Studies on the Structure of the Golgi Apparatus.
IV. Endostyle of Branchiostoma Indieum.1
By
M. K. Subramaniam, M.A., D.Se.,
Research Fellow, Department of Zoology, University of Madras.
With Plate 25 and 10 Test-figures.
INTBODUCTION.
IN a few recent papers an attempt has been made to elaborate
some of the suggestions made by Subramaniam and Gopala
Aiyar (1936 and 1937) regarding the evolution of the networklike Golgi apparatus, more especially, the relation between the
chromophilie and chromophobie portions. In spite of Bowen's
assertion (19266) that the ehromophobic portion has nothing
to do with the archoplasm or sphere, the suggestion is not
accepted by many English cytologists. In Part I of these studies
(Subramaniam and Ganapati, 1938) it was shown that the
centriole in L e c u d i n a was intra-nuclear, and had no sphere
substance during any of the stages. Hence the chromophobic
part of the Golgi apparatus in L e c u d i n a was shown to be
an inseparable component of the Golgi complex which had
nothing to do with the archoplasm. Further, Bowen's assertion
that the relation between the centrosome and the Golgi apparatus was only circumstantial was reiterated. Having thus removed the objections, if any, to a consideration of the concept
that the idiosome may form a core to the strands composing
the Golgi apparatus, it was shown in Part II of these studies
that a chromophobic core is actually differentiated, during active
secretory synthesis, in the liver cells of the common tree-frog
E h a c o p h o r u s (Subramaniam, 19386). The work on B r a n c h i o s t o m a was undertaken with a view (1) to study the
relation between the central and Golgi apparatus in a ciliated
1
An Abstract of this Paper was read before the Jubilee Session of the
Indian Science Congress held at Calcutta in January 1938.
430
M. K. SUBBAMANIAM
cell, where the parts of the central apparatus are supposed to
be associated with the cilia, and (2) also to see whether any hint
as to the mode of evolution of the network will be available.
It has to be mentioned even at the very outset that the Golgi
apparatus is not netlike but is constituted by discrete elements
in B r a n c h i o s t o m a .
MATERIAL AND METHODS.
Branchiostoma in di cum (Willey) occurs in large numbers
off the Madras coast (Prashad, 1934). Twice a month dredged
collections were brought to the laboratory and material was
fixed immediately in Champy, F.w.a., Nassonov, Mann, corrosive acetic, and Bouin mixtures. Fixation times were varied
and material fixed in Champy, Nassonov, and Mann's fluids
were osmicated for varying periods of time. It was found that
2 days' fixation in Nassonov and osmication at 36° 0. for 6 days
gave the best pictures of the Golgi apparatus. Though Champy,
Mann, and Nassonov material gave equally good pictures, the
Nassonov material was depended on for description as this
showed the least amount of shrinkage. Compared to Nassonov
fixed material, fixation by corrosive and Bouin were atrocious,
even though great care was taken in dehydration, embedding,
&c. There has been a complaint that osmicated material is
brittle and is very difficult to manipulate. Though this has
been my experience also, a little care reduces the risk of the
material becoming powdered. During the course of these investigations I succeeded in eliminating such a contingency.
The stage at which the material breaks up is in the paraffin cup.
It has been my experience that if the paraffin is allowed to
solidify in the cup, the later heating leads to convection currents
which accelerates the tendency of the material to break up. In
order to prevent the material from powdering, the paraffin in
the cup should be kept in the molten condition from the time
the material is put in to the time it is embedded. The fixed
material was stained with iron haematoxylin, Delafield's haematoxylin, acid fuchsin-methyl green, thionin, toluidin blue, &c.
It was found impossible to stain the mucous globules with any
of the counter stains in osmicated material. But they stand
STRUCTURE OF THK 60LGI APPARATUS
431
out with great clarity as unstained droplets especially in iron
haematoxylin stained material.
OBSERVATIONS.
S t r u c t u r e of t h e E n d o s t y l e .
Willey (1894) states that the endostyle is composed of high
columnar cells arranged in one layer, and that the tenuity of
the cells which are often placed at different levels gives rise to
the impression of cells occurring in several layers. This condition
was noted only occasionally in the case of non-glandular cells.
In the case of mucous gland-cells, it is true, the nuclei lie in the
basal half of the cells at different levels, but the impression that
it is composed of several layers is due to the fact that not all
the cells open to the outside at the same level. The gland-cells
are ciliated, and the cilia end in granules which have a regular
arrangement (Text-fig. 1). Text-fig. 1 represents the picture
obtained at one level alone, but if the objective is raised or
lowered slightly, more cilia and basal granules are seen in the
same area. Some gland-cells, in Bouin preparations, contain
iron haematoxylin staining granules scattered irregularly
through the cytoplasm, while in others the granules are few
and are limited in distribution to the area near the nucleus.
No centriole with a centrosphere was observed in a juxtanuclear position in any of the preparations. The groups of
gland-cells have a conical shape, the basal region being wider.
This conical appearance is accentuated when the eells are loaded
with mucus. In a perfect transverse section not all cells could
be traced from the base to the ciliated end. Hence it necessitated
more a study of the mucous regions than of individual cells,
in order to give the reader a correct idea of the secretory cycle.
Each group of gland-cells appears to be only in one particular
stage of activity, though the groups vary slightly among themselves. Thus one group may be at the beginning of secretory
activity while another may be loaded with mucous globules.
But no case was noticed of a mucous region having some cells
at the beginning of secretory activity and others loaded with
mucous globules. It may also be generally stated that while
482
M. K. SUBBAMANIAM
two of the groups of cells are loaded with mucus, the other
two were at the beginning of secretory activity. It appears
that by such an arrangement mucus could be secreted continuously if necessary. Another important observation made
is that excretion of the mucus takes place all together and not
TEXT-UG. 1.
The parts of the central apparatus are seen in relation with
Cilia. Bouin.
in individual globules. It is true that the individual cells may
excrete at slightly varying times, but the mucous globules of
a single cell are sent out as a whole (see Bowen, 1926 a).
The cells in Bouin material had an irregular vacuolated appearance. The material had to be overstained to see the nucleus
clearly. During the later stages of secretion the nucleus seems
to undergo some degenerative changes. In mitochondrial preparations the nucleus could be easily made out, while in
STBUCTURB OF THE GOMH APPARATUS
4B3
osmicated material it appears as a clear space, the wall of which
is often yellow, but with no structural details in the interior.
The cell-walls could be just made out in both mitochoHdrial
and Golgi preparations and are seen in the photomicrographs.
Mitochondria.
In Text-fig. 2 is shown a group of resting cells from a Champy
preparation. This figure was drawn from that region of the
animal containing the coecal prolongation in order to show that
there is no difference in the structure and shape of the mitochondria in cells from different regions of the pharynx itself.
No centriole was noticed in a juxta-nuelear position, and in
cells which had extruded their secretion 2-7 large mitochondria!
grains were observed. These counts were made from serial
sections where an attempt was made to follow out individual
cells. Such cells are shown at A in Text-fig. 2. Mo ehromophobie
components to these grains were noticed though the cells were
stained at varying intensities. Actual division stages of the
mitochondria were not observed, but some of the grains were
comparatively large in size suggesting the probability that this
enlargement is a prelude to division. In these resting cells
the mitochondria were not always collected together at the
basal region of the cell. They had a scattered distribution, as
can be seen in cells shown at B in Text-fig. 2. Apparently there
is an increase in the number of mitochondria as a prelude to the
secretory activity of the cell, and they are only very rarely
scattered uniformly throughout the cell. The usual condition
seen is collections of mitochondria in different regions of the
cell (Text-fig. 2 c). This scattering of the mitochondria! grains
is soon followed by rupture of these large grains into smaller
ones. A collection of these smaller grains is seen at D in Textfig. 2. When the large grains are breaking up into smaller ones
a change in the appearance of the cytoplasm takes place. It
appears as if the vacuoles in the cells have run together and
given rise to irregular stained areas. In Text-fig. 3 is shown the
breaking up of the mitochondrial grains. The pictures seen
suggest two different methods by which the smaller mitochondrial granules are formed. (1) A large grain breaks up into
NO. 323
Ff
434
M. K. STJBBAMAN1AM
B
D
TEXT-JIG. 2.
Large mitocliondria in resting cells. In cells at D these large
grains are shown breaking up. Champy.
ExPLANAraoir o r LBTTEEDTO.
<7i, Golgi body without a visible chromophobie area; git Golgi bodies
showing a central ohromophobic area; d.g., dividing Golgi
elements; g3, batonettes formed by the rupture of the chiomophilio rim due to increase in size of the mucus globules and
which seem to degenerate later; g.s., Golgi body containing
secretion in its chromophobie region; >%, large mitochondrial
grains; mt, small mitochondrial grains; m%, very small mitochondrial grains; n., nucleus; s., secretion.
STRUCTURE OP THE GOLGI APPARATUS
485
a number of smaller grains as in Text-fig. 2 B and Text-fig. 3,
•m2, and these by fragmentation give rise to the smaller mitochondria! granules (ms). (2) The large grains break up info
innumerable small mitochondrial granules (m3) without first
TEXT-FIG. 3.
Cells showing irregular areas due to presence of mucus together
with mitochondrial grains of different sizes. Champy.
breaking up into smaller grains. The distribution of the mitochondria is not uniform and they have no relation to the actual
secretory droplets. This is shown in Text-fig. 8, m., where the
mitochondria and Golgi apparatus could be seen side by side. In
Champy sections at the end of the secretory cycle the cells have
436
M. K. SUBRAMANIAM
the appearance shown in Text-fig. 4. Most of the mitochondria
seem to be extruded with the mucous secretion, and the few remaining ones begin to increase in size preparatory to the changes
in the cell leading to the production of new secretory granules.
TEXT-FIG. 4.
Cells at the final stages of secretion showing mucous globules which
have run together to form large vacuoles and very small mitochondria! grains. Champy.
Golgi A p p a r a t u s .
In Text-fig. 5 (photomicrograph 1, PI. 25) is shown a group of
raucous cells which exhibit the earlier stages of secretory activity.
STHTJCTTJBE OF THE GOBGI APPABATOS
487
At A are shown a few cells which I consider to be in the earliest
stages. The Golgi apparatus occurs as a few discrete grains
which have no definite position in the cell as can be seen in
the figure. In some cells all the grains lie on the lumen side of
TEST-FIG. 5.
Cells at the earliest stage of secretion. Some of the Golgi bodies
have a differentiated ehromophobie region. Stages suggesting
division are also seen. Nassonov.
the nucleus, and in some others they are on the other side of the
nucleus. Thus no ' Golgi zone' could be postulated in these cells.
Moreover, attempts at demonstrating a juxta-nuelear centrosome in these cells proved a failure, and it appears that the
granules arranged uniformly at the base of the cilia represent
the central apparatus. These basal granules which give attach-
438
M. K. SUBEAMANIAM
ment to the cilia may be seen in Text-fig. 1 and photomicrographs 2 and 4, PI. 25. The number of Golgi bodies in these early
non-secreting cells vary from 1-4. The shape of these bodies
is not perfectly spherical. Some are elongated and rod-like.
In Text-fig. 6 at d.g. are shown what apparently look like
TEXT-FIG. 6.
Same as Text-fig. 5 but drawn at a higher magnification.
division stages of Golgi bodies. The elongated rod-like structures mentioned above are seen almost throughout the secretory
cycle, and thus appear to be stages leading to the division of
these bodies. An increase in the number of these bodies takes
place and they also begin to grow in size. With the increase in
size a slight change takes place in their structure. The centre
of each Golgi body becomes brownish while the outer rim is
STRUCTUKE OF THE GOLGI APPARATUS
439
dark. When photographed at a magnification of 100x15
(photomicrograph 2, PI. 25) it is difficult to get a print in which
this differentiation can be made out. In this connexion Maedougald's (1936) observations in the cardiac muscle-fibres of
the developing chick are interesting. In the cells of the 18th day
chick the apparatus occurs in the longitudinal axis as two
filaments one on either side of the nucleus. This can be seen
in his figs. 14 and 15. In fig. 16 he shows a transverse section
of the muscle-cell. He remarks that the Golgi filament is not
a solid argentophile mass but ring-shaped in transverse section.
Whether this argentophobe region inside the strand is comparable to the chromophobic portion of the dictyosomes he
does not say, but he does not describe an idiosome in relation
with the outside of the filaments and networks observed by
him. The important point in this connexion is that the chromophobic region of the Golgi filaments is not figured by Macdougald in longitudinal sections of cells. In the mucous cells
of the endostyle which are preparing for the production of
mucous droplets the chromophobic area could be seen, and I
think it corresponds to the idiosome of the Golgi dictyosomes.
Photographing an object under an oil-immersion lens is difficult,
and I believe that the difficulty in getting a differentiation in
the negative is due to the fact that the difference in colour
between the chromophilic and chromophobic areas does not
give a sufficient contrast. Anyway, whether this state of affairs
is accepted or not at this stage, the chromophobic area inside
these bodies can be made out when secretion is actually beginning to appear. Even those bodies which have a differentiated
chromophobie region show appearances suggesting division.
The number of Golgi bodies increases enormously, and they
appear scattered irregularly through the cytoplasm. In not
all of them can a chromophobic differentiation be noticed. As
can be made out from Text-fig. 7 (photomicrograph 3, PL 25)
bodies without a duplex structure lie scattered among those
with a duplex structure. The thickness of the chromophilic rims
in bodies with a duplex structure varies in different regions of
its circumference. Some portions are thin and some others
thick. When the Golgi bodies are thus increasing in size and
440
M. K. SUBBAMANIAM
number, and becoming scattered through the cell, the mitochondrial grains break up into small granules. In Nassonov
slides stained with iron haematoxylin the Golgi elements and
mitochondria could be seen side by side.
TEXT-ITC. 7.
Increase in number and size of the Golgi bodies. Note their
distribution throughout the cytoplasm.
Thus having become irregularly scattered throughout the
cell, the Golgi bodies begin to secrete mucus. The first change
in the Golgi body is a loss of the brown colour of the chromophobie region as can be made out from photomicrograph 3,
PI. 25, and Text-fig. 7. Very peculiar shapes are assumed by
the Golgi elements during these stages (Text-fig. 7). Soon after
441
STBtTCTUBE OF THE GOLGI AFPABATUS
the idiosome has lost its brown colour an increase in size of the
vesicle takes place due to the formation of mucus. At A in
Text-fig. 8 is shown the beginning of the increase in size. During
this period the enlargement results in thin regions of the chromo-
-m.
TEXT-ITG. 8.
Mucus formation in the chromophobic part of the Golgi elements.
philic rims becoming still thinner. In Text-fig. 8 at B and c are
shown the appearance of Golgi elements loaded with mucus.
The different thickenings of the chromophil region in different
elements containing mucus are comparable, naturally, to those
of elements without any secretion inside. As secretion goes on
rupture of the rim takes place in the regions where it is thin,
leaving the thicker bits in contact with the mucous droplets
as shown in Text-fig. 8 at D, Text-fig. 9 at g3, and in photo-
442
M. K. SUBBAMANIAM
micrograph 4, PI. 25. The number of these bodies attached to
mucous globules varies from 1 to 6 though usually it is only
1 to 3. Secretion of mucus occurs in the chromophobic part of
the majority of the bodies. When the globule is fifly formed the
TEXT-FIG. 9.
Increase in size of the mucous droplets and the rupture of the
chromophilio rims.
Golgi elements attached to the mucous droplets begin to disintegrate. Finally, in gland-cells about to extrude their contents
(Text-fig. 10, photomicrograph 4, PI. 25), we get rows of mucous
globules with few Golgi bodies which do not show any duplex
structure. As mentioned before, the droplets are sent out
together and in this process some of the Golgi bodies may also
be carried to the pharyngeal lumen. One surmises that those
left in the cell begin to multiply once again.
STBUCTUBE OP THE GOLGI APPARATUS
443
DISCUSSION.
S t r u c t u r e of the (xolgi A p p a r a t u s .
In a previous contribution (Subramaniam, 1938 a) the views
regarding the existence of a network in vertebrates were ana-
TEXT-FIG. 10.
Cells loaded with mucous droplets. Golgi elements •without any
duplex structure and some batonettes in relation with the mucous
droplets are shown.
lysed, and it was shown that the Golgi apparatus cannot be
considered to have a netlike shape in a l l vertebrate tissues.
It was shown that the archoplasm described by many authors
(Woodger, 1925; Brambell, 1925; Eau and Ludford, 1925) was
not observed by them in Da Fano or Mann Kopsch preparations,
but that it was supposed to exist on the belief that the centrosphere and archoplasm were one and the same structure. Most
444
M. K. SUBRAMANIAM
of the authors who describe an archoplasm in relation with the
Golgi apparatus, basing their observations on a comparison of
the Da Fano or Mann Kopsch and ordinary histological preparations side by side, do not seem to hrve considered the
discrepancy that results from the above assumption that the
centrosphere and centriole should be in different regions of
the cytoplasm in cases where the centriole is intra-nuclear (King
and Gatenby, 1923) or associated with the cilia. In L e c u d i n a
(Subramaniam and Ganapati, 1938) it was shown that the
centriole is intra-nuclear and that the Golgi bodies in the cytoplasm differentiate a chromophohic area during some stages.
In B r a n c h i o s t o m a the other condition is illustrated. Components of the central apparatus are seen in relation with the
cilia, while the Golgi elements lying concentrated in the basal
and middle regions differentiate idiosomic areas in their interior.
This can be very clearly made out from the figures and photomicrographs. The evidence presented in L e c u d i n a and
B r a n c h i o s t o m a will convince any one that the suppositions
of King (1926) and Ludford and Gatenby (1921), who describe
separation of the chromophilic and chromophobic regions in the
sporoblast stage of H a p l o s p o r i d i u m c h i t o n i s and during
spindle formation in the maturation divisions, are erroneous.
The examples and results cited above strengthen my conviction that the chromophobic part of the Golgi apparatus can
have nothing to do with the centrosphere.
Turning now to the question of actual evidence regarding the
probable mode of formation of a net-like Golgi apparatus, the
investigations were as disappointing as the analysis of the papers
presented previously (Subramaniam, 1938 a). I am not, however, in a position to judge whether, like the presence of true
protonephridia, homologous with the protonephridia of the
Invertebrata (Goodrich, 1934, p. 671), the presence of the Golgi
apparatus in the form of discrete elements, is the retention of
an invertebrate characteristic by B r a n c h i o s t o m a .
F u n c t i o n of t h e Golgi A p p a r a t u s .
Bowen in 1926 suggested that the secretory theory of the
Golgi apparatus must depend on evidence on three points.
STITOCTUKE OF THE GOLGI APPARATUS
445
These are: (1) a demonstration of the general topographical
relationship of the Golgi apparatus and secretory granules such
as to allow the possibility of causal connexion between them;
(2) a more intimate knowledge of the finer structural features
of the Golgi apparatus in general; and (3) a critical demonstration of the relation of an individual secretory granule to the
material of the Golgi complex. These criteria are equally
applicable to mitochondria, and should have been applied by
those who postulate the origin of early secretion granules in
relation with the mitochondria. I shall take here the work of
Duthie (1934) as representing the school of thought which considers that the early secretion granules arise in relation to the
mitochondria, and show how inconclusive is the evidence.
Taking the first point of Bowen, viz. the general topographical
relationship between the secretory granules and mitochondria,
Dutbie describes the mitochondria asfilamentous,and that these
filaments lie transversely in the region above the nucleus and
longitudinally in the other regions. The Golgi apparatus is
supposed not to extend to the region above the nucleus thus
supporting a circumstantial relation between the mitochondria
and neutral red globules. He says: ' The question of the origin
of secretory granules in the pancreas has, however, been decided
by the study of changes occurring in the living cell by Hirsch
(1932). This method in the present author's experience shows
clearly that for the pancreas the early secretory elements are
found a t t h e cell b a s e 1 in fairly close topographical relationship to the mitochondria' (p. 44). As will be seen from his
Text-fig. 6 the mitochondria are scattered in the upper half of
the cell and the secretory or neutral red granules are supposed
to a r i s e o n l y a t t h e cell b a s e . This fact should lead
to the existence of two different kinds of mitochondria; those
in the basal region alone being responsible for the production
of neutral red globules, which differentiation Duthie does not
mention. The above-mentioned difficulty, therefore, reduces
the possibility of a topographical relation to one of little value.
Regarding the intimate knowledge of the structural details
of the mitochondria Hirsch and Duthie do not add much to
1
The spaced words are mine.
446
M. K. SUBRAMANIAM
the existing knowledge. Some of Duthie's descriptions require
elucidation. For example, in the sero-acinar cells of the rat's
submaxillary gland (Text-fig. 10, p. 33) he describes the mitochondria as filamentous bodies. In addition he descrit es certain
round deeply staining bodies also. He supposes these granular
bodies to be young secretory granules as similar mitochondria
are not found in the fresh condition. In Text-fig. 10 the mitochondria have no definite orientation, and naturally one expects
these bodies to be cut transversely also. This mistake in not
taking into consideration the possibility of the mitochondrial
filaments being cut transversely, detracts from the value of his
observations. The fact-that all neutral red staining granules
cannot be considered as prozymogen in fresh tissue may be
seen from the fact mentioned by Duthie that some of these are
insoluble in alkali. Hence in sections, in addition to the mitochondria being cut transversely, the above-mentioned granules
also lead to some difficulty in accepting the claims by Duthie.
Coming to the final point, viz. the critical demonstration of an
individual secretory granule with the mitochondria, Duthie's
and even Hirsch's (1931 and 1932) descriptions are not convincing. Janus green has a tendency to fade from the cell, and
in addition more details regarding the structure and position
of the Golgi apparatus are required before we can accept
Duthie's claims.
Turning now to the mucous secretion in the endostyle, it will
be apparent from the descriptions, drawing, and photomicrographs that there is no question of the mitochondria having
anything to do with the visible production of secretion globules
because these arise inside the Golgi elements. According to
Bowen's classification of the types of association between the
Golgi and secretory granules (1926), it appears that the endostyle glandular cells come under the category where the
granules remain in association with the Golgi apparatus till
they mature. This type of association has been found to occur
in the pelvic gland (Nassonov, 1923), and in the pancreas of
some salamanders (Bowen, 1924). Thus in B r a n c h i o s t o m a
the origin and relation between the mucous globules and the
Golgi gives evidence of the highest importance. A similar
STRUCTURE OF THE GOLGI APPARATUS
447
topographical relation between the Golgi and the developing
secretory granule has been described by Ludford (1925) during
fat formation in the tar tumours, and receatly in the Harderian
glands of the rat by Duthie (1934). To my knowledge I am the
first to present clear photomicrographic evidence of such a
connexion hetween the forming secretory droplets and the Golgi
elements in a somatic cell.
As far back as 1926 Bowen pointed out that the Golgi apparatus is not an autonomous structure, but part of the cellular
system as a whole. He conceived it merely as the immediate
focus of secretory synthesis, an activity to which the other parts
of the cell must contribute in one way or another. Thus the
materials for the synthetic operations must come from the
surrounding cytoplasm; and in their production the mitochondria, the general cytoplasmic background, and the nucleus
may be more or less involved. The final source is, of course,
from the food taken which is passed on to the blood. Thus while
the existence of this chain is fully realized, increasing evidence
has been accumulating that the Golgi apparatus is the immediate source of secretory synthesis. The wide distribution
of the mitochondria in gland-cells certainly indicate that they
play some part in cell economy. But the striking changes in
the Golgi apparatus in relation to secretion formation noted
by many workers, and clearly shown in the endostyle of
B r a n c h i o s t o m a , indicate that compared to mitochondria,
the Golgi apparatus present more critical, circumstantial, as
well as topographical evidence that it plays a more important
part. Similarly to the sero-acinar cells of the submaxillary and
the Harderian gland (Duthie, 1934 a, b) which form exceptions
to the ease reported in the pancreas, there may be some cells
in which the mitochondria may play a more important part.
Because they are exceptions more critical evidence is necessary
in order to establish that it is so. Only further work, in which
the shape and structure of the Golgi apparatus and mitochondria are given as much attention as the actual origin of
the secretion granules, can give us a clear idea as to the relative
roles played by the two cytoplasmic components in secretion.
448
M. K. SUBEAMANIAM
SUMMARY AND CONCLUSIONS.
The work on B r a n c h i o s t o m a was undertaken with a
view (1) to study the relation between the centi J and Golgi
apparatus in a ciliated cell where the parts of the central
apparatus are supposed to be associated with the cilia, and (2)
also to see whether any hint as to the mode of evolution of the
network can be obtained.
Each group of mucous secreting cells seems to be in a particular stage of activity. The variations in activity between the
individual cells in a particular group is very small. It was
observed that generally, while two groups of gland-cells are
loaded with mucus, the other two are at the beginning of the
secretory cycle.
The mitochondria, which appear as a few large grains in the
cells at the beginning of secretory activity, break up into small
granules before mucous droplets put in their appearance. Most
of these mitochondrial granules are extruded with the mucus.
The Golgi apparatus occurs as a few small grains having an
indefinite distribution. There is no ' Golgi zone', nor are these
elements collected around a centriole. These increase in size
and number and as they do so iii many of them the central
area becomes brownish. The Golgi elements finally get scattered
through the cell. This scattering is soon followed by the appearance of mucus in the chromophobic areas of the elements.
Increase in size of secretion droplets leads to rupture of the
osmiophilic portion in regions where it is thin, and mucous
globules with Golgi caps may be seen in the cytoplasm. Most
of the Golgi elements degenerate, and those left behind apparently become reorganized for the next secretory cycle.
Examples and results are cited to show that the chromophobic part has nothing to do with the centrosphere. It appears
difficult to decide whether the presence of Golgi apparatus in
the form of discrete elements is the retention of an invertebrate
characteristic by B r a n c h i o s t o m a .
The recent conclusions of Hirsch and Duthie are discussed,
and it is shown that unless equal importance is given in future
studies to the shape and structure of the Golgi apparatus and
STRUCTUEB OF THE GOLGI APPABATUS
449
mitochondria as is at present given to the function of these
eytoplasmic components we cannot accept the claims put forward by these authors. There appear to be exceptions to every
rule, and in order to accept these as exceptions critical evidence
is necessary.
ACKNOWLEDGEMENTS.
My thanks are due to Professor E. Gopala Aiyar for suggesting
the work, as also for his encouragement and criticisms of the
manuscript and results. My thanks are also due to the University of Madras for awarding me a Eesearch Fellowship.
REFERENCES.
Bowen, B . H., 1924.—"Possible Relation between the Golgi Apparatus
and Secretory Products", 'Amer. Journ. Anat.', 33.
1926a.—"Golgi Apparatus in Gland Cells, Parts I-IV", 'Quart.
Journ. Micr. Sci.', 70.
19266.—"The Golgi Apparatus—Its Structure and Functional
Significance", 'Anat. Eec.', 32.
Brambell, F. W. R., 1925.—"The Part played by the Golgi Apparatus
in Secretion and its subsequent reformation in the cells of the Oviducal
Glands of the Fowl", 'Journ. Roy. Micr. Soo.'
Duthie, B. S-, 1934a.—"Secretion of the Pancreas and Salivary Glands",
'Proc. Roy. Soc.', B, London, 114.
19346.—"Cytology of the Harderian Gland", 'Quart. Journ. Micr.
Sci.', 76.
Goodrich, E. S., 1934.—"Early Development of the Nephridia in Amphioxus. Part II. The Paired Nephridia", ibid., 76.
Hirsch, G. C , 1931.—"Theory of Fields of Restitution with special reference to the phenomena of Secretion", 'Biol. Rev. Camb. Phil. Soc.', 6.
1932.—"Lebendbeobachtung der Restitution des Sekretes im Pankreas. I. Teil. Restitution der Druse als gauzes nach Pilokarpinreizung", 'Zeit. Zellf. u. mikr. Anat.', 15.
King, S. D., 1926.—"Cytological Observations on Haplosporidium
(Minchinia) chitonis", 'Quart. Journ. Micr. Sci.', 70.
King, S. D., and Gatenby, J. B., 1923.—"Golgi Bodies of a Coccidian",
ibid., 67.
Ludford, R. J., 1925.—"General and Experimental Cytology of Cancer",
'Journ. Roy. Micr. Soc.'
Ludford, R. J., and Gatenby, J. B., 1921.—"Dictyokinesis in Germ Cells
or the Distribution of the Golgi apparatus during cell division", 'Proc.
Roy. Soc.', B, London, 92.
NO. 323
Gg
450
M. K. SUBRAMANIAM
Macdougald, T. J., 1936.—"Cytology of the Muscle. I. Golgi Apparatus
in the fibres of the Cardiac Muscle", 'Zeit. zellf. u. mikr. Anat.', 24.
Nassonov, D., 1923.—"Das Golgische Binnennetz und seine Beziehungen
zu der Sekretion", 'Arch. f. mikr. Anat.', 97.
Prashad, B., 1934.—"Collection of Cephaloehordates . . . in the Indian
Museum",'Rec. Ind. Mus.', 36.
Eau, A. S., and Ludford, R. J., 1925.—"Variation in form of Golgi Bodies
during the Development of Neurones", 'Quart. Journ. Micr. Sci.', 69.
Subramaniam, M. K., 1938a.—'Shape, Structure and Position of the
Golgi Apparatus in Vertebrate Cells.' (In press.)
19386.—"Structure of the Golgi Apparatus. I I . liver Cells of
Rhacophorus maculatus Gray", 'Proc. Ind. Acad. Sei.', 7.
1938c.—"Structure of the Golgi Apparatus. III. Mechanism of
Secretion of Golgi Bodies in the Intestinal Cells of Lumbriconereis",
'Proc. Ind. Acad. Sci.', 7.
Subramaniam, M. K., and Ganapati, P. N., 1938.—"Structure of the Golgi
Apparatus. I. Cytoplasmie Inclusions in the gregarine Lecudina
brasili", 'Cytologia', 9.
Subramaniam, M. K., and Gopala Aiyar, R., 1936.—"Possible Mode of
Evolution of the Network-like Golgi Apparatus of Vertebrate Somatic
Cells from Discrete Bodies of Invertebrates", 'La Cellule', 45.
1937.—"Shape and Structure of the Golgi Bodies in the Eggs of
Invertebrates with a Note on the Probable Mode of Origin of the Golgi
Network", 'Proc. Ind. Acad. Sci.', 5.
Willey, A., 1894.—"Amphioxus and the Ancestry of Vertebrates",
'Columbia Univ. Biol. N.Y.'
Woodger, J. H., 1925.—" Origin of the Germ Cells of the Fowl (Gallus
bankiva) studied by means of their Golgi Bodies", 'Quart. Journ.
Micr. Sci.', 69.
EXPLANATION OP PLATE 25.
All Photomicrographs were taken from Nassonov preparations at a
magnification of 100 x 15 with a bellows extension of 8 in.
Fig. 1.—The Golgi apparatus in cells at the beginning of the secretory
cycle.
Fig. 2.—Hypertrophy of the Golgi apparatus prior to secretory droplet
formation. The Golgi elements have increased in size and number and are
scattered throughout the cytoplasm.
Fig. 3.—Droplets of mucus being formed in the chromophobic areas of
the Golgi Bodies. In some cases the chromophilic rims have ruptured
and batonettes are seen sticking to mucus globules.
Fig. 4.—The cytoplasm loaded with droplets of mucus prior to their
extrusion. Some Golgi bodies are also seen.
Quart. Journ. Micr. Sci.
"•
• Subraman
FIG.
1
FIG.
2
Vol 81, N. S., PL 25
FIG.
FIG.
4
3