The large axons in Periplaneta are composed of bundles of osmiophobe strands about 0·5 µ thick which fan out into the body of the nerve-cell. These strands are here termed ‘neurofibrils’; it is suggested that the dictyosomes (Golgi bodies) are concerned in their secretion.

The dictyosomes are well stained by the osmium and ethyl gallate method. Each dictyosome surrounds or is applied to an unstained canal which runs into the cytoplasm and is believed to be continuous with the ‘neurofibrils’ as defined. There are all intermediate stages between thin fusiform sheaths of osmiophil material around the ‘neurofibrils’ and the thick rings or cuffs which form the dictyosomes. The clear canals through the dictyosomes are arranged in the main concentrically around the nucleus in the body of the cell; they commonly converge upon the nerve-fibre in the axon cone.

The purpose of this study is to consider possible connexions between the dictyosomes or Golgi bodies and the structural organization of insect axons. The neurones are exceedingly rich in lipids which undoubtedly play a major role in maintaining the cell structure. Earlier work has shown that the stabilization of lipid membranes by fixation with buffered osmium tetroxide, followed by visualization of the bound osmium with ethyl gallate (Wigglesworth, 1957, 1959 a, b), reveals structural details that are not readily seen by standard histological methods.

The last abdominal ganglion of the adult male cockroach, Periplaneta americana (L.), has been used for most of the observations, the third thoracic ganglion for some. Sections were cut at 0·5 to 2 µ. (Wigglesworth, 1959c). Some observations on the glial invaginations (Holmgren’s canals or ‘trophospongium’) of the large ganglion cells are being published elsewhere (Wigglesworth, 1960). The fine structure of cockroach ganglia as seen with the electron microscope is described by Hess (1958).

Structure of the large axons

The large axons of insects as seen in silver preparations appear to be made up of a number of neurofibrils which spread out in a fan-like manner in the cell-body to form a meshwork around the nucleus (Beams and King, 1932). But in silver preparations the cytological structure is so greatly impaired that the real existence of these neurofibrils has been a frequent subject of controversy.

The axons in Rhodnius as seen in sections stained with osmium and ethyl gallate (Wigglesworth, 1959b) have a rope-like structure and consist of unstained neurofibrils which fan out to form an investment of the nucleus on entering the cell-body. The existence of such neurofibrils is even more evident in the cockroach.

Giant axons arise from large cells in the last abdominal ganglion and run to the opposite side, forming a conspicuous chiasma where they intersect and cross over (fig. 1, A).

FIG. 1.

(plate). All a-ji sections, osmium tetroxide, and ethyl gallate.

A, giant axons crossing in mid-line, showing neurofibrils and mitochondria.

B, oblique section of axon showing neurofibrils.

C, transverse section of ganglion cells and base of axons, showing multiple invaginations by the glial cells.

D, oblique longitudinal section of base of axon, showing glial invaginations giving a false impression of neurofibrils.

E, ganglion cell showing rope-like neurofibrils in the axon traceable into the cell-body; dictyosomes converging on the axon.

F, dark type of ganglion cells with Golgi bodies ; the cell at top left shows neurofibrils from axon dispersing in the cell.

G, dark and pale types of ganglion cells ; the dark cells show clear canals (? neurofibrils) in the cytoplasm.

FIG. 1.

(plate). All a-ji sections, osmium tetroxide, and ethyl gallate.

A, giant axons crossing in mid-line, showing neurofibrils and mitochondria.

B, oblique section of axon showing neurofibrils.

C, transverse section of ganglion cells and base of axons, showing multiple invaginations by the glial cells.

D, oblique longitudinal section of base of axon, showing glial invaginations giving a false impression of neurofibrils.

E, ganglion cell showing rope-like neurofibrils in the axon traceable into the cell-body; dictyosomes converging on the axon.

F, dark type of ganglion cells with Golgi bodies ; the cell at top left shows neurofibrils from axon dispersing in the cell.

G, dark and pale types of ganglion cells ; the dark cells show clear canals (? neurofibrils) in the cytoplasm.

These axons are about 22 μ. thick. As seen in longitudinal section they consist of a great number of uniform unstained fibrils about 0·5 μ in thickness, which run a more or less wavy course along the axon. The matrix which fills the spaces between the fibrils is only slightly more osmiophil than the fibrils themselves, but this interfibrillar material stains sufficiently darkly to show up the white strands between. The whole structure resembles a bundle of cooked spaghetti with raisins in the form of mitochondria lying at intervals between the strands. Where the axon divides, a greater or smaller number of fibrils separate from the main bundle to form the branch. Where the axon comes off the cell-body the neurofibrils sometimes follow an oblique course which gives the whole fibre a rope-like appearance (fig. 1, E).

In transverse sections it is possible to see the fibrils very faintly as clearer holes in a very pale grey matrix. It is evident that the fibrils are not bounded by a continuous lipid membrane, which would define their limits much more clearly. It is doubtless for this reason that they do not show up well in electron micrographs.

There can be no doubt that these neurofibrils are constant structural components of the axons. They can be readily seen in smaller axons (fig. 1, B) down to a diameter of 3 μ or less; the individual fibrils remain about the same size throughout.

The axon is the product of the cell-body, and there is now good evidence that every axon is being continuously renewed by secretion from its ganglion cell (Koenig, 1958; Weiss, 1959). It would seem probable that there must be some organized structure within the cell to produce this degree of organization in the nerve-fibre.

The course of neurofibrils in the ganglion cell

When attempting to follow the neurofibrils into the body of the ganglion cell it is easy to be misled by the glial invaginations. As described elsewhere (Wigglesworth, 1960) these invaginations become progressively developed towards the base of the axon. Transverse sections at this level show these invaginations as radially disposed flanges or laminae extending far into the cell (fig. 1, c). Longitudinal sections in this region cut in a slightly eccentric plane can produce an appearance which superficially resembles a bundle of neurofibrils (fig. 1, D). But, as already pointed out, the neurofibrils are not bounded by lipid sheaths such as form the walls of the glial invaginations. They are far less conspicuous structures than these ‘pseudofibrils’.

None the less it is often possible to follow the true neurofibrils into the cell. They sometimes fan out evenly in all directions. Sometimes they follow a zigzag course. It is seldom possible to trace individual fibrils very far; they are soon lost among the granular contents of the cytoplasm (figs. 1, E, F; 2, A, G).

As pointed out by Hess (1958) the ganglion cells of the cockroach as seen with the electron microscope are of two sorts, dark cells and pale cells, depending on the density of the cytoplasm. These two kinds of ganglion cell are equally distinct after staining with osmium and ethyl gallate (figs. 1, G; 2, c). It is often difficult to detect the colourless neurofibrils in the cytoplasm of the pale cells; but in the dark cells they can readily be seen. The cell may sometimes appear filled with these fine, convoluted, unstained, worm-like structures (figs, 1, G; 2, c). If the cell is cut at right angles to the line of entry of the axon it is possible, by changing the plane of focus on a 2-µ section, to follow the fibrils up and down. They are not unlike the pore canals as seen in a tangential section of the cuticle, but follow a much less regular course.

The dictyosomes (Golgi bodies) and the neurofibrils

The object in following the neurofibrils in the body of the ganglion cell was to seek some visible structure which might be concerned in their secretion. It seems probable that the dictyosomes or Golgi bodies are involved.

The Golgi bodies are very evident in the cytoplasm of the ganglion cells as darkly staining objects varying from 0 · 75 μ to 3 · 0 μ across (fig. 1, F, G). They take the form of crescents, horse-shoes, parallel bars, or rings. For the most part they stain a deep brownish grey or black with osmium and ethyl gallate, which suggests that phospholipids are largely responsible for binding the osmium. But not uncommonly they have a thin outer shell which stains a uniform or granular blue black and is perhaps composed of triglycerides (fig- 3,A.

Every Golgi body is associated with a colourless unstained object or ‘vacuole’. The osmiophil substance may be applied to this in the form of a crescent, but usually surrounds it as a more or less complete ring. By focusing up and down it is generally possible to see that the colourless part of the Golgi body is a clear canal and that this becomes continuous with a similar canal which is soon lost in the cytoplasm. The dimensions of these canals are about the same as those of the neurofibrils.

It is difficult to avoid the impression that the Golgi bodies are in fact deeply staining cuffs around unstained canals which become continuous with the neurofibrils of the axons. The evidence in support of this interpretation may be formulated as follows :

(i) The Golgi bodies may sometimes include slightly larger vacuoles, but in nearly all of them the canals which they enclose are of the same size as the neurofibrils.

(ii) In large ganglion cells in which the Golgi bodies are very numerous, notably in the large pale cells which give rise to giant axons, the Golgi bodies have a characteristic orientation which agrees with the orientation of the neurofibrils. Throughout the greater part of the body of the cell they lie with the lumen, that is with the long axis, disposed concentrically around the nucleus (fig. 2, D); whereas in the axon cone they converge upon the base of the axon (figs. 1, E; 2, A, E; 3, B, C).

FIG. 2.

(plate). A, ganglion cells with Golgi bodies oriented towards base of axon; neurofibrils well seen in the cell to the left.

B, 4- μ section of ganglion cells; alcoholic Bouin and toluidine blue. Dictyosomes appear as clear spaces with RNA deposits between.

c, pale and dark ganglion cells; the dark cell to the right shows worm-like canals (?neuro- fibrils) in the cytoplasm ; the pale cell to the left shows dictyosomes oriented towards the axon.

D, pale cell showing concentric arrangement of dictyosomes.

E, similar cell to D with Golgi bodies oriented towards the axon (above).

F, ganglion cells 5 days after section of nerve, showing disorganization of axons.

G, corresponding cells on the normal side of the same insect.

FIG. 2.

(plate). A, ganglion cells with Golgi bodies oriented towards base of axon; neurofibrils well seen in the cell to the left.

B, 4- μ section of ganglion cells; alcoholic Bouin and toluidine blue. Dictyosomes appear as clear spaces with RNA deposits between.

c, pale and dark ganglion cells; the dark cell to the right shows worm-like canals (?neuro- fibrils) in the cytoplasm ; the pale cell to the left shows dictyosomes oriented towards the axon.

D, pale cell showing concentric arrangement of dictyosomes.

E, similar cell to D with Golgi bodies oriented towards the axon (above).

F, ganglion cells 5 days after section of nerve, showing disorganization of axons.

G, corresponding cells on the normal side of the same insect.

(iii) Where the neurofibrils are entering the base of the axon they are often enclosed in thin elongated cuffs of osmiophil material which taper away at both ends. Cuff-like structures of this kind may be seen radiating out from the base of the axon, and all intermediate stages between them and typical dictyosomes can be found (fig. 3, c). This suggests that the typical dictyosome results from the increasing concentration of this diffuse osmiophil covering into discrete points on the fibril. Such an interpretation would account for the great diversity in the form of the bodies in different ganglion cells.

FIG. 3.

A, series of typical dictyosomes from the group of dark ganglion cells shown in fig. 1, F, G. The bracketed groups below represent single dictyosomes as seen at three levels of focus.

B, detail of dictyosomes and mitochondria seen in the axon cone of the cell to the left of fig. 2, A. The axon lies to the right of the drawing.

c, selected dictyosomes from the pale cell shown in fig. 2, E. The axon lies beyond the upper part of the drawing. The dictyosomes have their actual orientation ; those above were in the axon cone, those below in the body of the cell.

FIG. 3.

A, series of typical dictyosomes from the group of dark ganglion cells shown in fig. 1, F, G. The bracketed groups below represent single dictyosomes as seen at three levels of focus.

B, detail of dictyosomes and mitochondria seen in the axon cone of the cell to the left of fig. 2, A. The axon lies to the right of the drawing.

c, selected dictyosomes from the pale cell shown in fig. 2, E. The axon lies beyond the upper part of the drawing. The dictyosomes have their actual orientation ; those above were in the axon cone, those below in the body of the cell.

(iv) An attempt was made to observe the changes in the Golgi bodies during axon regeneration. The crural nerve (nerve 5 of Pringle, 1939) runs immediately below the very thin cuticle at the base of the coxa. It is easy to make an incision at this point and to cut the nerve without injury to the tracheal system. This operation was performed on the left metathoracic leg in a dozen adult cockroaches about one month old.

The crural nerve is mainly a sensory nerve, but it does contain some 50 motor axons of 3 to 10 µ diameter. After section the sensitivity of the tarsus is lost and that leg is not used in walking. But, as Bodenstein (1957) has shown, regeneration takes place and motor activity is being recovered by 6 weeks after the operation.

As in Rhodnius (Wigglesworth, 1959b) the motor axons form two groups, with thin sheaths and with thick sheaths. Some at least of the motor axons of nerve 5 after entering the metathoracic ganglion run inwards and ventrally to end in a group of ganglion cells on the same side, which lie just a little anterior to the point of entry of the nerve, and are very readily compared with the corresponding cells of the opposite side.

These two groups of cells were compared at different periods after section of the nerve. At 5 days after section the ganglion cells on the operated side had their axons distended and apparently disorganized with no distinct neurofibrils visible (fig. 2, F). Within the cytoplasm of the cell-body the canals associated with the Golgi bodies were more dilated than on the control side.

At 3 weeks and 6 weeks after nerve section, when the movements of the muscles showed that regeneration was in progress, the cells on the operated side showed several differences from those on the control side. The nucleoli were greatly enlarged. The cytoplasm was more deeply stained, and close examination showed that this staining largely took the form of a diffuse dark sheath around the clear canals that are presumed to be neurofibrils. This darkening was associated with a darker staining of the Golgi bodies.

It looks as though during the regeneration of the axons there is an increase in the activity of the Golgi bodies.

Nucleic acid in the ganglion cell

It has been shown recently by Malhotra (1959) that cytoplasmic nucleic acid in the neurones of vertebrates has the same distribution as the Golgi apparatus and he concludes that in this material the Golgi structure results from the deposition of silver or osmium on the Nissl bodies. In the neurones of HelixBoyle (1937) found that the Nissl substance occurs in fine flocculent granules dispersed throughout the cytoplasm. A similar distribution of ribonucleic acid was observed by Shafiq and Casselman (1954) in the neurones of Locusta.

These results have been confirmed in Periplaneta. Ganglia were fixed with alcoholic Bouin solution and sections stained with toluidine blue, pyronin / methyl green, and gallocyanin before and after incubation with ribonuclease. RNA was present in the form of fine granules dispersed throughout the cytoplasm between the dictyosomes with a slight increase in density around the nucleus and possibly around the surface of the dictyosomes. The dictyosomes themselves appear as clear rounded spaces devoid of RNA (fig. 2, B).

According to current ideas the Golgi apparatus is concerned in some way with the later stages of the process of secretion in the cell. That has indeed been the opinion of classical cytologists for many years. The most suggestive evidence has been provided by a study, with the light microscope and with the electron microscope, of the cycle of secretion in the mammalian pancreas (Hirsch, 1958, 1959).

Whether the dictyosomes in the nerve-cells of invertebrates are homologous with the varied types of ‘Golgi apparatus’ in other cells is still a matter of controversy among cytologists (Baker, 1954; Cain, 1954); but the account of the dictyosomes in the neurones of Patella as seen in electron microscope sections (Lacy, 1957) agrees very well with the fine structure of the ‘Golgi apparatus’ in many other types of cells (compare Gatenby and Lutfy, 1956; Pollister and Pollister, 1957).

The secretory product of the neurone is the axon. There is now good evidence for a continuous secretion of axoplasm by the cell-body of the neurone (Koenig, 1958 ; Weiss, 1959). The large axons of the cockroach have been shown in the present paper to be highly organized at the light microscope level and made up of a great number of neurofibrils. It was therefore reasonable to expect some organized structure within the cell-body responsible for the secretion of these fibrils of axoplasm.

The histological evidence here presented suggests that the neurofibrils are traceable to the dictyosomes. There are all intermediate stages between cuffs of osmiophil material around unstained neurofibrils as they enter the base of the axon, and typical Golgi bodies of annular or horse-shoe form, the unstained core of which is continuous with unstained filaments that are thought to be neurofibrils. The orientation of the dictyosomes, with their unstained core or long axis concentric with the nuclear membrane in the body of the cell, and converging upon the base of the axon when they lie in the axon cone, reinforces this impression.

Fig. 4 is a schematic illustration of the relation that is inferred. Golgi bodies or a series of Golgi bodies are pictured as providing the matrix for individual neurofibrils. No method is available which will stain both axoplasm and Golgi bodies in the same preparation, and it is impossible to follow unstained filaments very far in the granular cytoplasm of the ganglion cells. It has therefore not been possible to make preparations which show irrefutably that every Golgi body is connected to a neurofibril, but that is what the evidence suggests.

FIG. 4.

Schematic drawing of ganglion cell showing glial invaginations and the suggested relation between dictyosomes and neurofibrils.

FIG. 4.

Schematic drawing of ganglion cell showing glial invaginations and the suggested relation between dictyosomes and neurofibrils.

It must be emphasized that these unstained strands which give a rope-like appearance to the large axons are structures of half a micron or so in thickness. They are quite distinct from the long slender filaments, 50-100 Å in diameter, which are so conspicuous in thin sections of axons examined with the electron microscope (Fernández-Morán, 1952; Vial, 1958; Wigglesworth, 1959b). These filaments of fibrous protein form less than a tenth of 1 % of the total axon volume (Schmitt, 1957), whereas the unstained neurofibrils here described (0 · 25 to 0 · 5 μ in diameter) make up the bulk of the axon as seen with the light microscope, and there is relatively little stainable material between.

It is this unstained strand, which must contain a large number of the widely spaced 100-Â filaments, that has been referred to throughout this paper as a neurofibril. In their studies on the structure of the mammalian neurone Palay and Palade (1955) showed more or less parallel arrays of the 100-Å filaments running in clear channels, some half a micron or less in width, between the Nissl bodies. It is channels of this type which would correspond with the neurofibrils discussed in this paper.

Baker
,
J. R.
,
1954
.
J.R. micr. Soc
.,
74
,
217
.
Beam
s.H. W.
, and
King
,
R. L.
,
1932
.
J. Morph-
,
53
,
59
.
Bodenstein
,
D.
,
1957
.
J. exp. Zool
.,
136
,
89
.
Cain
,
A. J.
,
1954
.
Ibid
.,
74
,
222
.
Fernández-Morán
,
H.
,
1952
.
Exp. Cell Res
.,
3
,
282
.
Gatenby
,
J. B.
, and
Lutfy
,
R. G.
1956
.
Nature
,
177
,
1027
.
Hess
,
A.
,
1958
.
J. biophys. biochem. Cytol
.,
4
,
731
.
Hirsch
,
G. C.
,
1958
.
Naturwiss
.,
45
,
349
.
Hirsch
,
G. C.
,
1959
.
Materia Medica Nordmark
,
11
,
3
.
Koenig
,
H.
,
1958
.
J. biophys. biochem. Cytol
.,
4
,
785
.
Lacy
,
D.
,
1957
.
Ibid
.,
3
,
779
.
Malhotra
,
S. K.
,
1959
.
Quart. J. micr. Sci
.,
100
,
339
.
Palay
,
S. L.
, and
Palade
,
G. E.
,
1953
.
J. biophys. biochem. Cytol
.,
1
,
69
.
Pollister
,
A. W.
, and
Pollister
,
P. F.
,
1957
.
Int. Rev. Cytol
.,
6
,
85
.
Pringle
,
J. W. S.
,
1939
.
J. exp. Biol
.,
16
,
220
.
Schmitt
,
F. O.
,
1957
.
J. cell. comp. Physiol
.,
49
(
Suppl
.),
165
.
Shafiq
,
S. A.
, and
Casselman
,
W. G. B.
,
1954
.
Quart. J. micr. Sci
.,
95
,
305
.
Vial
,
J. D.
,
1958
.
J. biophys. biochem. Cytol
.,
4
,
551
.
Weiss
,
P.
,
1959
.
Science
,
129
,
1290
.
Wigglesworth
,
V. B.
,
1957
.
Proc. roy. Soc. Lond. B
,
147
,
185
.
Wigglesworth
,
V. B.
,
1959a
.
Quart. J. micr. Sci
.,
100
,
285
.
Wigglesworth
,
V. B.
,
1959b
.
Ibid
.,
100
,
299
.
Wigglesworth
,
V. B.
,
1959c
.
Ibid
.,
100
,
315
.
Wigglesworth
,
V. B.
,
1960
.
(In the press
.)