Approximately 50% of a 50 kV beam of electrons can penetrate about 100 nm of a specimen with a density of 1 g/cm3 (the approximate density of most common biological specimens). Thus, it is not possible to study whole cells in a conventional 20-100 kV TEM. It is necessary to obtain specimen samples that are 50 nm thick or even thinner. Before this is done the sample must be stiffened so it is capable of bearing up against a sharp cutting edge sufficiently well such that very thin slices can be cut. Before this the specimen must be "killed" in such a way as to preserve faithfully every detail down to the molecular architecture just as it was in life. A critical factor in obtaining good experimental results is the isolation of the tissue in as close as possible to in vivo conditions. The term tissue is used in a liberal sense to include specimen samples whether they be from animal, plant, bacteria, algae, etc.
Most specimens can be excised from living tissue and gently cut into small pieces within a few minutes, thus minimizing trauma and the time that the tissue is deprived of its essential nutrients. Slicing of excised tissue is best done on a non-absorbant surface (e.g. Polyethylene slab) in a droplet of fixative. The tissue should generally be cut into pieces with at least one dimension smaller than 1 mm to assure rapid fixation of the entire piece of tissue.
Tissue can also be fixed in situ, for example in living anesthetized animals. This can greatly minimize the effects of cutting on unfixed tissues and of anoxia resulting from cutting off circulation.
Specimens such as protozoa, cell suspensions, algae, and bacteria can be prepared by being placed directly into fixative. If they are very small, the specimens can be centrifuged to form a pellet for easier handling immediately after the fixation. The speed of centrifugation necessary to form a pellet is dependent on the nature of the specimen.
a. Fixation
The main purpose of fixation is to maintain the original form of the specimen so it can resist the effects of subsequent steps in the preparative procedure. The aim is to preserve every detail of cellular ultrastructure right down to the molecular level exactly as it was in life the instant before cell death. The first problem is to obtain cells in their normal living condition, thus most tissues have to be obtained from a living organism. A variety of methods have been devised for getting fixative solution to the desired tissue. Because the time of penetration of the fixative into the tissue is critical, the tissue must generally be cut into very small pieces, preferably cubes of not more than 0.5 mm on a side.
1) Fixatives
The primary purpose of the fixative is to solidify the protein sol bounded by the cellular membranes and to preserve the spatial relationships of the various organelles as they were at the instant of fixation. It should also render insoluble all the other chemical constituents of the cell such as nucleic acids, nucleoproteins, carbohydrates and lipids. Some fixatives (such as osmium tetroxide) also provide electron contrast.
The detailed nature of the interactions of OsO4 with cellular components is not fully understood although it is known to react with unsaturated lipids and with tryptophan and histidine in proteins, thus cross-linking polypeptide chains together. OsO4 acts to stabilize cellular proteins which form the matrix of the cytoplasm. A number of factors, including pH, temperature and time of fixation are critical for good fixation. For example, the time of fixation is normally 30 to 90 minutes. If fixed for too short a period, there may be inadequate stabilization of the specimen, whereas, fixation for prolonged periods can occasionally lead to solubilization of some tissue components. OsO4 is usually prepared in a buffered solution at near neutral pH. It must be noted that various buffers produce different appearances in ultrathin sections of tissue fixed with OsO4 at identical concentration and pH. For example, differences in the preservation of nuclear chromatin, endoplasmic reticulum, mitochondria, and background cytoplasm have been observed.
Osmium is good for rendering lipid insoluble but has no effect on carbohydrates such as glycogen. It stabilizes protein sols by combining chemically with them and forming cross-links, not by precipitating them. It preserves lipids by forming addition compounds with unsaturated fatty acid chains. It is good at preserving the phospholipoprotein membrane skeleton of the cell. It preserves nucleic acids and carbohydrates poorly, but nucleoproteins well. Since it combines chemically with practically all the constituents of the cell, osmium metal remains behind in the fixed cell, attached firmly to the structures it stabilizes, and delineating them almost perfectly. It penetrates tissue very slowly due to the low diffusion rate of its large molecules.
Formaldehyde and glutaraldehyde are the most popular aldehydes used as fixatives. They preserve structure excellently but allow lipids to be completely extracted by the alcohol used as the dehydrating agent. They preserve glycogen well but provide no additional electron contrast. They penetrate tissue rapidly and have the property of stabilizing glycogen so its subsequent loss with OsO4 is prevented. They preserve structures such as microtubules which are poorly or not at all preserved by OsO4 and they stabilize nucleoproteins better than OsO4
Glutaraldehyde is often used in conjunction with OsO4 to give better specimen stability during subsequent dehydration, embedding, and electron microscopy. Two major advantages of aldehyde fixatives over OsO4 include:
- more rapid penetration of tissues resulting in less distortion and better preservation of structural relationships This is a strong oxidizing agent which preserves cell membrane structures and is particularly good in preserving myelin. It also acts as an electron stain by precipitating coarsely granular, electron dense material on the membranes. DNA and RNA-containing elements are not satisfactorily preserved, however. Thus, permanganate is better used as a selective rather than a general purpose fixative. The rate of penetration of permanganate is, like OsO4, slow in comparison with that of the aldehydes. The selective action of permanganate serves to make membranes stand out because of its strong reaction with the phospholipid component of the membrane. No one substance is likely to provide all the functions required by a fixative. Most modern work is therefore on material which has been fixed first in an aldehyde followed by a second fixation in OsO4. 2) Buffers and additives Effects due to osmotic pressure differences between the fixative solution and the cytoplasmic sol may cause disruptive effects (i.e. shrinkage and swelling) to the organelles. Thus, neutral salts or other inert substances are used in the fixative solution to make it isotonic with the tissue fluid and a buffer is used to maintain the pH of the fixative solution at the physiological value in spite of the chemical activity of the fixative. A commonly used buffer contains two components: sodium acetate to provide buffering capacity in the acid range and sodium barbiturate (Veronal or barbitone sodium) for the alkaline range. The necessity for using additives (e.g. Sodium chloride or sucrose) to increase osmotic pressure is a controversial matter, some saying the additives are essential and others claiming they are useless. 3) Fixation Technique A two-stage fixation procedure using 1-6% glutaraldehyde buffered with phosphate (0.05-0.1M, pH 6.8-7.6) followed by OsO4 is now almost universally used. Primary fixation may be performed (usually on ice) for 1-3 hours or more depending on the size and permeability of the cut tissue samples. After primary fixation, the unreacted glutaraldehyde is washed out since it combines with and reduces the osmium tetroxide causing unwanted precipitation in the tissue. Since OsO4 penetrates very slowly into fixed tissue, the blocks (chunks of tissue) must be made as small as possible before the secondary fixation (<0.5mm in the smallest dimension). The secondary fixation is normally carried out at room temperature with a 2% solution of OsO4 (buffered or unbuffered). Fixation steps are generally performed in a fume hood, especially when OsO4 is used since it is extremely volatile and toxic (instantly kills any epithelial cells it contacts). 4) Other fixation methods The most commonly used non-chemical method of fixation is that of freeze-substitution in which small pieces of the specimen are rapidly frozen by plunging them into isopentane cooled to liquid nitrogen temperature. To prevent ice crystal formation, the tissue is first soaked in cryoprotective agents such as glycerol, ethylene glycol, or dimethyl sulfoxide and is then rapidly frozen. Water is then removed by sublimation, with the specimen kept at about -20°C and under vacuum. The frozen-dried tissue is then embedded by infiltrating it with a plastic monomer (under vacuum) which is then polymerized. This technique is especially good for preserving enzymatic activity in the tissue sample, thus histochemical reactions can be performed on the sections to localize the sites of activity of certain enzymes. b. Dehydration The aim is to replace all the free water in the specimen with a fluid which is miscible both with water and with the embedding monomer. Agents such as ethyl alcohol, methyl alcohol, isopropyl alcohol, acetone, or the monomer of a water soluble plastic embedding medium are used to dehydrate the specimen. Ethyl alcohol is the most widely used because it does not harden the tissue and make it too brittle for subsequent ultrathin sectioning. Other agents can be used depending on the nature of the embedding material being used. For example, acetone is used when the polyester resin Vestopal is used. Inert compounds such as ethylene and propylene glycol can be used as effective dehydrating agents because they serve both to displace tissue water and stabilize the cell's macromolecular systems, thus resulting in retaining most cytoplasmic proteins and the fine structural relationships between them. The duration of the dehydration step is normally kept short to prevent extraction of tissue components and subsequent shrinkage. The specimen sample is transferred into mixtures of water and dehydrating fluid of decreasing water concentration. The usual routine calls for rapid dehydration: short treatment with 70% alcohol, then transfer to 95% and finally to 100% for about an hour. After the specimen has been transferred to 100% dehydrating agent, it may then be transferred to a fluid completely miscible with both alcohol and the resin monomer. Such a fluid is propylene oxide (1,2-epoxypropane: EPP) which is completely miscible with the epoxy resins almost invariably used for general embedding purposes. The processes of dehydration and infiltration should be carried out as rapidly as possible because all the reagents used are powerful lipid solvents, and remove a significant amount of lipid even after it has been fixed with OsO4. c. Block staining Further differential electron contrast can be added after osmium fixation and before the tissue is embedded. The washing and dehydrating fluids can be used to add further stains such as uranyl acetate, phosphotungstic acid, or potassium permanganate. d. Embedding In the early days of thin sectioning, paraffin waxes were used as embedding media but were found to be too soft to enable sections thinner than about 1 µm to be cut. The first embedding medium found suitable for electron microscopy was poly-butyl methacrylate. The majority of sectioning studies presently employ one of the epoxy resins. 1) Properties of an ideal embedding medium: - Soluble in ethanol (or acetone) before polymerization No one embedding medium possesses all of these properties. 2) Types of embedding media This penetrates tissues rapidly and is easily cut into sections in the thickness range 50-100 nm. Proper hardness of the tissue blocks is achieved by varying the ratio of a mixture of methyl methacrylate and n-butyl methacrylate monomers. Hardening of the blocks is the result of polymerization of the monomer molecules to form simple linear polymer chains. Among the problems encountered with methacrylate sections are: - Shrinkage - up to a 20% decrease in volume is encountered when methacrylate polymerizes The disadvantages of the methacrylates arise mainly from the fact that they form linear polymers which do not cross-link to form a stable three-dimensional structure. Gross tissue artifacts are caused due to shrinkage of the methacrylate when it polymerizes. Cell components are dragged apart giving a vacuolated appearance to the embedded tissue. Evaporation of the resin in the electron beam leads to collapse of the unsupported structures on top of one another causing gross distortion of the ultrastructure. The evaporated plastic also deposits on the polepieces of the objective lens leading to astigmatism that becomes uncorrectable. Evaporation of the methacrylate does, however, result in an increase in specimen contrast. Evaporation can be reduced by pre-irradiating the section at low beam intensity and by sandwiching the section between protective films of carbon. Aqueous solutions of these media are used immediately after fixation, both as dehydrating agents and then, in pure form as infiltration and embedding agents. Three water soluble agents include: Glycol methacrylate (2-hydroxyethyl methacrylate) - sections tend to stretch in the electron beam. Aquon - water soluble component of the widely used epoxy resin, Epon 812. It is of relatively low viscosity and sections fairly well. Durcupan - water soluble epoxy resin, which, when used without other added epoxies, gives very soft blocks. When sufficient additives are provided to improve its sectioning properties, most of the advantages associated with its water solubility appear to be lost. Vestopal W, an acrylic resin, uses an initiator and activator. If alcohol (not miscible with Vestopal) is used to dehydrate the specimen, then the alcohol must be replaced with styrene, which is miscible with both alcohol and Vestopal. Alternatively, acetone can be used to dehydrate the specimen. Vestopal-W has the disadvantages of being very viscous and immiscible with ethyl alcohol, but the sections are very stable in the electron beam. These are a family of thermosetting synthetic resins, which, when mixed with suitable curing agents and heated, polymerize irreversibly into cross-linked, yellow-brown solids. The resin has two types of chemically reactive groups: epoxide end groups and hydroxyl groups spaced along the length of the chain. If a mixture of resin, amine (to form long chain polymers), and anhydride (to form cross-bridges between resin molecules via the hydroxyl groups) is heated together, polymerization takes place in three-dimensions forming a stable, inert substance consisting of polyesters and polyethers very resistant to heat and solvents. The mechanical properties of the resulting polymer, which effect its cutting properties, are governed by the length of the hydroxyl-containing part of the resin chain, the chain length of the acid anhydride, and the proportion of amine. The mixed components are polymerized by heating to 60°C for 48 hours. The cutting properties of the final polymer depend on the components of the resin monomer mixture. The mixture consists of an epoxy resin, a hardener, an accelerator (controls the rate of hardening), and a plasticizer (controls hardness of the block). No polymerization damage is seen as with the methacrylates. Among the more popular epoxy resins in use are Araldite, Epon 812 (perhaps the most widely used embedding media), DER-334 (a Dow epoxy resin), and Maraglas. 3) Embedding procedure Tissue blocks are infiltrated with embedding medium in the same way as the dehydration step was performed. The final dehydrating fluid is replaced with a 50-50 mixture of dehydrating agent and resin mixture before infiltrating with 100% resin mixture. For example, with Epon 812, a stock mixture of the resin and the two hardeners is made up in the correct proportions, leaving out the catalyst. The catalyst is then added at the time the infiltration is done. The complete infiltration step usually requires several changes of embedding medium over a period of several hours. After infiltration, the sample is usually transferred with the tip of a wooden toothpick to the bottom of a gelatin drug capsule and overlaid with the embedding mixture. The block is then hardened in an oven for 2-4 days. Special polyethylene capsules (or rubber embedding molds) with the ends already pointed into truncated pyramids help simplify the subsequent trimming procedure. 4) Block hardness Much of the success in cutting ultrathin sections lies in matching the hardness of the embedded tissue to the hardness of the block of embedding resin. If the tissue is too hard relative to the polymerized resin, it may pull out of the soft block during sectioning. If the tissue is too soft, sectioning tends to be uneven. Epon 812 is very convenient because the hardness of the final block is controlled by the proportions of the two hardeners used in the monomer mixtures. e. Microtomy 1) Ultramicrotomes An ultramicrotome consists of a horizontal bar, to the front of which is attached the specimen holder (chuck) (Figs. II.13 and II.14). The bar is moved forward by means of an advance mechanism. A knife mount is positioned in front of the specimen. Sections are cut by repeatedly moving the specimen past the knife edge with a very small advance of the specimen towards the edge made between each successive cut. The thickness of the section is determined by the magnitude of its forward advance. The advance mechanism is either mechanical (hand operated or by motor drive) or thermal. There are several commercially available microtomes. All employ designs with a bypass mechanism to avoid specimen knife contact on the up or return stroke. Recent models employ an electromagnet to pull back the knife holder assembly by about 25 µm, enabling the specimen to bypass the knife without rubbing against it on the return stroke. To efficiently cut thin sections reproducibly, an ultramicrotome must meet the following requirements: - All movements must be free of vibrations of magnitude the order of 10 nm. The condition of the thin section that is cut depends largely on the response of the specimen block to the strains to which it is subjected while actually being cut. The ideal embedding medium is one which absorbs all extraneous strains elastically and recovers completely after the section is cut. 2) Block trimming The block is usually mounted in a chuck and the rounded end of the embedded block is trimmed freehand with a razor blade to a four-sided truncated pyramid of about 45° angle and about 0.5 mm square face (Figs. II.15-II.17). The smaller the area of the block face, the easier it is to cut very thin sections. Commercially available machines can be used to give precise trimming of the block. If the top and bottom edges of the block face are not exactly parallel, it may be impossible to obtain a ribbon of sections or the ribbon will curve over in one direction or another (see Fig. II.37 later) and cannot be mounted on a straight slot grid or holder. 3) Knives The characteristics of an ideal knife for ultrathin sectioning include: - The radius of curvature of the edge is considerably smaller than the thickness of the thinnest section required. Single crystal diamond knives approach this ideal. Originally, sharpened steel razor blades were used for thin sectioning but these knives had relatively short useful life spans and their edges were subject to corrosion by the trough liquid and room atmosphere. Glass knives have proven to be the most popular of the cutting edges because they are inexpensive, relatively easy to make and are convenient to use. A glass knife is generally triangular in shape, and is obtained by breaking a square piece of glass along a diagonal line scored into its surface. The cutting edge should be straight and even and the front surface (facing the specimen) very flat and smooth. The polished fractured edge of a diamond knife is harder and more wear resistant than glass, but it is not a better cutting edge. Also, the diamond knife must be cleaned before each use (Fig. II.18). Making good glass knives is the single most important step in obtaining good sections. The knives can be made by hand (only with much experience) or using a special machine which helps give reproducible results (i.e. length and position of score, depth of score, position of the bending force, amount of the bending force, and the rate at which it is applied) (Fig. II.19). The position of the knife, relative to the tissue block, is critical if uniformly thin sections are to be cut. The positional relationship of the knife edge to the specimen involves several angles defined as follows (Fig. II.20): Rake Angle - the angle between the line perpendicular to the front of the face of the block and the upper facet of the knife edge. Knife Angle - the angle subtended by the rear and upper facets of the knife edge. Also known as the bevel angle. Clearance Angle - the angle between the rear facet of the knife edge and the vertical plane of cutting. The usual knife angle is about 45° and the clearance angle is commonly adjusted to 2-5°. Thin sections have little physical strength. If they are cut on a dry knife, the static friction between the facet and the cut surface almost always causes the section to crumple up on itself as it is cut. Even so, the chances of picking up the section onto a metal grid undamaged and undistorted are very slim. The use of a trough or "boat" of liquid at the cutting edge provides a low friction surface for support of the freshly cut section. It also simplifies picking up the section on specimen support grids. The simplest way to prepare a trough is with the use of some waterproof tape (e.g.. silver tape) positioned so it doesn't overlap the cutting edge or the front face of the knife. Melted dental wax is used to seal the trough against leaks. Manufactured metal troughs are also available. For diamond knives, the trough is an integrally cast part of the holder (Figs. II.21-II.23). 4) Cutting thin sections The trimmed block is mounted in the microtome chuck. If the chuck is tightened too much the plastic will be squeezed out and flow towards the knife causing ultrathick sectioning. If the block is held too loosely it may vibrate and cause chatter marks on the sections. The block should also be set as far into the chuck as possible with little overhang to reduce chatter. The top and bottom of the block face should be set parallel to the knife edge (Figs. II.24 - II.26). In setting up the knife, the most important angle is the clearance angle, which should be small, but large enough so the block face does not contact the front face of the knife after the section is cut. The normal angle is between 1° and 5°. The softer the block, the greater the clearance angle needed. Once the block face is parallel to the knife edge, the trough is filled with water exactly to the level of the knife edge, leaving no meniscus. The knife is carefully advanced to within a few microns of the specimen block face (this takes experience to master without ending up too far away or so close that the knife edge hits the specimen). The automatic advance of the ultramicrotome is then activated to initiate the thin sectioning. Alternatively, on manual microtomes, each cycle of sectioning is initiated by the operator who determines the thickness of each section by setting the advance mechanism to the desired position. If everything has been set up correctly, a ribbon of thin sections should float from the knife edge into the trough (Figs. II.27-II.28). 5) Flattening compressed sections The stress in compressed sections (Fig. II.29) can often be relieved by exposing them to the vapor of a solvent such as xylene or ethylene dichloride, applied by holding a small brush dipped in the solvent close to the surface of the section as it floats on the water surface. Sometimes a little ethanol or acetone is added to the water in the trough to help release compression strains or distortions introduced during sectioning. Surface tension acting on the cut section is able to stretch out the section to the original dimension of the block face. 6) Mounting sections on grids The most common procedure for transferring the thin sections to the EM grid is to insert the grid, held with a pair of forceps, under the water surface and pick up the floating sections from below (Figs. II.30 and II.31). The fragile sections will generally wrinkle if they are picked up from above. 7) Section thickness All modern ultramicrotomes have automatic advance mechanisms. It is generally difficult if not impossible to cut sections of uniform thickness much thinner than 30-40 nm. The thickness of sections can be estimated from their color in reflected light. The difference in path length between the light reflected from the surface of the water beneath the section and the light reflected from the upper surface of the section gives rise to interference between the two light rays, which, if white light is used, causes reinforcement at certain visible wavelengths and hence makes the sections appear to be colored. The thinnest sections (<60 nm) appear dark grey to grey in color. Thicker sections appear silver (60-90 nm), gold (90-150 nm), or purple (150-190 nm). The quality of the section cut is primarily dependent on the quality of the knife edge and the dimensions and shape of the specimen block. A dull or chipped knife, inadequate clearance angle, and too large a block face may result in faulty sectioning. With a good knife edge that is properly positioned and with a well trimmed block, the quality of the section is dependent on the physical properties of the embedded block which effect its behavior when it is under strain during the actual cutting process. f. Section staining Thin sections are usually stained with solutions of heavy metal salts to enhance the scattering contrast of specimens by increasing the mass density differences of various components of tissues and cells, thus increasing the scattering of electrons outside the objective aperture. The metal ions of the staining solutions form complexes with certain components of cells, thus increasing their density. Often, such staining has little chemical specificity, but the contrast of components such as ribosomes and membranes is increased relative to their surroundings. The minimum thickness for visibility of a specimen is a function of the thickness of the object (in nm) times its density (g/cm3) and must be about 10 or greater. An unstained biological specimen has a density of about 1g/cm3, thus it must be at least 10 nm thick to be barely visible relative to its surroundings. A variety of stains are available, the most useful employing lead or uranium ions. The most commonly used heavy metal for section staining is lead (as lead citrate). 1% uranyl acetate in water or alcohol is the other most commonly used stain (alcohol appears to promote more rapid staining). Because of the different staining properties of uranium and lead, it is common practice to double-stain sections, first using warm uranyl acetate (30 min) followed by lead citrate (1-2 min). Other common stains include phosphotungstic acid and potassium permanganate and other lead (e.g. tartrate, cacodylate, citrate and ammonium acetate) and uranyl salts. The section is reacted with a solution of a heavy metal salt such as lead citrate by floating the sections mounted on the grid on the surface of a 50-100 µl droplet of the staining solution, usually for 10-60 minutes. The single droplet of staining solution is placed on the surface of a piece of Parafilm or dental wax and the grid is placed with the sections facing the surface of the drop. The grid is then removed and the sections washed with distilled water to prevent stain crystals from contaminating the surface of the section when it dries. g. Sectioning artifacts Unsatisfactory sections arise either from faulty embedding procedures or from faulty cutting. Section faults must be clearly distinguished from artifacts of faulty tissue preservation. Faulty sections are almost invariably due to faulty knives, improper knife angle, infiltration or polymerization, incorrect block hardness or lack of skill on the part of the operator (Figs. II.33-II.37). If sections cling to the upper edge of the block when cut and disappear down with it, the cause is probably due to the block pyramid being trimmed with a blunt razor blade. The pyramid sides must be mirror smooth or the ragged surface will cause tiny threads to remain between the section and upper edge of the block face. The same problem occurs if the block face has become wet. If the sections alternate between thin and thick, the cause may be a blunt knife, a bad match between block and tissue hardness, poor infiltration, or irregular polymerization due to inadequate mixing of the resin components. Skipping, in which a section is cut, then no section, etc., means that cut sections are twice the thickness desired. This may result from the advance being set too small. Chatter, or alternating lines of thick and thin areas of the section parallel to the cutting edge of the knife are caused by vibration of the knife or the block or both during sectioning (Fig. II.38). The block may be protruding too far out of the chuck or it may not be firmly gripped. The knife may also not be firmly gripped or the clearance angle may not be set correctly or the cutting speed may be set too high. Scratches running in the cutting direction arise from imperfections (chips) in the cutting edge of the knife. Either a new portion of the knife edge must be used or the knife should be replaced. 
Fig. II.13 The specimen arm in the LKB Ultrotome is moved up and down along the same path, but the knife is retracted during the upward stroke until the specimen block has cleared the knife edge (side view). (From Reid, p.227) 
Fig. II.14. The movements of the cutting cycle of the LKB Ultrotome. Retraction of the knife during the upstroke. Up-and-down movement of the specimen rod. Feed of the specimen block. All of these movements are confined to one plane, that of the paper. (From Sjostrand, p.247) 
Fig. II.15 Stages of trimming of a specimen block. The straight arrows indicate the directions of cutting with the trimming blade. (From Reid, p.278) 
Fig. II.16 Different methods of trimming a specimen block. (a) The sides are cut in a step. (b) The cuts are continued to the sides of the block. (From Reid, 5.5, p.279) 
Fig. II.17 The stages in obtaining ultrathin sections. Rough block trimming (1,2) is followed by the localization of the required tissue area in a 0.5 mm thick section stained for the light microscope (3). The block is retrimmed to the size and shape shown in (8), and a ribbon of grey sections is cut (5). The ribbon is divided u, using the finely pointed ends of mounted eyelashes, into short lengths which are 'parked' against the side of the trough (6). The grid is then introduced beneath the water surface, and each short length of sections is drawn on to a grid by means of an eyelash. The mounted sections (7) are then stained before examination in the electron microscope. (From Meek 1st ed., p.453) 
Fig. II.19. The LKB knife maker. (From Sjostrand, p.270) 
Fig. II.18 Cleaning a diamond knife with a soft wood stick. All movements must be made parallel to the kniofe edge. (From Reid, p.266) 
Fig. II.20 Knife and cutting angles. The clearance angle is tangential to the arc followed by the specimen. The knife angle does not necessarily bear any relationship to the actual cutting angle at the specimen. (From Meek 1st ed., p.448) 
Fig. II.21 (a) An adhesive tape trough waxed to the knife. (b) A metal trough waxed to the knife. The metal trough has a greater water surface area for manipulating ribbons of serial sections. (From Meek 1st ed., p.445)
Fig. II.22 A glass knife fitted with a metal trough. The back of the trough is gently warmed to melt the wax inside. (From Reid, p.263) 
Fig. II.23 A tape trough being applied to a glass knife. The razor blade is only used to cut the tape on the side furthest from the usable region of the knife edge. (From Reid,p.264)
Fig. II.24 The specimen block is oriented so that the upper and lower edges of the block face are parallel to the knife edge. The broken lines indicate the part of the face that is behind the knife when viewed from the front. (From Reid, p.290) 
Fig. II.25 To ensure that the whole of the block face is cut, the specimen holder is adjusted until the knife edge is parallel to its reflection seen in the mirror-like block face. (Viewed at a small angle from the vertical.) (From Reid, p.291)
Fig. II.26 When viewed from above, the upper edge of the block face should be parallel to the knife edge. (From Reid, p.293) 
Fig. II.27 Sections adhering together in ribbon formation while floating on a liquid surface. (From Reid, p.222) 
Fig. II.28 The appearance of the sections floating on the water surface in the trough when illuminating conditions are correct. (From Meek 1st ed., p.451) 
Fig. II.29 The stresses suffered by the section and the block during cutting. The actual cutting edge is curved, and fractures the block before paring the section off. The section may be considerably compressed, especially if the knife edge is blunt. (From Meek 1st ed., p.449)
Fig. II.30 (a) A specimen support grid held in a pair of tweezers. (b) The grid partly removed from the water with one end of the ribbon of sections adhering to it. (From Reid, p.315) 
Fig. II.31 A cluster of single sections being collected by raising the grid underneath them. (From Reid, p.316) 
Fig. II.32. Transferring ribbons of sections to staining solution and to film-covered one-hole micro-scope grid by means of a similar one-hole grid. A ribbon slightly shorter than the hole of the grid is broken off from the knife edge and moved to the center of the trough. (a) The ribbon is picked up by means of a one-hole grid with no supporting film, placed on the surface of the trough with the ribbon ion the center of the hole. The grid is picked up from the water surface by means of a pair of forceps. The ribbon stays in the center of the hole on the drop of water adhering to the grid. (b) The pick-up grid is placed on a drop of a staining solution in a Petri dish with wax-covered bottom. (c) Pick-up grid is then transferred to a similar grid with supporting film after proper washing by transferring the pick-up grid to a surface of distilled water. A water drop separates the two grids. (d) Removal of the water by means of a small piece of filter paper. (e) and (f) When the water is removed the ribbon sinks down through the hole of the pick-up grid and eventu-ally is collected on the supporting film of the second grid. It is important that the two grids are prevented from making contact over any large surface area by means of the pair of forceps holding the second grid. (From Sjostrand, p.292)
Fig. II.33. Displacement of the embedding medium due to plastic and elastic flow as a basis for splitting off of a thin surface layer of the specimen block. The reduction of the dimension of this surface layer in the direction of the cutting is indicated by arrows. (From Sjostrand, p.233) 
Fig. II.35. The theory of ultrathin sectioning. A zone of maximum deformation is indicated by the darkest stippling. (From Sjostrand, p.233) 
Fig. II.34. Theory of ultrathin sectioning. The displacement of the surface layer of the block introduced by the knife is partially reversible (elastic deformation). (From Sjostrand, p.233)
Fig. II.36 Variations in thickness seen as straight bands parallel to the knife edge are caused by vibrations. (From Reid, p.309) 
Fig. II.37 The upper and lower edges of the block face are not parallel to one another. (a) The ribbon of sections has curved towards the side of the trough and become obstructed. (right) When the top edge of the block is not parallel to the knife, the sections are only in contact with each other at one point and quickly float apart. (From Reid, p.301) 
Fig. II.38. Periodic variation of section thickness - "chatter" - due to vibrations of the block tip caused by the impact of the knife on the block tip. (From Sjostrand, p.235)