ABSTRACT
The fibres and the connective tissue ground-substance of Glossiphonia complanata have been studied morphologically and histochemically. The fibres, which run in the connective tissue-substance between the organs of the body, are cytoplasmic extensions of bipolar or multipolar cells. The average diameter of the fibres is between 1 and 2 μ, and a single fibre may exceed 100 μ in length. Each fibre is differentiated into a cortex and a medulla. The cortex contains arginine, and some acid and neutral mucopolysaccharides ; this suggests that it is collagenous. The medulla is an extension of the protoplasm of the cell-body.
The cell-body of the fibre has a basiphil cytoplasm which contains arginine, acid and neutral mucopolysaccharides, and much ribonucleic acid. Spherical lipochondria up to 4 μ in diameter are present in the cytoplasm, especially near the nucleus ; they contain chiefly phospholipid. Threadlike mitochondria are present in the cytoplasm of this cell.
The connective tissue ground-substance is present as a matrix between the cells and organs of the body. Histochemical tests showed that it contains small amounts of arginine, tyrosine, neutral and acid mucopolysaccharides. The refractive index of the ground-substance is about 1·35.
With plate (fig. 2)
THE strength and resilience of the connective tissues of leeches is well known; in some species, e.g. Glossiphonia complanata, it prevents easy isolation of the organs and cells of the body. Much of this strength is due to the presence in the ground connective tissue-substance of interlacing fibres ; these were noted and figured by Bourne (1884), Scriban (1910, 1923), Abeloos (1925), and Scriban and Autrum (1934) among others. No histo-chemical study of these fibres appears to have been published, though deductions as to their nature have been made on the basis of their general staining characteristics. It is the purpose of this paper to give the results of histo-chemical studies of these fibres, and on the substance which forms the intercellular matrix.
Material and Methods
Material was obtained from the fresh-water rhynchobdellid leech, G. complanata. Much of the work was carried out on fixed material, studied in sections, but fresh tissue was also examined with the phase contrast micro-cope and with the Baker interference microscope. Details of the histo-chemical tests are given in the appendix.
The connective tissue-fibres
General. In any transverse section of G. complanata stained with suitable techniques, a mass of fibres can be seen in the connective tissue between the muscles and other organs of the body ; these fibres were considered by Scriban and Autrum (1934) to arise from embryonic connective tissue-cells by a modification involving extension of processes from the cell-body. They state that the majority of these fibres originate from bipolar cells, though some, especially in the region surrounding the ganglion chain, have multipolar stellate cells as their source.
The fibres are embedded in the ground-substance of the connective tissue, and it is stated by these authors that their axes generally lie in the long axis of the body. From my study of sections cut transversely and also longitudinally, it appears that the general tendency is for the fibres to run obliquely along the body, so forming an interlacing system. This is represented very diagrammatically in fig. 1, where the fibre paths between muscle-bands are shown. In the figure the course of the fibres is represented for the sake of clarity as being much more regular than is the case (many of the fibres have been omitted for the same reason). In the deeper regions of the body, e.g. around the lateral coelomic sinuses, the fibres appear to be much more random in their arrangement, but there may be special concentrations arounci certain organs such as the testis sacs.
It has been noticed in the course of the present study that there is a tendency for oblique fibres running between the longitudinal muscle-blocks to cross over, so that fibres which had been running, for instance, below the dorso-median muscle-block may eventually be found beneath the dorsolateral muscles, and so on. This is seen in fig. 1, and also in fig. 2, A, which shows such a crossing in the lateral coelomic sinus region of the body; it may be that this is a device for ensuring the maximum possible strength consistent with the greatest degree of body flexibility, though the actual crossing over of fibres would not of itself give rise to increased mechanical strength. This would only be so if the fibres were attached to some part of the body structure, either at their ends or along their length, or if the fibres were attached to each other, or, alternatively, if they were totally bound up along the whole of their length with a surrounding matrix.
In Glossiphonia, the method of fibre termination has not been observed, so that no conclusion can be drawn about the first of these possibilities, but it is possible to say that no attachments along the fibre length have been seen. This point contrasts with the findings of Cowey (1952) for the fibres in the basement membrane system of the Nemertine Amphiporus. From his results, it is clear that in this animal the fibres (which seem to be collagenous) do actually make contact with other fibres. A geodetic system is thus formed, and this ensures that there is no slipping of the fibres relatively to each other in contraction or extension of the animal.
It seems likely, however, that in the case of Glossiphonia the third possibility holds good. The difficulty of extracting whole cells and fibres from the body points to the existence of some strong ground matrix, as do all the histochemical results outlined later in the paper. If the fibres are assumed to be in intimate contact with this ground-substance throughout their length, then this arrangement would be expected to confer great strength and flexibility on the animal.
The fibres do not appear to branch, although it is difficult to see how they terminate, on account of their small size and because it is rare for the whole length of any one fibre to be included in the plane of a section. This complicates direct measurement of fibre length, but by the use of thick celloidin sections it has been possible to follow individual fibres for distances of over 100 μ along their length (fig. 2, c). Isolated fibres have been obtained from material which has been macerated for several days, but in no experiment did they exceed 100 μ in length and they were usually much shorter. Measurements of fibre diameter, however, are relatively easy and the values obtained gave an average diameter of between 1 and 2 μ.
The connective tissue-fibres are not homogeneous, but are clearly differentiated into a medulla, which is an extension of the protoplasm of the cell-body along the axis of the fibre, and a cortex. This distinction is clearly shown in
fig-3, taken from Scriban and Autrum (1934). The fibres represented in the figure are from Pontobdella muricata and Hirudo medicinalis, but they resemble very closely those found in Glossiphonia. It is noticeable that, in the figure, the cortex of the fibre is very pronounced and is represented as Connective Tissue-fibres in the Leech, Glossiphonia complanata 33 extending unchanged over the body of the cell, whereas in Glossiphonia, though the fibre cortex does extend over the cell-body, it is not marked in this region, but is represented here by a thin layer only.
The cytoplasm of the cell-body extends down the fibre axis, and is shown in the figure as possessing a marked alveolar nature. In the present study this appearance was noticed only in material which had been fixed in a coagulant fixative ; in studies of unfixed tissues, or of tissues fixed in formaldehyde, the alveoli were not at all obvious; so it may be that they do not represent structure present in the living cell, but are fixation artifacts. The collagenous nature of the fibre cortex was inferred by most of the previous workers from the noted swelling of the fibre when treated with acetic acid, and from the general staining characteristics of the fibre. The fibres were found to give all the colour reactions appropriate to collagen (see appendix). Although these staining reactions afford very strong presumptive evidence that the substance of the fibres is, in fact, collagen, they do not constitute histochemical proof; so in order to try to establish the identity of this substance beyond doubt, a series of more precise histochemical tests was undertaken.
Histochemistry
Lipids. The cortex of the fibres did not give a positive reaction to any of the standard tests for lipids ; the axial protoplasm, however, sometimes contained small globules which resembled the lipochondria of the cell-body. It seems that the cortical material of the fibres contains no free lipids detectable with the histochemical tests in current use, and that the lipid which occurs in the medulla of the fibres is restricted to small lipochondria concentrated near the cell-body.
Amino-acids and proteins
Only three methods were used, namely, the Hg I nitrite test (Baker, 1956) for phenols, especially tyrosine, the same author’s modification of the Sakaguchi test for arginine and other guanidine derivatives, and the ‘coupled tetrazonium’ reaction (Danielli, 1947; Pearse, 1954).
With the Hg / nitrite test, negative results were obtained, the fibres not colouring at all, so that it seems probable that tyrosine does not form any appreciable part of these structures. Sakaguchi’s reaction, on the other hand, gave a fairly strong coloration of the fibres, although the colour was not so intense as that given by the nuclei in the section. This makes it seem very probable that arginine enters into the composition of these fibres, but it was not possible to decide to what extent the cortex or medulla of the fibre was responsible for this particular reaction. In view of the observation that the cell-body was only very feebly coloured in this test (implying a low concentration of arginine), it seems probable that the same would be true of the protoplasm forming the medulla of the fibre, so the intense reaction may provisionally be attributed to the cortical portion of the fibre.
The ‘coupled tetrazonium’ reaction was rather weakly positive for the fibres, but if the sections were treated with mild heat and benzoylation according to the directions given by Pearse (1954), before doing the reaction, then the fibres coloured much more strongly. In his book Pearse shows that this result is characteristic of collagen, among other substances, the application of heat being presumed to cause some physical change by which the reactive groups of tryptophane and certain other amino-acids are prevented from combining with benzoyl chloride but remain active to diazonium salts.
Though the amino-acid content of collagen has been known for a long time, very exact analyses have not been available until much more recently (e.g. Bowes and Kenten, 1949; Randall, 1953); as a result of such work it is known that there is much more arginine than tyrosine in collagen. This accords very well with the observed histochemical reactions, the Sakaguchi test for arginine being strongly positive, whereas the test for tyrosine is entirely negative. On the basis of the histochemical tests for amino-acids, therefore, it seems reasonable to suppose that some part of the fibre is composed of collagen.
Carbohydrates
When the PAS reaction was applied to the material, the results were rather indecisive, the fibres colouring only very slightly pink. As a control, sections were subjected to the PAS reaction after incubation in saliva; some of the positive reaction was abolished by this treatment, e.g. in the muscle-fibres, but the connective tissue-fibres were still faintly coloured by the Schiff’s reagent. It may be concluded that there is a small amount of carbohydrate present in association with the fibres, but the reaction was so feeble that it was impossible to say whether it was concentrated in the cortex of the fibre or, as proposed for collagen fibres in other material, actually on the surface of the fibre.
In order to see whether the positive reaction could be due to the presence of an acid mucopolysaccharide, sections were stained in toluidine blue and examined for the occurrence of metachromasy. It was found that the fibres showed a very slight degree of metachromasy, which was much more apparent in thick sections. Again there was some doubt as to whether the reaction was more intense in the actual fibres, or whether the reacting substance was located chiefly in the intercellular material which surrounds them. As it was possible that some of the polysaccharide existed in the form of a neutral mucopolysaccharide, the technique recommended by Lison (1953), involving treatment of the sections with concentrated sulphuric acid for a very short time before staining with toluidine blue, was used to check this point. After this technique, it was found that the intercellular material, as well as the fibres, became intensely metachromatic, so that it seems extremely probable that much of the carbohydrate is actually present in the form of a neutral mucopolysaccharide. If this is the case, treatment with chromic acid before staining should cause the appearance of metachromasy (Lison, 1953) but with this particular tissue this experiment was unsuccessful. It is concluded, however, on the basis of the other results, that there is a neutral mucopolysaccharide both in the connective tissue matrix, and either in or on the surface of the fibres.
As the experiments with toluidine blue suggested the presence of an acid mucopolysaccharide in small amounts, it was decided to measure the basiphilia of the tissues by means of the methylene blue extinction method (Pischinger, 1927; Dempsey and Singer, 1946; Pearse, 1954), as the capacity to bind methylene blue at a very low pH is usually indicative of sulphate groups, i.e. acid mucopolysaccharides, or of nucleic acids. In this experiment, the value observed for the fibres was between pH 4·6 and 5, so that this gave no evidence for the existence of an acid mucopolysaccharide. Although the methylene blue extinction method cannot be considered as giving an accurate measure of the isoelectric point of any tissue component, it serves to give some indication of its value, and when taken together with the other evidence may furnish useful information; it is, therefore, interesting to note that the figure of pH 4·6 to 5 is very close to the isoelectric point of collagen as quoted by Pearse (1954) in his table 8.
Miscellaneous tests
As the fibres are probably in active growth for a considerable period of the life of the animal, it was thought possible that there might be an accumulation of nucleic acid in the cytoplasm of the fibre. When the pyronin / methyl green technique of Jordan and Baker (1955) was tried, it was found that although a strongly positive reaction was given by the actual cell-body, there was no result for the fibre. This was also true for the Feulgen test for deoxyribonucleic acid, which was shown to give a positive result only in the nucleus of the fibre-cell.
In view of work by Moog and Wenger (1952) and Goreau (1956), drawing attention to the occurrence of mucopolysaccharide at sites of high phosphatase activity, it was thought possible that a similar condition might occur in Glossiphonia. Tests were tried for both acid and alkaline phosphatases, but no positive reaction could be found in the connective tissues ; so it seems that in this animal no relationship exists between the mucopolysaccharide content of the connective tissues and the occurrence of phosphatase.
Though the method of Weigert for staining elastin is not histochemically specific, it gives a good indication as to the presence or absence of elastic fibres. In this leech, the results of such a test are negative, a finding which is supported by the histochemical results already considered. Elastin contains little arginine, but more tyrosine, and has an isoelectric point much closer to neutrality than that of collagen, so that if elastin occurred in these fibres to any extent, the Sakaguchi reaction would not be expected to be strong, but the Hg / nitrite reaction for tyrosine would be expected to give a positive result. It thus seems reasonable to assume that the fibres of Glossiphonia do not contain any elastin; they do, however, give histochemical and other reactions which are entirely consistent with the hypothesis that they are mainly composed of collagen and probably contain, in addition, small quantities of both acid and neutral mucopolysaccharides.
The fibre cell-body
General. The structure of the cell-body of a connective tissue-fibre is Shown in diagrammatic form in fig. 4. The majority of these cells are bipolar, though as mentioned by Scriban and Autrum (1934), multipolar ones occur in the connective tissue around the nerve-cord. The actual cell-body is between 15 and 20 μ. long, and between 7 and 10 μ in diameter; its ends are drawn out into the long processes which form the connective tissue-fibres. The cortex of the fibre continues over the body of the cell, though it is very thin in this region. The cytoplasm of the cell is basiphil and stains easily with dye lakes, such as haemalum; in most fixed material the cytoplasm appears rather granular, but in no preparation was any structure observed which resembled the ‘surround’ noticed in the adipose cells of the same species.
The nucleus of the fibre-cell is oval, usually measuring 7 or 8 μ. along its length; there is usually a very well-developed nucleolus. Only two types of cytoplasmic inclusion, mitochondria and lipochondria, have been noticed in these cells. The mitochondria appear as straight or slightly curved rods, about 3 μ in length and 0·5 μ in diameter, scattered throughout the cytoplasm of the cell, with no apparent concentration in any particular region. The lipochondria are spherical, with diameters from 1 μ up to about 4 μ ; they occur throughout the cytoplasm, perhaps with some tendency to aggregate near the nucleus. Similar bodies also occur in the medulla of the fibrous processes of this cell.
Histochemistry
The histochemical tests carried out on these cells are listed in the appendix. The tests show two significant features; the cells possess much arginine, as revealed by the Sakaguchi test, and they show a pronounced degree of basiphilia. Some indication of this is given by the methylene blue extinction value for the cells, which is about pH 2·6, a figure very different from the value for the fibres. The pyronin / methyl green technique showed that this cytoplasm coloured strongly with pyronin, so it seemed probable that the basiphilia was due to the presence of ribonucleic acid. It has been demonstrated (Bradbury, 1956) that when slides are incubated in a bath of treated saliva, the basiphilia due to ribonucleic acid is abolished; in this cell, the treatment with saliva entirely prevented coloration of the cytoplasm by the pyronin, so it is concluded that the basiphilia is due to the presence of ribonucleic acid.
The usual tests for lipids were tried, but positive reactions were only obtained from the lipochondria. These were very strongly coloured by Baker’s acid haematein technique, as also were the mitochondria; they did not colour blue after pyridine extraction, so both inclusions appear to contain large quantities of phospholipid. As two types of fatty inclusions, one composed of neutral fat, and the other of phospholipid, were found in the adipose cells of Glossiphonia, it was decided to check whether a similar condition existed in this cell. When the acid haematein test is followed by Sudan IV, all lipids which are not coloured blue by the acid haematein will be revealed in red. In this cell, no fat drops coloured with the Sudan dye, so that it seems likely that only one series of fatty inclusions is present.
It has been noted in other material, e.g. the sympathetic neurone of the mammal (Casselman and Baker, 1955), cerebrosides are very often associated with the lipochondria. These compounds are not affected by fixation in cold acetone, but they are removed from the material if the acetone is kept near its boiling-point in a simple reflux apparatus. Any difference which is apparent between sections from material treated in these two ways and then coloured with Sudan black may be attributed to the presence of a cerebroside. In Glossiphonia it was found that no lipids could be detected after fixation in either hot or cold acetone, and it is therefore probable that in this cell the phospholipid is not in association with a cerebroside.
The ground-substance
The existence of some form of matrix between the cells and organs of the vertebrate body has been recognized for a considerable time, though some authors, e.g. Nageotte (1922), have questioned this view. Nageotte and his followers deny the existence of an amorphous substance, considering that the spaces between the components of the body are empty except for tissuefluids. Bensley (1934) has gone into the evidence for the existence of an intercellular ground-substance in vertebrates, and shows that there are very strong grounds for assuming its existence. Among other experiments, she measured its refractive index, demonstrated that it has an affinity for metallic mns, and found that it can be differentially stained with toluidine blue.
The ground-substance of the connective tissue of Glossiphonia appears perfectly homogeneous in unstained sections examined by phase contrast microscopy, but it may easily be recognized in sections by its uniform coloraron with standard histological techniques such as Masson or picro-indigocarmine. The ground-substance may also be revealed by histochemical tests such as the PAS reaction or by its differential staining with toluidine blue. As Bensley pointed out for the ground-substance of vertebrate connective tissue, there is a marked affinity for metallic ions and this property may be utilized to show very clearly the extent of the ground-substance in Glossiphonia. With osmium impregnation, this substance appears coloured a uniform brown and is seen to surround all the organs of the body; in it are embedded the muscles, connective tissue-fibres, pigment cells, adipose cells, the coelomic sinus system, and all the other structures of the body. Fig. 5 represents diagrammatically part of the animal seen in transverse section with the area occupied by the ground-substance indicated in stipple.
Histochemistry of the ground substance
It is seen from the appendix that all the tests for lipids are negative, as are those for enzymes and for nucleic acids.
The ‘coupled tetrazonium’ reaction was weakly positive, indicating the presence of a small amount of protein ; weak positive results were also given by Baker’s Hg / nitrite test and Sakaguchi’s test, so that some arginine and tyrosine are present, though neither seems to be very abundant.
The PAS reaction is weakly positive, with no decrease in intensity when the sections are treated with saliva before the test. This seems to rule out the possibility of a reaction due to glycogen. It suggests instead the presence of either an acid or a neutral mucopolysaccharide. When sections are stained with toluidine blue, a very weak metachromasy of the ground substance is noticed; treatment with concentrated sulphuric acid before staining with toluidine blue gives a strong metachromasy, as in the case of the fibres: this evidence supports the view that a neutral mucopolysaccharide is present. There was no increase in metachromasy nor in the intensity of the PAS reaction after treatment with chromic acid. Methylene blue extinction occurred between pH 5 and 6 ; this result does not give any significant information about the composition of the ground substance.
θ — the optical path difference expressed in degrees,
t — the thickness of the object in μ (in this case the thickness of the section),
n = the refractive index of the mounting medium, and
λ = the wavelength of the light used.
Measurements of the refractive index calculated from readings taken with the object mounted in water and from other readings with the object in glycerine were very similar; the mean of five measurements was 1 · 35, a figure very close to that obtained by Bensley (1934) for the connective tissue groundsubstance of vertebrates. Her method was to immerse the tissue in different media of known refractive indices and note when the tissue became difficult to distinguish from the background. The refractive index of the fluid in which this condition was fulfilled was taken as a close approximation to that of the tissue. The value measured for Glossiphonia is rather low, which suggests that the ground-substance does not contain much solid matter.
In an attempt to confirm the histochemical results on the carbohydrates by an alternative method, paper chromatography was used, though when used as described below it has the great drawback that the source of the extract is not precisely known. The experiments to be described show that the extracted substances are PAS positive and metachromatic, and show increase of metachromasy on oxidation. Histochemical study of sections showed that if glycogen is excluded, the only substances in the body which fulfilled all these conditions were in the cuticle, in the connective tissuefibres, in the connective tissue ground-substance, or in very small amounts in the adipose cells.
The possibility of glycogen being responsible was excluded by using starved animals. Furthermore, as the amount of material extracted was quite considerable, the adipose cells and the cuticle may be disregarded; and so, by exclusion, it seems that the substances subjected to paper chromatography must be derived from the fibres, or more probably from the ground-substance.
The extraction of the material was carried out by homogenizing two leeches in i ml of 10% aqueous calcium chloride solution, a solvent which was used by Meyer and Smith (1937) in their studies of the acid mucopolysaccharide, chondroitin sulphate. The aqueous extract was allowed to stand overnight at a temperature of 4° C before centrifuging in order to remove the cell debris. Spots of the supernatant liquid were placed on Whatman’s no. 1 paper, with samples of commercial chondroitin sulphate and heparin as comparison substances. A mixture of 0 · 25% propanol in M/15 phosphate buffer at pH 6 · 4 was used as a solvent, as recommended by Kerby (1953) for the separation of chondroitin sulphate and heparin. After the chromatogram had been run, the spots were developed either by spraying the paper with a 0 · 25% aqueous solution of toluidine blue, by performing the PAS reaction on the paper, or by oxidizing the substances with periodate before spraying with toluidine blue.
In the chromatograms developed with toluidine blue alone, the chondroitin sulphate and the heparin gave well-defined metachromatic spots, but there was no spot from substances present in the extract. When the paper was dried, however, a faint metachromatic spot appeared in a position which suggested that it was due to some substance present in the leech extract. No explanation can be put forward to account for the fact that this substance only showed metachromasy when the paper was dry; possibly the substance responsible may only have been present in extremely small amounts, or it may be the result of considerable structural change induced in the molecule of the substance by drying. The PAS reaction showed a spot, due to some substance in the extract, which coincided in position with that shown by the toluidine blue ; in both cases, the spots did not coincide with those due to the heparin or the chondroitin sulphate. After oxidation by means of periodate, the spot due to the extract was much more spread out, much larger, and appeared intensely metachromatic. This may be a result of the oxidation of a neutral mucopolysaccharide by the periodate, so giving a product which is metachromatic.
These experiments seem to confirm the presence of small amounts of an acidic, and also larger amounts of a neutral mucopolysaccharide in an extract from Glossiphonia. The location of these substances cannot be deduced from chromatography, but the histochemical work shows that both probably occur together in the ground-substance of the connective tissue and the fibres.
Discussion
It seems clear from a consideration of the results given in this paper that the fibres found in the connective tissue of Glossiphonia complanata differ very considerably in structure and arrangement from those of vertebrates. The fibres are sharply marked off into a cortex and a medulla; are clearly long extensions of the actual fibre-forming cells ; and retain a connexion with the cells which secrete them. These are all points of difference from the vertebrates. In addition, the fibres of Glossiphonia never seem to associate into large bundles, so commonly found in vertebrate animals, but remain single and interlace among each other in the ground-substance. There is a resemblance to vertebrates in that the fibres almost certainly contain collagen, together with mucopolysaccharides. Several problems still remain, however, in particular the questions of fibrogenesis, of the cellular state of the fibres, and the origin of the ground-substance.
Even in the vertebrates, the problem of fibrogenesis is still not solved; in particular there is still doubt whether the fibres are secreted intra- or extracellularly. Recent work published by Fitton Jackson (1955) suggests that even if the fibres are formed outside the cell, the actual process of formation is closely linked to events within the cell. In chick fibroblasts, granules have been observed in the protoplasm of the cell. These granules have been shown to consist of carbohydrates and some protein, and to possess alkaline phosphatase. It is suggested that these granules may be the precursors of the whole or part of the ground-substance, or possibly of the collagenous fibre itself. Extrusion of the granules into the ground-substance has not been noticed by this author but has been claimed by other workers, e.g. Stearns (1940). No such granules have been observed in the fibre cell-bodies of Glossiphonia, but a similarity does exist between the granules seen in the chick and the ground cytoplasm of the fibre-cells in Glossiphonia with respect to their chemical composition, so far as it can be determined. A conspicuous difference is that no phosphatase activity has been detected in the leech tissue.
With the methods at present available, it does not seem possible to determine whether the collagen forming the cortex of the fibre is actually deposited on the inner surface of the cell-membrane, or whether it is laid down on the outer side of the membrane, when the fibre might be considered to be extracellular.
Histochemical study of developing embryos of this leech could give mteresting results, and might perhaps throw some light on the problem of whether the fibre consists of some other substance before the collagen is deposited.
The ground-substance of the connective tissue shows similarities in chemical composition to that of vertebrates. Measurement of the refractive index suggests that the solid content is rather low. This may indicate that this ground-substance exists as a semi-fluid gel. Histochemical studies do not help much in deciding the origin of the ground-substance. It has been thought to arise in several ways, e.g. Bourne (1884) considered that it resulted from ‘ectoplastic modification of the corpuscles lying within it’. A more recent view is that the ground-substance is the result of secretion by the connective tissue-cells. As there are chemical similarities between the composition of the ground-substance and the fibre-producing cells, this does not seem entirely unreasonable. Future work with histochemical methods on the developing tissue may help us to understand this process more clearly.
ACKNOWLEDGEMENTS
I have pleasure in acknowledging my debt to Dr. J. R. Baker for supervising this work and for stimulating discussions. I also wish to thank Professor A. C. Hardy, F.R.S., in whose department the work was done, and Dr. P. C. J. Brunet for teaching me some essentials of chromatographic technique.
The investigation was carried out during the tenure of a Senior Hulme Scholarship of Brasenose College, Oxford, and of a grant from the Department of Scientific and Industrial Research.