The purpose of this investigation was to repeat as exactly as possible the original work on the ‘apparatus’ of Golgi in the Purkinje cells of the cerebellum of owls, and to re-investigate these cells by modem methods. The tawny owl, Strix aluco, was used instead of the closely-related ‘Strix flammea’ of Golgi. Golgi’s ‘osmio-bichromique’ technique of 1898 for silver impregnation was used successfully.

A reticulum corresponding to the basiphil Nets of Nissl can be seen in the living neurones by interference microscopy. The classical Golgi apparatus of the perikaryon is a deposit of silver or of osmium on this reticulum. The reticulum extends into the axons in the form of thin non-basiphil filaments, which are also blackened by Golgi methods.

These findings are in conformity with recent studies of the neurones of other vertebrates.

In a recent publication (1959) I showed that a three-dimensional reticulum could be seen by interference microscopy in the living neurones of vertebrates. This reticulum was homologized with the basiphil reticulum (Nets’) described by Nissl (1894 a) in these cells. Evidence was presented to show that the classical ‘Golgi apparatus’ in the perikaryon of the neurone of vertebrates was a deposit of silver or of osmium on this basiphil reticulum. In the axon, however, the ‘Golgi apparatus’ consists of non-basiphil filaments, which are extensions of the basiphil reticulum.

It was my purpose to extend this study to the site in which Golgi first described his ‘apparatus’, namely, the Purkinje cells of the cerebellum of owls. Golgi (1898) called his owls Strix flammea, by which he probably meant the barn-owl, Tyto alba. I have not been able to obtain this species, and have used instead the closely related tawny owl, Strix aluco (L.).

It would not appear that anyone else has examined the Purkinje cells of owls in the 60 years that have elapsed between Golgi’s investigations and my own.

Two live tawny owls were obtained from Mr. Robert Jackson, Holly Bank Nurseries, Grove Lane, Hale, Cheshire. These were killed by chloroform and the cerebellum was at once dissected out.

Freshly teased Purkinje cells mounted in saline with calcium chloride (Baker, 1944) were studied by positive phase-contrast and with the Baker interference microscope (× 100 double-focus objective).

Sections of material fixed in formaldehyde-calcium and postchromed (Baker, 1946) were coloured by Sudan black (Baker, 1949) and acid haematein (Baker, 1946) for the study of lipid inclusions.

Silver and osmium preparations were made by the methods of Aoyama (1929) and Mann-Kopsch (Weigl, 1910) respectively.

The greater part of the available material was used up in making preparations by Golgi’s original method, the ‘osmio-bichromique’ technique of 1898, which would not appear to have been tried by anyone during the last half century. Small pieces of the cerebellum were fixed in a mixture of 2 volumes of 3 % aqueous solution of potassium dichromate and 1 volume of 1 % osmium tetroxide. Pieces were taken out of this fluid at intervals of 1, 2, and 3 months, directly transferred to a 3% aqueous solution of copper acetate (for ‘rejuvenation’), and left for 2 or 3 days. The pieces were subsequently passed into a 3 % aqueous solution of potassium dichromate. In accordance with the suggestion of Golgi, repeated trials were made with different periods, from 3 days to a fortnight, in this fluid. Details of the after-treatment of these pieces were taken from an earlier paper by Golgi (1886). Golgi presumably cut sections by hand. Since this could not be done with the small pieces I used, sections were mostly cut by embedding in collodion (low viscosity nitrocellulose).

When the pieces are transferred from potassium dichromate to silver nitrate, a precipitate of silver chromate is formed (Golgi, 1886). A few preliminary washings were, therefore, given in a 0·5% solution of silver nitrate till no more precipitate appeared, and the tissue was then left in a 0·75% solution of silver nitrate for 24 to 48 h in daylight. (According to Golgi (1886), it makes no difference whether the tissue is left in the light or not.)

The best Golgi pictures were obtained in tissue fixed for 2 months, rejuvenated for 2 days, and treated with potassium dichromate for about a week. Cellular outlines were not very conspicuous in these preparations. They were better seen in sections of tissue fixed for 3 months, rejuvenated for 3 days, and treated with potassium dichromate for 4 to 6 days.

After silvering, the material was washed in several changes of 70% or 95% alcohol, and passed through absolute alcohol into a mixture of equal volumes of absolute alcohol and ether, and thence to collodion. The sections were cut at 15 or 20 μ. Frozen and paraffin sections were also made, but the results were not satisfactory.

Collodion sections were dehydrated in 95% alcohol and creosote and cleared in turpentine. They were mounted in dammar (in xylene), with or without coverglass. If the collodion is removed from the sections by soaking in absolute alcohol, they break up; it was, therefore, not removed.

Nissl bodies were shown by bleaching Mann-Kopsch or Aoyama preparations (see Malhotra, 1959) and staining in pyronine /methyl green (Jordan and Baker, 1955) or in cresyl violet (Fernstrom, 1958). Alternatively, pieces of cerebellum were fixed in Mann’s fluid with the addition of acetic acid (Baker, 1957), and stained in the same ways.

Sections of material fixed in Palade’s buffered osmium (Palade, 1952) and embedded in "-butyl methacrylate were cut at about 3 p., mounted in n-butyl methacrylate, and studied unstained by phase-contrast.

In an attempt to show the mitochondria of the Purkinje cell, pieces of cerebellum were fixed in Helly (1903) and postchromed; sections were stained by Metzner’s and Hirschler’s methods.

In living Purkinje cells studied by interference microscopy, a prominent reticulum can be clearly seen in the perikaryon. This reticulum, as described already in the neurones of other vertebrates (Malhotra, 1959), consists of massive bodies of varying sizes and shapes scattered throughout the cytoplasm, except at its extreme periphery. They are bigger and more concentrated round the nucleus. They do not seem to be sharply delimited from the ground cytoplasm. These massive bodies are connected with one another by thin strands, which show the same interference colour as the massive bodies. The strands may be straight or loop-shaped. There are also conspicuous, optically homogeneous spaces of very low refractive index in association with both the components of the reticulum. These ‘canalicular spaces’ are either elongate or irregularly rounded. They are bigger and more easily detectable than in the neurones of the mouse (Malhotra, 1959).

Apart from this reticulum, numerous small refringent globules (presumably lipid globules) are seen by interference microscopy, dispersed at random throughout the cytoplasm of the living Purkinje cells.

These refringent globules are better seen in living cells by phase-contrast microscopy. The canalicular spaces are also sharply outlined by phasecontrast. There are some indications of the presence of the strands of the reticulum, but the massive bodies cannot be seen in living neurones by phase-contrast.

All the structures described in the above paragraphs, except the massive bodies, can easily be seen in unstained sections of tissue that has been fixed in Palade’s buffered osmium and examined by phase-contrast. The massive bodies cannot be very clearly demonstrated in these preparations, probably because they are not sharply delimited. Small dark globules, which look like the lipid droplets, often appear to lie in short rows on the strands of the reticulum, but they are too small for reliable observation.

In Nissl preparations of the Purkinje cells, the basiphil component is found to exist in the form of large, irregular bodies of varying sizes and shapes (commonly called Nissl bodies), distributed in the cytoplasm except at its extreme periphery and in the axon hillock. They are bigger and sometimes more elongated round the nucleus. There are thread-like basiphil connections (Fäden of Nissl, 1894b) joining these Nissl bodies with one another to form a reticulum (Netz of Nissl, 1894 a). There are also canalicular spaces in association with the basiphil threads and with the Nissl bodies (fig. 1, B). This pattern of the basiphil reticulum agrees essentially with the reticulum seen in the living neurone by interference microscopy and also with the structures seen in unstained sections studied by phase-contrast. It is clear that the reticulum that is discernible in life by interference microscopy corresponds to the basiphil reticulum of routine Nissl preparations.

FIG. 1.

A, Purkinje cell from ‘Strix flammea’, reproduced from Golgi (1898). This is the earliest figure of Golgi’s ‘apparatus’. ‘Osmio-bichromique’ method. It is not certain whether the process in this figure is an axon or a dendrite. B, Purkinje cell from Strix aluco, showing basiphil reticulum in the perikaryon. Aoyama (desilvered), pyronine /methyl green. The nuclear membrane is covered by Nissl bodies, and is not clearly seen.

FIG. 1.

A, Purkinje cell from ‘Strix flammea’, reproduced from Golgi (1898). This is the earliest figure of Golgi’s ‘apparatus’. ‘Osmio-bichromique’ method. It is not certain whether the process in this figure is an axon or a dendrite. B, Purkinje cell from Strix aluco, showing basiphil reticulum in the perikaryon. Aoyama (desilvered), pyronine /methyl green. The nuclear membrane is covered by Nissl bodies, and is not clearly seen.

In Golgi preparations made by the method of Aoyama, the Golgi apparatus has a characteristic disposition, resembling that of the basiphil reticulum. The canalicular spaces are generally not seen in Aoyama preparations, presumably because they have been covered by a deposit of silver. If, however, these sections are bleached and studied after staining in basic dyes (Malhotra, 1959), the existence of these spaces is easily established. These spaces, covered with silver, account for some of the massive thickenings in Golgi preparations.

In Mann-Kopsch preparations (4 days’ osmication at 34° C) the Golgi pictures are mostly filamentous; the canalicular spaces are easily seen, associated with osmiophil threads. Large blackened lumps resembling Nissl bodies also occur in a few cells.

Both in silver and in osmium preparations the Golgi apparatus is sometimes seen to extend into the axon in the form of extremely fine filaments. There are often what appear to be lipid globules lying in rows on these filaments. More often than not the axon hillock appears to be empty in Golgi preparations. This led many cytologists of the past to believe that the Golgi apparatus does not extend into the axon (see Malhotra, 1959, for references); but Gatenby and his colleagues (1949, 1953) have also recently recorded the continuation of the Golgi apparatus into the axon of the neurones of vertebrates.

Suitable preparations made by the silver impregnation method of Golgi (1898; see p. 70) reveal that the Golgi picture resembles those produced by Aoyama’s method, and also resembles Golgi’s own figure (1898) depicting his ‘apparatus’ in the Purkinje cells of ‘Strix flammea’ (fig. 1, A). Sometimes, however, the impression is given that the dark objects are only a random deposit of silver, lying superficially in the cell. The nucleus and the cytoplasm of the periphery of the cell are almost always free from this impregnation. It seems unlikely, therefore, that the black material in these preparations is actually a random deposit of silver. It seems more likely that the initial deposit of silver is on the basiphil reticulum.

In Sudan black preparations, small lipid globules are seen scattered in the cytoplasm. They correspond in size and distribution to the refringent globules seen in living neurones by phase-contrast or interference microscopy. Similar globules react positively to the acid haematein test, which also darkens a reticular structure in some of the Purkinje cells. This structure appears to be the basiphil reticulum. It is remarkable that no reticulum has been seen in acid haematein preparations of the neurones of other vertebrates (Casselman and Baker, 1955; Malhotra, 1959). The reticulum of the Purkinje cell of the owl has an unusually great affinity for silver and osmium, and this fact may perhaps be correlated with a large phospholipid component, or with a phospholipid component that is easily ‘unmasked’. This may have been the reason why Golgi discovered his ‘apparatus’ in this particular cell of owls. The reliability of acid haematein test for phospholipid was not checked by pyridine extraction test (Baker, 1946), on account of lack of material.

No satisfactory mitochondrial preparations could be made either by Metzner’s or by Hirschler’s method of staining, because in these sections a reticular structure (which is almost certainly the basiphil reticulum) is stained and the mitochondria cannot be properly differentiated. This again suggests the presence of phospholipid in association with the reticulum.

I am greatly indebted to Dr. J. R. Baker, F.R.S., for supervising this work and for very helpful criticism, and to Professor Sir A. C. Hardy, F.R.S., in whose department the work was carried out. I thank Dr. Leonard Harrison Matthews, F.R.S., of the Zoological Society of London, for making this investigation possible by putting me in touch with a dealer who possessed tawny owls.

This work was done during tenure of a Post-Doctoral Fellowship 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.
,
1944
.
Quart. J. micr. Sci
.,
85
,
1
.
Baker
,
J. R.
1946
.
Ibid
.,
87
,
441
.
Baker
,
J. R.
1949
.
Ibid
.,
90
,
293
.
Baker
,
J. R.
1957
.
Ibid
.,
98
,
425
.
Casselman
,
W. G. B.
, and
Baker
,
J. R.
,
1955
.
Ibid
.,
96
,
49
.
Fernstrom
,
R. C.
,
1958
.
Stain Tech
.,
33
,
175
.
Gatenby
,
J. B.
,
Moussa
,
T. A.
, and
Dosekun
,
F.
,
1949
.
Cellule
,
53
,
13
.
El Banhawy
,
M.
, and
Gornall
,
J. I. K.
,
1953
.
Ibid
.,
55
,
137
.
Golgi
,
C.
,
1886
.
Arch. ital. Biol
.,
7
,
15
.
Golgi
,
C.
1898
.
Ibid
.,
30
,
60
.
Helly
,
K.
,
1903
.
Zeit. wiss. Mikr
.,
20
,
413
.
Jordan
,
B. M.
, and
Baker
,
J. R.
,
1955
.
Quart. J. micr. Sci
.,
96
,
177
.
Malhotra
,
S. K.
,
1959
.
Ibid
.,
100
,
339
.
Nissl
,
F.
,
1894a
.
Neur. Centralbl
.,
13
,
676
.
Nissl
,
F.
1894b
.
Ibid
.,
13
,
810
.
Palade
,
G. E.
,
1952
.
J. exp. Med
.,
95
,
285
.
Weigl
,
R.
,
1910
.
Bull, internat. Acad. Sci. Cracovie, Ser. B, 691
.