serial sagittal
sections varying in thickness from 20 to 200 n in the different specimens. A large
proportion of the specimens were stained m toto in alum cochineal and borax car-
mine before embedding; others were stained on the slides with hematoxylin and
eosin or similar stains. From 15 to 80 sagittal sections through the median part
of each embryo were used in this investigation, and usually one graphic recon-
struction was made of each specimen, although all of these are not illustrated.
They present a median profile view disclosing some structure — for instance, the
winding caudal end of the chorda dorsaUs or the sympathetic ganglia. In the
illustrations these are slightly schematicized in order to show distinctly their actual
relations. The graphic reconstructions were made by copying each section on
transparent paper from a projection apparatus. When the drawings under the
projection apparatus were completed the sheets were piled so that adjacent sec-
tions were accurately fitted one upon another. The desired parts of the sketch
on each sheet were then copied on another sheet of paper, due attention being
given to the form and relation of the component parts. This procedure was facili-
tated by the use of a glass table illuminated from below. In the case of cross-

AND OF THE CAUDAL END OF THE SPINAL CORD.

167

sections, a guide-line was established by marking upon each sheet two Unes, one
perpendicular to the other, thus forming a series of crosses which were exactly
superimposed throughout the entire pile. The individual sections were then
plotted off on millimeter paper by fitting the crosses to a chosen perpendicular
line, the distance between the sections being determined by the thickness of the
sections and the enlargement of the drawing. The enlargements with the projec-
tion apparatus were as follows: Embr.yos 4 to 16 mm., X70; embryos 17 to 38
mm., X50; embryos 39 to 52 mm., X30; embryos 67 to 125 mm., X20.

Table of Embryos Slwiivd.

Length
(CR) of
embryos

TnTn.
4 0
4.0

5.5

6.6
7.5
8.0

9.0

9.0

10.0

11.0
12.0

13.0
13.0
14 0
15.5

16.0
16.0

17.0
17.0
18.0
19.0
21 0

23 0
23.0

24 0
25.0
26 0
26.0
27.0
30.0
33.0
33.0
35.0

36 0

37 0
39 0
46.0

50 0

.50.0
52.0
67.0
80 0
100.0
125.0

Cata-
logue
No.

836

786

810

371
221
389

422

721

1197

544
852

485
643
940
390

406
43

576
991
432
431
837
453
382
632
584a
1008
405
875

75
145
211
199
449
972
362

95

184

96
448
1656
662
928
142

No. of somites.

28 -|-remnant.
30 +remnant.
37+ remnant.
. 36 + remnant, f
37+ remnant.
38+remnant.
38 +remnant.

38+remnant.

37

35+remnant.

38
37

37
37
36
35

36
37

35
34
34
33
35
35
34
35
34
34
34
34
34
35
34
34
32
34
34
35

34

35
34

34

No. of

spinal

ganglia.

32
31
32

32
32
32

32
33

33
33
33

32

31
32

31

31

32

32

32

32

31

31

32

31

31

31

31

31

31

30

31

30

30

30

30 \

29 /

29

30

29

Segmental level i
cloaca.

29tli so;;ment

29th s('-.;inent

30th segment

Between 30th and 31st
segments.

Extension of
caudal end of
spinal cord.

Tip of tail.

Do.
Do.
Do.

Between 31st and 32d

segments.
33d segment.
Between 32d and 33d

segments.

32d segment.

Between 33d and 34th
segments.

Between 34th and 35th
segments.

33d segment

33d segment

3.3d segment

33d segment

33d segment

35th segment

Below last segment. . .

35th segment

34th segment

34th segment

34th segment

34th segment.

34th segment

34th segment

Below last segment.

... Do

... Do

Do.

Tip of tail .

Do.
.Do.

.Do.
.Do.

Nearly to tip of tail

Do.
Do.

.Do.

Tip of tail.

Do..

Do. .

Nearly to tip of tail

Do

Do

Tip of tail

Nearly to tip of tail.

Tip of tail

Do

Do

Do

Below last vertebra

Do

Do

Level of demarcation

between main cord

and its more caudal,

atrophic portion.

33d segment.

33d segment.

(?)
33d segment.
32d segment.

Between 33d and 34th.
34th segment.

32d segment.
32d and 33d segments.
31st segment.
32d segment.
32d segment.
31st segment.
Between 31st and 32d.
Between 32d and 33d.
32d segment.
Between 31st and 32d.
Between 30th and 31st.
Between 31st and 32d.
32d segment.
Between 30th and 31st.
Between 30th and 31st.
31st segment. (?)
Between 29th and 30th
31st segment.
Between 29th and 30th.

(?)
Between 29th and 30th
segments.

168 DEVELOPMENT AND REDUCTION OF THE TAIL

SERIAL DESCRIPTION OF EMBRYOS.

Embryo No. 836, 4 mm. Greatest Length.

This embn^o represents the earliest stage studied in this investigation, and a dia-
grammatic profile reconstruction of it is shown in figure 1, plate 1. It was sectioned
transversely and a series of models of it were reconstructed under the direction of Professor
Evans. These were available for comparison and were of particular value in orienting
the sections in the lumbo-sacral region, where the caudal extremity curves so that it lies
transverse to the axis of the trunk. The embryo contains 28 somites and a long meso-
dermic remnant which extends about one-third of its length beyond the caudal end of the
spinal cord. The chorda dorsalis runs along the ventral surface of the spinal cord in close
apposition to it and terminates before reaching the caudal end of the central canal; whereas
the caudal gut extends farther down towards the tip of the tail, as indicated in figure 1.
The caudal extremity of the embryo consists of a mass of germinating cells into which the
ends of the spinal cord, the chorda dorsalis, and mesodermic remnant all merge, their out-
lines becoming entirely obliterated. The primordium of the caudal artery extends to the
tip of the tail.

Embryo No. 786, 4 mm. Greatest Length.

This specimen has 30 somites and a mesodermic remnant ; the latter consists of a long
cord joined to 4 globular masses, the last of which is slightly longer and thinner than the
others. The counting of the somites in small embryos is always very difficult, as pointed
out by Keibel, especially when the material is not well preserved. It was easy, however,
to make out the vesicula auditiva in the sagittal sections of this specimen, and caudo-
dorsal to it the glossopharyngeal nerve. The ganglionated vagus nerve in turn is situated
caudal to this, being surrounded by the internal jugular vein, which curves around its
caudal margin. The first cervical somite lies dorsal-caudal at a distance of 180 to 200 /u
from the vagus ganglion. The embryo has 3 occipital somites.

Embryo No. 810, 5.5 mm. Crown-Rxjmp Length.

A graphic reconstruction of the caudal end of embryo No. 810 is shown in figure 2,
plate 1, in which the structures are diagrammatically straightened out. In reahty the
caudal end is bent to the right and towards the front, its tip nearly reaching the right side
of the face. The first cer\'ical ganglion is about one-third the size of the second and lies
close to the Froriep ganglion, which is situated at the dorsal side of the bow-shaped trunk
of the accessory nerve and is particularly large on the left side. There are 32 spinal ganglia,
the last 2 being small. I was able to count on the right side 37 somites and a mesodermic
remnant ; on the left side only 36 somites and a remnant. In the cranial portion of the
mesodermic remnant there seem to exist potential somites, the outlines of which, however,
can not yet be made out. Its caudal part consists of a small, bell-shaped remnant. The
entire remnant is about the length and size of 4 somites. The mesoderm in the caudal
end of the embryo is primitive in character and can be seen dividing into its parietal and
visceral layers. Embedded in these can also be seen the caudal gut and a simple vascu-
lar plexus. The spinal cord, with its central canal, extends to the caudal end of the meso-
dermic renmant. The caudal gut, which extends from the cloaca to the end of the tail,
is beginning to disappear in this embryo at a point 200 n caudal to the cloaca and 657 m
from the extreme end of the tail between the thirtieth and thirty-first somites, as is indi-
cated in figure 2, plate 1. In figure 31, plate 2, these conditions are shown more in
detail. At the site of the beginning degeneration the ends are sharply pointed, but are
still connected by a cell-strand. The shorter, cranial portion of the caudal gut opens into
the cavity of the cloaca. The longer, caudal portion ends blindly at both extremities,
the caudal end merging into the mesodermic cell-mass and uniting with the caudal end of

AND OF THE CAUDAL END OF THE SPINAL CORD. 169

the chorda dorsalis. Each portion of the gut contains a distinct lumen. On the ventral
surface of the caudal region of the embryo, where the gut first begins to disappear, is a
small groove (indicated by X in fig. 31), which may represent the primordium of the sub-
caudal epithelial plate of Keibel.

Embryo No. 371, 6.6 mm. Crown-Rump Length.

This embryo has 37 somites and a mesodermic remnant. The thirty-seventh somite
lies close to the remnant, while the others are separated by narrow spaces into which blood
capillaries enter. The remnant is constricted at three points, the separations, however,
being incomplete. The caudal end of the medullary tube extends to the end of the tail,
as can be seen in figures 3 and 32. The caudal gut, which was distinctly recognized in the
5.5 nnm. specimen, has here undergone a more marked obliteration, and over its greater
part is left only a strand of cells which, to judge by their staining reactions, are probably
degenerating. The caudal portion of the gut contains a small lumen, however, and shows
no change from that noted in the preceding specimen. What constitutes the cranial end
of the shorter portion of the caudal gut in the younger embryo is here dilated and forms
part of the cloacal membrane. Ventral and parallel to the gut is a blood-vessel which
anastomoses with the middle sacral artery by means of numerous capillaries. The chorda
dorsalis runs within the substance of the vertebral column until it reaches the thirtieth
somite, when it emerges from the vertebral tissue and continues the remainder of its course
in the interval between the primitive vertebrae and spinal cord, until it finally loses itself
in the cell-mass at the end of the tail. The plexiform middle sacral artery follows a course
ventral to the vertebral column and can be traced to the tip of the cord.

Embryo No. 221, 7.5 mm. Crown-Rump Length.

Embryo No. 221 has 38 somites and a remnant, the latter showing constrictions at
three points. The caudal end of the embryo resembles that of a pig, as described by
Keibel in his work on the human embryo. The cloaca, which is situated at a level with
the thirty-first somite, is well developed, but as yet there has been no perforation of the
membrane. Below the chorda dorsalis is a short remnant of the caudal gut containing
a small lumen, as can be seen in figure 4, plate 1. At their caudal extremities the chorda,
spinal cord, and caudal gut appear to fuse together. The central canal of the spinal cord
is obliterated at its sacral portion, showing a pathological condition, and appears as a solid
mass of nervous tissue. The groove between the tail-bud and the cloaca, which in the
younger specimens indicated the first point of disappearance of the caudal gut, is here
.situated at the level of the thirty-first somite and is destined to later develop into the sub-
caudal epithelial plate. In this embryo there are 31 spinal ganglia with nerves.

Embryo No. 389, 8 mm. Crown-Rump Length.

As can be seen in figures 5 and 33, the caudal gut in embryo No. 389 still persists as
a group of cells inclosing a narrow lumen. This mass seems to fuse with the caudal ends
of the medullary tube and the chorda. The remnant of the caudal gut is surrounded by
a close network of capillaries, which possibly bear some relation to its absorption. At the
ventral wall of the spinal cord below the thirtieth somite can be seen two or three folds
indicated by X in figure 33. As such folds were not found in the younger specimens, I
assume that they begin at about this stage of development, and from now on I shall have
occasion to refer to them frequently. The winding chorda emerges from the vertebral
column at the thirty-fourth somite and continues its course to the end of the tail in the
interval between the vertebral column and the spinal cord. This embryo contains 38
somites and a mesodermic remnant. In the latter I was unable to make out any con-
strictions, only a mass of germinating cells. The spinal ganglia number 32.

170 DEVELOPMENT AND REDUCTION OF THE TAIL

Embryos No. 721 and No. 422, 9 mm. Crown-Rump Length.

As no embryos of this size in sagittal section were available, I have had to
make use of two that were cut in coronal sections, although in them the study of
the vertebral structures was much more difficult. In embryo No. 721 the caudal portion is
cut transversely, and in the sections through the end of the tail the caudal gut with its
round lumen can be distinctly recognized. The extreme end of the gut is surrounded by
a network of capillaries communicating with the middle sacral artery and vein. The cells
of the caudal gut seem to fuse with those of the spinal cord and the chorda dorsalis in the
caudal region. In this specimen there are 37 somites and a long remnant. I was able
to count 32 spinal ganglia, the thirty-second being small and without nerves. In embryo
No. 422 there are 38 somites with a remnant, and 32 spinal ganglia. The last one of these
also is small and contains no nerve-fibers. The specimen is so poorly preserved that the
caudal gut can not be clearly made out.

Embryo No. 1197, 10 mm. Crown-Rump Length.

In embryo No. 1 197 there are 35 primitive vertebra- of Remak, or scleromeres, although
the thirty-third and thirty-fourth appear to consist each of two parts fused together. This
fusion has occurred, apparently, at an earlier stage. At the caudal end of the vertebral
column the vertel)np have not completely developed, although the tissue is condensed,
showing that development is well under way. The chorda dorsalis is situated dorsal to
the primitive vertebra- in the caudal portion and is slightly contorted. Caudally it ter-
minates suddenly with a rounded end ventral to the neural tube at a level with the thirty-
fifth vertebra. There are 32 spinal ganglia with nerves, the last 2 nerves being quite
delicate. I have carefully examined each section of the caudal region in an effort to locate
a remnant of the caudal gut, but could find no trace of it.

Embryo No. 544, 11 mm. Crown-Rump Length.

In embryo No. 544 the caudal end is bent sharply to the right. In the graphic recon-
struction shown in figiu'e 34 it is represented as straightened out in order to show more
clearly the relations of the structures in this region. It is pictured in a more diagranmiatic
way in figin-e 6. Here 38 primitive vertebrae and a remnant are present, the latter differ-
ing from that described in the younger embryos. In those instances I have referred to
it as the mesodermic remnant, using the terminology of Keibel, whereas in this stage of
development (9, It), and 11 nun.) the mesodermic remnant has been gradually converted
into a non-vertebrated tail. The last scleromere or primitive vertebra, as shown in figure
34, is larger than the two or three more cranially situated ones. Two theories as to its
development present themselves for consideration: Its growth may be the result of (1)
fusion of the last two or three scleromeres which have become separated from the adjacent
somites; (2) the addition of cells from the mesodermic remnant. WTiile somewhat in doubt
as to which theory to accept, I am inclined to favor the latter, as this embryo contains 38
primitive vertebrae, corresponding to the maximimi number of somites found in the younger
embryos, and there is no condensed tissue or group of cells in the end of the tail to indicate
the primordium of a primitive vertebra.

Embryo No. 852, 12 mm. Crown-Rump Length.

In embryo No. 852 I found 37 scleromeres and a remnant; also 33 spinal ganglia, the
last 2 being small, with slender nerves. The last 3 nerves emerge at the same point to
form the caudal nerves, which run from about the thirty-second to the thirty-sixth sclero-
mere. The central canal of the spinal cord narrows between the thirty-fourth and thirty-
fifth scleromeres and on the ventral wall of this narrow part are a few folds. These are
indicated in figure 35. The chorda is almost entirely embedded in the tissue of the primi-

AND OF THE CAUDAL END OF THE SPINAL CORD. 171

tive vertebrae. In the younger specimens it emerges at the thirty-fifth scleromere, while
in this one it emerges at about the thirty-seventh and from thence runs ventral to the
spinal cord, touching the latter closely. The cloaca is situated at a level between the
thirty-third and thirty-fourth scleromeres.

Embryo No. 485, 13 mm. Crown-Rump Length.

This specimen is cut in transverse section and is therefore well suited for the study
of the spinal cord. There are 33 spinal ganglia, but at the thirty-third and thirty-second
complete nerve-fibers can not be made out. The thirty-third, in particular, comprises
such a small cell-group as to be hardly recognizable. There are 37 primitive vertebne
and a non-vertebrated tail portion 289 n long. As in the several specimens immediately
preceding it, the few caudal vertebrae are fused together, showing no distinct boundaries.
The non-vertebrated tail portion consists still of germinating mesenchymal cells, while
the more cranially situated scleromeres are gradually becoming converted into precar-
tilaginous tissue.

Embryo No. 643, 13 m.m. Crown-Ru.mp Length.

This specimen is cut in serial sagittal sections, but is, however, rather poorly preserved.
There are 37 primitive vertebrae, the last one being incomplete. Caudal to this is a long
non-vertebrated tail portion. There are 33 spinal ganglia, the last 2 having slender nerves
which, with the thirty-first, form the caudal nerves. These extend caudally along each
side of the tail.

Embryo No. 940, 14 mm. Crown-Rump Length.

This specimen is cut in transverse section, and it is therefore difficult to follow the
topography of some of the structures. I was able to count 36 primitive vertebrae, with a
remnant, and 33 spinal ganglia. The last two, especially the thirty-third, are small and
not provided with nerves. The twenty-ninth, thirtieth, and thirty-first spinal nerves extend
down along the sides of the cord to the level of the thirty-sixth vertebra. The central canal
of the spinal cord narrows at about the thirty-second vertebra, the portion caudal to
this being practically devoid of the mantle layer.

Embryo No. 390, 15.5 mm. Crown-Rump Length.

In embryo No. 390 the primitive vertebrae are differentiated into precartilaginous
tissue and between the vertebrae there is a small quantity of embryonic connective-tissue,
as indicated in figure 36. At this period the vertebral column consists of 35 precartilaginous
vertebrae; no additional segments can be recognized, although in the younger specimens
there were 38 somites and one very long mesodermic remnant. There is a long, irregular,
mesodermic cell-strand at the caudal end of the thirty-fifth segment, which, with the
spinal cord, extends to the end of the tail. Between the two lies the caudal end of the
chorda dorsalis, which, however, does not extend to the tip of the tail, as can be seen in
figures 8 and 36. About the thirty-fifth segment the chorda becomes embedded in the
substance of the vertebral column, the point at which it emerges being characterized by
a sharp curve in its course.

We are here confronted with the following questions: WTiat was the fate of the addi-
tional somites which could be so distinctly recognized in the earlier stages? And what
is the mesodermic cell-strand which extends from the last primitive vertebra? As to the
first, from the study of this material I am of the opinion that the last few vertebrae, which
have earlier developed from the sclerotomes, fuse together during the process of embryonic
development, thus forming the last vertebra in an embryo of this age. I am, however,
aware that the comparison of a series of embryos can not conclusively settle this question.
As concerns the cell-strand, it is my belief that it is formed of parts of the tissue which did

172 DEVELOPMENT AND REDUCTION OP THE TAIL

not go to make up the primitive vertebrae, and constitutes the primordium of the caudal
ligament. As is well known, the sclerotomes of the somites develop into not only primi-
tive vertebrae, but also into several other supporting tissues which form the framework of
the vertebral column. (Keibel and Mall, Human Embryology, I, page 331.)

The spinal cord narrows suddenly at the thirty-second vertebra, so that the portion
caudal to this, together with its canal, presents an appearance distinctly different from the
main body of the cord, as can be seen in figure 36. On account of its regressive appear-
ance, I shall hereafter refer to this part as the atrophic cord. One of its very character-
istic features is a slender, narrow canal, and it might be desirable to speak of this as the
narrow canal portion of the spinal cord.

Embryo No. 406, 16 mm. Crown-Rump Length.

A graphic reconstruction of embryo No. 406 is shown in figure 37, and a more dia-
grammatic sketch in figure 9. These show that the embryo has 36 cartilaginous vertebrae.
On the right side the thirty-second and thirty-third segments have fused together. The
last vertebra consists of two or three sections, each of which in an earlier stage probably
represented a comjjlete somite, these later fusing into one large segment. At the caudal
end of this is a group of undifferentiated mesodermal cells — the primordium of the caudal
ligament. The caudal end of the chorda dorsahs emerges at the thirty-sixth vertebra
and terminates abruptly between the vertebral column and the spinal cord. The spinal
cord narrows suddenly at the thirty-fourth vertebra. On the ventral wall of the canal in
this narrow or atrophic portion of the cord there are three or four folds (fig. 37). In some
of the sections can be seen a larger fold at the level of the twenty-ninth vertebra, which
hangs down to the level of the thirty-fourth, both sides adhering to the wall of the spinal
cord, thus forming a diverticulum, which is not shown in the illustration. The caudal
end of the emtjryo is bent sharply dorsal, the bent portion being marked off on the surface
by a shallow, circular furrow. The spinal cord extends to the end of the tail, conforming
to the shape of the bent portion. The extreme end contains a narrow cavity which rep-
resents the caudal end of the central canal; the canal is interrupted at the root of the tail,
where the cord appears to consist of solid nerve-tissue, as is indicated in figure 37. In this
specimen there are 31 spinal ganglia with distinct nerves.

Embryo No. 43, 16 mm. Crown-Rump Length.

This specimen has 37 cartilaginous vertebrae, the last being divided into three parts.
There are 32 spinal ganglia. A graphic reconstruction was made of this embryo, but it is
not illustrated in the figures.

Embryo No. 576, 17 mm. Crown-Rump Length.

This embryo has 35 cartilaginous vertebrae, the last consisting of three small pieces
fused together. The demarcation between these pieces can be more clearly recognized
in the lateral portions of the column than in the median plane; so in determining the com-
position of the last segment one must study carefully the more lateral line of sections. A
profile reconstruction of the embryo is shown in figure 38, and a more diagrammatic sketch
in figure 10. The tail, with the caudal end of the spinal cord, is bent sharply dorsalward.
The spinal cord narrows suddenly at the thirty-second vertebra and from this point down
the central canal, which extends the entire length of the cord, becomes much smaller
and rounder, while in the more cranial portion a transverse section of it would form an
elongated oval. The ventral wall of the canal in the atrophic portion presents several
transverse folds, as seems to be usually the case at this stage of reduction. The caudal
portion of the chorda dorsalis is convoluted and its end sharply retracted.

AND OF THE CAUDAL END OF THE SPINAL CORD. 173

Embryo No. 991, 17 mm. Crown-Rump Length.

The caudal end of embryo No. 991 is somewhat torn, but I feel reasonably sure that
it did not exhibit a long caudal process. At a level between the thirty-second and thirty-
third vertebrae the central canal of the spinal cord narrows suddenly, and on the ventral
wall of the atrophic portion of the cord are two folds. There are only 31 spinal ganglia,
the first cervical on each side being absent. The others are completely developed, and
even the thirty-first has its full complement of nerve-fibers. In embryo No. 576, just
described, this nerve was quite slender. The chorda dorsalis runs in a straight line through
the cartilaginous vertebral column and emerges from the thirty-fourth vertebra without
winding. This is the first specimen of the series that lacks the non-vertebrated tail. At
the point where the tail is found in younger embryos this has a small projection resembling
a tail-bud more than an actual tail. Between the last vertebra and this caudal projection
is a mesodermic cell-mass into which enter the plexif orm branches of the middle sacral artery
and vein.

Embryo No. 432, 18 mm. Crown-Rump Length.

Embryo No. 432 contains 34 cartilaginous vertebra?. The last is larger than the
thirty-third and on its lateral side shows three divisions, each consisting of young precar-
tilaginous tissue. The middle part, where the segments are partially fused, is made up
of cartilaginous tissue, and here the vertebra is incompletely divided into two segments —
cranial and caudal. A little to one side of the median line, therefore, it is possible to count 35,
and more laterally, 36 vertebrae. If the scleromeres of the several somites fuse together
and develop into the last cartilaginous vertebra, we may assume that the remaining
mesenchymal somites left in the caudal portion of the tail form the non-vertebrated
tail, and that in a more advanced stage of embryonic growth this substance develops into
the caudal ligament. The chorda dorsalis shows two branches at its caudal end; one is
formed at the thirty-second, the other at the thirty-fourth vertebra, and both follow a
dorsal course. The more caudal branch is the longer, and its pointed extremity, which
represents the caudal end of the chorda, adheres to the ventral wall of the atrophic por-
tion of the spinal cord, as shown in figure 39. The wall of the spinal cord is quite thick in
this region and at its upper portion are several folds. This condition is very interesting
on account of its possible mechanical relation to the chorda, because as the caudal end of
the chorda retracts it would tend to draw up the end of the spinal cord that is in the tail.

Embryo No. 431, 19 mm. Crown-Rump Length.

Embryo No. 431 has 33 cartilaginous vertebrae. The last is larger than the thirty-
second and consists of two segments. It would seem most probable, therefore, that it has
been developed by the fusion of two or more parts. The chorda dorsalis is considerably
distorted in the thirty-third vertebra, thus indicating a fusion of several primitive verte-
brae, and its caudal end adheres to the ventral wall of the spinal cord (fig. 13). The spinal
cord extends to the tip of the tail, and a short distance from its extremity the ventral wall
shows several folds. It appears possible that with the retraction of the chorda dorsalis
the neural tube is also drawn up, thus producing folds on its ventral wall. The caudal
end of the anterior spinal artery winds through these folds. The primordium of the ven-
triculus terminalis, between the wide and narrow parts of the central canal, is seen at the
level of the thirty-second vertebra. This embryo has 32 spinal ganglia, the thirty-second
being incomplete and without nerve-fibers; in all the others, however, the spinal nerves
are complete.

Embryo No. 837, 21 mm. Crown-Rump Length.

Embryo No. 837 has 35 cartilaginous vertebrae, but the last is very small and its
transition into cartilage is just beginning. It is surrounded by a voluminous mass of
precartilaginous tissue resulting from the fusion of the last few scleromeres. The caudal

174 DEVELOPMENT AND REDUCTION OF THE TAIL

end (if the chorda dorsahs projects from the extreme end of the vertebral column into the
region of the non-vertebrated tail and appears as if iii'e\'ioiisly it may have adhered to the
wall of the caudal end of the neural tube. In the thirty-fourth vertebra, and lietween
the thirty-fourth and thirty-third, the chorda shows a typical loop formation, while in its
main portion it is almost straight and is situated in the midline of the column. The cen-
tral canal of the spinal cord narrows sharply at the level of the thirty-second vertebra.
Its ventral wall shows a few folds in the region of the atrophic portion of the cord. The
spinal cord reaches to the tip of the tail and has a continuous canal thoughout. The
middle sacral artery and vein extend into the non-vertebrated portion of the tail.
Embryo No. 453, 23 mm. Crown-Rump Length.
In embryo No. 453 there are 35 vertebrae, the last consisting of precartilage tissue
which has not as yet developed into true cartilage. The chorda dorsalis is straight and
runs through the vertebral column in the midline. Its caudal end, however, shows 4
coils, as shown in figure 40, representing a profile reconstruction of the specimen. The
neural canal narrows at the thirty-first vertebra. At the level of the thirty-third vertebra
a sac-shaped cell-mass lies between the spinal cord and the vertebral column, separated,
however, from the wall of the former (fig. 40, diverticulum). This sac seems to have
resulted from a diverticulum of the ventral wall of the neural canal, the stalk of which has
been obliterated. The \entral wall of the atrophic portion of the spinal cord shows small
folds at the level f)f the thirty-fourth and thirty-fifth vertebra^. The caudal end of the
spinal cord expands slightly and its extreme end enters into the tail, which is now quite
reduced. The midtlle sacral artery conmiunicates with the anterior spinal artery by
means of a branch that curves about the tip of the vertebral column. The subcaudal
epidermal plate has nearly disappeared, while the post-anal swelling and the coccygeal
tubercle have become visible. There is a shallow furrow between the tail-end and the
primordium of the coccygeal tubercle, constituting a boundary between them. This
embryo has 32 spinal ganglia with nerves.

Embryo No. 382, 23 mm. Crown-Rump Length.
Embryo No. 382 has 34 cartilaginous vertebra?. The last three do not lie exactly in
a row in the median line, as can be seen in figures 15 and 4L The ventral portions of the
thirty-second and thirty-fourth vertebrae, and the dorsal portion of the thirty-third, have
become converted into cartilage, whereas the remainder of these two vertebrae still con-
sists of precartilage tissue, as in younger specimens. At the thirty-third vertebra the
chorda dorsalis gives off a short branch in a dorsal direction. Caudal to this the chorda
winds and finally disappears in the caudo-dorsal portion of the thirty-fourth vertebra.
Opposite the end of the chorda the wall of the spinal cord is so thickened as to give the
impression that the two might have been attached. It is to be regretted that in most of
the specimens the caudal end of the chorda dorsalis is torn. At the caudal end of the
last vertebra the middle sacral artery anastomoses with the anterior spinal artery through
a branch, similar to that mentioned in the last specimen. At the caudal end of the em-
bryo there is a bud-like structure of the skin. This proves to be a stunted tail, for at its
root can be recognized the sharp caudal end of the spinal cord and the terminal branches
of the middle sacral artery and vein. The central canal of the spinal cord narrows at the
thirty-second vertebra-, but the ventral wall of its atrophic portion does not exhibit the
folding that usually occurs in this region.

Embryo No. 632, 24 mm. Crown-Rump Length.
Embryo No. 632 is quite similar to No. 382, just described, except that it has no tail-
bud. There are 35 vertebra^ and 31 spinal ganglia. The caudal end of the spinal cord
is represented by a strand of loosely arranged cells which extends to the epidermis at a
point corresponding approximately to the root of the tail in a younger embryo, as can be
seen in figure 16.

AND OP THE CAUDAL END OF THE SPINAL CORD. 175

Embryo No. 584«, 25 mm. Crown-Rump Length.

A profile reconstruction of the caudal end of embryo No. 584a is shown in figure 42
and a simplified sketch is shown in figure 17. There are 34 vertebrae, the last being larger
than the thirty-third. Around the cartilaginous mid-portion of the last vertebra there
is considerable precartilaginous tissue, which has been formed by the fusion of several
scleromeres. At the thirty-first and thirty-second vertebrae the column curves ventrally.
The chorda dorsalis makes a loop in the thirty-fourth vertebra and gives off short branches.
The central canal narrows sharply at the level of the thirty-second vertebra, but expands
again at the dorsal portion of the thirty-third and thirty-fourth. This atrophic portion,
however, is not the primordium of the ventriculus terminalis. Unger and Brugsch com-
pare it with the sinus terminalis found in the amphibian embryo. It is my belief that it
represents the primordium of the coccygeal medullary vestige. The extremity of the
atrophic portion extends into the tip of the tail and is provided with a lumen throughout.
The caudal end of the chorda appears to have adhered to the ventral wall of the spinal
cord. There are 32 spinal ganglia, the thirty-second having slender nerves. The middle
sacral artery and \'ein enter into the tail, anastomosing through branches with the anterior
spinal arterj'.

Embryo No. 405, 26 mm. Crown-Rump Length.

Embryo No 405 has 34 vertebrae, as indicated in figure 18. The last one inclines
dorsally from the axis of the vertebral column. Between the thirtieth and thirty-first
vertebrae the axis of the ventral column shows a decided angle and below this point bends
ventrally, presenting the coccygeal curve which is characteristic of the adult. The chorda
dorsalis presents a spindle-shaped swelling at each intervertebral space and its caudal end
shows a loop-formation in the thirty-third vertebra. The spinal cord narrows at a level
between the thirtieth and thirty-first vertebrae; the atrophic portion, with its narrow
canal, is spiral at its caudal end and enters into the blunt tail-bud. On the dorsal surface
of this spiral part of the cord the epidermis is lacking; whether this is due to mechanical
injury or is a natural phenomenon could not be determined. From the ventral side two
branches of the anterior spinal artery enter. This embryo has 31 spinal ganglia com-
pletely supplied with nerves. The middle sacral artery anastomoses through a branch
with the anterior spinal artery.

Embryo No. 1008, 26 mm. Crown-Rump Length.

Embrj'o No. 1008 has 34 vertebrae, the last being larger than any of the others and
divided incompletely into two segments, as diagranmiatically shown in figure- 19. The
end of the chorda dorsalis presents a number of intricate coils and its caudal extremity lies
against the thick wall of the spinal cord. The spinal cord narrows at the thirty-second
vertebra, thus marking a boundary between the atrophic portion and the upper part of
the cord. The walls of the atrophic portion show two folds — one on the ventral, the
other on the dorsal wall. The former lies in the region of the thirty-fourth vertebra and
is similar to those already seen. The one on the dorsal wall, at a level between the thirty-
second and thirty-third vertebrae, is a diverticulum which projects caudo-dorsally and con-
tains a long, slender cavity continuous with the central canal. Such folds or diverticula
are seldom seen on the dorsal wall of the caudal end of the spinal cord. In the other
specimens studied folds were frequently encountered at this end of the cord, but always on
the ventral wall. On the ventral wall of the upper, wider part of the canal, at a level between
the twenty-ninth and thirtieth vertebrae, is another and much larger fold, extending down
about the length of two vertebrae. Both of its margins fuse with the ventral wall, thereby
forming a channel in the midUne of the fold which unites with the central canal. In this
specimen the spinal ganglia are 31 in number.

176 DEVELOPMENT AND REDUCTION OF THE TAIL

Embryo No. 875, 27 mm. C"rown-Rump Length.

In embryo No. 875 there are 34 vertebrae. The last is small and contains the wind-
ing part of the chorda dorsalis. The spinal cord narrows between the thirty-first and
thirty-second vertebrge, as shown in figure 20. Its caudal end expands sUghtly and the
extreme tip enters into the tail-bud. On the ventral wall of the central canal there are a
few small folds. Near the caudal end of the ^•ertebl•al column is a long, solid strand of
cells, similar in structure to the cells of the spinal cord, which may have become separated
from the latter at an earlier stage. Dorsal to the thirty-third and thirty-fourth vertebrae
is a small papilliform tail, which is non-vertebrated and contains the caudal end of the
vessels and a group of cells representing a remnant of the caudal end of the spinal cord.
There are 31 spinal ganglia with nerve-fibers. The coccygeal tubercle and post-anal
swelling are distinctly evident.

Embryo No. 7.5, 30 mm. Crown-Rump Length.

At the caudal end of embryo No. 75 there is a small papilliform tail containing a group
of cells which merge into the wall of the spinal canal, as shown in figure 43. The spinal
cord narrows suddenly at the mid-level of the thirty-second vertebra, and its atrophic por-
tion is further constricted at a level between the thirty-third and thirty-fourth vertebrae,
as indicated in figure 43 {constrict). The part below this constriction is the primordium of
the coccygeal medullary vestige and the upper part is destined in a later stage to undergo
retrogression, leaving a small cell-sac as a second coccygeal medullary vestige. There
are two large folds on the ventral wall of the spinal cord at a level with the thirty-first
vertebra. In the median plane they are triangular in shape and consist of ependymal
and mesenchymal cells that have been inverted, together with the wall. A large divertic-
ulum lies between these two folds. The space below the folds probably represents the
primordium of the ventriculus terminalis. Only the branches of the anterior spinal artery
enter into these folds. There are 34 cartilaginous vertebrae, and at thirty-first and thirty-
second vertebra? the column presents a typical curve. The chorda dorsalis shows a spindle-
shaped swelling between the vertebrae, and is much convoluted at the caudal end, as
seen in figure 43. There are 31 spinal gangUa; the nerves of the last pair are quite slender.

Embryo No. 145, 33 mm. Crown-Rump Length.

Embryo No. 145 has 35 vertebrae, as diagrammatically shown in figure 22. The
last one is situated on the dorsal side of the axis of the colunm, while the thirty-third and
thirty-fourth lean towards the ventral side. There are 31 spinal ganglia, the thirty-second
pair of nerves having no ganglia. In the caudal region there is a peculiarly shaped rem-
nant of the neural tube, possibly an anomaly of development, which is connected with the
main cord by a cell-strand. This cell-strand is directly continuous with the ependymal
layer of the primordium of the ventriculus terminalis, and may possibly be regarded as
the filum terminale. It emerges from the membranous sheath of the spinal cord, the more
cranial portion branching irregularly, while the caudal portion bends dorsally to enter
the minute tail-bud. The ends of the middle sacral artery and vein enter into the root
of the tail. In this embryo no coccygeal tubercle can be seen.

Embryo No. 211, 33 mm. Crown-Rump Length.

Embryo No. 21 1 has 34 vertebrae and 31 spinal ganglia. The vertebral column curves
ventrally at the thirty-first and thirty-second vertebrae. The caudal end of the chorda
dorsalis is undergoing regression and appears to be branching. The caudal end of the
spinal cord may be divided into three portions: (1) the primordium of the conus medul-
laris, which includes the primordium of the ventriculus terminalis; (2) the filum terminale;
(3) the coccygeal medullary vestige. The first extends about the length of the thirtieth

AND OF THE CAUDAL END OF THE SPINAL COlRD. l77

and thirty-first vertebrae, tapering gradually towards its caudal end. The ventriculus
terminalis, which is included in the conus medullaris, expands in the medial part dorso-
ventrally and transversely. The upper part of this cavity, which marks the entrance of
the central canal, narrows slightly; the caudal end narrows sharply and forms a canal
which terminates blindly at the end of the conus medullaris. The wall of the ventriculus
terminalis consists of gray and white substance and the cavity is lined with a layer of
ependymal cells. The dorsal wall is thicker than the ventral wall. The filum terminale
extends from the caudal end of the conus medullaris, without definite boundaries, to a
level between the thirty-second and thirty-third vertebra?. Its caudal end is represented
by a slender bundle of nerve-fibers, and in its cranial portion there is a strand of ependymal
cells. The large coccygeal medullary vestige is situated dorsal to the thirty-third and
thirty-fourth vertebrae and its wall is thrown into a number of folds. At the caudal end
it has two processes, one extending ventrally, the other dorsally. The latter enters into
a rounded eminence at the caudal end of the embryo which represents a tail-bud, termed
by Unger and Brugsch caudal tubercle. The post-anal swelling is well developed, while
the coccygeal tubercle is scarcely to be made out.

Embryo No. 199, 35 mm. Crown-Rump Length.
In number and development of its vertebrae embryo No. 199 is about the same as No.
972, description of which follows. The coccygeal vestige, however, shows greater expansion.

Embryo No. 449, 36 mm. Crown-Rump Length.
Embryo No. 449 contains only 32 vertebrae, the last one being the smallest, as indi-
cated in figure 24. The vertebral column shows a slight ventral curve at the point between
the thirtieth and thirty-first vertebrae. The chorda dorsalis exhibits no convolutions at
its caudal end. The spinal cord narrows at a level between the twenty-ninth and thirtieth
vertebrae and the ventral wall of the atrophic portion presents a few folds. The remnant
of the medullary tube was not found in the caudal region of this specimen. There are 31
spinal ganglia with nerves.

Embryo No. 972, 37 mm. Crown-Rump Length.

Embryo No. 972 has 34 vertebrae, the last two having fused together at the center.
The vertebral column curves ventrally at the thirtieth and thirty-first vertebrae, the curve
being so sharp that the thirty-second, thirty-third, and thirty-fourth vertebrae are situated
in a row nearly horizontal to the trunk, as can be seen in figure 44. Cranial to thirty-first
the chorda dorsalis expands between the vertebrae. The caudal end is winding and broken.

At the caudal end of the spinal cord one can recognize the primordia of the conus
medullaris, filum terminale, and coccygeal medullary vestige, as shown in figure 44. The
primordium of the ventriculus terminalis, which is included in the conus medullaris, appears
as a continuation of the central canal of the spinal cord without any line of demarcation,
and is situated at a level with the twenty-ninth vertebra. Its ventral wall is thinner
than the dorsal wall and shows a few small folds (fig. 44, x). The primordium of the filum
terminale extends from the caudal end of the conus medullaris, viz, at the level of the
under part of the thirtieth vertebra, to the middle of the thirty-second vertebra. It con-
tains an incomplete canal which is lined by a remnant strand of ependymal cells which
are directly continuous with the ependyma of the ventriculus terminalis. At its caudal
end is an ependymal strand which is directly continuous with the coccygeal medullary
vestige. In addition to this ependymal substance, there is a small bundle of nerve-fibers
along the ventral border of the filum terminale which extends into the white substance of
the cord above. The primordium of the coccygeal medullary vestige is situated dorsal
to the last two vertebra and contains a slender cavity. There are 30 spinal ganglia sup-
plied with complete nerves. The thirty-first ganglion has almost completely disappeared
on each side, leaving the nerves exposed.

178 DEVELOPMENT AND REDUCTION OF THE TAIL

Embryo No. 362, 39 mm. Crown-Rump Length.

Embryo No. 362 has 34 vertebrae, the last one being situated on the dorsal side of the
vertebral axis, as is diagrammatically shown in figure 26. At the thirtieth and thirty-
first vertebra? the column is bent ventrally. The caudal end of the chorda dorsalis winds
and branches in the last three vertebrae. There are 30 spinal ganglia with nerves, but the
thirty-first pair of nerves has no ganglia, their degeneration probably having occurred
before that of the nerves. The caudal portion of the spinal cord is divided into the conus
medullaris, filum terminale, and coccygeal medullary vestige. The ventriculus terminalis,
which is included in the conus medullaris owing to the folding of its walls, is subdivided
into two parts — an upper part, triangular in shape, and a lower, which is oblong and com-
municates with the upper by a narrow channel. The filum terminale reaches from the
caudal end of the conus medullaris to the ventral side of the coccygeal vestige, being
enveloped by the membrane of the spinal cord, the dura mater. This embryo presents a
small tail-bud at its caudal end, containing a group of cells which connects with the
coccygeal medullary vestige.

Embryo No. 95, 50 mm. Crown-Rump Length.

Although embryo No. 95 is recorded in the catalogue of the Carnegie Collection as
46 mm. crown-rump length, its state of de\'elopment more nearly corresponds with a 50
mm. embryo, and on this account I have used the latter measurement in the heading.
This specimen has 35 vertebrae. The last one is very small and partly fused with the one
above it. The column presents a ventral bend at the thirty-first vertebra, giving the typi-
cal coccygeal curve. The chorda dorsahs is disappearing in certain areas in the vertebral
bodies as far down as the thirtieth vertebra, but in each intervertebral space a fragment
remains. Caudal to the thirtieth vertebra the condition of the chorda remains the same
as in the younger specimens, and in the thirty-second it gives off a short dorsal branch.
The caudal end is more simple in form than in the younger stages, but I am inclined to
believe that at an earlier stage it too was winding, as one can see in the thirty-fifth vertebra
a few detached globules which probably at an earlier stage were continuous with the chorda
and with it formed a terminal looj).

At the caudal end of the spinal cord are two groups of cells connected by a cell-strand.
The more caudal one is situated dorsal to the thirty-fourth and thirty-fifth vertebrae; it is
somewhat larger than the other, is oblong in form and incloses an oval cavity — a fragment
of the central canal of the spinal cord. The other group of cells is situated dorsal to the
thirty-second and thirty-third vertebrae and incloses a long, narrow ca\'ity. The ventric-
ulus terminalis extends the length of two vertebrae — the twenty-ninth and thirtieth. At
this stage it has acquired its adult form. In none of the earlier specimens have I noted it
so perfectly developed, although embryos No. 449, 36 mm., and No. 199, 35 mm., show
a cavity at the caudal end of the central canal as the primordium of the ventriculus. In
this specimen the structure is cylindrical in shape, has six walls, and mea.sures 0.87 mm.
long, 0.23 mm. deep, and 0.52 mm. wide. The ventral wall is concave, the dorsal convex,
the sides slightly concave. The upper wall or ceiling is irregular and at the front presents
a long, narrow diverticulum directed cranio- ventral. Behind this diverticulum is a nar-
row channel which connects the ventriculus terminalis and the central canal of the s])inal
cord. The ^•entriculus terminalis is embedded in the nerve-fibers of the cord. The filum
terminale extends from the caudal end of the conus medullaris, at the level of the thirty-
first vertebra, to a point between the thirty-third and thirty-fourth vertebrae, close to the
column. It is covered by a membrane of the spinal cord ami passes through the ventral
side of the cell groups at the caudal end of the medullary tube. The pia mater covers
closely the whole surface of the spinal cord; it contains blood capillaries, and is visible at
the conus medullaris. The dura mater, which envelops loosely the pia mater, adheres

AND OF THE CAUDAL END OF THE SPINAL CORD. 179

to the wall of the vertebral canal as far as the midlevel of the thirty-first vertebra, at which
jinint it leaves the wall and unites with the caudal end of the conns medullaris. This
portion constitutes the primordiuni of the bursa dura:- matris. After the dura mater reaches
the conus medullaris it envelops the pia mater quite closely, both following a caudal course
and forming a sheath for the filum terminale. The point at which these membranes
terminate can not be definitely decided. It is probable that the pia mater extends nearly
to the end of the filum terminale between the thirty-third and thirty-fourth vertebrae.
The fibers of the dura mater appear to enter into the caudal and dorsal portions of the last
vertebra.

Embryo No. 184, 50 mm. Crown-Rump Length.

Embryo No. 184 has 34 vertebrae, the last one being the smallest, as is indicated in
figure 28. At the thirty-first vertebra the column presents a ventral curve, bringing the
thirty-second, thirty-third, and thirty-fourth vertebrae in about a horizontal row and at
right angles with the main column. The chorda dorsalis is disappearing in the 29 upper
vertebral bodies, but at the thirtieth and below there is no change from the earlier stages,
except that the chorda is relatively more slender. Its caudal end is bent caudo-dorsally
before terminating; from this point the caudal ligament takes its origin. The middle sadral
artery at this stage is a relatively delicate vessel, running from the ventral to the dorsal
side of the vertebral column, and curving about the apex of the thirty-fourth vertebra.
Its branches are plexiform, and in their meshes are groups of cells resembling neuroblast
cells. The caudal end of the spinal cord contains a large cavity representing the ventric-
ulus terminalis at a more advanced stage of development. The upper end of this cavity
connects with the central canal of the spinal cord; its lower end ternxinates in two horns,
the dorsal one of which is a blind pouch; the ventral horn is united with the caudal rem-
nant of the spinal cord by a strand of ependymal cells and many transverse folds. The
caudal remnant of the spinal cord consists of three separated portions. The first, which
is attached to the caudal end of the ventriculus terminalis by an ependymal cell-strand,
lies between the thirtieth and thirty-first vertebrae. This portion is embedded in nerve-
fibers. As in younger specimens, it incloses a narrow cavity interrupted about midway.
The second portion of the remnant is situated between the thirty-first and thirty-second
vertebrae and leans to the dorsal side of the filum terminale. It also contains a small
lumen. The third and largest portion is situated at the level of the thirty-third vertebra ;
its cavity is larger than the others and its caudal end enters into the caudal ligament.

The pia mater envelops the spinal cord and contains blood capillaries. It traverses
the course of the filum terminale, completely inclosing it, and appears to reach the dorsal
portion of the thirty-third vertebra, at which point the filum terminale ends. The dura
mater also covers the spinal cord over the pia mater. At the caudal end of the conus
medullaris, about the thirtieth vertebra, the dura mater adheres closely to the pia mater.
At the dorsal side of the thirty-third vertebra the fibers of the dura mater merge with the
fibers of the caudal ligament.

This embryo has .31 spinal ganglia on the right side and .30 on the left. The last
ganglion on either side is very small, being in process of retrogression. The right thirtieth
and thirty-first ganglia and the left thu-tieth are not located between the vertebrae, but
at the dorsal side of the upper vertebral bodies.

Embryo No. 448, 52 mm. Crown-Rump Length.

A profile reconstruction of the caudal end of embryo No. 448 is shown in figure 4.5
and a more diagranmiatic sketch is shown in figure 29. The embryo has 34 vertebrae, the
last of which is only three-fourths the size of the thirty-third. The last three have begun
to fuse, so that a section cut through the axis of the vertebral column shows one large
vertebral body representing the three vertebrae, as shown in figure 45. The vertebral

180 DEVELOPMENT AND REDUCTION OF THE TAIL

column is bent ventrally between the thirtieth and thirty-first vertebrae, forming an obtuse
angle and creating the typical coccygeal curve. Within the bodies of the vertebrae, from
the first to the twenty-ninth, the chorda dorsalis is disappearing, but a remnant still
remains in each intervertebral space. From the thirtieth to the thirty-fourth vertebrae it
continues without convolutions, but the caudal end is branched and winding, partially
disappearing at the dorso-caudal portion of the last vertebra close to the remnant of the
spinal cord. The spinal cord tapers to a point as the conus medullaris and proceeds as the
filum terminale from a level between the thirtieth and thirty-first vertebrae. Four por-
tions of the neural tube can be distinguished at the caudal end of the spinal cord: (1)
the sacral region of the spinal cord; (2) the conus medullaris and its contained ventriculus
terminalis; (3) the filum terminale; (4) a remnant. The first consists of the ependymal
zone, the mantle zone which contains the germinating nerve-cells, and the marginal zone,
as is typical for the cord as a whole. The conus medullaris extends from the twenty-eighth
to the thirtieth vertebrae, tapering gradually. In this region there is a large cavity, which
in a median sagittal section shows four walls. Through the front of the upper wall the
ventricle joins with the central canal of the spinal cord. The lower wall is narrow and
from it extend two ependymal cell-strands. The longer of these goes straight downward
to the first cell-group of the remnant of the medullary tube, through the axis of the conus
medullaris. The shorter strand can be seen at the corner between the lower and ventral
wall in figure 45. At the ventral wall is a diverticulum, the entrance to which appears
as a narrow stalk consisting of a solid cord of ependymal cells and connecting with the
ependymal cells of the cavity. This diverticulum is divided into two parts which are united
by a cell-strand — a small upper sac and a larger lower sac. The ventral walls of both
sacs are situated close to the surface of the conus medullaris, but do not open into it. The
lower part of the conus medullaris consists chiefly of nerve fibers of the spinal cord, and
here the central canal is entirely obliterated, leaving a long strand of ependymal cells.
The conus medullaris extends to a point between the thirtieth and thirty-first vertebrae,
and from there continues as the filum terminale, which extends to the last vertebra, skirt-
ing close along the dorsal side of the vertebral cohunn. At the dorsal side of the filum
terminale there are two remnants of the primitive neural tube. One of these is situated
just dorsal to the apex of the conus medullaris. It contains a slender lumen, the remains
of the central canal of the spinal cord. The other remnant (fig. 45, ves. m. co.) is situated
dorso-caudal to the thirty-third and thirty-fourth vertebrae. It is oblong in shape and
likewise contains a cavity, somewhat larger, which represents a remnant of the central
canal. Its caudal end is sharp and fuses with the caudal ligament. The latter is not so
distinct in this specimen as in the younger ones.

The caudal end of the sympathetic nerve-trunk lies between the middle sacral artery
and vein, the three passing along the ventral side of the thirty-third and thirty-fourth verte-
brae, where they curve around the apex of the last vertebra. The caudal ligament forms at
the caudal end of the thirty-fourth vertebra and extends dorso-cranial to the coccygeal
vestige. The caudal portions of the sympathetic trunks unite ventral to the thirtieth
vertebra and become as one. After the union of these cords two additional ganglia can
be seen— one at the thirty-first, the other at the thirty-third vertebra. From the latter
the sympathetic nerve-trunk follows the midline of the vertebral column and curves around
the last vertebra to the dorsal side, as shown in figure 45. The condition of the dorso-
caudal portion of the nerve-trunk can not be clearly recognized.

The pia mater covers entirely the surface of the spinal cord and is rich in blood capil-
laries. It also envelops that portion of the filum terminale containing the cell-groups
which connect with the ependjTiial cells of the ventriculus terminalis. The dura mater
traverses the wall of the vertebral canal enveloping the spinal cord and its covering of pia
mater. In the caudal region of the spinal cord there does not appear to be a distinct space

AND OF THE CAUDAL END OF THE SPINAL CORD. 181

between the pia mater and dura mater and hence the arachnoid membrane is not visible
at this point. A short distance from this, however, where the membranes envelop the
conus medullaris, there is a marked space between the two membranes and here the arach-
noid can be fairly well made out, forming a fibrous network of embryonic connective-
tissue. At the level of the caudal third of the thirtieth vertebra where the filum terminale
begins, the dura mater fuses with the pia mater and the two become separated from the
wall of the vertebral canal and extend caudalward. The second group of cells, which
lies caudo-dorsal to the thirty-third vertebra, does not seem to be covered by the pia mater
or dura mater, these membranes having disappeared a short distance above.

Embryo No. 1656, 67 mm. Ceown-Rump Length.

There are 34 vertebrae in embryo No. 1656, the last being the smallest. At the thirty-
first and thirty-second the vertebral column shows a ventral curve, the angle being sharper
than in the younger specimens. The vertebrse are separated by embryonic tissue which
is to develop at a later stage into intervertebral fibro-cartilage. This separation becomes
progressively more marked above the thirtieth vertebra. Between the vertebrae which
still lie close together is a small sjiace where the chorda dorsalis coils as it emerges from the
vertebral bodies in the median line. Several of these coils can be seen in figure 46, which
is a profile reconstruction through the caudal end of the embryo. The blood-vessels enter
the vertebral bodies from the ventral and dorsal side.

In the conus medullaris there are two medullary ventricles. The more cranially
situated one is somewhat smaller than the other, measuring 0.55 by 0.25 by 0.33 mm.
Its form, as seen in the sagittal plane, can be recognized in figure 46 (vent. t. cran.). The
lower cavity is oblong in shape, measures 1.1 by 0.3 by 0.36 nmi., and presents a canal-
like appendage 1.7 mm. in length, as seen in figure 46 {Append.). This appendage tapers
to a point and continues as a cell-strand. Toward the caudal end of the strand, in the
path of the filum terminale, are two small groups of cells which represent the remnants
of the ependymal cells of the medullary tube (fig. 46, Re. epend.).

The phenomenon of dedifferentiation at the caudal end of the spinal cord is well
shown in this specimen. The appendage of the lower cavity was a complete ventriculus
terminalis at the first stage ; the main body of the cavity was a complete one at the second
stage, and the upper cavity is the ventriculus terminalis at the present stage, thus showing
a progressive upward trend. The gray substance which primarily existed around the ven-
triculus terminalis has now disappeared as the result of degeneration, and the caudal end
of the central canal has gradually enlarged. The caudal end of the lower cavity, however,
is becoming gradually narrow because the caudal portion of the conus medullaris, which
contains the ventriculus terminaUs, has also gradually become atrophied and lost its cell-
like substances. The septum between the two ca\aties is a remnant of the gray substance
of the spinal cord, in which the degeneration is not yet complete.

The filum terminale follows a downward course from the end of the conus medullaris
and nerve-fibers can be recognized as far down as the caudal portion of the thirty-second
vertebra. In the caudal region are found two cell-groups representing remnants of the
neural tube ; one, which lies between the thirty-second and thirty-third vertebrae, contains
no lumen, and the epithelial cells are undergoing degeneration. The other is situated
dorsally between the thirty-third and thirty-fourth vertebrae and incloses a small lumen.

The membranes of the spinal cord are more easily made out in this specimen than in
the younger ones. The dura mater is separated from the periosteum of the vertebral
bodies, especially at the ventral wall of the vertebral canal, by a dense plexus of blood-
vessels, connective-tissue, and small spaces. This separation occurs at a level between
the twenty-seventh and twenty-eighth vertebrae, and the dura mater becomes adherent
to the conus medullaris between the twenty-eighth and twenty-ninth vertebrae, following an

182 DEVELOPMENT AND REDUCTION OF THE TAIL

oblique course from the periphery to the center of the vertebral canal. There is thus laid
out the early form of the dural sac. Outside of this sac the fibers are separated into tufts
which run parallel and caudalward. In the space between the dural sac and the conus
meduUaris the arachnoid membrane can be seen developing. The pia mater envelops
closely the spinal cord and supports the blood-vessels; between the twenty-fifth and twenty-
eighth vertebrae it is separated from the dura mater and the arachnoid by a still wider space .

Embryos No. 662, 80 mm. Crown-Rump Length; No. 928, 100 mm. Crown-Rump Length;

No. 142, 125 mm. Crown-Rump Length.

As the investigation with embr>'os Nos. 662, 928, and 142 was not ver>^ satisfactory,
I shall not attempt to give its results in detail at this time. I have, however, made a special
study of the coccygeal medullary vestige because of its importance in comparison with
the same structure in younger specimens. In the 80 and 100 mm. embryos the coccygeal
vestige is very small and its contained cavity narrower than in the younger specimens.
In the 125 nun. embryo (negro) the structure is well developed and shows one long offshoot
stretching under the epidermis at the sacral region. It is quite different in form and
condition from the case reported by Tourneux (Precis d'embrj^ologie humain), and there-
fore does not present the loop formed by a more deeply situated limb {segment coccygien
direct) and a more superficial limb (segment coccygien refleche). In this embryo the
coccygeal vestige contains a slender cavity.

DEVELOPMENT AND REDUCTION OF THE TAIL.

In considering the process of reduction of the tail I should like, in the first
place, to refer to the important study of this condition in mammals made by Braun
(1882), w'hose conclusions in general are as follows:

(1) The tail of the mammalian embrj'o consists of two portions — a vertebrated part
and a non-vertebrated part, the latter situated caudal to the former.

(2) The non-vertebrated part appears usually in the form of a thread at the end of
the vertebrated tail, and consequently may be designated the caudal filament {Schwanz-
faden). Being usually thinner than the tail itself, it is consequently sharply marked off
from the latter.

(3) The vertebrated part of the tail can again be subdivided into two parts according
to whether it projects from the body or not. The projecting part is designated as tail,
although it has long been well known that this is directly continuous with the sacral verte-
brae. The relative size of the internal and external tail varies, and hence we meet with
long-tailed, short-tailed, and tailless mammals.

(4) The caudal filament is a transitory' structure, although for a time it contains the
end of the spinal cord, the chorda dorsalis, and the caudal gut. These structures undergo
resorption and the last tissue to persist is the epidermis, the caudal thread for a time per-
sisting of only epidermis cells.

(5) The caudal gut originally extends into the tail; before being resorbed it separates
into fragments which disappear, the last to persist being the part constricted off at the tip
of the tail.

(6) The chorda dorsalis always projects beyond the caudal vertebrae, where it separates
into forked processes or curls in irregular loops. This part disappears completely.

(7) The spinal cord originally extends to the tip of the tail. The latter, however,
soon exceeds it in length, when it terminates at the base of the caudal filament. It was
possible to show in sheep embryos that the ascensus meduUae is due not alone to the over-
growth of the vertebrae, but that also there is degeneration and absorption of the caudal
end of the spinal cord, to which in part the formation of the filum terminale owes its origin.

AND OF THE CAUDAL END OF THE SPINAL CORD. 183

It would appear, therefore, that in his subdivision of the mammahan embryonic
tail Braun included in the caudal filament that portion which lay between it and
the vertebrje. JNIy own position on the subject is briefly this: Is the caudal fila-
ment, through all the stages of mammalian embryonic life, one and the same thing as
the non-vertebrated tail? That an intermediate portion exists between the two was
apparently not recognized by Braun, but in man it constitutes a most important
factor in the reduction of the tail vertebrse. After detailed investigation with the
material at hand, numbering about 40 embryos ranging from 4 to 50 mm. in
length, I was able to divide the caudal structure as follows: (1) the vertebrated
portion; (2) the mesodermic end portion. In somewhat older embryos the first
is subdivided into a proximal portion with persisting vertebrae, and a portion from
which the primitive vertebra? have disappeared (lost-vertebrae portion). This
point can better be understood by referring to the embryos themselves.

In embryo No. 221, 7.5 mm. long, the tail contains 38 somites and a long
mesodermic remnant. The somites, which later develop into precartilaginous
vertebrae, are well defined by the presence of small blood capillaries between them.
On the dorsal surface of the tail the boundaries of the somites can be recognized
distinctly as transverse shallow grooves. In the last somite, which is in contact
with the mesodermic remnant, the boundary is not nearly so clear as in the others
and would probably have disappeared altogether in the retrogressive process had
the embryo hved. In this specimen the tail is entirely a vertebrated tail, as each
somite is capable of development into a vertebra. The long mesodermic remnant
at the caudal end, although separated by segmentation from the mesodermal sheet,
evidently would not have developed into precartilaginous tissue. This part I
have differentiated from the vertebrated tail as a mesodermic remnant, using the
term employed by Keibel. I was able to distinguish in this embryo, therefore,
two divisions of the tail — a long somitic portion, the vertebrated tail, and a short
mesodermic portion, the caudal end of which may be compared with the caudal
filament but not with the lost-vertebra? tail. In this stage the non-vertebrated
portion has not as yet developed — that is, the portion in which the somites or precar-
tilaginous vertebrae have disappeared. The last somite, however, shows signs of
disappearing, and after a time, therefore, the lost-vertebrae portion will appear in
place of the last somite. In other words, the reduction phenomenon has begun
in the last somite; this progresses in the tail from one somite to the other, each
losing its distinct boundaries, the blood capillaries fusing and disappearing.

In somewhat older specimens (8 mm., fig. 33), the last somite is larger than
the iireceding one and evidently rejjresents the fusion of two pieces — the thirty-
eighth somite and the mesodermic remnant. At the 1 1 mm. stage the lost-vertebrae
portion of the tail becomes well developed (fig. 34). In the 12 and 15 mm. em-
bryos the boundaries of the thirty-sixth, thirty-seventh, and thirty-eighth somites
have become indistinct. In the 15.5 mm. specimen these three somites are con-
verted into a cord which extends to the end of the tail (fig. 36, s(r. cell). This
cord consists of embryonic cells wliich at an earher stage of development existed
in the somites as sclerotomes. In the median portion of the cell-strand are three

184 DEVELOPMENT AND REDUCTION OF THE TAIL

or four segments, and the last vertebra also shows two or three divisions. At this
stage three types of vertebrae can be recognized at the caudal end of the vertebral
column: (1) the vertebrae which have developed from the sclerotomes into precar-
tilaginous or primitive vertebrae; (2) the incomplete primitive vertebrae, or the
parts of the thirty-sixth and thirty-seventh sclerotomes which form the last verte-
brae; (3) the cell-strand formed by the fusion of the last two somites and perhaps
the thirty-sixth as well. I have not been able to determine whether or not the
mesodermic remnant has merged into this strand. This mesodermic cell-strand —
the primordium of the caudal ligament — is diagrammatically shown in figures 36
and 37 (str. cell). In the 16, 17, and 18 mm. embryos the last vertebra is larger
than the more proximally situated ones and consists of two or three pieces united
in the median plane; 35 vertebrae, developed into precartilage or cartilage tissue,
were found in the 15.5, 17, and 18 iimi. embryos, and 36 in the 16 mm. embryos.
In those 17 mm. and larger the last two or three primitive vertebrae were usually
found to be fusing at the center of the column, while in the lateral parts they show
the divisions quite distinctly.

His did not find any extra vertebrae in the tail, but many other authors have
recognized from 2 to 4. Perhaps the material upon which His based his studies
did not include specimens in the same stages of development as my 15.5, 17, and 18
mm. embryos, in which the last vertebra consisted of two or three pieces. The 21
mm. embryo also represents the typical condition at this stage, the tail showing
extra vertebrae. In this embryo can be clearly demonstrated a short tail consisting
of two portions, such as has been described by His, the extreme non-vertebrated
or, more correctly speaking, the lost- vertebrae portion, and the vertebrated por-
tion. I am sure that embryos of this age never present a caudal filament
homologous with that of other mammals, and I can not therefore agree with His,
Braun, Keibel, linger, and others, who describe the non-vertebrated portion of
the tail in the human embryo as a caudal filament, since this portion at an earUer
period contained somites capable of development into primitive vertebrae. Ecker,
who studied human and mammalian embryos, asserted that the human embryo
never has the caudal filament such as is the rule for mammalia. I could not recog-
nize clearly a caudal filament in any of my specimens. In one of the three embryos
(6.5, 7.5, and 8 mm.) which showed a portion of the caudal gut in the end of the
tail, the caudal end was demarcated by a bend, and this might have been mistaken
for a caudal filament. I am of the opinion that in the 8 and 10 mm. embryos the
portion of the tail beyond the pointed end of the vertebral column, as shown in
figiu-e 34, can be compared with the caudal filament in mammals, but is not the
true caudal filament described by Braun. It consists of the caudal end of the
mesodermic remnant and contains the end of the neural canal and, in the 8 mm.
embryo (fig. 33), a part of the caudal gut. I beheve, also, that the non-vertebrated
portion of the mammalian tail, which is not included in the caudal filament, is
homologous with the non-^'ertebrated tail of the human embryo.

Having compared the non-vertebrated portion of the mammalian tail with
that of human embryos, I have concluded that in the former the reduction process

AND OF THE CAUDAL END OF THE SPINAL CORD. 185

occurs at an early stage, just as it does in the human embryo. This theory is
based upon two facts: First, the last caudal vertebra is larger than the next
proximally situated one, as in the cow embryo described by Braun, its increased
size being due to the fusion of the last three pieces. This phenomenon represents
the reduction process in the lower segments at a certain stage. Second, the num-
ber of vertebrae in the tail of sheep embryos, as asserted by Braun, is variable, and
this variation must be due to a stronger or weaker effect of reduction.

In the 19 and 23 mm. specimens there is a very short tail with a caudal tubercle.
In the 19 mm. embryo the vertebrae of the tail have fused together into one large
vertebra — the thirty-third — in which can be recognized two or three pieces. The
caudal end of the chorda dorsahs within this vertebra shows several coils, indi-
cating a fusion of the last few vertebrae. The lost-vertebrae portion of the tail is
represented in this specimen by the caudal extremity which contains the ends of
the neural canal, the middle sacral artery and vein, and the chorda dorsalis. The
latter adheres to the ventral wall of the neural canal and it appears as if the neural
tube is retracted cranialward. In the 23 mm. embryo the development of the
caudal end is farther advanced. The furrow between the tail root and the primitive
anus becomes gradually more shallow, and the vertebral portion of the tail is em-
bedded in the embryonic tissue which will later develop into the coccygeal tubercle.
This shortening of the tail is evidently brought about by three factors: (1) fusion
of the last few caudal vertebrae; (2) rapid growth of the aUmentary canal and its sur-
rounding structures; (3) the flexion of the caudal portion of the vertebral column.

(1) The disappearance of the last few caudal vertebrae by fusion, leaving only
the winding end of the chorda dorsahs, is a well-estabhshed proof of the compres-
sion and final disappearance of the caudal vertebrae and the chorda dorsalis which
was within them, linger and Brugsch took the view that in spite of the presence
of an external tail one could still speak of the formation of a coccygeal tubercle,
inasmuch as the segments of the caudal region, which in their most caudal portion
are already reduced, have begun to show a moderate, ventrally directed curve in
their axis, which is eventually to be the coccygeal eminence. They point out
that two factors are of importance in the formation of the coccygeal eminence:
(a) the fusion (reduction) of the most caudal segments; (b) the bending in the axis
of the caudal vertebrae. In the 25 and 27 mm. embryos the lost-vertebrae portion
of the tail becomes rounded off and is shown as the caudal tubercle. Its extremity
appears as a bud-shaped appendage and contains the caudal ends of the spinal
cord, with its central canal, and of the middle sacral artery and vein. This bud-
Uke appendage is called by many authors the caudal filament, but this is incorrect
for the reason, as stated above, that it represents only a part of the lost- vertebrae
portion of the tail which was primarily the vertebral portion, and therefore could
never be considered as the caudal filament described by Braun.

(2) The area between the vertebral column and the rectum, especially the
root of the tail, increases rapidly in a caudo-ventral direction. The caudal region
of the rectum also extends down, its growth being in proportion to that of the
vertebral column (figs. 39, 40, and 43). It is the beUef of many authors— Rosen-

186 DEVELOPMENT AND REDUCTION OF THE TAIL

berg, Ecker, Keibel, and others — that the tail and the coccygeal tubercle in human
embryos become shorter and finally disapj^ear by an increase in volume of the
caudal soft tissues, muscular tissue, subcutaneous connective-tissue, etc., which
surround the caudal part of the ^'ertebral column.

At an earlier stage the swelling between the primitive anus and the root of the
tail is called the post-anal swelling. Keibel asserts that in embryos 11 mm. and
larger the root of the tail is separated from the ventral trunk by double j^lates of
epithelial cells which lie between it and the anus. Therefore, by means of these
plates, which consist of two sheets of epidermal cells connected ventrally to the
post-anal epidermis and dorsally to the ventral side of the tail, the tail-root is dis-
tinctly marked off. Following Keibel's idea, Tourneux speaks of it as depression
sous candale de Vintecjument externe. Unger and Brugsch describe the stages of
disappearance of this post-anal swelling in embryos 25 and 45 mm. long. In my
specimens it is quite clear. In the 13 and 14 mm. embryos these plates are visible,
but in the 45 mm. specimen they have disappeared. After observing the specimens
in the various stages, my conclusions on this point are as follows: At a certain
stage (13 mm. and older) the digestive tube grows more rapidly than the vertebral
canal, so that the depression gradually straightens out. At this time the caudal
region of the digestive tract — viz, the cloacal region — develops faster than the
caudal end of the vertebral column, which constitutes the caudal end of the
internal tail. Moreover, the mesodermic tissue between the primitive anus and the
root of the tail develops rapidly and gradually bulges downward. By the swell-
ing of the caudo-ventral region of the tail-root the fold of epidermis, or so-called
epithelial plate, is stretched by degrees and at last disappears. The growing of
the coccygeal tubercle would also aid in this process. In his paper Keibel asserts
that the mesodermic tissue between the primitive anus and the tail-root grows
luxuriantly at certain stages and bulges downward. He terms this swelling die
■postanalen wulst (post-anal swelling). In this way the epithelial plates disappear.
This epidermal plate between the anus and tail-root moves gradually caudal ward.
In the 12 mm. embryo it is situated at the level of the thirty-third vertebra, and in
the 46 mm. specimen has moved down to the level of the thirty-fourth vertebra.
The caudal end of the rectum — viz, the caudal end of the digestive tract, and perhaps
that of the genito-urinary organs as well — has likewise moved caudalward.

(3) Originally the caudal portion of the vertebral column is nearly a straight
line, but in embryos about 20 mm. long the axis of the column shows a distinct
ventral flexion at about the level of the thirtieth or thirty-first vertebra. There
Is a second flexion which is dorsal at the caudal end of the vertebral column,
between the vertebrated portion and the lost-vertebrse portion of the tail, which
is seen in younger embryos. The caudal remainder of the lost-vertebrse tail has,
therefore, moved to the dorsal side of the vertebral column, being joined to the
last vertebra by bands of embryonic connective-tissue. These bands are the so-
called caudal Ugament. In these embryos the ridge or epidermal plate between
the coccygeal tubercle and the rectum has become shallow almost to the point of
disappearance, as shown in figures 40 and 42. In the 30, 33, and 39 mm. embryos
the lost-vertebree portion of the tail has almost entirely disappeared from the sur-

AND OP THE CAUDAL END OF THE SPINAL CORD. 187

face of the skin. In the first embryo the tail remnant is surrounded by a minute
fvuTow, while in the 33 mm. specimen it appears as a rounded eminence; and finally,
at the 39 mm. stage, the remnant of the tail is represented by a small papilla.
These remnants contain groups of cells from the primitive neural canal. The
apex of the caudal conical eminence, the caudal tubercle, according to linger and
Brugsch, in which the cell-strand of the neural canal enters, is a part of the lost-
vertebrse tail, or so-called non- vertebra ted tail. The various stages in the reduc-
tion of the tail as shown on the skin surface do not present the same appearance in
every embryo; but on section evidences of its reduction and disappearance are
invariably found dorsal to the caudal end of the vertebral column — that is, dorsal
to the coccygeal tubercle and in the median line of the embryo. I have seen no
case in which the remnant of the tail is situated just at the caudal end nearly in
line with the extended axis of the vertebral column — namely, at the top of the
coccj'geal tubercle. At this stage the caudal ligament is well developed and con-
sists of bands of connective-tissue. The curve of the vertebral column is quite
jM-ominent at the thirtieth, thirty-first, and thirty-second vertebrae. This flexion
of the caudal portion of the column begins at about the 25 mm. stage, although
sometimes it does not appear until the 33 mm. stage. These embryos show a small
tail at the caudo-dorsal end. In the 30 mm. embryo (fig. 43) and the 39 mm.
embryo, where the tail is disajjpearing from the surface of the skin, this curving
of the vertebral column becomes more marked than in the younger specimens.

Concerning the disappearance of the tail in the human embryo, I am of the
ojjinion that, while the lost-vertebne portion of the tail disappears from the skin
surface, a few vertebrse of the tail fuse with the one above, u.sually the thirty-fourth,
a part of which disappears by dedifferentiation; and that the caudal jjortion of the
column, which consists of the thirty-first, thirty-second, thirty-tliird, and thirty-
foiu'th vertebrae, is bent to the ventral side, sinking into the embryonic tissue
between the rectum and the coccygeal tubercle. After the tail entirely disappears
there appears outside of the ventral region of the tail root a blunt conical eminence
known as the coccygeal tubercle, or eminentia coccygeal^. This tubercle is a tempo-
rary swelling formed by bulging of the caudal end of the vertebral column and the
addition of embryonic tissue contained in the lost- vertebrae tail at an earlier stage.
The tubercle disappears at some time between the 33 and 52 mm. stage, while the
caudal end of the vertebral column, the so-called internal tail (after Braun), sinks
deeper into the soft tissues which surround and envelop it (figs. 44 and 45).

In describing the embryonic tail in mammals, Braun divides it into internal
tail {die innere Schwanz) and external tail (aussere Schwanz). This theoretical
arrangement may be the better one. In the human embryo, at least in my speci-
mens, it can be clearly demonstrated. In an embryo of 21 mm. one can recognize
the external tail, which may be divided into two portions — vertebrated and non-
vertebrated. In the 27, 30, and 39 mm. specimens the thirty-first, thirty-second,
thirty-third, and thirty-fourth vertebrae belong to those of the internal tail. I
agree with other authors that the human embryo has a true tail at a certain stage
of its development and that the second coccygeal vertebrae and those caudal to it
in the adult are the true tail vertebrae in the philogenetic sense.

188 DEVELOPMENT AND REDUCTION OF THiE tAlL

CHORDA DORSALIS.

In the 4 mm. embryo the chorda dorsahs hes close to the ventral side of the
neural tube, but cranial to the twenty-first segment it is separated from the tube
by the tissue of the primitive vertebrae. At this stage it forms a long, narrow tube,
its caudal end consisting of only a small cell-strand which terminates in a cell-mass
above the caudal extremities of the neural tube and caudal gut. In the 5 mm.
embryo, which is shown in figure 31, the chorda cranial to the thirty-third somite
is separated from the ventral side of the neural tube, while caudal to the thirty-
third the two are contiguous. The chorda terminates in the mesodermic remnant,
being covered by the ventral wall of the neural tube, and at its caudal end is united
with that of the caudal gut by a cell-strand. In specimens 7.5 to 11 mm. the
greater part of the chorda dorsahs cranial to the thirty-fourth or thirty-fifth somite
is embedded in the primitive vertebral column and shows considerable winding.
Caudal to the thirty-first somite the chorda is placed between the neural canal
and the primitive vertebral column. In passing down through the column it shows
a series of segmental undulating curves — that is, it alternately bends ventrally and
dorsally. The dorsal bends occur at the foci of vertebral formation which eventu-
ally become the intervertebral spaces. In older embryos— 12 to 14 mm. — this
segmental undulation of the chorda gradually disappears. In the 12 mm. embryo
(as shown in figure 35) the chorda is more completely embedded in the column,
although here its terminal portion emerges to he in the space between the spinal
cord and the tissue of the column. As the embryo advances in age this bending of
the chorda gradually decreases, until at about the 18 mm. stage it becomes straight
in its main portion, while the caudal part, which was hitherto straight, now becomes
curved, the first indication of the formation of undulations (compare figs. 35, 36, 37,
and 39) , which, however, are probably not segmental like those above described, but
are a phenomenon of the process of reduction in the caudal primitive vertebra?.

When we compare the 7.5, 8, and 11 mm. embryos with those from 15 to 19
mm. it is easy to see that at first the few caudally situated primitive vertebrae—
the scleromeres — fuse together, and the chorda which is within them becomes
convoluted and recedes cranialward. The winding portion of the chorda is, there-
fore, situated in the last vertebra which has developed by the fusion of several
vertebrae. This condition remains the same up to the 18 mm. or even more ad-
vanced stage, and finally, in the 23 nmi. stage the chorda presents a spiral appear-
ance, as shown in figure 40. Braun found this same process in sheep and other
mammaUan embryos, the end of the chorda projecting caudalward from the last
vertebra. Ecker also noted this projection of the chorda in his mammalian mate-
rial. These authors recognized the winding or branched end of the chorda in the
caudal filament or in the extreme end of the tail and concluded that this was its
primitive state. In human embryos, however, as mentioned above, at the earliest
stage when the chorda reaches the extreme end of the tail, its caudal end is straight
and shows no winding until the reduction of the tail begins. In embryos from 15
to 23 mm. the caudal end projects almost caudalward from the last vertebra which
has been formed by the fusion of several vertebra?. His did not find such a condi-

AND OF THE CAUDAL END OF THE SPINAL CORD. 189

tion in human embr^yos, although it is usual. According to Braun, the occurrence
of a free, naked end of the chorda is due to the disappearance of the last primitive
vertebrae by which it had previously been surrounded. In the human embryo the
caudal end of the chorda was never surrounded by primitive vertebrae, but was
situated between the neural canal and the primitive vertebral column. In the
25 mm. (fig. 42) and the 30 mm. embryo (fig. 43) the coil-hke appearance of the
chorda is typical, and these stages, therefore, are the clearest of any thoughout
embryonic hfe. Later on, for example, in the 37 mm. embryo (fig. 44), the caudal
end of the chorda is disappearing, leaving a few remnants which have become
separated from the main chordal strand. This degenerative fragmentation and
partial absorption of the terminal portion of the chorda results in a great variety
of forms. Very seldom is the caudal end branched in the earlier stages. From the
39 mm. stage the chorda becomes gradually reduced and is finally converted into
a more simple form, as shown in figures 45 and 46. While the short caudal por-
tion shows the above-mentioned variations, the main strand changes but slightly.
After it becomes straightened some portions of it which lie in the intervertebral
fibro-cartilage show spindle-shaped swellings, as shown in figures 39 and 42 (18
to 25 mm.). At last, in the 50 and 52 mm. embryos, the parts embedded in the
vertebral bodies disappear, leaving small remnants in the intervertebral spaces.
Frequently these remnants show visible coils, as can be seen in figure 46. These
remain often until a later stage. The disappearance of the chorda below the thir-
tieth vertebra occurs later than that of the main strand, and we can therefore still
recognize it in the 67 mm. embryo as a continuous cord through the caudal vertebral
bodies.

DEVELOPMENT OF THE CAUDAL END OF THE SPINAL CORD.

In the 4 mm. embryo the caudal end of the neural tube fuses with the solid
mass of mesodermal cells which extends to the ventral side of the tail. The caudal
ends of the chorda dorsalis and caudal gut also merge with this cell-mass. In the
5.5 to 7.5 mm. embryos the caudal end of the neural tube, with its central canal,
extends to the apex of the tail and merges into the mesodermal cell-mass, entirely
losing its boundaries. In the 8 mm. specimen a difference can be plainly recog-
nized between the caudal portion of the spinal cord and the portion that lies cranial
to the thirty-second somite. Caudal to this level the central canal is distinctly
narrower. Thus it may be divided into two portions, an upper, wider canal and
a caudal narrow or atrophic canal. The former constitutes the main part of the
central canal of the spinal cord. On cross-section it is oval in shape and its walls
show no folds. The caudal part is narrower in its dorso-ventral diameter than is
the main canal, and therefore on section presents a more rounded form. Some-
times a large fold is found between the two parts and in the 11 and 12 mm. speci-
mens can be seen the primordia of other folds on the walls of the atrophic canal,
especially on the ventral side, as shown in figures 34 and 35. The distinction
between the atropine portion of the spinal cord and the main part is quite marked
in the 15.5, 16, 17, and 19 mm. embryos. The caudal end of the wider canal
expands transversely, and where it narrows into the atrophic canal constitutes the

190

DEVELOPMENT AND REDUCTION OF THE TAIL

initial form of the ventriculus terminalis. Tlic portion of the si)inal cord which
incloses the ventriculus is the i)riniitive conus meduUaris.

The ventriculus terminalis was found by Argutinsky in a 45 nmi. embryo; by
Brugsch and Unger in a 25 mm. embryo, and what may be considered as its primor-
dium is already apparent in my specimens of 11 and 12 mm. respectively, as shown
in figures 34, 35, and 36. It can be seen from stage to stage retreating cranialward,
while the atrophic canal gradually lengthens. In the 18, 23, and 25 mm. embryos
the latter expands noticeably in the median or caudal region, as shown in figures
39, 40, and 42. This was also observed by linger and Brugsch, who, however,
did not regard it as the primordium of the ventriculus terminalis, but rather as a
homologue of the sinus terminahs of the amphibians, w^hich develops at the caudal
end of the central canal of the spinal cord. In the majority of my sjDecimens from
18 to 30 mm. the caudal end of the atrophic canal shows diverticula such as those
descrii^ed by Unger and Brugsch. In such embryos the spinal cord becomes

temporarily longer than the vertebral col-
umn. It seems probable, therefore, that
in the wall of the atrophic portion of the
cord the ependymal cells increase rapidly
by proliferation, and perhaps also by the
migration of other ependymal cells from
the more caudal part of the tail, which
is in process of regression. By reason of
these two processes folds develop in this
wall, such as are well shown in text-figure 1 .
In these embr^yos the caudal end of the
chorda dorsahs seems to exert an attrac-
tion upon the caudal end of the medullary
tube, thus drawdng it into a more cranial
position (fig. 39). In embryos of 30, 33,
35, and 37 mm. the atrophic canal is longer
and narrower than m the slightly younger

specimens, but the caudal end is still di-

lated. At several points the canal has be-
come so narrowed that its central cavity
is obhterated and gradually becomes con-
verted into a cell-strand, as shown in fig-
ure 44.

At the stage where the embryo has
entirely lost its external tail the spinal
cord is about the same length as the ver-
tebral column, as shown clearly in the 30 and 37 mm. specimens. From this time
on the vertebral column increases relatively in length, although there is no cessa-
tion of growth of the spinal cord. In the 37 mm. embryo a l^undle of nerve-fibers
(i. e., marginal zone) is visible on the ventral side of the atrophic cord, as shown in

Text-fiuvre 1.

Sagittal section through caudal end of a human embryo
18. mm. long (Carnegie Collection, No. 432. slide 22,
row 1, sec. 3), enlarged 50 diameters. The wall of
the atrophic portion of the spinal cord is thinner
and represents a much younger stage of develoi>-
nient than that of the main part of spinal cord.
From this time on it undergoes gradual regressive
changes with obliteration of the central canal and is
eventually replaced by a simple fibrous strand.

AND OF THE CAUDAL END OF THE SPINAL CORD. 191

figure 44 (fil. t.), and represents a prinutive filum terminale. Tliis structure extends
caudalward from the apex of the primitive conus medullaris at a level between the
twenty-ninth and tWrtieth vertebrae. In the 33 mm. embryo the cranial portion of
the atrophic canal and its ependymal lining disappear and the bundle of nerve-fibers
remains in the sheath of the dura mater. The dilated portion of the atrophic canal
remains as the coccygeal vestige. The process of dedifferentiation of the caudal end
of the spinal cord, viz, the reduction of the cranial end of the atrophic canal, advances
step by step, so that what appeared as a long, narrow tube in the 37 mm. embryo
is divided into several parts in the 50 imn. specimen, each part containing a cavity,
the more cranially situated part being joined to the ventriculus terminahs by a cell-
strand which is destined also to disappear in the course of development (figs. 44, 45,
and 46). The ventriculus terminalis, which had only begun in the 37 mm. speci-
men, has developed completely in one of 50 nmi. Its formation is the result of
the gradual constricting of the expanded area of the central canal which marks the
division between its upper wider and lower atrophic portions, and a separate cavity
is thus formed. Complete fusion of the margins, however, does not occur, a nar-
row channel being left which connects the two portions of the central canal. In
their thesis Brugsch and Unger have described in detail this process of reduction
of the central canal. In my specimens the phenomenon is first noted in the 39
mm embryo. In the 39, 46, 50, and 52 mm. specimens the conus medullaris,
ventriculus terminahs, and filum terminale are quite distinct, and the remnant
of the neural tube caudal to the filum terminale persists as the coccygeal medullary
vestige. The form of the ventriculus terminalis varies in the different specimens,
while that of the conus medullaris is much the same in all. In the specimens
above enumerated the ventriculus terminalis is situated at about the level of the
twenty-ninth or thirtieth vertebra, the position of both it and the conus medullaris
gradually becoming more cranial as the result of the fact that the growth of the
vertebral column becomes progressively more rapid than that of the spinal cord. In
these embryos, especially the 46 mm. specimen, the membranes of the spinal cord —
the dura mater and pia mater — are visible. At a level with the upper border of
the thirty-first vertebra, in the 46 mm. embryo, the dura mater may be seen to leave
the wall of the vertebral canal for the caudal end of the conus medullaris, which
marks the beginning of the filum terminale, thus forming a sheath for the latter.
As a rule, the coccygeal medullary vestige is on the dorsal side of the last two
vertebrae. It is situated in the connective-tissue surrounding the vertebrae and
does not adhere to the epidermis. Tourneux and Hermann discovered the caudal
remnant of the spinal cord in a 37 mm. embryo and termed it the vestiges medul-
laires coccygietis. Tourneux advanced the theory that the slightly enlarged caudal
tip of the neural tube is closely united in the deep layers of the skin. Toward the
end of the third month the spinal colunrn, developing more rapidly than the soft
parts, draws along the part of the neural tube adherent to it, the extreme tip of
which remains attached to the skin. As a result of this the terminal or coccygeal
portion of the neural tube becomes bent in the form of a loop, the more deeply
situated Umb being termed segment coccygien direct, and the more superficial one

192 DEVELOPMENT AND REDUCTION OF THE TAIL

segment coccygien reflechc. In mj' specimens I did not find such to be the case,
nor was it noted by Unger and Brugsch. I beheve, therefore, that the condition
noted by Tourneux is of rare occurrence. In the 39 nun. embryo may be seen a small
papiUiform tail, at the root of which is a group of cells representing the remnant of
the spinal cord. The caudal end of the coccygeal medullary vestige appears to
adhere to the epidermis, but in reality does not, although Tourneux and Hermann
found that it did adhere in their case.

Concerning the development of the coccygeal medullary vestige from the rem-
nant of the neural tube, I am led to the following conclusions:

(1) The expanded caudal end of the neural tube in an embryo in which the
tail has disappeared is the primordium of the coccygeal vestige (figs. 40, 42, and 43).

(2) In addition to the coccygeal vestige there frequently occurs a similarly
formed epithelial sac situated in a more cranial position.

(3) The caudal end of the coccygeal vestige merges into the caudal hgament,
as believed by Brugsch.

(4) In the younger specimens the fibers which persist as the filum terminale
always he ventral to the epend>Tnal cells, which become the coccygeal vestige.

(5) The middle sacral artery and vein extend to and curve around the apex
of the coccygeal vestige.

(6) Only in rare cases is the coccygeal vestige lacking. In my entire series
of specimens, from 4 to 125 imn., in only one did I actually fail to find it (fig. 24).

(7) In specimens from 4 to 100 mm. the coccygeal vestige is not adherent to
the epidermal layer of the skin.

(8) The coccygeal vestige continues to grow after the 100 mm. stage.

In the 67 mm. embryo, as can be seen in figure 46, the ventriculus terminalis
occupies a more cranial position than in the younger specimens; the conus medul-
laris has become relatively more slender and the filum terminale longer, the latter
disappearing caudal to the thirty-second vertebra, two remnants of the neural tube
being left. The membranes of the spinal cord are here also considerably further
developed than in the younger specimens. At this stage the arachnoid mem-
brane Ues between the dura mater and pia mater. At the upper border of the
twenty-seventh vertebra the dura mater leaves the wall of the vertebral canal
for the filum terminale, forming a sheath and reaching the filum terminale at a
level between the twenty-eighth and twenty-ninth vertebra?. In younger speci-
mens, for example in the 46 nun. embryo, this separation occurs at a level with
the thirty-fii-st vertebra. Therefore, the caudal end of the dural sac, as well as
the spinal cord, recedes cranialward. This phenomenon is an evidence of the
relatively more rapid growth of the vertebral column. What, then, is the cause
of the lengthening of the filum terminale? In the 33 mm. embryo I could recog-
nize distinctly, below the conus meduUaris, a bundle of fibers representing a primi-
tive filum terminale. In the 37, 39, and 50 mm. embryos this reaches almost to
the caudal end of the neural tube or the coccygeal medullary vestige. In the 37
and 39 nun. specimens it extends farther caudalward than in any of the others
(fig. 44) . It is my opinion, therefore, that the filum terminale consists at an early

AND OF THE CAUDAL END OF THE SPINAL CORD. 1^3

stage of nerve-fibers, especially those from the ventral portion of the spmal cord,
although von KoUiker does not hold this view. After the development of the ven-
triculus terminahs the caudal portion of the conus medullaris is converted into the
filum terminale by the ventriculus terminalis and conus medullaris moving cranial-
ward. This is due to the fact that the gray substance which Ues next to the ven-
triculus terminahs is undergoing degeneration and the caudal end of the central
canal is gradually excavated, while the caudal end of the ventriculus terminalis
narrows by degrees, losing its cellular substance. The relative lengthening of the
filum terminale, therefore, is due to the growth of the nerve-fibers with their sheath
of dura mater and pia mater, and in part also to the gradual addition of tissue
from the caudal portion of the conus medullaris, which has become converted into
the tissue of the filum terminale.

ABNORMALITIES OF THE CAUDAL END OF THE SPINAL CORD.
(a) Embryo No. 40.5, 26 mm. ; (b) embryo No. 145, 33 mm. ; (c) embryo No. 449, 36 mm.

In the first two specimens the caudal tip of the neural tube is spiral. It is
probable that this part is covered with a layer of epidermis, although I could not
discover it and therefore conclude that it was injured in the preparations. In
the first specimen the caudal end of the spinal cord enters into the tail-bud. Two
branches of the anterior spinal artery penetrate between the coils of the spinal
cord. In the second specimen the coil of the caudal end of the spinal cord forms
the summit of the papillary tail and terminates at the ventral side of its root.
The third specimen has only 32 vertebra? and no remnant of the neural tube.

SPINAL GANGLIA.

It is very difficult to locate the first cervical ganghon at a very early stage
of embryonic develoi)ment, particularly if the specimen is poorly preserved. This
structure is frequently found in close apposition to the Froriep ganghon on the
trunk of the spinal accessory nerve. Sometimes it is poorly developed and re-
sembles a Froriep ganglion, except for the fact that it hes always on the ventral
side of the trunk of the accessory nerve, while the Froriep ganglion lies on the
dorsal side. The first cervical is smaller than the others, and the second in turn
is smaller than the third. In embryo No. 991 (17 mm.) both first cervical ganglia
are lacking. In embryos from 5 to 10 mm. there are in most cases 32 pairs of
ganglia; from 12 to 14 mm. there may be 33 pairs. When the number is 33 the
last caudal ganghon is usually very small and has no nerve. In embryos from 15
to 33 mm. the number is usually 31. I have frequently found the thirty-second
spinal nerve without a ganglion, the latter having degenerated. In embryos from
35 to 67 nma. long, and older, there are usually only 30 ganglia; occasionally there
may be 31, but the last is usually undergoing degeneration.

SYMPATHETIC GANGLIA.

In embryos 33 mm. long, and older, the caudal ends of the paired sympathetic
ganglionated nerve-trunks join together at the upper plane of the ventral side of
the thirtieth vertebra, as shown in text-figure 2. At the point of union there is

194

DEVELOPMENT AND REDUCTION OF THE TAIL

usually found a ganglion; another occurs approximately at a point between the
thirty-first and thirty-second vertebrae. From the latter ganglion the single nerve-
trunk follows the course of the middle sacral artery and vein, running between
them, and emerges dorsally from the caudal end of the ver-
tebral colunm, where the coccygeal medullary vestige and
caudal ligament curve about the apex of the column. This
nerve consists of a large bundle of non-medullated nerve-
fibers, but the structure of its caudal end can not at this
stage be made out distinctly. At a level between the
thirty-third and thirty-fourth vertebrae, or perhaps a little
above, there is a small group of cells representing a sym-
jiathetic ganglion. At this point are frequently found nu-
merous jjlexiform branches of blood-vessels enmeshing this
group of cells. This richly vascularized cell-group may be
the primordiiun of the glandula sacralis.

-Vert

-Ganq.symp-

-Carl Intv

Text-figure 2.

Vpntr.ll view of caudal portion of
vertebral column, showing char-
acteristic arrangement of sym-
pathetic ganglion-strands in hu-
man embryos between 46 and
67 mm. long. Cart, inlv., inter-
vertebral fibrocartilage ; Gang,
si/mp., ganglionic node; Veil.,
body of \ertebra.

SUMMARY.

(1) The human embryo possesses a true tail composed
of ]jrimitive vertebrae and the caudal ends of the spinal
cord, chorda dorsalis, and middle sacral artery and vein.

(2) The longest and most completely developed tail
among the specimens examined by me was found in a 7.5
mm. embryo. This was 1.2 mm. in length.

(3) The human embryo does not possess a caudal filament homologous with
that of other mammals.

(4) The reduction of the tail, especially of the primitive vertebrae, begins when
the embryo has reached a length of about 8 or 9 mm.

(5) Prior to this the tail consists of two parts: a proximal longer part (the
vertebrated tail), wliich has well-formed somites, and a caudal shorter part which
contains only a mesodermic remnant.

(6) In embryos from 25 to 27 mm. the tail is reduced to a small papilla, in
which are contained the caudal ends of the spinal cord and the middle sacral artery
and vein, and into wliich the end of the caudal ligament enters. As a rule this
tail-bud is not situated directly at the caudal extremity of the vertebral column,
but slightly dorsal to it. The vertebrated portion of the external tail has retracted
into the soft tissues and has thus become an internal taU, whereas the lost-vertebrae
tail projects temporarily and finally it also disappears.

(7) At the time the division of the external tail takes place two eminences
appear at the caudal region; one ventral (coccygeal tubercle or Steisshocker), the
other dorsal (caudal tubercle or Kaudalhocker). The first is due to the pushing
up of the caudal end of the internal-tail vertebrae, formerly situated in the root
of the external tail and constituting the vertebral portion of it in younger embryos.
The second is usually shaped like a small papilla and by some authors is called
the tail-bud or caudal filament, although the latter, as stated above, is entirely a

AND OF THE CAUDAL END OF THE SPINAL CORD. 195

misnomer. Sometimes it appears as a rounded eminence and was therefore termed
caudal tubercle by Unger and Brugsch.

(8) The tail-like appendage that occasionally persists in adults may possibly
-be explained as a persistent caudal tubercle that did not undergo the normal reduc-
tion. It must be granted, however, that in none of the cases reported has osseous
or cartilaginous tissue been found.

(9) When the embryo reaches 30 to 35 mm. the tail has usually entirely dis-
appeared, although the time of disappearance is quite variable. Thus, the tail
was found to have disappeared in the 24 mm. embryo, while in one 39 mm. long
it still persisted.

(10) In embryos above 40 mm. in length that have lost the external tail I
have designated as the internal tail the portion caudal to the thirtieth vertebra,
for three reasons: (o) the curve of the vertebral column occurs at the thirty-
first and thirty-second vertebrae; (h) below the twenty-ninth vertebra the spinal
ganglia disappear at about the same time as the external tail; (c) the sympathetic
ganglion strands unite between the thirtieth and thirty-first vertebrae.

(11) The disappearance of the canal of the caudal gut had already begun in a 5.5
mm. embryo and in a 6.5 mm. specimen the caudal gut had become converted into a
long cell-strand, excej)! for a short caudal portion. In embryos 7.5, 8, and 9 mm. the
remnant of the caudal gut, inclosing a small cavity, was still found in the end of the
tail. In those 10 mm. and older the caudal gut had entirely disappeared.

(12) In the very youngest specimens the medullary tube reaches to the extreme
tip of the tail.

(13) In those 11, 12, and 15.5 mm. long the medullary tube can be divided
into two parts at the level of the thirty-second vertebra: a cranial part, having
a wide central canal, and an atrophic caudal part with a narrow canal. This
distinction becomes quite marked in the 15.5 mm. specimen. The canal at the
junction of these two parts is slightly enlarged transversely and constitutes the
primordium of the ventriculus terminalis.

(14) The atrophic portion of the spinal cord gradually becomes more slender,
although its caudal end remains unchanged for some time or in some instances
shows temporary enlargement. It later subdivides, the cranial end forming the
cell-strand of the filum terminale and the caudal end developing uito the coccygeal
medullary vestige.

(15) The caudal end of the wider part of the spinal cord develops into the
conus medullaris and its lumen constitutes the ventriculus terminalis.

(16) In the 46 mm. embryo the ventriculus terminalis is perfectly developed.

(17) The conus medullaris and the ventriculus terminalis recede cranialward
as the result of two processes: (a) the growth of the vertebral column, which is
more rapid than that of the spinal cord ; (6) the degeneration of the gray substance
which forms the upper wall of the ventriculus, thus causing the cavity to enlarge
and gradually move upward while its caudal end narrows.

(18) The extent of the coiling of the chorda dorsaUs in its various stages
indicates the extent of fusion of the last primitive vertebrae.

196

DEVELOPMENT AND REDUCTION OF THE TAIL.

(19) In embryos from 12 to 14 mm. the spinal ganglia are 33 in nvmiber, but
at about this period reduction begins in the more caudally situated ones. Thus
in an embryo of 67 mm. there are but 29 ganglia.

In conclusion, I take this opportunity to acknowledge my indebtedness to
the late Professor Franklin P. Mall for the privilege of using the valuable material
in the collection of human embryos belonging to the Carnegie Institution of
Washington. I also wish to thank Dr. George L. Streeter for his kind assistance
in the preparation of this paper.

BIBLIOGRAPHY.

Aroutinsky, p. Ueber die Gestalt und die Enstehungsweise

des Ventriculus terminalis und iiber das Filum

terminale des Riickenmarkes bei Neugeborenen.

Arch. f. mikrosk. Anat., 1898, lii, 501-.53O.
Bardeen, C. R. Studies of the development of the human

skeleton. The development of the lumbar, sacral

and coccygeal vertebrae. .\mer. Jour. Anat., 1905,

IV, 266-302.
Bartels, Max. Die geschwanzten Menschen. Arch. f.

Anthropol., 1884, xv, 45-131.
Berry, J. Babv with a tail. Memphis Medical Journal,

1894, XIV, p. 105.
Braun, M. Entwickelungsvorgange am Schwanzende bei

einigen Siiugethieren mit Beriicksichtigung der

Verhaltnisse beim Menschen. Arch. f. Anat. u.

Physiol. (Anat. Abth.), 1882, 207-241.
Brugsch, T. Zur Frage der Schwanzbildung beim Men-
schen. Zeitschr. f. Heilk., 1907, xxviii (.\bth. f.

path.), 155-161.
and E. Unger. Die Entmcklung des Ventriculus

terminalis beim Menschen. .\rch. f. mikrosk.

Anat., 1903, lxi, 220-232.
f'LARKE, J. L. Further researches in the gray substance

of the spinal cord. Philosoph. Trans. Royal Soc.

London. 1859, r.XLix.
Darwin, C. R. The Descent of Man. 1871, I.
Dickinson, A. A child with a tail. Brooklyn Medical

Journal, 1894, viii, p. 568.
EcKER, A. Icones Physiologicce. Erliluterungsgeschichte.

Leipzig, L. Voss, 1851-59.
. Der Steisshaarwirbel, die Steissbeinglaze, und das

Steissbeingrubschen wahrscheinliche Ueberbleibsel

embryonaler Formen in der .Steissbeingegend beim

ungeborenen und der Wachsenen Menschen. Arch.

f. Anthropol., 1880, xii, 129.
. Beitriige zur Kenntniss der ausseren Formen

jiinster menschlicher Embrj'onen. .\rch. f. Anat.

u. Physiol. (Anat. Abth.), 1880. 403^06.
. Besitzt der menschliche Embryo einen Schwanz?

Arch. f. .\nat. u. Physiol. (.\nat. Abth.), 1880, 421.
FoL, H. Sur la queue de I'embryon humain. Comptes

rend. 1885, Janvier-Juin, 1469-1472.
Gerlach, L. . Ein Fall von Schwanzbildung bei einem

menschlichen Embryo. Morph. Jahrb., 1880, vi,

106-124.
Harrison, R. G. On the occurrence of tails in man. with

description of the case reported by Dr. Watson.

Johns Hopkin.s Hosp. Bull.. 1901, xii, p. 96.
His, Wilhelm. (a) .\natomie menschlicher Embryonen.

Leipzig, 1880, i, p. 89.
. (6) Ueber den .Schwanzthiel des men.sch!ichen

Embryo. Arch. f. Anat. u. Physiol. (Anat. Abth.),

1880, 431-440.

Keibel, Franz. Ueber den Schwanz des menschlichen
Embryo. Arch. f. Anat. u. Phvsiol. (Anat. Abth.),

1891, 356-388.

Keibei.. F., and F. P. Mall. Human embryology. J. B.

Lippincott, Philadelphia, 1910.
Kohlbrijgge, J. H. F. Schwanzbildung und SteissdruBe

des Menschen und das Gesetz der Ruchschlagsver-

erbung. Naturk. Tijdschr. Nederl. Indie. Batav.,

1898, LVii, 163-195.
v. Koliiker, a. Gundriss der Entwicklungsgeschicht des

Menschen und der hoheren Thiere. 2d ed., 1884.

Leipzig, Engelmann.
KoNSTANTiowiTscn, W. Zux Frage der Schwanzbildung

beim Menschen. Zeitschr. f. Heilk. path. Anat.,

1907, xxviii.
Krause, W. Der Ventriculus terminalis des Riickenmarka.

Arch. f. mikrosk. .\nat., 1875, xi, 216.
Pyatnitski, Ivan I. On the question of the formation of a

tail in man, and of human tails in general, according

to data from the literature and personal researches.

A. Muchnik, St. Petersburg, 1892, pp. 96.
Roden.Kcker, Geokg. Ueber den S.^ugetierschwanz mit

besonderer Berilchsichtigung der caudalen Anhange

des Menschen. Freiburg, 1898, pp. 39.
Rosenberg, E. Ueber die Entwickelung der Wirbelsaule

und des Centrale carpi des Menschen. Morph.

Jahrb., 1876, i, 81-198.
. Ueber eine primitive Form der wirbelsaule des

Menschen. Morph. Jahrb., 1899, xxvii, 118.
ScHAEFFEH. OsKAR. Beitr.ag zur .\ctiologie der Schwanz-

bildungen beim Menschen. Arch. f. Anthropol.,

1892, XX, 189.

ScHWARZ. Beitrag zur Kenntniss des geschwanzten Men-
schen. Munchen med. Wochschr., 1912, lix, 928.

TouRNEUX, F., and G. Hermann. .Sur la persistance de
vestiges medullaries eoccygiens pendant toute la
periode foetale chez I'homme et sur le role de ces
vestiges dans la production des tumeurs sacro-
coccygienues congenitales. Jour, de I'Anat. et de
la Phy.siol., 1887, xxiii, 498-529.

Toi'HNEL'X, F. Sur la structure et sur le developpement
du fil terminal de la moelle chez I'homme. C. R.
hebd. de la Soc. de Biol., ser. 9, T. 4, 1892, ,340-343.

Unger, E., and T. Brugsch. Zur KenntnLss der fovea und
fistula sacrococcygea caudalis und der Entwicklung
des ligamentum caudale beim Menschen. .4rch.
f. mikrosk. .\nat.. 1903, lxi, 151-219.

ViRrHow, R. .Schwanzbildung beim Menschen. Berlirj
klin. Wochenschr.. 18.84, x.xi, p. 8.

Waldeyer, W. Die Caudalanhiinge des Menschen. Sitz.-
bericht. Akad. Wi.ss.. Berlin, 1896, p. 775.

Watsun, W. T. The exhibition of a three-months infant
with a caudal append.age. Johns Hopkins Hosp.
Bull., 1900, XI.

DESCRIPTION OF PLATES.

(The description of Plate 1 will be found on the plate itself.)

The figures on Plates 2 to 4 represent profile reconstructions of the caudal region in a series of human embryos
selected from the Carnegie Collection. The different structures are indicated by the following abbrevia-
tions:

A. s. m. Arteria sacralis media.

Append. Caudal extension of the terminal ventricle.

Caud. non-v. Part of tail containing no vertebrae.

Ch. Chorda dorsalis.

CI. Cloaca.

Constrict. Constriction marking off remnant of spinal cord

which is to form the coccygeal medullary vestige.
Con. med. Conus medullaris.
Dura. Dura mater.
Fil. t. Filum terminale.

Gang. symp. Sympathetic ganglionated cord.
Int. cau. Intestinum caudale.
Lig. cau. Ligamentum caudale.
Med. sp. Medulla spinalis.

Med. sp. atr. Atrophic portion of the spinal cord.
Mem. cl. Membrana cloacalis.

Pia. Pia mater.

PI. vase. Plexus vasculosus.

Re. epend. Remnant mass of ependymal cells.

Rect. Rectum.

Re. int. cau. Remnant of caudal gut.

Rud. cau. Rudiment of tail.

Str. cell. Stria cellularis.

Coccygeal tubercle.

Vena sacralis media.

Ventriculus terminalis.

cran. Cranial portion of terminal ventricle.

cau. Caudal portion of terminal ventricle.
X. Characteristic folds in ventral wall of spinal oord.
XX. Furrow on surface corresponding to level at which the
caudal gut begins to disappear.

CO.

Tub
V. s. m
Vent. t.
Vent. t.
Vent. t.

Plate 2.

Fig. 31. Embryo Xo. 810, .5..5 mm., enlarged .31. .5 diameters. The caudal gut already shows a constriction separating
it from the cloaca. The lines along the dorsal margin of the spinal cord represent the boundaries of the
myotomes.

Fig. 32. Embryo No. 371, 6.6 mm., enlarged 31. .'i diameters. The segmental levels in this specimen are determined
by the sclerotomes. It will be noted that the tail has attained nearly its maximum development, and
as compared with the more cranial parts it will hereafter gradually take on a more atrophic appearance.

Fig. 33. Embryo No. 389, 8 mm., enlarged 31. .5 diameters. The coccygeal portion of the spinal cord is aheady dis-
tinctly narrower than the main cord.

Fig. 34. Embryo No. 544, 11 mm., enlarged 31.5 diameters. In this specimen the caudal gut has disappeared. The
vertebrated and non-vertebrated portions of the tail are clearly demarcated.

Fig. 35. Embryo No. 852, 12 mm., enlarged 31.5 diameters. The non-vertebrated portion of the tail is here rela-
tively much shorter than in the previous specimen.

Fig. 36. Embryo No. 390, 15.5 mm., enlarged 31.5 diameters. A rudiment of the tail persists as a small elevation
(rud. cau.). The caudal end of the vertebral column terminates in a fibrous strand of cells. The termi-
nal three sclerotomes may be regarded as having been converted into this .strand, or they may have fused
into one irregular vertebra.

Pl.\te 3.

Fig. 37. Embryo No. 406, 16 ram. crown-rump length, enlarged 31.5 diameters. The slender atrophic portion of
the spinal cord is clearly demarcated from the remainder of the cord owing to the fact that it retains its
earlier embryonic form. The point at which its narrow canal opens into the main canal corresponds to
the future terminal ventricle.

Fig. 38. Embryo No. 576, 17 mm. crown-rump length, enlarged 22.5 diameters. As compared with the last specimen,
the caudal region has undergone marked reduction and resembles the condition that will be seen in
embryos about 20 mm. long.

Fio. 39. Embryo No. 432, 18 mm. crown-rump length, enlarged 22.5 diameters. It will be noted that embryos of
about this size show the tendency toward a sharp dorsal retroflexion of the caudal rudiment. The charac-
teristic thinness and wrinkUng of the wall of the atrophic portion of the spinal cord is also well represented
in this embryo.

Fig. 40. Embryo No. 453, 23 mm. crown-rump length, enlarged 22.5 diameters. In this specimen, just ventral to
the junction of the atrophic part with the remainder of the cord, is a mass of cells which appeared to form
a diverticulum, although a communication between its lumen and the central canal of the cord could not
be clearly made out.

Fig. 41. Embryo 382, 23 mm. crown-rump length, enlarged 22.5 diameters. The relative narrowness of the lumen
of the atrophic portion of the spinal cord is a characteristic preliminary to its transition into the filum
terminale. Near the caudal tip, at the point marked X, is the seat of fusion with the chorda dotsalis.

Fig. 42. Embryo No. 584a, 25 mm. crown-rump length, enlarged 22.5 diameters. A point of fusion with the chorda
(}orsaUs, marked X, can be seen in this specimen somewhat similar to that in figure 41.

197

198 DEVELOPMENT AND REDUCTION OF THE TAIL.

Plate 4.

Fig. 43. Embryo \o. 7.5, 30 mm. crown-rump length, enlarged 22. .5 diameters. The relative thinning-out of the
walls of the conus medu,llaris results in a tendency to their being thrown into large irregular folds. In
the re^^ion of the conus medullaris the regressive tendency sets in after that region has attained propor-
tions larger than those seen in the more caudal part. It consequently becomes expressed by a thinness
of the walls, producing a transparent terminal ventricle in contrast to the obliteration seen in the filum
terminate.

Fig. 44. Embryo No. 972, 37 mm. crown-rump length, enlarged 18 diameters. This specimen shows the transition
of the atrojjhic sjjinal cord into a fibrous filum terminale. The terminal portion retains its lumen and
persists as the coccygeal medullary vestige.

Fig. 4.T. Embryo No. 448, 52 mm. crown-rump length, enlarged 13.5 diameters. At this time the terminal ventricle,
the filum terminale, and the coccygeal medullary vestige are distinctly marked off from each other, and
their general adult characteristics attained.

Fig. 46. Embryo No. 1656, 67 mm. crown-rump length, enlarged 9 diameters. The regressive condition of the walls
of the terminal ventricle are expressed by their relative thinness and their irregularity. The filum termi-
nale is almost entirely converted into a solid fibrous strand in which traces of ependymal masses can be
found. The membranes of cord can be seen and present an arrangement that closely simulates that of
the adult.

The figures on Plate I are designed to show diagrammatically the relations of the caudal end of the spinal cord and the vertebral
column in a series of human erahryos varying from 4 mm. to 67 mm. long, and are so arranged that the segmental levels, as
indicated at the left, correspond throughout, the thirty-fourth segment being emphasized by a heavy line. In the two younger
stages the segments were detennined by the myotomes; the remainder were determined by the sclerotomes, or the bodies of the
vertebrte. In making the diagram it was found necessary to make all the segments of the same width; in reality the more
caudal ones are relatively much narrower. Also, the individual segments become wider in the older stages, whereas in the
diagram they are kept at the same width. There is thus introduced an axial distortion which should be kept in mind in
studying the figures. In figures 1 to 15 the surface profile of the caudal region is indicated and the cloacal membrane is
shown by a wider Hue. In figures 1 to 5 the caudal gut is shown by a broken Une. In figures 4 and 5 it will be noted that a
remnant of the gut is still present, though its communication with the cloaca is interrupted. Early stages in the formation of
the filum t«rminale are shown in figures 25 to 30. The figures in this plate are all based upon profile reconstructions made
from the following embryos:

Fig.

Embryo.

Fig.

Embryo.

F...

Emhryo.

1

Stf. 830

11

No. 432

21

No. 75

2

810

12

837

22

145

3

371

13

431

23

21 J

221

14

453

24

449

389

15

3S2

25

972

10

632

20

7

862

17

SS4o

27

3B0

18

40S

28

1N4

■106

10

1008

29

44H

10

570

20

875

30

16CU

KUNITOMO

Sfrcell. /■ Med.sp.alr.
Rud.cau.

KUNITOMO

PLATE 3

Med.sp.air

KUNITOMO

PLATE 4

Rectf

Tub. CO.

lied.sp.atr.
Constrict.

A s.m

Ves m.co.

ril.t.(Med.sp.dr)

Med.sp

Vent.t.
Con.med.

ndyrna.

L/qCau. Ves. m.co.

Ves, m.co.

^

MBI. WHOl LIBK-^'V

UH laKX P

ilffk