Ultrastructure of the Invasion of Eimeria magna Sporozoites into Cultured Cells*†‡

SYNOPSIS. Monolayers of bovine kidney cells were overlaid with Eimeria magna sporozoites and observed with phase-contrast optics until penetration of the cells by the parasites had begun. Cells and penetrating parasites were fixed with glutaraldehyde and OsO₄-containing ruthenium red, dehydrated, and embedded in situ. Cells being penetrated were selected for study in the electron microscope. The lack of intracellular staining with ruthenium red and intact plasmalemmas of cells being penetrated, was accepted as evidence that the sporozoites did not disrupt the plasma membranes. The sporozoite caused invagination of the host cell plasmalemma until the parasite was entirely within the cell, after which the invagination was sealed off by short pseudopodia enclosing the sporozoite within a membrane-lined vacuole inside the cell. Often myelin-forms, apparently of host cell origin, were seen in the space between the sporozoite and the cell.     

Ruthenium-red staining of skeletal and cardiac muscles

Summary The effects of ruthenium red (RR) on amphibian and mammalian skeletal muscles and mammalian myocardium were examined. In skeletal muscle cells, a discrete pattern of staining can be brought about within the lumina of the terminal cisternae (junctional sarcoplasmic reticulum [SR]) by sequential exposure to RR and OsO₄. After prolonged immersion in RR solution, formation of pentalaminar segments (zippering) occurs at various points along the longitudinal (network) SR tubules. Zippering can be elicited in skeletal SR at any stage of preparation prior to postfixation with OsO₄. By means of dispersive X-ray analysis, both ruthenium and osmium were seen to be deposited in skeletal muscle junctional SR, and ruthenium was detected in the myoplasm as well. In skeletal muscles whose T tubules were ruptured by exposure to glycerol, the pattern of SR staining and zippering resulting from ruthenium-osmium treatment was not affected. These findings indicate that RR is capable of passage across the sarcolemma of skeletal muscle and that this passage does not occur solely under conditions in which the plasma membrane is damaged. In contrast, RR does not opacify or modify any region of the SR of cardiac muscle. However, after this treatment, randomly distributed opaque bodies, composed of parallel lamellar structures, appear throughout the myocardial cells. A few of these bodies are associated with lipid droplets, but the rest are of unknown origin. The failure of the SR of cardiac muscle to stain after exposure to ruthenium dye (even though this material enters these cells) suggests that the chemical composition of cardiac SR is significantly different from that of skeletal muscle SR.Supported in part by PHS grant HL-11155 (to N.S.) and American Heart Grant-in-Aid 78-753 (to M.S.F.). The authors are grateful to Drs. David Harder and Lawrence Sellin for their assistance with the preparation of frog skeletal muscle, to Dr. S.K. Jirge for his helpful suggestions and discussions, and particularly to Dr. Kenneth R. Lawless and Ms. Ann Marshall of the Department of Materials Sciences, University of Virginia School of Engineering, and Col. John M. Brady of the United States Army Institute of Dental Research, Walter Reed Army Medical Center, for their help with, and for the use of, the X-ray analysis equipment     

Ultrastructural differentiation of the epithelium and mucous glands of the esophagus in the chick embryo

Esophagi of White Leghron chick embryos, six days to hatching, were fixed in 5% glutaraldehyde in a cacodylate-sucrose buffer, postfixed in 1% OsO₄, dehydrated and embedded in Araldite. Tissues were sectioned with glass knives and viewed with a Philips EM 100C. The epithelium in early stages of development is characterized by small intercellular spaces, few cell processes and few desmosomes. In contrast the differentiated epithelium contains numerous cell processes, large intercellular spaces and numerous desmosomes. Mucin appeared in the mucous glands at 17 days.     

Ultrahistochemical study on the ruthenium red surface staining

Summary The cell coat picture effect which is usually obtained with the conventional RR method, that is with the RR/OsO₄ coupled reaction, is investigated. In this first paper, each of conceivable events which might take place between RR, OsO₄ and cell surface membrane is discussed or studied. Various tests are carried out on ascites Ehrlich carcinoma cells and Zajdela ascites hepatoma cells treated with numerous chemical reagents, as also on a few pure proteins. The set of data supports the concept that the staining pattern is due to the combination in surface membranes of RR with a colloidal-like form of OsO₂. The latter might occur during the formation of stable cyclic osmic acid diesters between OsO₄ and membrane unsaturated lipids. A possibility by which the resulting marker is though also to be in a colloidal-like state is put forward. A next report will deal with this problem.     

Penetration rates of osmium tetroxide with different fixation vehicles

Summary The presence of any substance other than OsO₄ in osmium tetroxide fixation mixtures causes a decrease in the penetration rate of the fixative. Fixation times should therefore be prolonged whenever a fixation mixture containing osmotically active substances is used.With I Figure in the Text     

Relative speed of fixation of glutaraldehyde and osmic acid in plant cells measured by grana appearance in chloroplasts

Summary Brief fixation in a mixture of glutaraldehyde and OsO₄ caused stacked chloroplast grana membranes in leaf cells of wheat, barley, tobacco, maize, cowpea, pigweed or bean plants to distend and vesiculate. Fixation with glutaraldehyde followed by OsO₄ prevented this fixation artifact. In a fixative mixture, OsO₄ apparently reacted with cell contents before glutaraldehyde.     

Ultrastructure of cardiac muscle of Trochosa terricola Thor., Pardosa amentata Clerck,P. pullata Clerck,and Pisaura mirabilis Clerck (Araneae: Lycosidae,Pisauridae)

Summary The ultrastructure of the cardiac muscle in some araneae has been investigated. The sarcolemma invaginates at the Z band and may extend into the cells through several myofibrils. Numerous T-tubules are given off both from sarcolemmal invaginations and also directly from the cell surface. In passing through the Z band, the luminal diameter of T-tubules greatly increases. Dyadic and a few triadic couplings are found mainly at the A-I level. Peripheral couplings were not seen.The ruthenium red solutions employed were prepared as described by Myklebust (1975), but the fixative contained 2% sucrose. The hearts were fixed for 3 1/2h in the ice cold solution, washed in buffer and post-fixed for 1 1/2h in a ruthenium red/OsO₄ solution. Dehydration and embedding was performed as described above. Sections were cut on a Reichert Om U2 microtome, stained with lead citrate (Reynolds, 1963) and examined in a Philips 300 electron microscope at 80 kV.Semithin sections for light microscopy were stained with toluidine blue (Mercer and Birbeck, 1972).I am much obliged to cand. real Erling Hauge for identifying the specimens used in the investigation proportion of relaxed sarcomeres. The fixed pieces were washed in a buffered sucrose solution and post-fixed in a 1% OsO₄ solution buffered with sodium cacodylate for 1 h. Following washes in a compatible buffer and in distilled water, the tissue was stained en bloc for l h in a 2% aqueous uranyl acetate solution. The tissue was dehydrated through cold acetone and embedded in Epon 812.     

Tissue fixation and osmium black formation with nonvolatile octavalent osmium compounds

Summary Several compounds of osmium^VIII, including potassium osmiamate and coordination complexes of OsO₄ with ammonia and various heterocyclic nitrogen compounds, have been synthesized and characterized. They have also been evaluated as substitutes for OsO₄ in postfixation of biological specimens and in light and electron microscopic cytochemical methods resulting in osmium black formation.The most useful of these osmic compounds, a molecular addition complex of hexamethylenetetramine (methenamine) with OsO₄, has a negligible vapor pressure of OsO₄. It has the molecular formula C₆H₁₂N₄.2OsO₄ and has been designated osmeth. Although it has only limited solubility, aqueous solutions of the compound (or of OsO₄) can be rapidly prepared by dissolution in a minimal amount of dimethylformamide and subsequent dilution with distilled water or buffer. Although stable in the solid state, the complex in solution undergoes partial dissociation releasing OsO₄, and the odor of OsO₄ becomes apparent.Such solutions of osmeth are (0.25%) considerably less concentrated with respect to OsO₄ than solutions (1–2%) ordinarily employed for ultrastructural preservation or in cytochemical studies. Osmeth has limited value for postosmication after glutaraldehyde fixation because the generation (release) of OsO₄ appears to be slow. Adequate osmication of tissue blocks exists only at the surface, but effective osmication can be achieved throughout tissue sections. In cytochemical reactions resulting in the formation of osmium blacks, the osmeth solutions are as effective as OsO₄ solutions of equivalent concentrations. Our findings indicate that OsO₄ solutions of less than 1% may be satisfactorily utilized in many cytochemical studies.Osmeth is safer and more convenient to handle than OsO₄ because small amounts may be solubilized as needed. It should be considered as a substitute for OsO₄ in ultrastructural cytochemistry.This investigation was supported by NIH research grant number DE 02668 from the National Institute of Dental Research and by NIH grant number RR 05333 from the Division of Research Facilities and ResourcesVisiting Professor, Dental Research Center, University of North Carolina at Chapel Hill, Jan.-May, 1975. Supported in part by USPHS Grant HD 09209     

PERMEABILITY OF INTESTINAL CAPILLARIES : Pathway followed by Dextrans and Glycogens

The pathway followed by macromolecules across the wall of visceral capillaries has been studied by using a set of tracers of graded sizes, ranging in diameter from 100 A (ferritin) to 300 A (glycogen). Polysaccharide particles, i.e. dextran 75 (mol wt ~75,000; diam ~125 A), dextran 250 (mol wt 250,000; diam ~225 A), shellfish glycogen (diam ~200 A) and rabbit liver glycogen (diam ~300 A), are well tolerated by Wistar-Furth rats and give no vascular reactions ascribable to histamine release. Good definition and high contrast of the tracer particles were obtained in a one-step fixation—in block staining of the tissues by a mixture containing aldehydes, OsO₄ and lead citrate in phosphate or arsenate buffer, pH 7.4, followed by lead staining of sections. The glycogens and dextrans used move out of the plasma through the fenestrae and channels of the endothelium relatively fast (3–7 min) and create in the pericapillary spaces transient (2–5 min) concentration gradients centered on the fenestrated sectors of the capillary walls. The tracers also gained access to the plasmalemmal vesicles, first on the blood front and subsequently on the tissue front of the endothelium. The particles are temporarily retained by the basement membrane. No probe moved through the intercellular junctions. It is concluded that, in visceral capillaries, the fenestrae, channels, and plasmalemmal vesicles, viewed as related parts in a system of dynamic structures, are the structural equivalent of the large pore system.     

CYTOCHEMICAL STAINING OF MULTIVESICULAR BODY AND GOLGI VESICLES

To investigate the origin and nature of vesicles found within multivesicular bodies (mvb), the cytochemical staining properties of mvb vesicles were compared with those of other cytoplasmic vesicles, i.e. those associated with the Golgi complex and endocytic vesicles found near the apical cell surface. Rat epididymal tissue was stained in unbuffered OsO₄ for 40–48 hr, and the distribution of stain was compared to that of reaction products for acid phosphatase (AcPase) to mark lysosomal vesicles, or thiamine pyrophosphatase (TPPase) to mark certain Golgi vesicles, or infused with peroxidase (HRPase) to demonstrate endocytic vesicles. Mvb vesicles were stained only by OsO₄; AcPase, TPPase, and HRPase reaction products stained the mvb matrix. OsO₄ also stained certain vesicles along the convex surface of the Golgi complex. The findings suggest that mvb vesicles in epididymal epithelium are not lysosomes and are not involved in protein uptake. The majority of these vesicles have cytochemical reactions in common with vesicles located along the convex surface of the Golgi complex and may be derived therefrom. A minority are derived from the mvb-limiting membrane.     

Endocytic vesicles and surface invaginations in cultured vascular endothelium: a morphometric comparison

Ruthenium red staining of plasma membrane glycoproteins of confluent cultured arterial endothelial cells revealed that the limiting membrane of many apparently discrete cytoplasmic vesicles was continuous with the plasmalemma. Surface invaginations accessible to ruthenium red appeared as vesicles when sectioned out of the plane of attachment to the cell surface, Morphometric analysis of ruthenium red-positive (RR+) and ruthenium red-negative vesicles (RR-) indicated that 47.2% of the total apparent vesicle population was RR+ and that those infoldings accounted for 19.6 +/- 1.4% of the cell surface in transverse sections. Whereas 14.9% of the true vesicles (ruthenium red-negative) were coated vesicles, only 1.1% of RR+ "vesicles" were coated pits. These studies show that although many deep infoldings of the cell surface may be misinterpreted as vesicles, almost all are uncoated. The existence of discrete coated vesicles (independent of coated pits) in vascular endothelium in vitro is readily apparent.     

Fine Structure of Bacillus subtilis : I. Fixation

The fine structure of Bacillus subtilis has been studied by observing sections fixed in KMnO₄, OsO₄, or a combination of both. The majority of examinations were made in samples fixed in 2.0 per cent KMnO₄ in tap water. Samples were embedded in butyl methacrylate for sectioning. In general, KMnO₄ fixation appeared to provide much better definition of the boundaries of various structures than did OsO₄. With either type of fixation, however, the surface structure of the cell appeared to consist of two components: cell wall and cytoplasmic membrane. Each of these, in turn, was observed to have a double aspect. The cell wall appeared to be composed of an outer part, broad and light, and an inner part, thin and dense. The cytoplasmic membrane appeared (at times, under KMnO₄ fixation) as two thin lines. In cells fixed first with OsO₄ solution, and then refixed with a mixture of KMnO₄ and OsO₄ solutions, the features revealed were more or less a mixture of those revealed by each fixation alone. A homogeneous, smooth structure, lacking a vacuole-like space, was identified as the nuclear structure in a form relatively free of artifacts. Two unidentified structures were observed in the cytoplasm when B. subtilis was fixed with KMnO₄. One a tortuous, fine filamentous element associated with a narrow light space, was often found near the ends of cells, or attached to one end of the pre-spore. The other showed a special inner structure somewhat similar to cristae mitochondriales.     

Differentiation of Cells in the Hypophysis and in Pancreatic Islets

Deparaffinized, 3-5μ, sections are brought to water, oxidized 3.5 min in an equal-parts mixture of 0.3% H₂SO₄ and 0.3% KMnO₄, and decolorized with 4% K₂S₂O₅. Nuclei are stained with Gomori's (1939) chromium-hematoxylin, and cell granules with Cason's (1950) mixture. The eosinophilic cells of the hypophysis and the alpha cells of pancreatic islets (of Langerhans) stain carmine red; basophilic and beta cells stain dark blue. Heidenhain's _susa is the most suitable fixative for hypophysis, Bouin's fluid for pancreas; but a satisfactory result is obtainable after formalin-sublimate or plain formalin. Besides studying the ratio of the cell types in the hypophysis or in pancreatic islets, it is possible to estimate the granule content of the cells. The method works on human autopsy material provided fixation of hypophysis occurs within 24 hr, and. pancreas, 12 hr post mortem, and it is suitable also for quite fresh organs.     

Osmium tetroxide footprinting of a scaffold attachment region in the maizeAdh1 promoter

Osmium tetroxide (OsO₄) reacts with the thymine residues of double-stranded DNA, but thymines that are unpaired or under torsional stress are hyperreactive. Although OsO₄ hyperreactivity has been primarily utilized to identify Z-DNA structures in supercoiled plasmids, OsO₄ will also identify other torsional perturbations of DNA. In this study, OsO₄ was used to footprint an AT-rich region (between –780 and –500) of the maizeAdh1 promoter. Hyperreactive sites were identified bothin vitro andin vivo in an area that coincides with AT motifs similar to those found in scaffold attachment regions. Further, the region of OsO₄ hyperreactivity lies within a fragment of DNA that is associated with the nuclear scaffold in histone-depleted nuclei.     

The Cell Surface of Trypanosoma musculi Bloodstream Forms. I. Fine Structure and Cytochemistry*†

SYNOPSIS. An extracellular surface coat was observed at the fine-structural level on the outer lamina of the pellicular and flagellar membranes of intact Trypanosoma musculi bloodstream forms. The surface coat had a mean width of 9.2 nm, and was composed of a somewhat electron dense, uneven, fibrillar-like matrix. Brief trypsin treatment of living blood forms completely removed the cell surface coat. Several cytochemical methods applicable to electron microscopy were used to detect the presence and distribution of carbohydrates in the trypanosome's surface coat and pellicular membrane. The polycationic dye compounds employed were: ruthenium red, ruthenium violet, Alcian blue chloride, and lanthanum nitrate. Electron-dense stain reaction products, indicative of polysaccharides, were evident in the surface coat of cells treated with these dyes, which also agglutinated both living and glutaraldehyde fixed cells. Like the surface coat, the pellicular membrane of trypsinized cells gave strong positive staining reactions with the several dyes, indicating the presence of membrane bounded carbohydrates, and living and glutaraldehyde-fixed trypsinized cells were agglutinated with the polycationic stains. Bloodstream forms were treated with the enzymes, α-amylase, dextranase, and neuraminidase. No obvious morphologic difference, however, was apparent between the surface coat of untreated cells and those subjected to treatment with any of the various glycoside hydrolase enzymes. Further, these enzymes had no apparent gross effect on the staining affinity of the surface coat for the several polycationic dyes. Cationized ferritin was used to visualize the negative cell surface charge of T. musculi bloodstream forms. Large quantities of cationized ferritin were bound in the surface coat matrix. Glycoside hydrolase enzyme treatments had no apparent effect on the amount of ferritin bound in the surface coat. Cationized ferritin was bound also to the outer lamina of the pellicular membrane in trypsinized cells, which had quantitatively less ferritin bound per surface unit area than bloodstream forms untreated by the enzyme. Living and glutaraldehyde-fixed cells were agglutinated with cationized ferritin. The results obtained in the various experiments indicated that polyanionic polysaccharides were constituent terminal ligands of the surface coat matrix and pellicular membrane in T. musculi bloodstream forms.     

Collagen Formation by Fibroblasts of the Chick Embryo Dermis

This investigation has sought to determine the relation between collagen fiber and fibroblast during fibrogenesis. Toward this end the surfaces of chick fibroblasts grown under in vitro conditions have been examined with the electron microscope after fixation in OsO₄. Supplementary information has been obtained from thin sections of fibroblasts fixed in situ during phases of fiber production. The evidence provided by these studies and by various conditions of the experiments indicates that the unit fibrils of collagen form in close association with the cell surface. They were never observed within the cell. When these unit fibrils form in bundles it appears as though templates of some nature, possibly coinciding with stress fibers within the cell cortex, influence the polymerization of the fibrils out of material available at the cell surface. From here the fibrils and bundles of them are shed into the intercellular spaces and there grow to limited diameters by accretion of materials from the general milieu.     

Neue Befunde zur Beschaffenheit des basalen Labyrinthes im Nierentubulus

Zusammenfassung Normale Rattennieren werden von der Aorta her mit Ferritin-Lösung perfundiert. Bei einer Versuchsgruppe wird die Nierenvene während der Perfusion abgeklemmt und die Ferritin-Infusion unter Überdruck kurzfristig fortgesetzt, wobei eine deutliche Volumenzunahme des Organes eintritt. Noch während der Perfusion wird durch Auftropfen von Osmiumtetroxydlösung auf die Nierenoberfläche fixiert.Die elektronenmikroskopische Untersuchung subkapsulär gelegener Tubuluskonvolute führt zu folgenden Ergebnissen: 1. In den Nieren ist nach Ferritin-Infusion unter Überdruck eine größere Ferritin-Menge durch die Wand der peritubulären Kapillaren in das Interstitium und in das interzelluläre Labyrinthsystem des Tubulusepithelwalles — bis an die Zonula adhaerens der Schlußleistenkomplexe — vorgedrungen. Damit erweist sich das Labyrinth-system als einheitlicher Diffusionsraum. 2. Regelmäßig ist in das Labyrinth des distalen Tubulus mehr Ferritin eingeströmt als in das des proximalen, wobei eine Erweiterung der Labyrinthspalten gegenüber den Kontrollen nicht festzustellen ist. 3. Bei Ferritin-Perfusion ohne Erhöhung des hydrostatischen Druckes in den peritubulären Kapillaren bleibt der Einstrom in das Labyrinth aus.In intravitalfixierten Kontrollnieren zeigen die interzellulären Labyrinthspalten des proximalen und des distalen Tubulus eine ziemlich konstante, aber unterschiedliche Weite (proximal 19,8±2,9 nm, distal 32,7±2,3 nm). Mikrodensitometrisch wird innerhalb der Labyrinthspalten die gleiche durchschnittliche Dichte registriert wie im Grundplasma des intravitalfixierten Tuhulusepithels. Es wird angenommen, daß das Labyrinthsystem eine organische Interzellularsubstanz enthält, deren funktionelle Bedeutung weiterer Klärung bedarf.

---

Summary Using a solution of ferritin an arterial perfusion was carried out in rat kidneys, until the blood had left the renal vascular system. Then the renal vein was clamped and the arterial infusion with ferritin was continued. During this procedure we intensified the pressure of the arterial infusion, until the volume of the kidney has been increased significantly (about factor 1.5). Under these conditions the kidney cortex was fixed in situ by dripping of OsO₄-solution on the surface of the organ. Electron microscopic studies of subcapsular convoluted tubules led to following results:1. Numerous ferritin molecules have left the lumen of the peritubular capillaries passing the pores of the endothelium, both the endothelial and the epithelial basement membrane and finally they have gained access to the intercellular spaces (basal labyrinth) of the tubular epithelial wall. By this the ferritin molecules could be located directly below the zonula adhaerens of the junctional complex. 2. Constantly more ferritin rushed into the labyrinth of distal than into that of proximal tubules. The labyrinth spaces containing ferritin are not widened in comparison with these of control tubules. 3. In additional perfusion experiments without ligature of the renal vein and without increased hydrostatic pressure no inrush of ferritin into the basal labyrinth had been observed. Therefore it can be concluded that the inrush of ferritin into the intercellular labyrinth is a consequence of an elevated hydrostatic capillary pressure during the infusion.In control kidneys fixed in vivo using isotonic OsO₄-solution (by dripping on the surface of the kidney or by tubular microperfusion) the intercellular spaces of both proximal and distal convoluted tubules show a rather constant, but different width (proximal 19.8±2.9 nm, distal 32.7±2.3 nm). Morphological and microdensitometric findings demonstrate within the intercellular labyrinth spaces the same average density as in the groundplasm of the intravital fixed tubular epithelium. It is suggested that the labyrinth system contains an organic intercellular substance the functional significance of which is not yet clear and must be further investigated.

---

---

Auf die Befunde an den peritubulären Kapillaren bei Ferritin-Durchtritt soll im Rahmen dieser Arbeit nicht näher eingegangen werden.     

Permeability of the foetal capillary endothelium of the guinea-pig placenta to haem proteins of various molecular sizes

Summary Haem proteins of different molecular sizes were perfused into the foetal circulation of the guinea-pig placenta to study the permeability of the foetal endothelium.The smallest molecules tested, microperoxidase (a_e 1.0 nm) and cytochrome C (a_e 1.5 nm), readily penetrated the endothelium; tracer-reaction product was found in the subendothelial space of the capillaries. However, there was no uptake of these two tracers into the syncytiotrophoblast layer of the placenta. An intermediate-sized molecule, myoglobin (a_e 1.7 nm), produced only a weak reaction product in the subendothelial space even when perfused at high concentration. The largest molecule tested, haemoglobin (a_e 2.8 nm), did not penetrate the foetal endothelium at any of the concentrations employed.The foetal capillary endothelium thus provided a barrier to protein penetration from the foetal circulation, dependent on molecular size. There was evidence that the site of this barrier was located in the lateral intercellular spaces between the endothelial cells.The syncytiotrophoblast of this haemomonochorial placenta provided an almost absolute barrier to protein penetration from the foetal circulation. As other workers have described maternal-to-foetal transmission of proteins across this layer in the guinea-pig, a working hypothesis of the role of endothelium and syncytiotrophoblast in maternal/foetal protein exchange is discussed.     

The Use of OsO4 Fixative for Feulgen-Stained Preparations

OsO₄ has many advantages over Carnoy's fixative mixture for the Feulgen nuclear staining in the protozoan Tokophrya infusionum. While Carnoy's fluid used prior to the Feulgen reaction produces shrinkage of the macronucleus and coarse clumping of its chromatin bodies, OsO₄ preserves faithfully the size and shape of the macronucleus and its chromatin material. This finding seems to be of special importance in view of the fact that electron microscopy relies on OsO₄ fixation. The satisfactory preservation of structured detail in Feulgen-stained preparations is of importance for the correlation of histochemical and morphological information.     

A histological method for the visualization of the intercellular permeability barrier in mammalian stratified squamous epithelia

Summary Mammalian epidermis and oral epithelia possess an intercellular permeability barrier which is located in the superficial region of the tissue. This study reports a staining reaction which appears to demonstrate a histological correlate of this functional property. Specimens of ear skin, palate, buccal and oesophageal mucosa and of cornea and bladder were obtained from adult rabbits and rats, bisected and either incubatedin vitro with 2.5% horseradish peroxidase as a tracer or fixed and processed for light microscopy and stained with a modification of Hart's elastin stain. Examination of specimens prepared by each procedure showed a complementary staining pattern in the intercellular spaces of the stratum corneum or in the superficial region of the non-keratinized tissue. In the epidermis and oral and oesophageal epithelia, the region which excluded the tracer stained with the modified elastin stain. In contrast, the corneal and bladder epithelia neither excluded the tracer nor showed intercellular staining. This relationship between staining of the intercellular space and the exclusion of tracer suggests that the intercellular material in the superficial region of epithelia may be chemically altered to form a barrier substance, possibly as the result of the discharge of the contents of the membrane-coating granules which are present in all the epithelia examined except the cornea and bladder.