Some extensions of the simple theory of phase-contrast microscopy are considered. It is emphasized that transparency, rather than thickness, is the limiting factor for the successful employment of the method. Certain transparent insect larvae (Chaoborus, Chironomus) can be observed in the living state by phase-contrast illumination.

The statement that the method is of no value for the examination of fixed and stained sections is based on consideration of an ideal physical case. In practice the method may be a valuable adjunct to routine examination of such material. Examples are given of the application of phase-contrast microscopy to normal and pathological stained sections.

The purpose of this paper is to present a preliminary report on certain less obvious applications of phase-contrast microscopy which may be of some biological interest. The principles of this method of microscopy are now too well known to necessitate any detailed description (Zernike, 1942; Burch and Stock, 1942).

Briefly, the method sets out to render transparent objects visible by converting changes in phase, which cannot be appreciated by the human eye or by the photographic plate, into changes of amplitude, which can be so appreciated. The theory, as worked out by Zernike (1942), shows that for the case of a grating composed of non-absorbing bars of different refractive indices, the final image is exactly that which would ordinarily be produced by a grating composed of absorbing bars, provided that the changes in phase produced by the grating are small. Perhaps the most remarkable feature of Zernike’s theory, as applied to a grating, is that it shows the enormous superiority of phasecontrast illumination over other methods for examining transparent material. Without going into details, the image produced by an absorbing grating can be represented by the sum of an infinite number of sine and cosine waves. The image of a transparent grating can be represented by a different series of waves. Ideally we should like to convert this second series into the first, and this is exactly what phase-contrast illumination does! All the other common methods of observation—oblique illumination, central and oblique dark ground illumination, or observation with reduced substage aperture produce images which are represented by different sine and cosine series, which while in some cases showing certain similarities to the ‘ideal’ series, in general fall far short of it. Roughly speaking, we may say that the effect of phase contrast illumination is as if we stained the object with a dye which stained each point with an intensity proportional to the product of its thickness and refractive index. To this extent, and subject to the performance of the microscope, phase-contrast illumination gives a true representation of what is actually present. All other methods commonly employed to render transparent objects visible produce an image which is often a mere caricature of the object.

It must be pointed out that this almost exact theory has only been worked out for very special objects, such as gratings, the elements of which produce very small changes of phase. It is reasonable to ask how small must these changes be, and what happens if they become large. It is not easy to give an exact answer to these questions since the mathematics involved is rather complicated. Fortunately, we can tackle the whole question in a much more general way, and without assuming any particular type of object.

Following Zernike’s treatment the object (or the image) can be represented by a vector diagram as in Text-fig. 1a.

Each point is represented in phase and amptitude by a vector. Areas which absorb light without affecting its phase will be represented by vectors lying along the line OM. Areas which absorb no light but merely alter its phase (i.e. transparent objects) will be represented by points on the circumference of a circle. Now according to the Abbe theory of the microscope, the final image is produced by interference between the central undiffracted image of the object and a series of diffraction images. In order to obtain phase contrast it is necessary to change the phase of the central pencil relative to the phases of the diffracted pencils. In practice this is usually accomplished by inserting a phase-retarding or phase-advancing strip into the back focal plane of the objective, and an illuminating diaphragm of corresponding shape in the substage. The phase strip is usually constructed to produce a 90° change of phase in the central image. The reason for this is that the exact theory for the case of a grating requires such a phase change in order to transform a ‘phase-grating’ into an ‘amplitude grating’. In practice, however, and in dealing with a more general type of object, this phase change need not be exactly 900. Indeed in some cases it is advantageous to produce a phase change other than 90°.

Returning to our vector diagram (Text-fig. 1) the effect of the central pencil alone can be represented by a vector which is the average of all the vectors representing the object. Let us suppose it to be represented by the line OC. The effect of the phase strip is to rotate this vector through 900, or in other words to shift the origin of the system from O to O’ (positive phase contrast) or O” (negative phase contrast). We can now see that non-absorbing points, which are represented by vectors such as P and Q on the circumference of the circle, will appear to have very different intensities when the origin is shifted to O’ or O”. Whereas with normal illumination the intensities of such points were represented by (OP)2 and (OQ)2, i.e. equal and therefore showing no contrast, with phase-contrast illumination the intensities become (O’P)2 and (O’Q)2 or (O”P)2 and (O”Q)2. Restricting ourselves to positive phase contrast (i.e. origin at O’) we see that highly refracting points will appear darker, less refracting points lighter, than the mean illumination.

Now suppose the object contains many points which produce large changes in phase. The effect of such points will be represented by vectors OPχ, OP2, OP?, OPit &c., around the circumference of the circle. If these points are sufficiently numerous the point representing the average of the whole system will approximate to O. Thus phase-contrast illumination would be of no value for such an object. At the same time we observe that the intensity of a point producing a phase change θ will be the same as the intensity for points producing phase changes of 2nπ-\-θ, i.e. an extremely refractile region may not always appear any darker than a less refractile region.

In practice, according to Zernike (1942), phase-contrast illumination is most useful for objects which produce phase changes of less than 45° (one-eighth of a wave-length) since objects thicker than this can be seen reasonably well by other methods. In most cases this means that phase contrast is best used with very thin objects. Single cells such as protozoa or tissue cultures are ideally suited for this method of examination. The excellent films produced by Michel (on the development of grasshopper spermatocytes) and by Hughes (on the growth of cells in tissue cultures) have already shown the enormous potentialities of the method for this type of material. Unfortunately, the insistence on the use of thin objects has rather tended to obscure the point that it is not thickness per se that is important, but the amount of phase change. Phase-contrast illumination may be quite useful for examining relatively thick but very transparent objects. Now the larva of Chaoborus (the phantom larva) has long been known for its transparency, and it was thought that it would provide good material for examination by phase-contrast illumination. The larva is usually several millimetres in diameter but can be flattened somewhat by mounting in water under a cover slip. At the same time this immobilizes the insect, but if carried out gently does not appear to harm it, for the latter will swim away quite normally when released after examination. Now it must be admitted that many of the cells and tissues of Chaoborus can be observed by the familiar method of reducing the substage iris to a very small diameter and altering the focus. As a matter of fact this is one method (though a very poor one) of producing phase contrast. Resolving power is lost owing to the small substage aperture, the method only works properly when the object is slightly out of focus, diffraction effects are accentuated, and the image is a very poor representation of what is actually present. With properly adjusted phase-contrast illumination the results are very different. It is possible to state with reasonable confidence that we can now observe living cells in the living intact multicellular organism with greater clarity, and with a greater chance of seeing what is actually present, than ever before. Pl. I, fig. 1, shows a low-power general view of part of a living Chaoborus larva, focused on two ganglia with the intervening nerve-cord. Pl. I, figs. 2 and 3, are higher power views of single ganglia, taken with the in. objective. These photographs show one large nerve-cell on each side of the ganglion. The nuclei and nucleoli are clearly seen and it will be observed that the cytoplasm is highly granular. These granules are particularly well seen in Pl. I, fig. 3. It is not yet possible to establish their nature with certainty. At first sight they might be taken for Nissl granules, and indeed their appearance in Pl. I, figs. 2 and 3, closely resembles that in fixed sections stained with Borrels methylene blue, a recognized Nissl stain. It is possible, however, that the granules seen in the living cells may be part of the Golgi apparatus, which is often very easily seen in other living cells by phase-contrast illumination. Another region of some interest is shown in Pl. 1, fig. 4. It is necessary to speak with caution of such material as the tissues of living insect larvae, but as a result of prolonged observations I am of the opinion that Pl. I, fig. 4, is a photograph of a living motor nerve-ending in a muscle-fibre. (All the muscles of Chaoborus larvae are composed of single fibres.) The pyramidal accumulation of sarcoplasm containing one or more nuclei with what appears to be a nerve-fibre entering at the apex corresponds to the description of the ‘Doyères hillock’ given in the entomological literature (see Morison, 1927). Pl. II, fig. 5, shows another view of these end-plates taken at the same magnification but from a smaller specimen. Two end-plates are seen (they lay on slightly different planes and a compromise focus had to be chosen in order to show both on the same photograph) and they appear to be interconnected by two nerve-fibres which seem to emerge from a common junction, at which there appears to be some sort of cell, possibly a peripheral nerve-cell, but more probably a neurilemmal cell. Numerous sarcoplasmic granules are also seen. Other insect larvae are also suitable for study by this method, e.g. young mosquito and Chironomus larvae (Pl. II, fig. 6). The giant salivary gland chromosomes in the intact Chironomus larva can often be seen quite well. Numerous possibilities exist for the study of secretory and excretory activity of cells in insect larvae. It is very much to be hoped that suitable vertebrate material may be found. I have carried out preliminary observations on the tissues of a tadpole’s tail and the results were quite promising. Unfortunately, these observations were made rather late in the spring when frog tadpoles were very scarce and rather large. The transparent tails of very young tadpoles should be quite suitable. The mesenteries of small vertebrates have not on the whole been found satisfactory, though very clear views of the capillary endothelium have sometimes been obtained. It is possible that a very thin transparent chamber inserted in the ear of a rabbit may prove satisfactory.

It is generally stated that phase-contrast illumination is of no value for examining stained objects. This opinion seems to have been based mainly on theoretical consideration of an over-simplified ideal case. In Text-fig. 1 we saw that Zernike represented all absorbing points as lying on the line OM, and all phase-changing points as lying on the circumference of the circle. When the origin is shifted from O to O’, the absorbing points on OM become roughly equidistant from O’ and therefore lose contrast. Now a stained section will contain points like U and V (Text-fig. 1b) which both absorb and produce changes of phase. The intensities of such points will be proportioned to (O’U)2 and (O’V)2 and they will appear to have good contrast, whereas with ordinary illumination (origin at O) they will appear equally stained. We thus have the theoretical possibility that phase-contrast illumination may after all be of some value for examining stained objects. To put the matter very crudely we should expect to lose much of the contrast due to the staining, but we might gain a different contrast due to differences in refractive index in both the object and the stains. (In a private communication Dr. Burch has suggested that stains may have anomalous refractive indices in the region of their absorption bands, and that it would be interesting to examine stained sections by phase-contrast illumination using monochromatic light of different wave-lengths, in order to vary staining contrast.) It must be admitted at once that this simple theory is far from adequate. Unfortunately, the exact treatment bristles with difficulties, and, indeed, it is unlikely that the problem will ever be solved for any but very special cases. In the circumstances it was judged best to try the method in practice. A wide range of routine normal and pathological material was examined, but a systematic examination of different stains and mounting media has not yet been made.

There are several ways in which phase-contrast illumination may be of value for examining stained sections. Perhaps its most obvious use is to render visible details of structure which have not been properly stained by the dyes employed. Such cases occur especially in certain histochemical reactions. Text-fig. 2a shows a section of bone marrow stained by the Feulgen technique. Only the nuclei are visible. Text-fig. 2b shows the identical field as seen by phase-contrast illumination. The cytoplasm as well as the nuclei of the cells is now rendered visible.

Other examples of the way in which unstained or poorly stained details can be made clearer are shown in Pls. II and III, figs. 7-10. It should be pointed out that it is often extremely difficult to demonstrate the superiority of phasecontrast illumination over ordinary illumination by means of photographs. The reason for this is that one automatically tries to get the best possible photograph out of any given material. Details which are seen with difficulty through the microscope can be rendered clearer by filters, differences in exposure, development, and printing. Strictly speaking one ought to treat both phase-contrast and ordinary pictures identically, but this is very difficult as the former require longer exposures. A direct comparison under the microscope is far more convincing than any series of photographs. Were it not for a natural tendency to ‘load the dice’ against oneself the photographic superiority of phase-contrast would be even more clear-cut.

Neurological material offers a particularly interesting field for study by means of phase-contrast microscopy. It is well known that scarcely any two methods of staining nervous tissues will produce the same result. As regards the various methods of silver staining, one may almost say that the same method in the hands of two different workers will produce different results. It is, therefore, of some interest to know what an entirely unstained nervo cell looks like. The answer is seen in Pl. Ill, figs, 11, 1z.

These cells were almost invisible by ordinary illumination, even when the substagc iris was almost closed, and were quite impossible to photograph, except by phase-contrast illumination. The appearance is more or less what one might expect to see if it were possible to use a mixture of silver staining and Nissl staining. It is the purpose of this paper to describe the possible applications of a new technique rather than to discuss results, but for the present we may say that the nerve-cell, free from staining artifact (“but not, of course, free from fixation artifact), may show appearances suggestive of Nissl granules, Golgi apparatus, boutons terminaux, and numerous extremely fine cell processes. A further investigation on frozen-dried material and tissue cultures is being undertaken.

Returning, however, to the question of stained nervous tissue, in general

Barer—Some Applications of Phase-contrast Microscopy 497 we may say that any given stain is fairly specific for either nerve-cells or nerve processes. Thus toluidin blue, methylene blue, and cresyl violet stain nuclei and Nissl granules but not nerve-fibres, whereas the reverse is on the whole true of most silver stains. We thus have the possibility of rendering the missing element visible by phase-contrast illumination.

Pl. IV, fig. 13, shows the Purkinje cells of the cerebellum seen in a Bielschowsky preparation. The nerve-fibres are very clear but the cells seem almost devoid of detail. Pl. IV, fig. 14, shows the same field by phase-contrast illumination. The nuclear structure is clearly visible. Conversely it is often possible to render the nerve-fibres more visible in Nissl preparations by means of phase-contrast. Finally, it is necessary to mention the various myelin stains, used for studying nerve-tracts. These usually leave the nervecells unstained and the latter can be rendered visible by phase-contrast illumination. Text-figs, 1a, 1b show this effect in a Weigert-Pal preparation (× 1 50) of a transverse section of the spinal cord (monkey).

The main disadvantage of the method is that it very often renders too much detail visible. Most stains are to a certain extent selective, but phase-contrast illumination may reveal so many minute fibrils that it is very easy to miss the wood for the trees. A considerable amount of further study will be necessary before any exact evaluation of the method as applied to neurological material can be made.

Many types of granules are extremely easily seen by means of phasecontrast illuminatiön. Pl. Ill, figs. 9, 10, shows a section of liver from a case of pernicious anaemia. Iron-containing pigment granules (haemosiderin) are present but very difficult to see by ordinary methods. Phase contrast reveals the presence of myriads of these granules with remarkable clarity. Even such normally well-seen pigment granules as melanin are rendered clearer by phase contrast. Pl. IV, figs. 15,16, from a section of a melanoma of the skin illustrate this point. The method has also been used with success for demonstrating secretory granules. It should be mentioned that all the photographs of fixed material shown were from routine paraffin sections, 8-12 µ thick, mounted in Canada balsam.

It is necessary to emphasize once more that it is not the purpose of this paper to discuss results in detail, but rather to suggest certain applications of phase-contrast microscopy which require further study. As regards the examination of living cells and tissues anyone who has ever seen such material by properly adjusted phase-contrast illumination would readily acknowledge its superiority over the older methods. The main problem here is to find suitable material. Isolated cells such as bacteria, protozoa, tissue cultures, or spermatozoa are of course ideal, but multicellular organisms are in many ways more interesting and important. Insect larvae and other transparent creatures such as certain medusae may yield useful information, but it would be very valuable to be able to extend the range of available material, especially to vertebrate tissues. The examination of stained sections by phase-contrast illumination is still in its infancy. Time and further investigation alone will show its uses and limitations. All that can be said at present is that it may prove to be a useful adjunct to the usual methods for examination of certain types of material. It is not possible to predict what will be seen by its aid. In some cases certain details may be rendered more clearly visible, in others the method will be found of no value. One practical point should be emphasized : the appearance of a stained section under the low power ( in. objective) is often bizarre and disappointing. The investigator should not be disturbed by this but should always proceed to examine the section with the in. objective, when the appearance is often greatly improved. I am at present unable to offer any satisfactory explanation for this phenomenon, but recent experiments suggest that appearances can be greatly improved by proper choice of colour filters. One last word: it is of course very important that any new method of microscopy should be received with caution. It must be admitted that there are many theoretical and practical points about phasecontrast illumination that require much further study. Nevertheless the theoretical basis underlying the method is sufficiently well understood to enable one to say that the image seen by its use bears at least as close a relation to what is actually present as that seen by more orthodox methods. Indeed, it is necessary to point out that the theory underlying the observation of stained objects by ordinary illumination is very far from complete. The theory underlying the use of oblique or central dark ground illumination is even less perfectly understood, but this has not prevented the extensive use of these methods. The proper course to pursue for the present is to use the method cautiously as an adjunct to other routine methods and to seek out its special applications and limitations.

I wish to thank Professor W. E. Le Gros Clark, F.R.S., for his encouragement and advice. I have had several valuable discussions on histological matters with Drs. A. Brodal, P. Glees, and G. Bourne. Mr. A. W. Dent and Mr. L. G. Cooper have provided technical assistance. Finally, I wish to express my special gratitude to Dr. C. R. Burch, F.R:S., for many stimulating discussions on all aspects of phase-contrast microscopy.

Burch
,
C. R.
, and
Stock
,
J. P. P.
,
1942
.
J. sci. Inst
.,
19
,
71
.
Morison
,
G. D.
,
1927
.
Quart. J. micr. Sci
.,
71
,
395
.
Zernike
,
F.
,
1943
.
Physica
,
9
,
686
, 974.

All the figures are untouched photomicrographs taken with the Cooke, Troughton, and Simms phase-contrast equipment. Green filter. Orthochromatic film.

PLATE I

Fig. 1. Low power view (phase contrast × 15o) of two segments of a living Chaoboms larva. Focused at the level of the nerve-cord and ganglia. Two large nerve-cells are visible in the lower ganglion.

Fig. 2. Higher power view (× 500) of a single ganglion showing two nerve-cell nuclei and nucleoli.

Fig. 3. Another example (× 500) showing the granules in the cytoplasm of the large nervecell on the right.

Fig. 4. Motor nerve-ending (phase contrast × 500) in striated muscle-fibre of a Chaoborus larva. Note the nucleus.

PLATE I

Fig. 1. Low power view (phase contrast × 15o) of two segments of a living Chaoboms larva. Focused at the level of the nerve-cord and ganglia. Two large nerve-cells are visible in the lower ganglion.

Fig. 2. Higher power view (× 500) of a single ganglion showing two nerve-cell nuclei and nucleoli.

Fig. 3. Another example (× 500) showing the granules in the cytoplasm of the large nervecell on the right.

Fig. 4. Motor nerve-ending (phase contrast × 500) in striated muscle-fibre of a Chaoborus larva. Note the nucleus.

PLATE II

Fig. 5. Another example of motor nerve-endings; same magnification as Fig. 4 (× 500) but from a smaller specimen. The two muscle-fibres lay on slightly different levels and a compromise focus was chosen. Note the sarcoplasmic granules and the apparent inter-connexion of the nerve-fibres.

PLATE II

Fig. 5. Another example of motor nerve-endings; same magnification as Fig. 4 (× 500) but from a smaller specimen. The two muscle-fibres lay on slightly different levels and a compromise focus was chosen. Note the sarcoplasmic granules and the apparent inter-connexion of the nerve-fibres.