1. This article discusses current electron microscope specimen techniques (particularly thin sectioning) and various special aspects of protoplasmic structure. It is not a general review and deals only with subjects studied in this laboratory. Dimensions specified are approximate only.
2. As regards sectioning (which is inevitable for the study of fine cell structure in most complex tissues, by whatever microscope) electron microscope practice has two considerable advantages over light microscope practice: (a) osmium tetroxide solutions alone (without addition of dichromate, &c.) are excellent for cytoplasmic fixation; (b) after cutting, sections receive no further treatment (such as de-waxing and re- and de-hydration). Moreover, criticisms regarding vacuum drying of electron microscope specimens are irrelevant to electron microscope studies of sections. For high resolution, sections must be cut at 005 µ or less and examined without removal of embedding medium. Such sections are obtained by embedding in medium-hard plastics and cutting on a glass knife. Specially sharpened blades can be used instead of glass, but it is doubtful whether waxes can be substituted for plastics.
3. Many, possibly all, animal and plant flagella contain two similar central subfibrils surrounded by a ring of nine fibrils different in size and chemical composition from the central pair. As far as is known at present, mammalian sperm are unique (a) in containing a second concentric ring of nine sub-fibrils and (b) in possessing a double-helix sheath round the axial sub-fibrils of the tail. Bacterial flagella consist of single fibrils each equivalent to one of the component sub-fibrils of a multi-fibrillar flagellum. They often occur in bunches, but so far no intermediates have been found between these bunches and the sheathed ‘9 + 2’ flagella of animals and plants. Vertebrate striated muscle consists of sub-fibrils which are (very roughly) 100 Å thick, 250 Å centre to centre and which in resting amphibian muscle have a 400 A periodicity; in cross sections these sub-fibrils are packed solid (not in hollow cylinders) and in a fairly regular array.
4. Nuclear, cell, and mitochondrial membranes appear double in cross sections (150-300 Å thick); this may be due to the dissolving away of internal lipoid leaving two outer protein sheets, but none of these membranes is thin enough to contain simply a bimolecular lipoid layer. Electron microscope studies of striated cell borders confirm that in some sites there may be distinct filaments of variable length and in others closed ducts, or rods, covered distally by a smooth membrane. One border described contains distinct filaments which join basal mitochondria. It is not yet certain whether the complex internal ‘double-membranes’ of sectioned mitochondria arise from tubes, or paired sheets, or both.
5. When sectioned after freezing-drying or buffered osmic or formal fixation, the cytoplasm of many protein secreting cells in vertebrates is full of double membranes, like those of mitochondria, and of equally uncertain origin. Sections of many other cells show similar structures, varying in thickness from 75 Å--600 Å. There is some evidence that they are associated with cytoplasmic ribonucleic acid.