Those who have studied the neurones of Crustacea have given very- divergent accounts of the ‘Golgi apparatus’ of these cells. Poluszynski (1911) described only short, straight or slightly sickle-shaped threads; he considered them as the Golgi apparatus. Monti (1915) believed that the Golgi apparatus of this cell was constituted by the mitochondria. She thought that they were long, wavy threads, which were blackened by the Golgi methods and formed a reticulum of the type seen in the neurone of vertebrates. Ross (1922) described a framework of ‘trophospongium’ connected with the surface of the cell. He thought this was additional to a reticular ‘Golgi apparatus’.

In view of these and other diverse opinions on the Golgi apparatus, I thought it desirable to reinvestigate the cytoplasmic inclusions of this cell and to try to interpret the above findings in the light of my own results.

Neurones from the thoracic and abdominal ganglia of the prawn, Leander serratus, and the crayfish, Astacus fluviatilis, were investigated. The prawns were obtained from the Marine Biological Laboratory, Plymouth, and kept until wanted in an aquarium. They were fed on Mytilus. The crayfish were obtained from the Surrey Trout Farm, Haslemere, Surrey. They were also kept in an aquarium and fed on cod.

Methods

1. Phase-contrast and polarizing microscopy for the study of unstained, living neurones.

2. Vital dyes. Neutral red, dahlia, and Janus green. Two drops of a 0·5% aqueous stock solution diluted with 2·5 ml of sea-water for the neurones of Leander.

3. Buffered osmium. Tissue was fixed in buffered osmium (Palade, 1952) or buffered osmium with the addition of calcium chloride (Baker, 1958) and embedded in n-butyl methacrylate as for electron microscopy. Sections were cut at about 3 μ and mounted in the monomer or in glycerol (Baker, 1960); these were studied by phase-contrast or interference microscopy.

4. Mitochondrial techniques. Material fixed in Helly (1903) for about 24 h (with postchroming) or Altmann or ‘NH4-Altmann’ (Baker and Luke, 1959) was stained by the method of Hirschler (1927) or Metzner (1928).

5. Nissl techniques. Tissue fixed in Zenker (1894; 3 h) or Helly was stained in pyronine /methyl green (Jordan and Baker, 1955) or cresyl violet (Fern- strom, 1958) or basic fuchsine. Ribonuclease (Bradbury, 1956) and trichloro- acetic acid (Pearse, 1954) were used for control.

6. Golgi techniques. The osmium impregnation methods of Kopsch (1902), Mann-Kopsch (Weigl, 1910), and Kolatchev (1916; after fixation in Champy, 1911) were used. The Mann-Kopsch method was also used on neurones that had previously been vitally coloured with neutral red. The colouring of the neurones was checked before transferring the tissue to the fixative. Post- osmication lasted 24 h in this case.

Ganglia of Leander were also fixed in a solution of osmium tetroxide in sea-water and post-osmicated as usual. When this material is embedded in paraffin, the results are not satisfactory. If, however, the tissue is embedded in methacrylate, the finished preparations are good. The details of this technique are as follows.

(i) Fix in a 1% solution of osmium tetroxide in sea-water (1 ml of a 2% aqueous osmium tetroxide and 1 ml of sea-water with the addition of 3·5 g of sodium chloride per 100 ml) for 2 to 3 h. Fixation for longer periods is not satisfactory.

(ii) Wash thoroughly in running water to get rid of the salts present in the sea-water; give a last wash in distilled water.

(iii) Post-osmicate in a 2% aqueous solution of osmium tetroxide at 34° C for 3 to 4 days.

(iv) Wash in running water for 6 h.

(v) Dehydrate and embed in w-butyl methacrylate.

(vi) Cut 3 μ sections and attach them to the slides like paraffin sections (Baker, 1960).

(vii) Dissolve the polymer in ethyl acetate.

(viii) Mount in Canada balsam.

Silver impregnation methods of Golgi (1898, dichromate/osmium; 1908, arsenious acid), Veratti (Golgi, 1900), Ramón y Cajal (1914), and Aoyama (1929) were used.

7. Material fixed in Helly, Lewitsky-saline (Baker, 1956a), and Carnoy (1886) was stained in iron haematoxylin.

8. Rawitz’s ‘inversion staining’ technique (Baker, 1959; Przełęcka, 1959).

9. Tissue fixed in Helly and embedded in paraffin was stained with Mas- son’s tricolor stain (Pantin, 1948) or xylidine red alone.

10. Tissue fixed in Zenker and embedded in paraffin was stained with Mallory’s phosphomolybdic acid /haematoxylin (1938).

The details of the techniques used for histochemical investigation of the neurones of Leander are given in the appendix. The neurones of Astacus were not studied histochemically.

Most of the previous work has been done on the neurones of Astacus fluviatilis. I hope to repeat these observations as far as possible. But these cells are difficult to fix well, especially the large cells. This difficulty of fixation has also been experienced by other workers on the neurones of Crustacea (Monti, 1915; Ross, 1922). For this reason I extended my study to the neurones of the prawn, Leander serratus; and I found them rather convenient for this investigation.

The cytoplasmic inclusions of the neurones of Leander will first be con- sidered in detail. The differences shown by those of Astacus will then be mentioned.

Morphology

Neurone of Leander

A wide variation occurs in the size of the nerve-cells in a ganglion. They can conveniently be grouped into small, medium-sized, and large neurones. The cytoplasmic inclusions discussed below are very similar in all three categories of cells, and a single description will suffice.

It will be helpful to the reader to start by mentioning briefly the kinds of cytoplasmic inclusions present in this cell, and then to relate in turn the appearances seen after the application of particular techniques.

Fig. 1 provides a summary of the results obtained. Four kinds of cytoplasmic constituents can be demonstrated. These are:

FIG. 1.

Diagrammatic illustration of the cytoplasmic inclusions of the neurone of Leander.

FIG. 1.

Diagrammatic illustration of the cytoplasmic inclusions of the neurone of Leander.

(1) Spherical or almost spherical bodies, which often show a binary structure; that is to say, the surface is distinct from the interior.

(2) Mitochondria in the form of short rods and granules.

(3) Nissl substance, diffusely distributed.

(4) Hollow tape-like objects, which penetrate into the neurone from its surface, and ramify in the form of filaments in the cytoplasm. These correspond to the ‘trophospongium’ of Ross (1922). They appear to be the invaginations of the cell membrane. The contents of the cells surrounding the neurone flow into these invaginations.

In the living neurone studied by phase-contrast microscopy, refringent spherical or almost spherical bodies are the most prominent of the cytoplasmic constituents. They are dispersed at random. They vary considerably in size up to about 2 or 212 μ in diameter. These spherical bodies often show a dark surface and a pale interior. This appearance suggests that lipid is probably localized at the surface. At times the dark edge seems to be unevenly spread round the pale globules. They seem to be very delicate structures, as they tend to become distorted under the pressure of the coverslip. These refringent spheres also extend into the axon, but here they are small and optically homogeneous.

In addition, phase-contrast reveals in the living neurone long, wavy, filamentous structures throughout the cytoplasm. These are very thin and are not always clearly seen. Sometimes tape-like structures which appear like in- vaginations from the cell membrane are also seen.

No pigment is seen in the living neurone.

The cytoplasmic objects that are coloured with neutral red, while the cell is still alive, correspond in size and distribution with the spherical bodies observed in the unstained neurone by phase-contrast. When such cells are examined by direct microscopy the bodies dyed with neutral red appear homogeneous (fig. 2, A). Sometimes the impression is given that the neutral red globules are more intensely coloured at the periphery than internally. If these preparations are studied by phase-contrast microscopy, most of the neutral red globules show a dark surface, such as is seen in the unstained cell. This shows that neutral red does not interfere with the optically visible structure of the bodies that are vitally dyed by it. Small globules coloured with neutral red are also seen in the axon.

FIG. 2.

(plate). A, neurone of Leander vitally dyed with neutral red. B, neurone of Leander; Altmann/Metzner; note chromophil crescents associated with chromophobe material, and also ‘trophospongial’ prolongation (tr) from the surface of the cell into the cytoplasm of the neurone. C, neurone of Leander; Aoyama (with postchroming) / acetylated Sudan black; note the binary structure of the sudanophil spherical bodies (life-like), and ‘trophospongium’ (tr). D, neurone of Astacus; Kolatchev. The scale on fig. c also applies to B and D.

FIG. 2.

(plate). A, neurone of Leander vitally dyed with neutral red. B, neurone of Leander; Altmann/Metzner; note chromophil crescents associated with chromophobe material, and also ‘trophospongial’ prolongation (tr) from the surface of the cell into the cytoplasm of the neurone. C, neurone of Leander; Aoyama (with postchroming) / acetylated Sudan black; note the binary structure of the sudanophil spherical bodies (life-like), and ‘trophospongium’ (tr). D, neurone of Astacus; Kolatchev. The scale on fig. c also applies to B and D.

Living neurones dyed in Janus green and studied by direct microscopy show minute granules, granules lying in small rows, and small rods (fig. 3). These are presumably mitochondria. Janus green does not colour the very long filamentous structures seen in the unstained cell by phase-contrast; they do not appear to be mitochondria.

FIG. 3.

Neurone of Leander vitally dyed with Janus green. The ‘trophospongium’ is not coloured.

FIG. 3.

Neurone of Leander vitally dyed with Janus green. The ‘trophospongium’ is not coloured.

Dahlia does not colour any of the cytoplasmic inclusions visible in the living cell by phase-contrast; the dye is uniformly dispersed in the ground cyto- plasm.

Thin sections of ganglia fixed in buffered osmium (Palade, 1952) and studied by phase-contrast (Baker, 1960) do not generally show the spherical bodies that are so obvious in the living cell. Instead, objects shaped like the letter U and V, and like bananas and crescents are seen. Even rings are also found, but they are very rare. Each of these bodies shows the double structure seen in the spherical bodies by phase-contrast, that is to say, a dark sheath associated with a much lighter area. The latter is often greatly distorted. These are the appearances that are commonly called ‘Golgi apparatus’ or ‘dictyosomes’ in the neurones of invertebrates (Thomas, 1947; Moussa, 1950; Beams and others, 1953; Shafiq, 1953, 1954; Young, 1953, 1956; Gressonand others, 1956; Malhotra, 1956, 1957; Chou, 1957 a, b; also see Lacy, 1957, and Baker, 1959 for references). Sometimes small, dark, curved rods or almost straight rods are also seen in such preparations. These do not show any material associated with them and appear homogeneous under phase-contrast. If, however, these sections are examined by interference microscopy, a thin layer of some material having a very low refractive index is discernible in association with the objects, which appeared homogeneous under phase-contrast. It is conceivable that the material in association with these bodies was originally bigger and was disrupted during the preparation of the tissue, leaving behind its traces which could not always be seen by phase-contrast.

Neurones fixed in buffered osmium and studied by phase-contrast often also show long, slightly wavy or straight objects, which look like narrow tapes, at or near the surface of the cell, where they are connected with the cell mem- brane. In the interior of the cell thin, delicate filaments are seen. The latter belong to the same category of objects as the long, fine filaments seen in the living cell by phase-contrast. These filaments are in fact the ramifications of the tapes but no connexions between the two objects have been observed in these preparations.

None of the silver impregnation methods has given satisfactory results. Sometimes, however, I have seen long, blackened filaments in the medium-sized neurones after Ramón y Cajal’s method (fig. 4, A). It is highly unlikely that these filaments are mitochondria. They do not conform to the description of the objects that are vitally dyed by Janus green (see p. 79), nor do they resemble the objects believed to be mitochondria in fixed preparations (see p. 83). These are more likely to be parts of the framework formed by the branching of the tape-like structures.

FIG. 4.

Neurones of Leander from. Ramón y Cajal preparations.

FIG. 4.

Neurones of Leander from. Ramón y Cajal preparations.

Another kind of cytoplasmic inclusion can sometimes be made out in favourable Ramón y Cajal preparations. This is constituted by the spherical bodies. Some of these are very badly distorted, while the others are in the form of crescents, as seen in buffered osmium neurones. The latter do not always show a chromophobe interior. The spherical bodies are also seen to almost retain the shape they had in life (fig. 4, B); but cells showing such bodies are extremely rare. Some of these argentophil bodies lie in contact with the blackened filaments described in the above paragraph; they give the impression of being enlarged localized regions on the filaments (fig. 4, A). There is no evidence that this is a genuine relationship. It seems to be the result of fixation.

In Mann-Kopsch preparations the objects which readily reduce osmium tetroxide correspond to those seen in the material fixed in buffered osmium and studied by phase-contrast. They are better demonstrated here than in the corresponding silver preparations. Osmiophil material is localized at the surface of the osmiophil sphere. But very often the osmiophil material is seen as large saucers or bananas with the osmiophobe material spread along its length. Such objects measure up to about 7 μ in length. When the osmiophil bodies are ring-shaped (in optical section) they measure up to 3 · 5 μ in diameter. These osmiophil bodies conform in distribution and binary structure with the spherical bodies seen in the unstained neurone, or one vitally dyed by neutral red; but the biggest of these do not exceed 2 · 5 μ. The Mann-Kopsch method presumably swells them. A similar swelling of similar bodies has been described by myself (1956) in the neurones of Schistocerca gregaria.

The osmiophil bodies are smaller in neurones that had been vitally dyed in neutral red and then subjected to the Mann-Kopsch method. The osmiophil material in these cells also is located at the surface of the spherical bodies, as in routine Mann-Kopsch preparations. This experiment shows that neutral red does not interfere with the osmiophil material of these objects. It will be re- membered that a similar sheath could be seen at the surface of the globules dyed with neutral red and studied by phase-contrast (see p. 79).

Spherical bodies that retain their life-like form are rarely met with after Mann-Kopsch treatment. Such objects are, however, common in the neurones that had been fixed in osmium tetroxide in sea-water and then post-osmicated (p. 76). This experiment indicates that the spherical bodies, which often burst open into crescents after fixation in mercuric chloride /osmium, remain stable when osmium tetroxide in sea-water is substituted as the fixative.

The small spherical bodies are blackened all through after post-osmication. Such globules are again more often seen in the neurones fixed in osmium tetroxide in sea-water and post-osmicated than by the Mann-Kopsch method. Fixation in osmium tetroxide in sea-water followed by post-osmication is best of all the ‘Golgi’ impregnation techniques used on the neurones of Leander. The objects that are responsible for the ‘Golgi’ appearances, that is to say, the neutral red-staining spherical bodies, are preserved in a more life-like form than by other ‘Golgi’ methods. The darkening of the general cytoplasm is also much less than by the Mann-Kopsch method.

If pieces of ganglia that had been in a 2% solution of osmium tetroxide for about 12 h (according to Kopsch’s method) are teased in glycerol and studied by phase-contrast, the spherical bodies of the living cell are represented by bright spheres having a thin dark sheath at the surface. Some of these are distorted and look like bright curved or straight objects with a dark edge. Some of the tissue was left in osmium tetroxide for about a fortnight and studied after sectioning. Such neurones reveal objects similar to those seen in Mann- Kopsch preparations, but the spherical or almost spherical bodies seen after short osmication are often not seen. These seem to have opened up into rods and crescents with or without the remnants of the chromophobe sphere. Neurones prepared in this way are very much shrunk and the osmiophil material is only slightly darkened.

The tape-like structures are also sometimes seen in osmium tetroxide preparations, but they are much less darkened than the spherical bodies. The ultimate branching of these tape-like objects is not clear in these preparations.

The objects described in osmium preparations are very clearly demon- strated in Rawitz preparations. In properly differentiated cells the chromo- phil component is intensely coloured while the chromophobe sphere is light pink. Sometimes there is little differentiation into these two parts, presumably owing to insufficient differentiation. These preparations also clearly show narrow tape-like structures extending from the surface of the cell into the cytoplasm. These are also intensely coloured by basic fuchsine (after mordanting in tannic acid and potassium antimony tartrate). In this they resemble the surface of the spherical bodies, which are now mostly in the form of crescents and rods. The elongate structures appear to be hollow, though they are tape-like near their attachment to the surface of the cell.

Both the kinds of objects described in the above paragraph are coloured alike by acid fuchsine (Metzner) or alcoholic haematoxylin (Hirschler) or iron-haematoxylin after fixation in Helly (with postchroming) or Altmann (fig. 2, B). When the elongate structures are in contact with the cell surface there is no difficulty in recognizing their true identity. But if the section does not happen to pass through this region they can easily be mistaken for the long filamentous mitochondria. I have been able to study these structures very clearly in neurones fixed in Helly (with post-chroming) and stained in iron haematoxylin (fig. 5, A, B). In suitable sections the branching of these elongate tape-like structures and their ramification can easily be followed. The branching is scarce; and the ultimate branches appear to be thin filaments at the limit of resolution of the ordinary microscope (fig. 5, A), but they may in fact be canalicular like those at the surface of the cell. Ross (1915, 1922) has also described similar delicate threads as parts of the ‘trophospongium’ forming a framework in the cytoplasm of the neurones of Cambarus. A few long, delicate filaments corresponding to these fine branches have also been seen in favourable preparations of neurones fixed in Lewitsky-saline and stained in iron haematoxylin (fig. 6).

FIG. 5.

Neurones of Leander from Helly (with postchroming) /iron haematoxylin preparations, showing the branching of ‘trophospongial’ structures in the cytoplasm of the neurone in A.

FIG. 5.

Neurones of Leander from Helly (with postchroming) /iron haematoxylin preparations, showing the branching of ‘trophospongial’ structures in the cytoplasm of the neurone in A.

FIG. 6.

Neurone of Leander from Lewitsky-saline /iron haematoxylin preparations.

FIG. 6.

Neurone of Leander from Lewitsky-saline /iron haematoxylin preparations.

Lewitsky-saline/haematoxylin preparations also sometimes show spherical bodies with a thin chromophil ring or crescent covering the chromophobe interior (fig. 6). These are sometimes distorted, and then look like bright spaces. The chromophil material in such cases lies along their edges. There is no doubt that these are the same objects as are seen in ‘Golgi’ preparations, but they are not swollen now.

Both the kinds of objects, namely, the structures corresponding to the spherical bodies of the living cell and tape-like extensions from the cell surface into the cytoplasm, are coloured alike by xylidine red (Masson) after fixation in Helly followed by postchroming. Mallory’s phosphomolybdic acid /haem- atoxylin stains the tape-like extensions, but their further branching is not clearly revealed.

The spherical objects do not resist paraffin embedding after fixation in Carnoy.

The mitochondria of these cells are very difficult to study in fixed preparations. The long filamentous branches of the tape-like structures are very easy to confuse with long filamentous mitochondria, but such mitochondria seem not to exist in the neurone of Leander. In suitable Metzner preparations minute granules, granules lying in short rows, and small rods are seen. These appear to be the mitochondria: they resemble the objects that are vitally dyed by Janus green (fig. 3). In collaboration with Dr. Meek, I have confirmed by electron microscopy that the mitochondria of these cells are mostly small granules and short rods (1959).

In neurones fixed in Zenker or Helly and stained in pyronine /methyl green (Jordan and Baker, 1955) or cresyl violet (Fernstrom, 1958) or basic fuchsine, there is an intense general staining in the cytoplasm; but there are no obvious Nissl bodies of the kind so commonly seen in the neurones of vertebrates (see Malhotra, 1959, 1960 for references). If the sections are treated with ribonuclease (Bradbury, 1956) or trichloroacetic acid (Pearse, 1954) before staining, all the basiphilia is extracted. The Nissl substance is to be regarded as uniformly dispersed.

Neurone of Astacus

The cytoplasm of the large neurones is often fixed as very coarsely granular, whereas the small and medium-sized neurones adjoining the large cells are better, but not ideally, fixed in the same section.

The cytoplasmic inclusions of the neurone of Astacus are similar in general to those of Leander. Only the points of difference are discussed here.

The spherical bodies are generally smaller than in Leander, varying from less than 1μto 1·5 μ; but in a few cells they are bigger and occasionally they are equal in size to the biggest spherical bodies in Leander. The osmiophil bodies vary in size in an exactly corresponding way (fig. 2, D). In contrast with the saucers and bananas seen in Leander after osmium impregnation techniques, the osmiophil material is usually in the form of bowls or shallow cups. The chromophobe material is more clearly seen than in Leander.

In a few small cells in Ramón y Cajal preparations long, wavy threads are seen. These are disposed in a net-like fashion (fig. 7), and resemble the loose net of Monti (1914, 1915). Similar structures are also seen in Metzner preparations, and might have been interpreted as mitochondria, if I had not studied the neurones of Leander. These filamentous objects resemble the branches of the tape-like structures, which spring from the surface of the cell. But such a connexion with the cell membrane has never been observed in Astacus, probably because the peripheral part of the cytoplasm is nearly always damaged by being pulled away from the surrounding tissue, and in the process the cell membrane is torn. The filamentous objects seem to branch much more freely than in Leander. I have seen similar filaments disposed in a net-like manner in the small neurones fixed in buffered osmium and studied by phase-contrast.

FIG. 7.

Small neurones of Astacus from Ramón y Cajal preparations, showing the Golgi net of Monti (1915). The nuclear membrane is shown in dotted line.

FIG. 7.

Small neurones of Astacus from Ramón y Cajal preparations, showing the Golgi net of Monti (1915). The nuclear membrane is shown in dotted line.

It is very difficult to say anything definite about the mitochondria of this cell. Vital dyeing with Janus green has not been successful and therefore there is no control. Nevertheless, the objects which are seen in Metzner or Hirschler preparations in the form of minute granules and short threads may be mito- chondria (fig. 8). It is difficult to be sure, because some of these threads may be fine branches of the tape-like structures which are connected with the surface of the cell.

FIG. 8.

Neurone of Astacus from Altmann/Hirschler preparations. Note small rod-like objects which are probably mitochondria.

FIG. 8.

Neurone of Astacus from Altmann/Hirschler preparations. Note small rod-like objects which are probably mitochondria.

No birefringent material was observed in the cytoplasmic inclusions of the neurones of Astacus.

Histochemistry

If the ganglia of Leander serratus are subjected to the standard Sudan black technique of Baker (1949) after fixation in formaldehyde/calcium with post- chroming (Baker, 1946), the cytoplasmic structures that are coloured corre- spond in size and distribution to the spherical bodies seen in the living cell. The colouring is feeble and is confined to a thin layer at the surface of these bodies. Colouring with Sudan black is very feeble after fixation in Lewitsky- saline. The acid haematein (AH) test is negative for these bodies after fixation in formaldehyde /calcium.

If Aoyama’s formaldehyde-cadmium mixture is used as fixative instead of formaldehyde /calcium in the acid haematein test, the results are strikingly different. In sections of the tissue fixed in this way, postchromed as for the AH test, and coloured with Sudan black, the colour is considerably intensified at the surface of the binary spherical bodies. Acetylated Sudan black (Cassel- man, 1954, 1959), which is a more specific test for lipids, brings out the spherical bodies clearly. They appear sharply delimited by a thin sudanophil layer, which is usually entire (fig. 2, c), but sometimes only partial. Sometimes these objects are distorted into irregular structures. Their surface now gives also a true AH-positive reaction (positive after AH; negative after pyridine extraction). This shows the phospholipid nature of the surface of these bodies. This is in accordance with the refringent appearance seen by phase-contrast. Though Aoyama’s silver impregnation method has not given satisfactory results, nevertheless if the Sudan black technique is used on material fixed in Aoyama’s fluid (with postchroming), the spherical bodies are generally pre- served in a life-like or almost life-like condition.

The above results suggest that the phospholipid which is localized at the surface of the spherical bodies is ‘masked’; that is, it probably exists as lipo- protein complex. It does not respond to routine tests for phospholipid unless it is first liberated by the use of some ‘unmasking’ agent. Though tests for cytoplasmic proteins (Sakaguchi for arginine and Hg/nitrite for tyrosine) are negative for the spherical bodies, it is difficult to exclude the possibility of their presence in these bodies.

The spherical bodies also react positively to the PAS test. After digestion in saliva or distase the PAS reaction is still positive, but weaker. This suggests the presence of a carbohydrate. If a ganglion is fixed in cold acetone and sections are treated by the PAS test, the spherical bodies react positively; the reaction is weaker if hot acetone in a Soxhlet apparatus is used as fixative. Colouring with acetylated Sudan black is also feebler after extraction with hot acetone than with cold acetone. The facts suggest the presence of cerebroside (Casselman and Baker, 1955), and would account in part for the positive PAS reaction.

When Cain’s Nile blue test is applied to the neurones the surface of the spherical bodies is blue. This indicates the acidic nature of this part.

The spherical bodies reacted negatively to all the other histochemical tests listed in the appendix.

There are certain resemblances between the trophospongial framework and the spherical bodies in their responses to histochemical tests, but there are also the following important differences.

(1) Colouring of the trophospongial framework after Sudan black is very feeble even when an unmasking fixative is used. Whereas this framework is positive to the standard acid haematein test, the spherical bodies are negative.

(2) The trophospongial framework does not show any appreciable PAS- positive material.

(3) The spherical bodies do not resist paraffin embedding after fixation in Carnoy.

(4) The trophospongial framework does not colour in life with neutral red.

Beyond the inclusions listed earlier (p. 77), no other distinct structure can be demonstrated in the cytoplasm of the neurones of Leander or Astacus by any of the techniques used in this study.

The spherical bodies seen in the unstained living cell are coloured by neutral red. After colouring, the binary appearance of the spheres cannot be easily seen by direct microscopy. It is not clear whether neutral red colours the surface of these spheres or their interior or both. The work of Young (1953, 1956) on the neurones of cephalopods shows that the surface of the lipid droplets is more deeply coloured by neutral red. In the neurones of Lumbricus it is the lipid material at the surface of the ‘binary spheroids’ that takes up the dye (Thomas, 1954). Hirsch (1939) has also recorded that in the intestine of Ascaris neutral red colours the ‘externum’ (which is also blackened by osmium tetroxide), while in the pancreas of the white mouse (Hirsch, 1959) neutral red sometimes colours the ‘externum’ and sometimes the ‘inter- num’. It is, therefore, likely that in those lipid globules that exhibit a clear differentiation into a lipid sheath and a non-lipid interior, neutral red is capable of colouring both the components. It is, however, uncertain what determines the uptake of this basic dye, which is avoided by almost everything except lipid droplets in the cytoplasm of most animal cells. It is probable that the phosphoric group of the phospholipids attracts the dye at the surface. The colouring of the internum is not explained. It may perhaps be due to the presence of acidic mucopolysaccharide, which is thought often to occur in the interior of such bodies, or to the presence of ribonucleoprotein (compare Dustin, 1947).

Experiments with neutral red on the neurones of Leander show that colouring by this dye does not interfere with the optically visible structure of the bodies it colours.

The objects commonly called ‘Golgi apparatus’ or ‘dictyosomes’ in the neurones of invertebrates (see p. 80 for references) are commonly represented by spherical bodies in life. During the process of fixation the spherical bodies often break open into rods and crescents (Baker, 1958; Chou and Meek, 1958). These represent the lipid surfaces of the spheres, and are the basis of reduction of osmium tetroxide or silver nitrate. This is in conformity with the findings of Young (1953, 1956), Shafiq (1953), Chou (1957 a, b), Chou and Meek (1958), and Baker (1959), who regard the ‘Golgi apparatus’ of the neurones of invertebrates as represented by neutral red globules in the living cell. I also arrived at this conclusion in studies of the neurones of insects and molluscs (1956, 1957). Chou and Meek (1958) have found that the lipid globules, which are often seen as crescents and rods in the electron micrographs of the neurones of Helix aspersa after fixation in buffered osmium, appear as circles (as in life) if calcium chloride is added to the fixative (Baker, 1958). Similar results have been obtained with the neurones of Leander when osmium tetroxide in sea-water is used as a fixative instead of Mann’s fluid (in Mann-Kopsch). This may perhaps be due to the presence of calcium in sea-water.

In addition to the spherical bodies, there is another structure in the neurones of Crustacea which has been labelled in the past as the ‘Golgi apparatus’. This is the ‘trophospongium’ of Holmgren (Ross, 1915, 1922; Welsh and Schallek, 1946). It originates as narrow hollow structures at the surface of the cell, which branch into fine threads, ramifying throughout the cytoplasm (Ross, 1922). Recently Wigglesworth (1959) has described very similar structures in the neurones of the hemipteran, Rhodnius prolixus. He regards them as branching invaginations of the surface membrane, into which the glial cytoplasm extends. Hess (1958) also observed similar invaginations in electron micrographs of the neurones of the cockroach, Periplanet a americana.

Monti (1915) and Migliavacca (1929) mistook the fine thread-like branches of the ‘trophospongium’ for mitochondria, because they are also stained by mitochondrial techniques. They thought that the Golgi network was formed by the deposition of silver on these threads. Monti and Migliavacca did not realize that these threads were parts of the ‘trophospongium’.

Moreover, Monti, being a pupil of Golgi, did not believe in the existence of ‘Golgi apparatus’ in the form of separate elements. She considered that the short, straight or slightly sickle-shaped bodies seen by Poluszynski (1911) were artifacts of prolonged reaction. Poluszynski was in fact looking at the lipid sheaths of the spherical bodies, which had burst open into rods and crescents. He also observed a life-like or almost life-like appearance of the spherical bodies in the form of ‘rings’ and also of ‘granules’ which had blackened all through, but he regarded them as swollen threads, which, he believed, were the real form of the Golgi apparatus. Migliavacca considered that the ‘rings’ were a particular developmental stage of the mitochondria. If they had controlled their results with the study of the living cell and had used vital dyeing, they would not have arrived at these conclusions.

Ross (1922) felt certain that the filamentous Golgi network seen by Monti (1915) in the small and medium-sized neurones of Crustacea was quite inde- pendent of the mitochondria. Although he observed and described in detail the delicate filaments of the ‘trophospongium’, he thought that the Golgi net- work was a separate body, and did not realize that what he saw in Golgi pre- parations was in fact the branching threads of the ‘trophospongium’. Ross, however, thought that the objects described by Poluszynski (1911) were probably components of the trophospongial framework. This does not seem to be true.

With regard to the large cells, Ross stated that no one previously had succeeded in showing a typical reticular Golgi apparatus; and he also could not do so, as these cells are very difficult to fix well and do not often show the ramification of the trophospongium.

Parat (1928) claimed that the Golgi appearances were produced by the deposition of osmium or silver on the periphery of the ‘neutral red vacuoles’. This is certainly true, and there is no doubt that he was dealing with spherical bodies. Since the modern techniques of histochemistry were unknown to Parat, he could not have realized that his ‘neutral red vacuoles’ had a sheath of lipid material. His conclusion that the neutral red vacuoles correspond to the trophospongial system of Holmgren was purely hypothetical and unfounded on fact.

Dornesco (1934) and Lacroix (1935) claimed the existence of neutral red ‘vacuoles’ independent of the osmiophil and argentophil bodies (‘dictyosomes’). My results do not support this. Moreover, Dornesco reported that the neutral red vacuoles in the neurones of Potambius run together with the passage of time. I have never observed this in the neurones of Leander while the cell is alive.

I have not encountered in the neurones of Leander or Astacus the small, compact and localized net described by Monti (1914, 1915) and Migliavacca (1929) in the neurones of Astacus and Homarus.

My results do not confirm the findings of Lacroix (1935) on the existence of distinct Nissl bodies in the neurones of Astacus. My observations show that the Nissl substance is uniformly dispersed throughout the cytoplasm in the neurones of Crustacea. This is in agreement with the observations of Ross (1922). Recently Moussa and Banhawy (1959) have also recorded that the Nissl bodies in the various developmental stages of Schistocerca gregaria are in the form of granules which are evenly distributed throughout the cytoplasm of the neurones of the thoracic ganglia. The evidence suggests that while the Nissl substance of the neurones of vertebrates is aggregated into lumps (Palay and Palade, 1955; Hess, 1955; Palay, 1956; Palade, 1958; Barton and Causey, 1958) that of invertebrates is more uniformly dispersed through the cytoplasm (De Robertis and Bennett, 1954; Hess, 1958). Schultz and his colleagues (1956) have made comparable observations on the sensory neurones of the spinal cord of the lamprey. They saw no Nissl bodies and in conformity with this they found no endoplasmic reticulum in electron micrographs. Chou and Meek (1958) working with the neurones of Helix aspersa did not encounter in their electron micrographs any aggregations of the endoplasmic reticulum suggestive of lumps of Nissl bodies.

Four kinds of cytoplasmic inclusions can be recognized in the neurones of Leander serratus and Astacusfluviatilis. These are (i) spherical or almost spherical bodies, which often show a differentiated cortex and medulla; (ii) mitochondria, in the form of minute granules and short rods; (iii) Nissl substance, uniformly dispersed; (iv) ‘trophospongial’ structures, which are connected with the surface of the cell, and ramify in the form of delicate filaments throughout the cytoplasm.

Neutral red colours the spherical bodies in life; it does not seem to interfere with their optically visible structure. The spherical bodies often burst open into rods and crescents; these correspond to what other authors have called ‘Golgi apparatus’ or ‘dictyosomes’. The term ‘Golgi apparatus’ has also been applied by certain authors to the ‘trophospongial’ structures.

Histochemical study reveals that the surfaces of the spherical bodies, which are blackened by osmium tetroxide or silver nitrate in the Golgi methods, respond to tests for phospholipid after an ‘unmasking’ fixative has been used. The evidence also suggests the presence of cerebroside (galactolipid) in these bodies.

I am grateful to Dr. J. R. Baker, F.R.S., for suggesting and supervising this investigation. I wish to thank Professor Sir A. C. Hardy, F.R.S., in whose department the work was done; Dr. Blanche-P. Clayton for translating some French and German papers; and Mr. Q. Bone for translating papers by Monti published in Italian.

This work was done during the tenure of a Postdoctoral Research Fellow- ship of the Panjab University, India, and a travel grant in the ‘Commonwealth University Interchange Scheme’ from the British Council.

Aoyama
,
F.
,
1929
.
Zeit. wiss. Mikr
.,
46
,
489
.
Baker
,
J. R.
,
1946
.
Quart. J. micr. Sci
.,
87
,
441
.
Baker
,
J. R.
,
1947
.
Ibid
.,
88
,
115
.
Baker
,
J. R.
,
1949
.
Ibid
.,
90
,
293
.
Baker
,
J. R.
,
1956a
.
Ibid
.,
97
,
621
.
Baker
,
J. R.
,
1956b
.
Ibid
.,
97
,
161
.
Baker
,
J. R.
,
1958
.
J. Histochem. Cytochem
.,
6
,
303
.
Baker
,
J. R.
,
1959-
J- Roy. micr. Sci
.,
77
,
116
.
Baker
,
J. R.
,
1960
.
Cytological technique, 4th ed. London (Methuen) (in the press)
.
Baker
,
J. R.
, and
Luke
,
B. M.
,
1959
.
Unpublished observations
.
Barton
,
A. A.
, and
Causey
,
G.
,
1958
.
J. Anat
.,
92
,
399
.
Beams
,
H. W.
,
Sedar
,
A. W.
, and
Evans
,
T. C.
,
1953
.
Cellule
,
55
,
293
.
Bignardi
,
C.
,
1940
.
Boll. Soc. ital. Biol, sperim
.,
15
,
594
. Quoted from Chou, 19576.
Bradbury
,
S.
,
1956
.
Quart. J. micr. Sci
.,
97
,
323
.
Cain
,
A. J.
,
1947
.
Ibid
.,
88
,
383
.
Cain
,
A. J.
,
1949
.
Ibid
.,
90
,
75
.
Carnoy
,
J. B.
,
1886
.
Cellule
,
3
,
1
.
Casselman
,
W. G. B.
,
1954
.
Quart. J. micr. Sci
.,
95
,
321
.
Casselman
,
W. G. B.
1959
.
Histochemical technique. London (Methuen)
.
Casselman
,
W. G. B.
and
Baker
,
J. R.
,
1955
.
Quart. J. micr. Sci
.,
96
,
49
.
Champy
,
C.
,
1911
.
Arch. d’Anat. micr
.,
13
,
55
.
Chou
,
J. T.-Y.
,
1957a
.
Quart. J. micr. Sci
.,
98
,
47
.
Chou
,
J. T.-Y.
1957b
.
Ibid
.,
98
,
59
.
Chou
,
J. T.-Y.
and
Meek
,
G. A.
,
1958
.
Ibid
.,
99
,
279
.
De Robertis
,
P.
, and
Bennett
,
H. S.
,
1954
.
Fine structure of cells. Groningen (Noordhoff)
.
Dornesco
,
G. Th
.,
1934
.
Bull. Hist, appl
.,
11
,
190
.
Dustin
,
P.
,
1947
.
Symp. Soc. exp. Biol
.,
1
,
114
.
Fernstrom
,
R. C.
,
1958
.
Stain Tech
.,
33
,
175
.
Golgi
,
C.
,
1898
.
Arch. ital. Biol
.,
30
,
60
.
Golgi
,
C.
,
1900
.
Anat. Anz. Verb. Anat. Ges
.,
14
,
164
.
Golgi
,
C.
,
1908
.
Arch. ital. Biol
.,
49
,
269
.
Gomori
,
G.
,
1952
.
Microscopic histochemistry. Chicago (University Press)
.
Gresson
,
R. A. R.
,
Threadgold
,
L. T.
, and
Stinson
,
N. E.
,
1956
.
Cellule
,
58
,
7
.
Helly
,
K.
,
1903
.
Zeit. wiss. Mikr
.,
20
,
413
.
Herxheimer
,
G. W.
,
1901
.
Dtsch. med. Wschr
.,
36
,
607
.
aHess
,
A.
1958
.
J. biophys. biochem. Cytol
.,
4
,
731
.
Hirsch
,
G. C.
,
1939
.
Protoplasma monographien, 18. Berlin (Borntraeger)
.
Hirsch
,
G. C.
1959
.
Handbuch der Biologie, I, Heft 9/10
.
Hirschler
,
R.
,
1927
.
Zeit. wiss. Mikr
.,
44
,
216
.
Jordan
,
B. M.
, and
Baker
,
J. R.
,
1955
.
Quart. J. micr. Sci
.,
96
,
177
.
Kolatchev
,
A.
,
1916
.
Arch, russes d’Anat. d’Hist. d’Embriol
.,
1
,
383
.
Kopsch
,
F.
,
1902
.
Sitzungsber. Kgl. Preuss. Akad. wiss. zu Berlin
,
39
,
929
.
Lacroix
,
P.
,
1935
.
Cellule
,
44
,
251
.
Lacy
,
D.
,
1957
.
J. biophys. biochem. Cytol
.,
3
,
779
.
Lewitsky
,
G. A.
,
1931
.
Bull. appl. Bot. Genet. Plant Breeding
,
27
,
175
.
Lison
,
L.
,
1953
.
Histochimie et cytochimie animales. Paris (Gauthier-Villars)
.
Malhotra
,
S. K.
,
1956
.
Quart. J. micr. Sci
.,
97
,
177
.
Malhotra
,
S. K.
,
1957
.
Ibid
.,
98
,
65
.
Malhotra
,
S. K.
,
1959
.
Ibid
.,
100
,
339
.
Malhotra
,
S. K.
,
1960
.
Ibid
.,
101
,
69
.
Malhotra
,
S. K.
, and
Meek
,
G. A.
,
1959
.
Unpublished observations
.
Mallory
,
F. B.
,
1938
.
Pathological technique. Philadelphia (Saunders)
.
Mc Manus
,
J.
F. A
.,
1948
.
Stain Tech
.,
23
,
99
.
Metzner
,
R.
, and
Krause
,
R.
,
1928
.
In Abderhalden’s Handbuch der biologischen Arbeits- methoden. Abt. V, Teil
,
2
,
1
Halfte, 325. Quoted from Chou, 1957a.
Migliavacca
,
A.
,
1929
.
Arch. ital. Anat. Embriol
.,
27
,
351
.
Monti
,
R.
,
1914
.
Atti della Reale Acad. Naz. Lincei Rendiconti
,
23
,
172
.
Monti
,
R.
1915
.
Arch. ital. Anat. Embriol
.,
14
,
1
.
Moussa
,
T. A.
,
1950
.
J. Morph
.,
87
,
27
.
Moussa
,
T. A.
, and
Banhawy
,
M.
,
1959
.
Nature
,
184
,
915
.
Palade
,
G. E.
,
1952
.
J. exp. Med
.,
95
,
285
.
Palade
,
G. E.
,
1958
.
Frontiers in Cytology. New Haven (Yale University Press)
.
Palay
,
S. L.
,
1956
.
Neurochemistry. London (Cassell)
.
Palay
,
S. L.
and
Palade
,
G. E.
,
J. biophys. biochem. Cytol
.,
1
,
69
.
Pantin
,
C. F. A.
,
1948
.
Notes on microscopical technique for zoologists. Cambridge (University Press)
.
Parat
,
M.
,
1928
.
Arch. Anat. micr
.,
24
,
73
.
Pearse
,
A. G. E.
,
1954
.
Histochemistry, theoretical and applied. London (Churchill)
.
Poluszynski
,
G.
,
1911
.
Bull, intern. Acad. Sc., Cracovie, B, no volume number, 104
.
PrzelĘcka
,
A.
,
1959
.
Quart. J. micr. Sci
.,
100
,
231
.
RamÓn y Cajal
,
S.
,
1914
.
Trab. Lab. Invest. Biol. Univ., Madrid
,
12
,
127
.
Ross
,
L. S.
,
1915
.
J. comp. Neur
.,
25
,
523
.
Ross
,
L. S.
1922
.
Ibid
.,
34
,
37
.
Schultz
,
R. L.
,
Berkowitz
,
E. C.
, and
Pease
,
D. C.
,
1956
.
J. Morph
.,
98
,
251
.
Shafiq
,
S. A.
,
1953
.
Quart. J. micr. Sci
.,
94
,
319
.
Shafiq
,
S. A.
1954
.
Ibid
.,
95
,
305
.
Thomas
,
O. L.
,
1947
.
Ibid
.,
88
,
445
.
Thomas
,
O. L.
1954
.
Cellule
,
56
,
229
.
Weigl
,
R.
,
1910
.
Bull. Internat. Acad. Sci. Cracovie, Ser. B, 691
.
Welsh
,
J. H.
, and
Schallek
,
W.
,
1946
.
Physiol. Revs
.,
26
,
447
.
Wigglesworth
,
V. B.
,
1959
.
Quart. J. micr. Sci
.,
100
,
299
.
Young
,
J. Z.
,
1953
.
Ibid
.,
94
,
399
.
Young
,
J. Z.
1956
.
Endeavour
,
15
,
5
.
Zenker
,
K.
,
1894
.
Miinch. med. Woch
.,
41
,
532
.